Researcher Profile and Settings

Affiliation

• Faculty of Information Science and Technology　Media and Network Technologies　Information Media Science and Technology

Job Title

• Professor

URL

http://www-lmd.ist.hokudai.ac.jp

J-Global ID

• 200901012583237794

Research Interests

• image generation   action analysis   multi-spectrum analysis   machine learning   deep learning   CT   PET   X-ray image   SNS   electron microscope   image recognition   image restoration   Web mining   supre-resolution   image coding   medical image   satellite image   infrastructure   multimedia processing   EEG   NIRS   MRI   visualization   big data   IoT   AI   information retrieval   genetic algorithm   texture   noise reduction   ディジタル フィルタ   Fuzzy inference   sports video   ウェーブレット   music   quantization   model identification   semantic analysis   テキスト処理   image processing   signal processing   画像検索   

Research Areas

• Informatics / Intelligent informatics
• Informatics / Intelligent robotics
• Life sciences / Basic nursing
• Informatics / Web and service informatics
• Informatics / Sensitivity (kansei) informatics
• Informatics / Human interfaces and interactions
• Informatics / Database science
• Manufacturing technology (mechanical, electrical/electronic, chemical engineering) / Measurement engineering
• Manufacturing technology (mechanical, electrical/electronic, chemical engineering) / Control and systems engineering

Educational Organization

•  Bachelor's degree program　School of Engineering
•  Master's degree program　Graduate School of Information Science and Technology
•  Doctoral (PhD) degree program　Graduate School of Information Science and Technology

Academic & Professional Experience

• 2022/07 - Today Hokkaido University
• 2020/10 - Today Hokkaido University Vice President
• 2020/04 - Today 北海道大学 大学院情報科学研究院 研究院長
• 2020/04 - Today 北海道大学 大学院情報科学研究院 学院長
• 2006/04 - Today Hokkaido University Graduate School of Information Science and Technology
• 2022/04 - 2023/03 Hokkaido University Education and Research Center for Mathematical and Data Science Director
• 2017/07 - 2021/03 HOKKAIDO UNIVERSITY Education and Research Center for Mathematical and Data Science Director
• 2018/04 - 2020/03 Hokkaido University Graduate School of Information Science and Technology
• 2017/04 - 2020/03 北海道大学 総合ＩＲ室副室長
• 2013/04 - 2020/03 Hokkaido University
• 2017/04 - 2019/03 Hokkaido University Front Office for Human Resource Education and Development
• 2004/04 - 2006/03 Hokkaido University Graduate School of Information Science and Technology
• 1997/04 - 2004/03 Hokkaido University Graduate School of Engineering
• 1994/01 - 1997/03 Hokkaido University School of Engineering
• 1989/08 - 1993/12 Hokkaido University Research Institute for Electronic Science

Education

• 1986/04 - 1988/03  北海道大学
• 1982/04 - 1986/03  Hokkaido University  School of Engineering

Association Memberships

• INFORMATION PROCESSING SOCIETY OF JAPAN   日本学術会議   米国IEEE   電子情報通信学会   映像情報メディア学会   日本工学アカデミー   

Research Activities

Published Papers

• Zero-Shot Traffic Sign Recognition Based on Midlevel Feature Matching

  Yaozong Gan, Guang Li, Ren Togo, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Sensors 23 (23) 9607 - 9607 2023/12/04 

  Traffic sign recognition is a complex and challenging yet popular problem that can assist drivers on the road and reduce traffic accidents. Most existing methods for traffic sign recognition use convolutional neural networks (CNNs) and can achieve high recognition accuracy. However, these methods first require a large number of carefully crafted traffic sign datasets for the training process. Moreover, since traffic signs differ in each country and there is a variety of traffic signs, these methods need to be fine-tuned when recognizing new traffic sign categories. To address these issues, we propose a traffic sign matching method for zero-shot recognition. Our proposed method can perform traffic sign recognition without training data by directly matching the similarity of target and template traffic sign images. Our method uses the midlevel features of CNNs to obtain robust feature representations of traffic signs without additional training or fine-tuning. We discovered that midlevel features improve the accuracy of zero-shot traffic sign recognition. The proposed method achieves promising recognition results on the German Traffic Sign Recognition Benchmark open dataset and a real-world dataset taken from Sapporo City, Japan.
• Visual Emotion Recognition Through Multimodal Cyclic-Label Dequantized Gaussian Process Latent Variable Model.

  Naoki Saito 0006, Keisuke Maeda, Takahiro Ogawa 0001, Satoshi Asamizu, Miki Haseyama

  Journal of Robotics and Mechatronics 35 (5) 1321 - 1330 2023/10
• Self-supervised learning for gastritis detection with gastric X-ray images.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  International Journal of Computer Assisted Radiology and Surgery 18 (10) 1841 - 1848 2023/10
• Off-Screen Sound Separation Based on Audio-visual Pre-training Using Binaural Audio

  Masaki Yoshida, Ren Togo, Takahiro Ogawa, Miki Haseyama

  SENSORS 23 (9) 4540 - 4540 2023/05 

  This study proposes a novel off-screen sound separation method based on audio-visual pre-training. In the field of audio-visual analysis, researchers have leveraged visual information for audio manipulation tasks, such as sound source separation. Although such audio manipulation tasks are based on correspondences between audio and video, these correspondences are not always established. Specifically, sounds coming from outside a screen have no audio-visual correspondences and thus interfere with conventional audio-visual learning. The proposed method separates such off-screen sounds based on their arrival directions using binaural audio, which provides us with three-dimensional sensation. Furthermore, we propose a new pre-training method that can consider the off-screen space and use the obtained representation to improve off-screen sound separation. Consequently, the proposed method can separate off-screen sounds irrespective of the direction from which they arrive. We conducted our evaluation using generated video data to circumvent the problem of difficulty in collecting ground truth for off-screen sounds. We confirmed the effectiveness of our methods through off-screen sound detection and separation tasks.
• Multimodal Natural Language Explanation Generation for Visual Question Answering Based on Multiple Reference Data

  He Zhu, Ren Togo, Takahiro Ogawa, Miki Haseyama

  ELECTRONICS 12 (10) 2023/05 

  As deep learning research continues to advance, interpretability is becoming as important as model performance. Conducting interpretability studies to understand the decision-making processes of deep learning models can improve performance and provide valuable insights for humans. The interpretability of visual question answering (VQA), a crucial task for human-computer interaction, has garnered the attention of researchers due to its wide range of applications. The generation of natural language explanations for VQA that humans can better understand has gradually supplanted heatmap representations as the mainstream focus in the field. Humans typically answer questions by first identifying the primary objects in an image and then referring to various information sources, both within and beyond the image, including prior knowledge. However, previous studies have only considered input images, resulting in insufficient information that can lead to incorrect answers and implausible explanations. To address this issue, we introduce multiple references in addition to the input image. Specifically, we propose a multimodal model that generates natural language explanations for VQA. We introduce outside knowledge using the input image and question and incorporate object information into the model through an object detection module. By increasing the information available during the model generation process, we significantly improve VQA accuracy and the reliability of the generated explanations. Moreover, we employ a simple and effective feature fusion joint vector to combine information from multiple modalities while maximizing information preservation. Qualitative and quantitative evaluation experiments demonstrate that the proposed method can generate more reliable explanations than state-of-the-art methods while maintaining answering accuracy.
• Boosting automatic COVID-19 detection performance with self-supervised learning and batch knowledge ensembling.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Comput. Biol. Medicine 158 106877 - 106877 2023/05
• Canonical Correlation Analysis Introducing Label Dequantization for Visual Emotion Recognition

  SAITO Naoki, MAEDA Keisuke, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  電子情報通信学会論文誌D 情報・システム J106-D (5) 337 - 348 1880-4535 2023/05/01 

  Supervised multi-view canonical correlation analysis via cyclic label dequantization (sMVCCA-CLD) for visual emotion recognition is presented in this paper. In the CCA approach, the dimension of latent common space is limited to the minimum dimension among those of all features. The dimension of label features i.e., the number of classes for label information, tends to be lower than those of the other features. Then the dimension of the latent common space constructed by CCA becomes lower. Therefore, there is a possibility of misssing important information that is necessary for the estimation from the latent common space due to the dimensionality constraint. To overcome this constraint, sMVCCA-CLD increases the dimension of the label features by the label dequantization process, and estimates the canonical correlation between multi-view features. In addition, sMVCCA-CLD performs the label dequantization considering that the emotions are represented by a cyclic model, e.g., Plutchik's and Mikel's wheels. Consequently, the construction of the latent common space for the accurate recognition of emotions becomes feasible.
• COVID-19 detection based on self-supervised transfer learning using chest X-ray images.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Int. J. Comput. Assist. Radiol. Surg. 18 (4) 715 - 722 2023/04
• Estimation of Degradation Degree in Road Infrastructure Based on Multi-Modal ABN Using Contrastive Learning.

  Takaaki Higashi, Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 23 (3) 1657 - 1657 2023/02
• Diversity Learning Based on Multi-Latent Space for Medical Image Visual Question Generation.

  He Zhu, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 23 (3) 1057 - 1057 2023/02 

  Auxiliary clinical diagnosis has been researched to solve unevenly and insufficiently distributed clinical resources. However, auxiliary diagnosis is still dominated by human physicians, and how to make intelligent systems more involved in the diagnosis process is gradually becoming a concern. An interactive automated clinical diagnosis with a question-answering system and a question generation system can capture a patient's conditions from multiple perspectives with less physician involvement by asking different questions to drive and guide the diagnosis. This clinical diagnosis process requires diverse information to evaluate a patient from different perspectives to obtain an accurate diagnosis. Recently proposed medical question generation systems have not considered diversity. Thus, we propose a diversity learning-based visual question generation model using a multi-latent space to generate informative question sets from medical images. The proposed method generates various questions by embedding visual and language information in different latent spaces, whose diversity is trained by our newly proposed loss. We have also added control over the categories of generated questions, making the generated questions directional. Furthermore, we use a new metric named similarity to accurately evaluate the proposed model's performance. The experimental results on the Slake and VQA-RAD datasets demonstrate that the proposed method can generate questions with diverse information. Our model works with an answering model for interactive automated clinical diagnosis and generates datasets to replace the process of annotation that incurs huge labor costs.
• Canonical Correlation Analysis Introducing Label Dequantization for Visual Emotion Recognition

  斉藤直輝, 前田圭介, 小川貴弘, 浅水仁, 長谷山美紀

  電子情報通信学会論文誌 D(Web) J106-D (5) 1881-0225 2023
• Few-Shot Personalized Saliency Prediction Using Tensor Regression for Preserving Structural Global Information.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2307.02799 2023
• Personalized Content Recommender System via Non-verbal Interaction Using Face Mesh and Facial Expression.

  Yuya Moroto, Rintaro Yanagi, Naoki Ogawa, Kyohei Kamikawa, Keigo Sakurai, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ACM Multimedia 9399 - 9401 2023
• Gromov-Wasserstein Autoencoders.

  Nao Nakagawa, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICLR 2023
• Proficiency-level Estimation Using Heterogeneous Features via Label Dequantized CCA.

  Tatsuki Seino, Naoki Saito 0006, Takahiro Ogawa 0001, Satoshi Asamizu, Miki Haseyama

  ICCE-Taiwan 813 - 814 2023
• Parameter-efficient Tuning of a Pre-trained Model via Prompt Learning in Cross-modal Retrieval.

  Huaying Zhang, Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-Taiwan 811 - 812 2023
• Binaural Audio Generation with Data Augmentation from 360° Videos.

  Masaki Yoshida, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-Taiwan 795 - 796 2023
• Prediction of Shoot Events by Considering Spatio-temporal Relations of Multimodal Features.

  Ryota Goka, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-Taiwan 793 - 794 2023
• Shoot Event Prediction in Soccer Considering Expected Goals Based on Players' Positions.

  Ryota Goka, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-Taiwan 449 - 450 2023
• Estimation of Amyloid-β Positivity Using QSM Images Considering Age Information.

  Tsubasa Kunieda, Ren Togo, Noriko Nishioka, Yukie Shimizu, Shiro Watanabe, Kenji Hirata, Keisuke Maeda, Takahiro Ogawa 0001, Kohsuke Kudo, Miki Haseyama

  ICCE-Taiwan 165 - 166 2023
• A Medical Domain Visual Question Generation Model via Large Language Model.

  He Zhu, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-Taiwan 163 - 164 2023
• Defense Against Black-Box Adversarial Attacks Via Heterogeneous Fusion Features.

  Jiahuan Zhang, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Binauralization Robust To Camera Rotation Using 360° Videos.

  Masaki Yoshida, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Learning Graph Laplacian from Intrinsic Patterns via Gaussian Process.

  Koshi Watanabe, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Estimation of Visual Contents from Human Brain Signals via VQA Based on Brain-Specific Attention.

  Ryo Shichida, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Class-Aware Shared Gaussian Process Dynamic Model.

  Ryosuke Sawata, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Improving Dropout in Graph Convolutional Networks for Recommendation via Contrastive Loss.

  Hiroki Okamura, Keisuke Maeda, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1 - 5 2023
• Multi-Label Classification in Anime Illustrations Based on Hierarchical Attribute Relationships.

  Ziwen Lan, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 23 (10) 4798 - 4798 2023
• Prediction of Shooting Events in Soccer Videos Using Complete Bipartite Graphs and Players' Spatial-Temporal Relations.

  Ryota Goka, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 23 (9) 4506 - 4506 2023
• Material Compound-Property Retrieval Using Electron Microscope Images for Rubber Material Development.

  Rintaro Yanagi, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  IEEE Access 11 88258 - 88264 2023
• Hierarchical Multi-Label Attribute Classification With Graph Convolutional Networks on Anime Illustration.

  Ziwen Lan, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  IEEE Access 11 35447 - 35456 2023
• SpectralMAP: Approximating Data Manifold With Spectral Decomposition

  Koshi Watanabe, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE ACCESS 11 31530 - 31540 2169-3536 2023 

  Dimensionality reduction is widely used to visualize complex high-dimensional data. This study presents a novel method for effective data visualization. Previous methods depend on local distance measurements for data manifold approximation. This leads to unreliable results when a data manifold locally oscillates because of some undesirable effects, such as noise effects. In this study, we overcome this limitation by introducing a dual approximation of a data manifold. We roughly approximate a data manifold with a neighborhood graph and prune it with a global filter. This dual scheme results in local oscillation robustness and yields effective visualization with explicit global preservation. We consider a global filter based on principal component analysis frameworks and derive it with the spectral information of the original high-dimensional data. Finally, we experiment with multiple datasets to verify our method, compare its performance to that of state-of-the-art methods, and confirm the effectiveness of our novelty and results.
• Summarizing Data Structures with Gaussian Process and Robust Neighborhood Preservation

  Koshi Watanabe, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, ECML PKDD 2022, PT V 13717 157 - 173 2945-9133 2023 

  Latent variable models summarize high-dimensional data while preserving its many complex properties. This paper proposes a locality-aware and low-rank approximated Gaussian process latent variable model (LolaGP) that can preserve the global relationship and local geometry in the derivation of the latent variables. We realize the global relationship by imitating the sample similarity non-linearly and the local geometry based on our newly constructed neighborhood graph. Formally, we derive LolaGP from GP-LVM and implement a locality-aware regularization to reflect its adjacency relationship. The neighborhood graph is constructed based on the latent variables, making the local preservation more resistant to noise disruption and the curse of dimensionality than the previous methods that directly construct it from the high-dimensional data. Furthermore, we introduce a new lower bound of a log-posterior distribution based on low-rank matrix approximation, which allows LolaGP to handle larger datasets than the conventional GP-LVM extensions. Our contribution is to preserve both the global and local structures in the derivation of the latent variables using the robust neighborhood graph and introduce the scalable lower bound of the log-posterior distribution. We conducted an experimental analysis using synthetic as well as images with and without highly noise disrupted datasets. From both qualitative and quantitative standpoint, our method produced successful results in all experimental settings.
• Recallable Question Answering-Based Re-Ranking Considering Semantic Region for Cross-Modal Retrieval

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE OPEN JOURNAL OF SIGNAL PROCESSING 4 1 - 11 2023 

  Question answering (QA)-based re-ranking methods for cross-modal retrieval have been recently proposed to further narrow down similar candidate images. The conventional QA-based re-ranking methods provide questions to users by analyzing candidate images, and the initial retrieval results are re-ranked based on the user's feedback. Contrary to these developments, only focusing on performance improvement makes it difficult to efficiently elicit the user's retrieval intention. To realize more useful QA-based re-ranking, considering the user interaction for eliciting the user's retrieval intention is required. In this paper, we propose a QA-based re-ranking method with considering two important factors for eliciting the user's retrieval intention: query-image relevance and recallability. Considering the query-image relevance enables to only focus on the candidate images related to the provided query text, while, focusing on the recallability enables users to easily answer the provided question. With these procedures, our method can efficiently and effectively elicit the user's retrieval intention. Experimental results using Microsoft Common Objects in Context and computationally constructed dataset including similar candidate images show that our method can improve the performance of the cross-modal retrieval methods and the QA-based re-ranking methods.
• Interpretable Visual Question Answering Referring to Outside Knowledge.

  He Zhu, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2303.04388 2023
• Text-Guided Image Manipulation via Generative Adversarial Network With Referring Image Segmentation-Based Guidance.

  Yuto Watanabe, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  IEEE Access 11 42534 - 42545 2169-3536 2023 

  This study proposes a novel text-guided image manipulation method that introduces referring image segmentation into a generative adversarial network. The proposed text-guided image manipulation method aims to manipulate images containing multiple objects while preserving text-unrelated regions. The proposed method assigns the task of distinguishing between text-related and unrelated regions in an image to segmentation guidance based on referring image segmentation. With this architecture, the adversarial generative network can focus on generating new attributes according to the text description and reconstructing text-unrelated regions. For the challenging input images with multiple objects, the experimental results demonstrate that the proposed method outperforms conventional methods in terms of image manipulation precision.
• Cross-Modal Image Retrieval Considering Semantic Relationships With Many-to-Many Correspondence Loss.

  Huaying Zhang, Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  IEEE Access 11 10675 - 10686 2169-3536 2023 

  A cross-modal image retrieval that explicitly considers semantic relationships between images and texts is proposed. Most conventional cross-modal image retrieval methods retrieve the target images by directly measuring the similarities between the candidate images and query texts in a common semantic embedding space. However, such methods tend to focus on a one-to-one correspondence between a predefined image-text pair during the training phase, and other semantically similar images and texts are ignored. By considering the many-to-many correspondences between semantically similar images and texts, a common embedding space is constructed to assure semantic relationships, which allows users to accurately find more images that are related to the input query texts. Thus, in this paper, we propose a cross-modal image retrieval method that considers semantic relationships between images and texts. The proposed method calculates the similarities between texts as semantic similarities to acquire the relationships. Then, we introduce a loss function that explicitly constructs the many-to-many correspondences between semantically similar images and texts from their semantic relationships. We also propose an evaluation metric to assess whether each method can construct an embedding space considering the semantic relationships. Experimental results demonstrate that the proposed method outperforms conventional methods in terms of this newly proposed metric.
• Similar interior coordination image retrieval with multi-view features

  Ren Togo, Yuki Honma, Maiku Abe, Takahiro Ogawa, Miki Haseyama

  International Journal of Multimedia Information Retrieval 11 (4) 731 - 740 2192-6611 2022/08/26 

  This paper presents a novel similar image retrieval method for interior coordination. Interior coordination is very familiar; however, it is still an abstract and difficult concept. Even if we are involved in coordination every day, it does not mean we can become professional coordinators. By realizing the retrieval that can provide similar interior coordination images from a query room image, inspiring users' ideas for interior coordination becomes feasible. In the proposed method, we extract image features specialized for interior coordination and realize similar interior coordination image retrieval. We employ multi-view features: object-based, color-based, and semantic-based features, in the feature extraction phase. The extracted features are used to calculate similarity between the query image and the database images for the retrieval. We conducted experiments using a sophisticated real-world interior coordination image dataset. Furthermore, we qualitatively and quantitatively evaluated the effectiveness of the proposed method.
• Brain Decoding of Multiple Subjects for Estimating Visual Information Based on a Probabilistic Generative Model

  Takaaki Higashi, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  SENSORS 22 (16) 6148 - 6148 2022/08 

  Brain decoding is a process of decoding human cognitive contents from brain activities. However, improving the accuracy of brain decoding remains difficult due to the unique characteristics of the brain, such as the small sample size and high dimensionality of brain activities. Therefore, this paper proposes a method that effectively uses multi-subject brain activities to improve brain decoding accuracy. Specifically, we distinguish between the shared information common to multi-subject brain activities and the individual information based on each subject's brain activities, and both types of information are used to decode human visual cognition. Both types of information are extracted as features belonging to a latent space using a probabilistic generative model. In the experiment, an publicly available dataset and five subjects were used, and the estimation accuracy was validated on the basis of a confidence score ranging from 0 to 1, and a large value indicates superiority. The proposed method achieved a confidence score of 0.867 for the best subject and an average of 0.813 for the five subjects, which was the best compared to other methods. The experimental results show that the proposed method can accurately decode visual cognition compared with other existing methods in which the shared information is not distinguished from the individual information.
• Defect Detection of Subway Tunnels Using Advanced U-Net Network

  An Wang, Ren Togo, Takahiro Ogawa, Miki Haseyama

  SENSORS 22 (6) 2330 - 2330 2022/03 

  In this paper, we present a novel defect detection model based on an improved U-Net architecture. As a semantic segmentation task, the defect detection task has the problems of background-foreground imbalance, multi-scale targets, and feature similarity between the background and defects in the real-world data. Conventionally, general convolutional neural network (CNN)-based networks mainly focus on natural image tasks, which are insensitive to the problems in our task. The proposed method has a network design for multi-scale segmentation based on the U-Net architecture including an atrous spatial pyramid pooling (ASPP) module and an inception module, and can detect various types of defects compared to conventional simple CNN-based methods. Through the experiments using a real-world subway tunnel image dataset, the proposed method showed a better performance than that of general semantic segmentation including state-of-the-art methods. Additionally, we showed that our method can achieve excellent detection balance among multi-scale defects.
• Microscopy and biomimetics: the NanoSuit® method and image retrieval platform

  Takahiko Hariyama, Yasuharu Takaku, Hideya Kawasaki, Masatsugu Shimomura, Chiyo Senoh, Yumi Yamahama, Atsushi Hozumi, Satoru Ito, Naoto Matsuda, Satoshi Yamada, Toshiya Itoh, Miki Haseyama, Takahiro Ogawa, Naoki Mori, Shuhei So, Hidefumi Mitsuno, Masahiro Ohara, Shuhei Nomura, Masao Hirasaka

  Microscopy 71 (1) 1 - 12 2050-5698 2022/01/29 

  Abstract This review aims to clarify a suitable method towards achieving next-generation sustainability. As represented by the term ‘Anthropocene’, the Earth, including humans, is entering a critical era; therefore, science has a great responsibility to solve it. Biomimetics, the emulation of the models, systems and elements of nature, especially biological science, is a powerful tool to approach sustainability problems. Microscopy has made great progress with the technology of observing biological and artificial materials and its techniques have been continuously improved, most recently through the NanoSuit® method. As one of the most important tools across many facets of research and development, microscopy has produced a large amount of accumulated digital data. However, it is difficult to extract useful data for making things as biomimetic ideas despite a large amount of biological data. Here, we would like to find a way to organically connect the indispensable microscopic data with the new biomimetics to solve complex human problems.
• Popularity-Aware Graph Social Recommendation for Fully Non-Interaction Users.

  Nozomu Onodera, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Proceedings of the 4th ACM International Conference on Multimedia in Asia(MMAsia) 30 - 5 2022
• Affective Embedding Framework with Semantic Representations from Tweets for Zero-Shot Visual Sentiment Prediction.

  Yingrui Ye, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Proceedings of the 4th ACM International Conference on Multimedia in Asia(MMAsia) 6 - 7 2022
• Visual Sentiment Prediction Using Cross-Way Few-Shot Learning Based on Knowledge Distillation.

  Yingrui Ye, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  2022 IEEE International Conference on Image Processing(ICIP) 3838 - 3842 2022
• Human-Centric Image Retrieval with Gaze-Based Image Captioning.

  Yuhu Feng, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  2022 IEEE International Conference on Image Processing(ICIP) 3828 - 3832 2022
• Few-Shot Personalized Saliency Prediction with Similarity of Gaze Tendency Using Object-Based Structural Information.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  2022 IEEE International Conference on Image Processing(ICIP) 3823 - 3827 2022
• Gaussian Distributed Graph Constrained Multi-Modal Gaussian Process Latent Variable Model for Ordinal Labeled Data.

  Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  2022 IEEE International Conference on Image Processing(ICIP) 3798 - 3802 2022
• GCN-Based Multi-Modal Multi-Label Attribute Classification in Anime Illustration Using Domain-Specific Semantic Features.

  Ziwen Lan, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  2022 IEEE International Conference on Image Processing(ICIP) 2021 - 2025 2022
• Trend Prediction of Students' Mock Examination Results Using Matrix Completion.

  Yutaka Yamada, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  11th IEEE Global Conference on Consumer Electronics(GCCE) 891 - 892 2022
• Shoot Event Prediction from Soccer Videos by Considering Players' Spatio-Temporal Relations.

  Ryota Goka, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  11th IEEE Global Conference on Consumer Electronics(GCCE) 406 - 407 2022
• Refinement of Gaze-based Image Caption for Image Retrieval.

  Yuhu Feng, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  11th IEEE Global Conference on Consumer Electronics(GCCE) 272 - 273 2022
• Boosting Automatic COVID-19 Detection Performance with Self-Supervised Learning and Batch Knowledge Ensembling.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2212.09281 2022
• COVID-19 Detection Based on Self-Supervised Transfer Learning Using Chest X-Ray Images.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2212.09276 2022
• Union-set Multi-source Model Adaptation for Semantic Segmentation.

  Zongyao Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2212.02785 2022
• RGMIM: Region-Guided Masked Image Modeling for COVID-19 Detection.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2211.00313 2022
• Rubber Material Retrieval System using Electron Microscope Images for Rubber Material Development.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  MMAsia 44 - 3 2022
• Disentangled Image Attribute Editing in Latent Space via Mask-Based Retention Loss.

  Shunya Ohaga, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  MMAsia 25 - 7 2022
• Assessment of Image Manipulation Using Natural Language Description: Quantification of Manipulation Direction.

  Yuto Watanabe, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 1046 - 1050 2022
• Improving Model Adaptation for Semantic Segmentation by Learning Model-Invariant Features with Multiple Source-Domain Models.

  Zongyao Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 421 - 425 2022
• A Multimodal Interpretable Visual Question Answering Model Introducing Image Caption Processor.

  He Zhu, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 777 - 778 2022
• Cross-modal Image Retrieval Considering Semantic Relationships with Object Information.

  Huaying Zhang, Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 775 - 776 2022
• Free-viewpoint Sports Video Generation Based on Dynamic NeRF Considering Time Series.

  Masato Kawai, Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 408 - 409 2022
• Content-based Image Retrieval Using Effective Synthesized Images from Different Camera Views via pixelNeRF.

  Yuki Era, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 404 - 405 2022
• Analysis of Relationships between Visual Cognitive Contents and Response of Each Brain Region via Visual Question Answering.

  Ryo Shichida, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 402 - 403 2022
• GCN-based Collaborative Filtering Considering Personality Bias.

  Hiroki Okamura, Keisuke Maeda, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 278 - 279 2022
• Prediction of Amyloid-β Positivity Using QSM Images Based on Bootstrap Your Own Latent.

  Tsubasa Kunieda, Ren Togo, Noriko Nishioka, Yukie Shimizu, Shiro Watanabe, Kenji Hirata, Keisuke Maeda, Takahiro Ogawa 0001, Kohsuke Kudo, Miki Haseyama

  GCCE 137 - 138 2022
• Cross-platform Recommendation Considering Common Users' Preferences Based on Preference Propagation GraphNet.

  Kazuki Yamamoto, Keisuke Maeda, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 135 - 136 2022
• Union-Set Multi-source Model Adaptation for Semantic Segmentation.

  Zongyao Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ECCV (29) 579 - 595 2022
• Trial Analysis of the Relationship between Taste and Biological Information Obtained While Eating Strawberries for Sensory Evaluation.

  Keisuke Maeda, Ren Togo, Takahiro Ogawa 0001, Shin-ichi Adachi, Fumiaki Yoshizawa, Miki Haseyama

  Sensors 22 (23) 9496 - 9496 2022
• Distress Detection in Subway Tunnel Images via Data Augmentation Based on Selective Image Cropping and Patching.

  Keisuke Maeda, Saya Takada, Tomoki Haruyama, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 22 (22) 8932 - 8932 2022
• Compressed gastric image generation based on soft-label dataset distillation for medical data sharing.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Comput. Methods Programs Biomed. 227 107189 - 107189 2022
• Dataset Complexity Assessment Based on Cumulative Maximum Scaled Area Under Laplacian Spectrum.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2209.14743 2022
• Compressed Gastric Image Generation Based on Soft-Label Dataset Distillation for Medical Data Sharing.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2209.14635 2022
• Dataset Distillation using Parameter Pruning.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2209.14609 2022
• Dataset Distillation for Medical Dataset Sharing.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2209.14603 2022
• Gromov-Wasserstein Autoencoders.

  Nao Nakagawa, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2209.07007 2022
• TriBYOL: Triplet BYOL for Self-Supervised Representation Learning.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2206.03012 2022
• Self-Knowledge Distillation based Self-Supervised Learning for Covid-19 Detection from Chest X-Ray Images.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  CoRR abs/2206.03009 2022
• Generating Captions of Imagined content from Human Brain Activities Applying An Image Captioning Model.

  Saya Takada, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  4th IEEE Global Conference on Life Sciences and Technologies(LifeTech) 614 - 615 2022
• Knowledge-Guided Sequential Recommendation with Reinforcement Learning Using Empirical Distribution Function.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  4th IEEE Global Conference on Life Sciences and Technologies(LifeTech) 187 - 188 2022
• Transformer Based Multimodal Scene Recognition in Soccer Videos.

  Yaozong Gan, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICME Workshops 1 - 6 2022
• Scene Retrieval in Soccer Videos by Spatial-temporal Attention with Video Vision Transformer.

  Yaozong Gan, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-TW 453 - 454 2022
• Multi-scale Defect Detection from Subway Tunnel Images with Spatial Attention Mechanism.

  An Wang, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-TW 305 - 306 2022
• Action Classification Based on LSTM Using First and Third Person Videos of Engineers Inspecting Bridges.

  Tsuyoshi Masuda, Keisuke Maeda, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICCE-TW 46 (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37)) 303 - 304 1342-6893 2022
• TriBYOL: Triplet BYOL for Self-Supervised Representation Learning.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 3458 - 3462 2022
• Divergence-Guided Feature Alignment for Cross-Domain Object Detection.

  Zongyao Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 2240 - 2244 2022
• Self-Knowledge Distillation based Self-Supervised Learning for Covid-19 Detection from Chest X-Ray Images.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 1371 - 1375 2022
• Regularization Meets Enhanced Multi-Stage Fusion Features: Making CNN More Robust against White-Box Adversarial Attacks.

  Jiahuan Zhang, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 22 (14) 5431 - 5431 2022 

  Regularization has become an important method in adversarial defense. However, the existing regularization-based defense methods do not discuss which features in convolutional neural networks (CNN) are more suitable for regularization. Thus, in this paper, we propose a multi-stage feature fusion network with a feature regularization operation, which is called Enhanced Multi-Stage Feature Fusion Network (EMSF(2)Net). EMSF(2)Net mainly combines three parts: multi-stage feature enhancement (MSFE), multi-stage feature fusion (MSF2), and regularization. Specifically, MSFE aims to obtain enhanced and expressive features in each stage by multiplying the features of each channel; MSF2 aims to fuse the enhanced features of different stages to further enrich the information of the feature, and the regularization part can regularize the fused and original features during the training process. EMSF(2)Net has proved that if the regularization term of the enhanced multi-stage feature is added, the adversarial robustness of CNN will be significantly improved. The experimental results on extensive white-box attacks on the CIFAR-10 dataset illustrate the robustness and effectiveness of the proposed method.
• Controllable Music Playlist Generation Based on Knowledge Graph and Reinforcement Learning.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 22 (10) 3722 - 3722 2022
• Learning intra-domain style-invariant representation for unsupervised domain adaptation of semantic segmentation.

  Zongyao Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Pattern Recognit. 132 108911 - 108911 2022
• Dataset complexity assessment based on cumulative maximum scaled area under Laplacian spectrum.

  Guang Li 0008, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  Multimedia Tools and Applications 81 (22) 32287 - 32303 2022
• User-centric multimodal feature extraction for personalized retrieval of tumblr posts.

  Kazuma Ohtomo, Ryosuke Harakawa, Takahiro Ogawa 0001, Miki Haseyama, Masahiro Iwahashi

  Multimedia Tools and Applications 81 (2) 2979 - 3003 2022
• Chain centre loss: A psychology inspired loss function for image sentiment analysis.

  Yun Liang 0014, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Neurocomputing 495 118 - 128 2022
• Generative Adversarial Network Including Referring Image Segmentation For Text-Guided Image Manipulation.

  Yuto Watanabe, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 4818 - 4822 1520-6149 2022 

  This paper proposes a novel generative adversarial network to improve the performance of image manipulation using natural language descriptions that contain desired attributes. Text-guided image manipulation aims to semantically manipulate an image aligned with the text description while preserving text-irrelevant regions. To achieve this, we newly introduce referring image segmentation into the generative adversarial network for image manipulation. The referring image segmentation aims to generate a segmentation mask that extracts the text-relevant region. By utilizing the feature map of the segmentation mask in the network, the proposed method explicitly distinguishes the text-relevant and irrelevant regions and has the following two contributions. First, our model can pay attention only to the text-relevant region and manipulate the region aligned with the text description. Second, our model can achieve an appropriate balance between the generation of accurate attributes in the text-relevant region and the reconstruction in the text-irrelevant regions. Experimental results show that the proposed method can significantly improve the performance of image manipulation.
• Human Emotion Recognition Using Multi-Modal Biological Signals Based On Time Lag-Considered Correlation Maximization.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 4683 - 4687 2022
• Distributed Label Dequantized Gaussian Process Latent Variable Model for Multi-View Data Integration.

  Koshi Watanabe, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 4643 - 4647 1520-6149 2022 

  In this paper, we present a novel method for multi-view data analysis, distributed label dequantized Gaussian process latent variable model (DLDGP). DLDGP can integrate multi-view data and class information into a common latent space. In the previous multi-view methods, the dimension of label features transformed from the class information is much smaller than those of the other modalities, which causes a dimensionality-limitation problem in the latent space. DLDGP extends the dimension of the label features by a distributed label dequantization scheme. Additionally, DLDGP calculates correlation between different classes by encoding class information into distributed features. DLDGP can correctly capture the relationship between multi-view data and obtain the latent features with high expression ability. Experimental results show the effectiveness of our method by using the open dataset.
• Variational Bayesian Graph Convolutional Network for Robust Collaborative Filtering.

  Nozomu Onodera, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICASSP 3908 - 3912 2022
• Time-Lag Aware Latent Variable Model for Prediction of Important Scenes Using Baseball Videos and Tweets.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 22 (7) 2465 - 2465 2022
• Refining Graph Representation for Cross-Domain Recommendation Based on Edge Pruning in Latent Space.

  Taisei Hirakawa, Keisuke Maeda, Takahiro Ogawa 0001, Satoshi Asamizu, Miki Haseyama

  IEEE Access 10 12503 - 12509 2022
• Interactive Re-ranking via Object Entropy-Guided Question Answering for Cross-Modal Image Retrieval.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  ACM Trans. Multim. Comput. Commun. Appl. 18 (3) 68 - 17 1551-6857 2022 

  Cross-modal image-retrieval methods retrieve desired images from a query text by learning relationships between texts and images. Such a retrieval approach is one of the most effective ways of achieving the easiness of query preparation. Recent cross-modal image-retrieval methods are convenient and accurate when users input a query text that can be used to uniquely identify the desired image. However, in reality, users frequently input ambiguous query texts, and these ambiguous queries make it difficult to obtain desired images. To overcome these difficulties, in this study, we propose a novel interactive cross-modal image-retrieval method based on question answering. The proposed method analyzes candidate images and asks users questions to obtain information that can narrow down retrieval candidates. By only answering questions generated by the proposed method, users can reach their desired images, even when using an ambiguous query text. Experimental results show the proposed method's effectiveness.
• Deterioration Level Estimation Based on Convolutional Neural Network Using Confidence-Aware Attention Mechanism for Infrastructure Inspection.

  Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  Sensors 22 (1) 382 - 382 2022
• Domain Adaptive Cross-Modal Image Retrieval via Modality and Domain Translations

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E104A (6) 866 - 875 0916-8508 2021/06 

  Various cross-modal retrieval methods that can retrieve images related to a query sentence without text annotations have been proposed. Although a high level of retrieval performance is achieved by these methods, they have been developed for a single domain retrieval setting. When retrieval candidate images come from various domains, the retrieval performance of these methods might be decreased. To deal with this problem, we propose a new domain adaptive cross-modal retrieval method. By translating a modality and domains of a query and candidate images, our method can retrieve desired images accurately in a different domain retrieval setting. Experimental results for clipart and painting datasets showed that the proposed method has better retrieval performance than that of other conventional and state-of-the-art methods.
• Database-adaptive Re-ranking for Enhancing Cross-modal Image Retrieval.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  MM '21: ACM Multimedia Conference 3816 - 3825 2021
• Interpretable Representation Learning on Natural Image Datasets via Reconstruction in Visual-Semantic Embedding Space.

  Nao Nakagawa, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  2021 IEEE International Conference on Image Processing(ICIP) 2473 - 2477 2021
• Cross-Domain Recommendation Method Based On Multi-Layer Graph Analysis With Visual Information.

  Taisei Hirakawa, Keisuke Maeda, Takahiro Ogawa 0001, Satoshi Asamizu, Miki Haseyama

  ICIP 2688 - 2692 2021
• Time-Lag Aware Multi-Modal Variational Autoencoder Using Baseball Videos And Tweets For Prediction Of Important Scenes.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 2678 - 2682 2021
• Segmentation-Aware Text-Guided Image Manipulation.

  Tomoki Haruyama, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 2433 - 2437 2021
• Few-Shot Personalized Saliency Prediction using Person Similarity based on Collaborative Multi-Output Gaussian Process Regression.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 1469 - 1473 2021
• Interest Level Estimation via Multi-Modal Gaussian Process Latent Variable Factorization.

  Kyohei Kamikawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 1209 - 1213 2021
• Deep Metric Network Via Heterogeneous Semantics for Image Sentiment Analysis.

  Yun Liang 0014, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 1039 - 1043 2021
• Correlation-Aware Attention Branch Network Using Multi-Modal Data For Deterioration Level Estimation Of Infrastructures.

  Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  ICIP 1014 - 1018 2021
• User Background Information-Aware Music Recommendation with Reinforcement Learning-Based Knowledge Graph Exploration.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  IEEE International Conference on Consumer Electronics-Taiwan(ICCE-TW) 1 - 2 2021
• Degradation Level Estimation of Road Structures via Attention Branch Network with Text Data.

  Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  IEEE International Conference on Consumer Electronics-Taiwan(ICCE-TW) 1 - 2 2021
• Cross-view Self-supervised Learning via Momentum Statistics in Batch Normalization.

  Guang Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  IEEE International Conference on Consumer Electronics-Taiwan(ICCE-TW) 1 - 2 2021
• Triplet Self-Supervised Learning for Gastritis Detection with Scarce Annotations.

  Guang Li, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 787 - 788 2021
• Defense Against Image Captioning Attacks via A Robust and Stable Recurrent Neural Network.

  Jiahuan Zhang, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 785 - 786 2021
• Text-Guided Image Manipulation for Desired Region Using Referring Image Segmentation.

  Yuto Watanabe, Ren Togo, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 661 - 662 2021
• Multi-label Image Recognition Based on Multi-modal Graph Convolutional Networks Using Captioning Features.

  Ziwen Lan, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 46 (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37)) 273 - 274 1342-6893 2021
• A Trial of Fine-grained Classification of Expert-novice Level Using Bio-signals While Inspecting Subway Tunnels.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 204 - 205 2021
• Listener Recommendation for Artist Based on Knowledge Graph and Reinforcement Learning.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 202 - 203 2021
• Movie Rating Estimation Based on Weakly Supervised Multi-modal Latent Variable Model.

  Koshi Watanabe, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 195 - 196 2021
• Detection of Off-screen Sound Based on Loss Function of Self-supervised Audio-visual Spatialization.

  Masaki Yoshida, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 193 - 194 2021
• Visual Sentiment Prediction Using Few-shot Learning via Distribution Relations of Visual Features.

  Yingrui Ye, Yuya Moroto, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 191 - 192 2021
• Multi-class Similar Scene Retrieval in Soccer Videos: A Scene Confusion Reduction Method Based on Combination of Long and Short Frame Sequences.

  Yaozong Gan, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 117 - 118 2021
• Sports Action Detection Based on Self-Supervised Feature Learning and Object Detection.

  Tsuyoshi Masuda, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 54 - 55 2021
• Analysis of Social Trends Related to COVID-19 Pandemic Utilizing Social Media Data.

  Taisei Hirakawa, Keisuke Maeda, Takahiro Ogawa 0001, Satoshi Asamizu, Miki Haseyama

  GCCE 43 - 44 2021
• Estimating Imagined Images from Brain Activities via Visual Question Answering.

  Saya Takada, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 35 - 36 2021
• Action Classification from Egocentric Videos Using Reinforcement Learning-based Pose Estimation.

  Shunya Ohaga, Ren Togo, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 9 - 10 2021
• Graph Analysis-based Recommendation via Entity Embeddings Using Wikipedia.

  Nozomu Onodera, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  GCCE 5 - 6 2021
• Reliable Estimation of Deterioration Levels via Late Fusion Using Multi-View Distress Images for Practical Inspection.

  Keisuke Maeda, Naoki Ogawa, Takahiro Ogawa 0001, Miki Haseyama

  Journal of Imaging 7 (12) 273 - 273 2021
• Feature Integration Through Semi-Supervised Multimodal Gaussian Process Latent Variable Model With Pseudo-Labels for Interest Level Estimation.

  Kyohei Kamikawa, Keisuke Maeda, Takahiro Ogawa 0001, Miki Haseyama

  IEEE Access 9 163843 - 163850 2021
• [Paper] Personalized Recommendation of Tumblr Posts Using Graph Convolutional Networks with Preference-aware Multimodal Features

  Kazuma Ohtomo, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama, Masahiro Iwahashi

  ITE Transactions on Media Technology and Applications 9 (1) 54 - 61 2021
• A Proposal and Evaluation of a New Method Incorporating Indicators of Blood Flow and Resistance to Improve the Accuracy of Continuous Blood Pressure Estimation Using PPG

  川上健, 川上健, 川上健, 住友和弘, 菅野厚博, 小川貴弘, 南重信, 長谷山美紀

  電気学会論文誌 E 141 (6) 1341-8939 2021
• User-selectable Event Summarization in Unedited Raw Soccer Video via Multimodal Bidirectional LSTM

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 9 (1) 42 - 53 2186-7364 2021 

  A new method that generates user-selectable event summaries from unedited raw soccer videos is presented in this paper. Since there are more unedited raw soccer videos than broadcasted/distributed soccer videos and unedited videos have various viewers, it is necessary to analyze these videos for meeting the demands of various viewers. The proposed method introduces a multimodal CNN-BiLSTM architecture for analyzing unedited raw soccer videos. This architecture extracts candidate scenes for event summarization from unedited soccer videos and classifies these scenes into typical events. Finally, our method generates user-selectable event summaries by simultaneously considering the importance of candidate scenes and the event classification results. Experimental results using real unedited raw soccer videos show the effectiveness of our method.
• Cross-domain Recommendation Based on Multilayer Graph Analysis Using Subgraph Representation

  Taisei Hirakawa, Keisuke Maeda, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama

  INTERNATIONAL WORKSHOP ON ADVANCED IMAGING TECHNOLOGY (IWAIT) 2021 11766 0277-786X 2021 

  This paper presents cross-domain recommendation based on multilayer graph analysis using subgraph representation. The proposed method constructs two graphs in source and target domains utilizing user-item embedding and trains link relationships between the users' embedding features on each above graph via graph convolutional networks considering subgraph representation. Thus, the proposed method can obtain features with high representation ability, and this is the main contribution of this paper. Then the proposed method can estimate the user's embedding features in the target domain from those in the source domain and recommend items to users by using the estimated features. Experiments on real-world e-commerce datasets verify the effectiveness of the proposed method.
• Interest Estimation Method Based on 2D Pose Features on Security Camera

  Yuki Honma, Ren Togo, Maiku Abe, Takahiro Ogawa, Miki Haseyama

  INTERNATIONAL WORKSHOP ON ADVANCED IMAGING TECHNOLOGY (IWAIT) 2021 11766 0277-786X 2021 

  This paper proposes a customer interest estimation method using security camera to meet the demand of the retail industry. In the field of retail industry, it is considered that the understanding of customers' interests in the real store can be used for various marketing activities such as the product development and the layout of the store. Then, it is important to pay attention to customers' behavior in the real store. Their behavior is often recorded by the cameras installed in the store for security purposes. A method for estimating their interests from the videos of the security camera is presented in this paper. The novelty of our method is three-fold. Firstly, the experimental data of subjects in our group were taken by using the security camera already installed in the real store. Secondly, we used a pre-trained posture estimation model and treated the results as the features to be trained by a two-layer neural network model. Finally, a professional have annotated the subjects' interests. The effectiveness of our method was confirmed by comparing with benchmark supervised machine learning models.
• Preliminary study of AI-assisted diagnosis using FDG-PET/CT for axillary lymph node metastasis in patients with breast cancer

  Zongyao Li, Kazuhiro Kitajima, Kenji Hirata, Ren Togo, Junki Takenaka, Yasuo Miyoshi, Kohsuke Kudo, Takahiro Ogawa, Miki Haseyama

  EJNMMI RESEARCH 11 (1) 2191-219X 2021/01 

  Background To improve the diagnostic accuracy of axillary lymph node (LN) metastasis in breast cancer patients using 2-[F-18]FDG-PET/CT, we constructed an artificial intelligence (AI)-assisted diagnosis system that uses deep-learning technologies. Materials and methods Two clinicians and the new AI system retrospectively analyzed and diagnosed 414 axillae of 407 patients with biopsy-proven breast cancer who had undergone 2-[F-18]FDG-PET/CT before a mastectomy or breast-conserving surgery with a sentinel lymph node (LN) biopsy and/or axillary LN dissection. We designed and trained a deep 3D convolutional neural network (CNN) as the AI model. The diagnoses from the clinicians were blended with the diagnoses from the AI model to improve the diagnostic accuracy. Results Although the AI model did not outperform the clinicians, the diagnostic accuracies of the clinicians were considerably improved by collaborating with the AI model: the two clinicians' sensitivities of 59.8% and 57.4% increased to 68.6% and 64.2%, respectively, whereas the clinicians' specificities of 99.0% and 99.5% remained unchanged. Conclusions It is expected that AI using deep-learning technologies will be useful in diagnosing axillary LN metastasis using 2-[F-18]FDG-PET/CT. Even if the diagnostic performance of AI is not better than that of clinicians, taking AI diagnoses into consideration may positively impact the overall diagnostic accuracy.
• A Note on Detection of Sports Action Based on Temporal Cycle Consistency Learning

  Tsuyoshi Masuda, Ren Togo, Takahiro Ogawa, Miki Haseyama

  INTERNATIONAL WORKSHOP ON ADVANCED IMAGING TECHNOLOGY (IWAIT) 2021 11766 0277-786X 2021 

  This paper presents a method for action detection based on Temporal Cycle Consistency(TCC) Learning. The proposed method realizes the action detection of flexible length segments based on a frame-level action prediction technique. We enable calculation of similarities for spatio-temporal features based on TCC to detect target actions from input videos. Finally, our method determines temporal segments by smoothing the frame-level action detection result. Experimental results show the validity of the proposed method.
• Interior Coordination Image Retrieval with Object-Detection-Based and Color Features

  Ren Togo, Takahiro Ogawa, Miki Haseyama

  INTERNATIONAL WORKSHOP ON ADVANCED IMAGING TECHNOLOGY (IWAIT) 2021 11766 0277-786X 2021 

  This paper presents a new interior coordination image retrieval method using object-detection-based and color features. Interior coordination requires consideration of objects' positional information and the overall atmosphere of the room simultaneously. However, similar image retrieval methods considering the coordination characteristics have not been proposed. In the proposed method, we extract different types of features from interior coordination images and realize the similar interior coordination image retrieval based on our newly derived features.
• Self-Supervised Learning for Gastritis Detection with Gastric X-Ray Images.

  Guang Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  CoRR abs/2104.02864 2021
• Soft-Label Anonymous Gastric X-ray Image Distillation.

  Guang Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  CoRR abs/2104.02857 305 - 309 2021
• IR Questioner: QA-based Interactive Retrieval System.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  ICMR '21: International Conference on Multimedia Retrieval(ICMR) 611 - 614 2021 

  Image retrieval from a given text query (text-to-image retrieval) is one of the most essential systems, and it is effectively utilized for databases (DBs) on the Web. To make them more versatile and familiar, a retrieval system that is adaptive even for personal DBs such as images in smartphones and lifelogging devices should be considered. In this paper, we present a novel text-to-image retrieval system that is specialized for personal DBs. With the cross-modal scheme and the question-answering scheme, the developed system enables users to obtain the desired image effectively even from personal DBs. Our demo is available at https://sites.google.com/view/ir-questioner/.
• Human Emotion Estimation Using Multi-Modal Variational AutoEncoder with Time Changes.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2021 IEEE 3RD GLOBAL CONFERENCE ON LIFE SCIENCES AND TECHNOLOGIES (IEEE LIFETECH 2021) 67 - 68 2021 

  A human emotion estimation method via feature integration using multi-modal variational autoencoder (MVAE) with time changes is presented in this paper. To utilize multi-modal information such as gaze and brain activity data including some noises, the proposed method newly introduces MVAE into the human emotion estimation. Furthermore, the proposed MVAE can consider the changes in bio-signals with time and reduce the effect of noises caused in bio-signals by using the probabilistic variation. Experimental results with that of some state-of-the-art methods indicate that the proposed method is effective.
• Music Playlist Generation Based on Graph Exploration Using Reinforcement Learning.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2021 IEEE 3RD GLOBAL CONFERENCE ON LIFE SCIENCES AND TECHNOLOGIES (IEEE LIFETECH 2021) 2021 53 - 54 2021 

  Spreading of music streaming platforms that use playlists to make recommendations, automatic playlist generation has been actively researched. Recently, it has been reported that playlists that have high diversity and smooth track transitions increase user satisfaction. Our previous method that used a two-dimensional space as a reinforcement learning environment has achieved these demands, but there remains the problem that the content of multi-dimensional acoustic features cannot be retained accurately. To solve this problem, in this paper, we present a new method of music playlist generation based on reinforcement learning using a graph structure constructed from multi-dimensional acoustic features directly. The new playlist generation provides greater diversity and smoother track transitions than the previous method. Experimental results are shown for verifying the effectiveness of the proposal method.
• Question Answering from Brain Activity Data via Decoder Based on Neural Networks.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2021 IEEE 3RD GLOBAL CONFERENCE ON LIFE SCIENCES AND TECHNOLOGIES (IEEE LIFETECH 2021) 51 - 52 2021 

  We build a model that can estimate what subjects recognize from functional magnetic resonance imaging (fMRI) data via a visual question answering (VQA) model. The VQA model can generate an answer to a question about an image. We convert fMRI signals into image features via an fMRI decoder based on the relationship between the fMRI signals and the image features extracted from the gazed image. Then this allows the VQA model to answer a visual question from the fMRI signals measured while the subject is gazing at the image. Though brain decoding, which interprets what humans recognize, has become overwhelmingly popular in neuroscience, they often suffer from the small datasets of brain activity data. To overcome the small size of datasets of fMRI signals, we introduce an fMRI decoder based on neural networks that have a high expressive ability. Even when we do not have enough fMRI signals, the proposed method derives the answer to what a person is looking at from fMRI signals. Experimental results on several datasets show that our method allows us to answer a question about gazed images from fMRI signals.
• Cross-Domain Semi-Supervised Deep Metric Learning for Image Sentiment Analysis.

  Yun Liang 0014, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 4150 - 4154 2021
• Feature Integration via Semi-Supervised Ordinally Multi-Modal Gaussian Process Latent Variable Model.

  Kyohei Kamikawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 4130 - 4134 2021
• Multi-Modal Label Dequantized Gaussian Process Latent Variable Model for Ordinal Label Estimation.

  Masanao Matsumoto, Keisuke Maeda, Naoki Saito 0006, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 3985 - 3989 2021
• Semantic-Aware Unpaired Image-to-Image Translation for Urban Scene Images.

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 2150 - 2154 2021
• Classification of Expert-Novice Level Using Eye Tracking And Motion Data via Conditional Multimodal Variational Autoencoder.

  Yusuke Akamatsu, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 1360 - 1364 2021
• Estimation of Visual Features of Viewed Image From Individual and Shared Brain Information Based on FMRI Data Using Probabilistic Generative Model.

  Takaaki Higashi, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 1335 - 1339 2021 

  This paper presents a method for estimation of visual features based on brain responses measured when subjects view images. The proposed method estimates visual features of viewed images by using both individual and shared brain information from functional magnetic resonance imaging (fMRI) data when subjects view images. To extract an effective latent space shared by multiple subjects from high dimensional fMRI data, a probabilistic generative model that can provide a prior distribution to the space is introduced into the proposed method. Also, the extraction of a robust feature space with respect to noise for the individual information becomes feasible via the proposed probabilistic generative model. This is the first contribution of our method. Furthermore, the proposed method constructs a decoder transforming brain information into visual features based on collaborative use of both estimated spaces for individual and shared brain information. This is the second contribution of our method. Experimental results show that the proposed method improves the estimation accuracy of the visual features of viewed images.
• Human-Centered Favorite Music Classification Using EEG-Based Individual Music Preference Via Deep Time-Series CCA.

  Ryosuke Sawata, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP) 1320 - 1324 2021
• Rubber Material Property Prediction Using Electron Microscope Images of Internal Structures Taken under Multiple Conditions.

  Ren Togo, Naoki Saito 0006, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Sensors 21 (6) 2088 - 2088 2021 

  A method for prediction of properties of rubber materials utilizing electron microscope images of internal structures taken under multiple conditions is presented in this paper. Electron microscope images of rubber materials are taken under several conditions, and effective conditions for the prediction of properties are different for each rubber material. Novel approaches for the selection and integration of reliable prediction results are used in the proposed method. The proposed method enables selection of reliable results based on prediction intervals that can be derived by the predictors that are each constructed from electron microscope images taken under each condition. By monitoring the relationship between prediction results and prediction intervals derived from the corresponding predictors, it can be determined whether the target prediction results are reliable. Furthermore, the proposed method integrates the selected reliable results based on Dempster-Shafer (DS) evidence theory, and this integration result is regarded as a final prediction result. The DS evidence theory enables integration of multiple prediction results, even if the results are obtained from different imaging conditions. This means that integration can even be realized if electron microscope images of each material are taken under different conditions and even if these conditions are different for target materials. This nonconventional approach is suitable for our application, i.e., property prediction. Experiments on rubber material data showed that the evaluation index mean absolute percent error (MAPE) was under 10% by the proposed method. The performance of the proposed method outperformed conventional comparative property estimation methods. Consequently, the proposed method can realize accurate prediction of the properties with consideration of the characteristic of electron microscope images described above.
• Detection of Important Scenes in Baseball Videos via a Time-Lag-Aware Multimodal Variational Autoencoder.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Sensors 21 (6) 2045 - 2045 2021 

  A new method for the detection of important scenes in baseball videos via a time-lag-aware multimodal variational autoencoder (Tl-MVAE) is presented in this paper. Tl-MVAE estimates latent features calculated from tweet, video, and audio features extracted from tweets and videos. Then, important scenes are detected by estimating the probability of the scene being important from estimated latent features. It should be noted that there exist time-lags between tweets posted by users and videos. To consider the time-lags between tweet features and other features calculated from corresponding multiple previous events, the feature transformation based on feature correlation considering such time-lags is newly introduced to the encoder in MVAE in the proposed method. This is the biggest contribution of the Tl-MVAE. Experimental results obtained from actual baseball videos and their corresponding tweets show the effectiveness of the proposed method.
• Deterioration level estimation via neural network maximizing category-based ordinally supervised multi-view canonical correlation.

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  Multim. Tools Appl. 80 (15) 23091 - 23112 1380-7501 2021 

  A deterioration level estimation method via neural network maximizing category-based ordinally supervised multi-view canonical correlation is presented in this paper. This paper focuses on real world data such as industrial applications and has two contributions. First, a novel neural network handling multi-modal features transforms original features into features effectively representing deterioration levels in transmission towers, which are one of the infrastructures, with consideration of only correlation maximization. It can be realized by setting projection matrices maximizing correlations between multiple features into weights of hidden layers. That is, since the proposed network has only a few hidden layers, it can be trained from a small amount of training data. Second, since there exist diverse characteristics and an ordinal scale in deterioration levels, the proposed method newly derives category-based ordinally supervised multi-view canonical correlation analysis (Co-sMVCCA). Co-sMVCCA enables estimation of effective projection considering both within-class divergence and the ordinal scale between classes. Experimental results showed that the proposed method realizes accurate deterioration level estimation.
• Disentangled Representation Learning in Real-World Image Datasets via Image Segmentation Prior.

  Nao Nakagawa, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 110880 - 110888 2169-3536 2021 

  We propose a novel method that can learn easy-to-interpret latent representations in real-world image datasets using a VAE-based model by splitting an image into several disjoint regions. Our method performs object-wise disentanglement by exploiting image segmentation and alpha compositing. With remarkable results obtained by unsupervised disentanglement methods for toy datasets, recent studies have tackled challenging disentanglement for real-world image datasets. However, these methods involve deviations from the standard VAE architecture, which has favorable disentanglement properties. Thus, for disentanglement in images of real-world image datasets with preservation of the VAE backbone, we designed an encoder and a decoder that embed an image into disjoint sets of latent variables corresponding to objects. The encoder includes a pre-trained image segmentation network, which allows our model to focus only on representation learning while adopting image segmentation as an inductive bias. Evaluations using real-world image datasets, CelebA and Stanford Cars, showed that our method achieves improved disentanglement and transferability.
• Detection of Important Scenes in Baseball Videos via Bidirectional Time Lag Aware Deep Multiset Canonical Correlation Analysis.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 84971 - 84981 2169-3536 2021 

  A novel method for detection of important scenes in baseball videos based on correlation maximization between heterogeneous modalities via bidirectional time lag aware deep multiset canonical correlation analysis (BiTl-dMCCA) is presented in this paper. The proposed method enables detection of important scenes by collaboratively using baseball videos and their corresponding tweets. The technical contributions of this paper are twofold. First, since there are time lags between not only "tweets and corresponding multiple previous events" but also "events and corresponding multiple following posted tweets", the proposed method considers these bidirectional time lags. Specifically, the representation of such bidirectional time lags into the derivation of their covariance matrices is newly introduced. Second, the proposed method adopts textual, visual and audio features calculated from tweets and videos as multi-modal time series features. Important scenes are detected as abnormal scenes via anomaly detection based on a generative adversarial network using multi-modal features projected by BiTl-dMCCA. The proposed method does not need any training data with annotation. Experimental results obtained by applying the proposed method to actual baseball matches show the effectiveness of the proposed method.
• Distress Image Retrieval for Infrastructure Maintenance via Self-Trained Deep Metric Learning Using Experts' Knowledge.

  Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 65234 - 65245 2169-3536 2021 

  Distress image retrieval for infrastructure maintenance via self-trained deep metric learning using experts' knowledge is proposed in this paper. Since engineers take multiple images of a single distress part for inspection of road structures, it is necessary to construct a similar distress image retrieval method considering the input of multiple images to support determination of the level of deterioration. Thus, the construction of an image retrieval method while selecting an effective input from multiple images is described in this paper. The proposed method performs deep metric learning by using a small number of effective images labeled by experts' knowledge with information about their effectiveness and a large number of unlabeled images via a self-training approach. Specifically, an end-to-end learning approach that performs retraining of the model by assigning pseudo-labels to these unlabeled images according to the output confidence of the model is achieved. Thus, the proposed method can select an effective image from multiple images that are input at the retrieval as a query image. This is the main contribution of this paper. As a result, the proposed method realizes highly accurate retrieval of similar distress images considering the actual situation of inspection in which multiple images of a distress part are input.
• Text-Guided Style Transfer-Based Image Manipulation Using Multimodal Generative Models.

  Ren Togo, Megumi Kotera, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 64860 - 64870 2169-3536 2021 

  A new style transfer-based image manipulation framework combining generative networks and style transfer networks is presented in this paper. Unlike conventional style transfer tasks, we tackle a new task, text-guided image manipulation. We realize style transfer-based image manipulation that does not require any reference style images and generate a style image from the user's input sentence. In our method, since an initial reference input sentence for a content image can automatically be given by an image-to-text model, the user only needs to update the reference sentence. This scheme can help users when they do not have any images representing the desired style. Although this text-guided image manipulation is a new challenging task, quantitative and qualitative comparisons showed the superiority of our method.
• Perceived Image Decoding From Brain Activity Using Shared Information of Multi-Subject fMRI Data.

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 26593 - 26606 2169-3536 2021 

  Decoding a person's cognitive contents from evoked brain activity is becoming important in the field of brain-computer interaction. Previous studies have decoded a perceived image from functional magnetic resonance imaging (fMRI) activity by constructing brain decoding models that were trained with a single subject's fMRI data. However, accurate decoding is still challenging since fMRI data acquired from only a single subject have several disadvantageous characteristics such as small sample size, noisy nature, and high dimensionality. In this article, we propose a method to decode categories of perceived images from fMRI activity using shared information of multi-subject fMRI data. Specifically, by aggregating fMRI data of multiple subjects that contain a large number of samples, we extract a low-dimensional latent representation shared by multi-subject fMRI data. Then the latent representation is nonlinearly transformed into visual features and semantic features of the perceived images to identify categories from various candidate categories. Our approach leverages rich information obtained from multi-subject fMRI data and improves the decoding performance. Experimental results obtained by using two public fMRI datasets showed that the proposed method can more accurately decode categories of perceived images from fMRI activity than previous approaches using a single subject's fMRI data.
• Supervised Fractional-Order Embedding Multiview Canonical Correlation Analysis via Ordinal Label Dequantization for Image Interest Estimation.

  Masanao Matsumoto, Naoki Saito 0006, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE Access 9 21810 - 21822 2169-3536 2021 

  Supervised fractional-order embedding multiview canonical correlation analysis via ordinal label dequantization (SFEMCCA-OLD) for image interest estimation is presented in this paper. SFEMCCA-OLD is a CCA method that realizes accurate integration of features including low-dimensional ordinal label features. In general, since information is lost due to a limitation of the number of classes, i.e., the dimension of ordinal label information is smaller than those of other features, derivation of highly accurate integration of features is difficult. In SFEMCCA-OLD, the dimension of the ordinal label information can be increased by estimation of the canonical correlation between multiview features. We call this approach ordinal label dequantization. In addition, by introducing a fractional-order technique, our method can calculate optimal projections for noisy data such as real data. Experimental results show that the accuracy of SFEMCCA-OLD for image interest estimation is better than that of recent CCA-based methods.
• Chronic atrophic gastritis detection with a convolutional neural network considering stomach regions

  Misaki Kanai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  WORLD JOURNAL OF GASTROENTEROLOGY 26 (25) 3650 - 3659 1007-9327 2020/07 

  BACKGROUND The risk of gastric cancer increases in patients withHelicobacter pylori-associated chronic atrophic gastritis (CAG). X-ray examination can evaluate the condition of the stomach, and it can be used for gastric cancer mass screening. However, skilled doctors for interpretation of X-ray examination are decreasing due to the diverse of inspections. AIM To evaluate the effectiveness of stomach regions that are automatically estimated by a deep learning-based model for CAG detection. METHODS We used 815 gastric X-ray images (GXIs) obtained from 815 subjects. The ground truth of this study was the diagnostic results in X-ray and endoscopic examinations. For a part of GXIs for training, the stomach regions are manually annotated. A model for automatic estimation of the stomach regions is trained with the GXIs. For the rest of them, the stomach regions are automatically estimated. Finally, a model for automatic CAG detection is trained with all GXIs for training. RESULTS In the case that the stomach regions were manually annotated for only 10 GXIs and 30 GXIs, the harmonic mean of sensitivity and specificity of CAG detection were 0.955 +/- 0.002 and 0.963 +/- 0.004, respectively. CONCLUSION By estimating stomach regions automatically, our method contributes to the reduction of the workload of manual annotation and the accurate detection of the CAG.
• 胃X線画像を用いたAIによるH.pylori感染識別と今後の展望

  藤後 廉, 小川 貴弘, 間部 克裕, 加藤 元嗣, 長谷山 美紀

  日本消化器がん検診学会雑誌 (一社)日本消化器がん検診学会 58 (2) 127 - 127 1880-7666 2020/03
• Multimodal Important Scene Detection in Far-view Soccer Videos Based on Single Deep Neural Architecture

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 8 (2) 89 - 99 2186-7364 2020 

  The details of the matches of soccer can be estimated from visual and audio sequences, and they correspond to the occurrence of important scenes. Therefore, the use of these sequences is suitable for important scene detection. In this paper, a new multimodal method for important scene detection from visual and audio sequences in far-view soccer videos based on a single deep neural architecture is presented. A unique point of our method is that multiple classifiers can be realized by a single deep neural architecture that includes a Convolutional Neural Network-based feature extractor and a Support Vector Machine-based classifier. This approach provides a solution to the problem of not being able to simultaneously optimize different multiple deep neural architectures from a small amount of training data. Then we monitor confidence measures output from this architecture for the multimodal data and enable their integration to obtain the final classification result.
• Important Scene Detection Based on Anomaly Detection using Long Short-Term Memory for Baseball Highlight Generation

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TAIWAN) 1 - 2 2381-5779 2020 

  This paper presents an important scene detection method based on anomaly detection using a Long Short-Term Memory (LSTM) for baseball highlight generation. In order to deal with multi-view time series features calculated from tweets and videos, we adopt an anomaly detection method using LSTM. LSTM which can maintain a long-term memory is effective for training such features. Introduction of LSTM into important scene detection of baseball videos is the biggest contribution of this paper. Experimental results show high detection performance by our method.
• Interpretable Convolutional Neural Network Including Attribute Estimation for Image Classification

  Kazaha Horii, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 8 (2) 111 - 124 2186-7364 2020 

  An interpretable convolutional neural network (CNN) including attribute estimation for image classification is presented in this paper. Although CNNs perform highly accurate image classification, the reason for the classification results obtained by the neural networks is not clear. In order to provide interpretation of CNNs, the proposed method estimates attributes, which explain elements of objects, in an intermediate layer of the network. This enables improvement of the interpretability of CNNs, and it is the main contribution of this paper. Furthermore, the proposed method uses the estimated attributes for image classification in order to enhance its accuracy. Consequently, the proposed method not only provides interpretation of CNNs but also realizes improvement in the performance of image classification.
• Image Retrieval Based on Supervised Local Regression and Global Alignment with Relevance Feedback for Insect Identification

  Keisuke Maeda, Susumu Genma, Takahiro Ogawa, Miki Haseyama

  ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 8 (3) 140 - 150 2186-7364 2020 

  A method for image retrieval based on supervised local regression and global alignment (sLRGA) with relevance feedback for insect identification is presented in this paper. Based on the novel sLRGA, which is an extended version of LRGA, the proposed method estimates ranking scores for image retrieval in such a way that the neighborhood structure of a feature space of the database can be optimally preserved with consideration of class information. This is the main contribution of this paper. By measuring the relevance between all of the images and the query image in the database, sLRGA realizes accurate image retrieval. Furthermore, when positive/negative labels to retrieved images are given by users, the proposed method can improve image retrieval performance considering the query relevance information via use of both relevance feedback and sLRGA. This is the second contribution of this paper. Experimental results show the effectiveness of the proposed method.
• Estimation of Person-Specific Visual Attention via Selection of Similar Persons

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TAIWAN) 1 - 2 2381-5779 2020 

  This paper presents a method for estimation of person-specific visual attention based on estimated similar persons' visual attention. For improving the estimation performance of person-specific visual attention, the proposed method uses the dataset including the large number of images and corresponding gaze data of many persons not including the target person and trains an estimation model based on deep learning. By using the estimated visual attention of similar persons for the target image, the proposed method estimates the visual attention of the target person with the small amount of gaze data. Experimental results show that the proposed method is effective for estimation of person-specific visual attention.
• A Method for Player Importance Prediction from Player Network Using Gaze Position Estimated by LSTM

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 8 (3) 151 - 160 2186-7364 2020 

  A novel method for player importance prediction from a player network using gaze positions estimated by Long Short-Term Memory (LSTM) in soccer videos is presented in this paper. By newly using an estimation model of gaze positions trained by gaze tracking data of experienced persons, it is expected that the importance of each player can be predicted. First, we generate a player network by utilizing the estimated gaze positions and first-arrival regions representing players' connections, e.g., passes between players. The gaze positions are estimated by LSTM that is newly trained from the gaze tracking data of experienced persons. Second, the proposed method predicts the importance of each player by applying the Hypertext Induced Topic Selection (HITS) algorithm to the constructed network. Consequently, prediction of the importance of each player based on soccer tactic knowledge of experienced persons can be realized without constantly obtaining gaze tracking data.
• An Estimation Method of Candidate Region for Superimposing Information Based on Gaze Tracking Data in Soccer Videos

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TAIWAN) 1 - 2 2381-5779 2020 

  A novel method estimating candidate regions for superimposing information in soccer videos based on gaze tracking data is presented in this paper. The proposed method generates a likelihood map based on visual attention regions based on the gaze tracking data and detection results of objects such as players and soccer goals in soccer videos. Candidate regions for superimposing information are estimated by using the likelihood map. Experimental results show that the proposed method realizes effective candidate region estimation.
• Image Retrieval with Data Augmentation of Sentence Labels Based on Paraphrasing

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TAIWAN) 1 - 2 2381-5779 2020 

  Text-based image retrieval is a fundamental study in the field of information retrieval. Recent text-based image retrieval methods employ deep neural networks (here-inafter referred to as deep neural TBIR) to retrieve a desired image from a sentence query and achieve the state-of-the-art performance in TBIR. To improve the retrieval performance of the deep neural TBIR method further, it is essential to prepare diverse sentence labels in training data. However, it takes a lot of effort to prepare diverse sentence labels in training data. To address this problem, we propose a novel deep neural TBIR method with data augmentation of the sentence labels in training data. Experimental results show the effectiveness of the proposed method.
• Interactive re-ranking for cross-modal retrieval based on object-wise question answering.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  MMAsia 2020: ACM Multimedia Asia(MMAsia) 37 - 7 2020
• Similar scene retrieval in soccer videos with weak annotations by multimodal use of bidirectional LSTM.

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  MMAsia 2020: ACM Multimedia Asia(MMAsia) 27 - 8 2020
• Quantitative Analysis of Engineer's Skill Using Wearable Sensor Data while Inspecting Highway Bridge.

  Genki Suzuki, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2nd IEEE Global Conference on Life Sciences and Technologies(LifeTech) 111 - 112 2020
• Estimating Viewed Images with Natural Language Question Answering from fMRI Data.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2nd IEEE Global Conference on Life Sciences and Technologies(LifeTech) 99 - 100 2020
• Distress Level Classification of Road Infrastructures via CNN Generating Attention Map.

  Naoki Ogawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2nd IEEE Global Conference on Life Sciences and Technologies(LifeTech) 97 - 98 2020
• Interest Estimation for Images Using Eye Gaze-based Visual and Text Features via DLPCCA.

  Masanao Matsumoto, Naoki Saito 0006, Takahiro Ogawa, Miki Haseyama

  2nd IEEE Global Conference on Life Sciences and Technologies(LifeTech) 3 - 4 2020
• Mvgan Maximizing Time-Lag Aware Canonical Correlation for Baseball Highlight Generation.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO WORKSHOPS (ICMEW) 1 - 6 2330-7927 2020 

  This paper presents multi-view unsupervised generative adversarial network maximizing time-lag aware canonical correlation (MvGAN) for baseball highlight generation. MvGAN has the following two contributions. First, MvGAN utilizes textual, visual and audio features calculated from tweets and videos as multi-view features. MvGAN which adopts these multi-view features is the effective work for highlight generation of baseball videos. Second, since there is a temporal difference between posted tweets and the corresponding events, MvGAN introduces a novel feature embedding scheme considering a time-lag between textual features and other features. Specifically, the proposed method newly derives the timelag aware canonical correlation maximization of these multi-view features. This is the biggest contribution of this paper. Furthermore, since MvGAN is an unsupervised method for highlight generation, a large amount of training data with annotation is not needed. Thus, the proposed method has high applicability to the real world.
• Generation of Viewed Image Captions From Human Brain Activity Via Unsupervised Text Latent Space.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2521 - 2525 1522-4880 2020 

  Generation of human cognitive contents based on the analysis of functional magnetic resonance imaging (fMRI) data has been actively researched. Cognitive contents such as viewed images can be estimated by analyzing the relationship between fMRI data and semantic information of viewed images. In this paper, we propose a new method generating captions for viewed images from human brain activity via a novel robust regression scheme. Unlike conventional generation methods using image feature representations, the proposed method makes use of more semantic text feature representations, which are more suitable for the caption generation. We construct a text latent space with unlabeled images not used for the training, and the fMRI data are regressed to the text latent space. Besides, we newly make use of unlabeled images not used for the training phase to improve caption generation performance. Finally, the proposed method can generate captions from the fMRI data measured while subjects are viewing images. Experimental results show that the proposed method enables accurate caption generation for viewed images.
• Multimodal Image-to-Image Translation for Generation of Gastritis Images.

  Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2466 - 2470 1522-4880 2020 

  We present a new multimodal image-to-image translation model for the generation of gastritis images using X-ray and blood inspection results. In clinical situations, clinicians estimate the prognosis of the target disease by considering multiple inspection results. Similarly, we take a multimodal approach in the task of gastric cancer risk prediction. Visual characteristics of the gastric X-ray image and blood index values are highly related in the evaluation of gastric cancer risk. If we can generate a prediction image from blood index values, it contributes to the clinicians' sophisticated and integrated diagnosis. Hence, we learn a model that can map non-gastritis images to gastritis images based on the blood index values. Although this is a challenging multimodal task in medical image analysis, experimental results showed the effectiveness of our model.
• Image Retrieval With Lingual And Visual Paraphrasing Via Generative Models.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2431 - 2435 1522-4880 2020 

  A new approach that improves text-based image retrieval (hereinafter referred to as TBIR) performance is proposed in this paper. TBIR methods aim to retrieve a desired image related to a query text. Especially, recent TBIR methods allow us to retrieve images considering word relationships by using a sentence as a query. In these TBIR methods, it is necessary to uniquely identify a desired image from similar images using a single query sentence. However, the diverse expressive styles for a query sentence make it difficult to uniquely identify a desired image. In this paper, we propose a novel TBIR method with paraphrasing on multiple representation spaces. Specifically, by paraphrasing a query sentence on lingual and visual representation spaces, the proposed method can retrieve a desired image from various perspectives and then it can uniquely identify a desired image from similar images. Comprehensive experimental results show the effectiveness of the proposed method.
• Variational Autoencoder Based Unsupervised Domain Adaptation For Semantic Segmentation.

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2426 - 2430 1522-4880 2020 

  Unsupervised domain adaptation, which transfers supervised knowledge from a labeled domain to an unlabeled domain, remains a tough problem in the field of computer vision, especially for semantic segmentation. Some methods inspired by adversarial learning and semi-supervised learning have been developed for unsupervised domain adaptation in semantic segmentation and achieved outstanding performances. In this paper, we propose a novel method for this task. Like adversarial learning-based methods using a discriminator to align the feature distributions from different domains, we employ a variational autoencoder to get to the same destination but in a non-adversarial manner. Since the two approaches are compatible, we also integrate an adversarial loss into our method. By further introducing pseudo labels, our method can achieve state-of-the-art performances on two benchmark adaptation scenarios, GTA5-to-CITYSCAPES and SYNTHIA-to-CITYSCAPES.
• Important Scene Detection Of Baseball Videos Via Time-Lag Aware Deep Multiset Canonical Correlation Maximization.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 1236 - 1240 1522-4880 2020 

  This paper presents a new important scene detection method of baseball videos based on correlation maximization between heterogeneous modalities via time-lag aware deep multiset canonical correlation analysis (Tl-dMCCA). The technical contributions of this paper are twofold. First, textual, visual and audio features calculated from tweets and videos are adopted as multi-view time series features. Since Tl-dMCCA which utilizes these features includes the unsupervised embedding scheme via deep networks, the proposed method can flexibly express the relationship between heterogeneous features. Second, since there is the time-lag between posted tweets and the corresponding multiple previous events, Tl-dMCCA considers the time-lag relationships between them. Specifically, we newly introduce the representation of such time-lags into the derivation of their covariance matrices. By considering time-lags via Tl-dMCCA, the proposed method correctly detects important scenes.
• Soft-Label Anonymous Gastric X-Ray Image Distillation.

  Guang Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 305 - 309 1522-4880 2020 

  This paper presents a soft-label anonymous gastric X-ray image distillation method based on a gradient descent approach. The sharing of medical data is demanded to construct high-accuracy computer-aided diagnosis (CAD) systems. However, the large size of the medical dataset and privacy protection are remaining problems in medical data sharing, which hindered the research of CAD systems. The idea of our distillation method is to extract the valid information of the medical dataset and generate a tiny distilled dataset that has a different data distribution. Different from model distillation, our method aims to find the optimal distilled images, distilled labels and the optimized learning rate. Experimental results show that the proposed method can not only effectively compress the medical dataset but also anonymize medical images to protect the patient's private information. The proposed approach can improve the efficiency and security of medical data sharing.
• Estimation Of Visual Contents Based On Question Answering From Human Brain Activity.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 61 - 65 1522-4880 2020 

  We propose an estimation method for free-form Visual Question Answering (VQA) from human brain activity, brain decoding VQA. The task of VQA in the field of computer vision is generating an answer given an image and a question about its contents. The proposed method can realize answering arbitrary visual questions about images from brain activity measured by functional Magnetic Resonance Imaging (fMRI) while viewing the same images. We enable estimating various information from brain activity via a unique VQA model, which can realize a more detailed understanding of images and complex reasoning. In addition, we newly make use of un-labeled images not used in the training phase to improve the performance of the transformation, since fMRI datasets are generally small. The proposed method can answer a visual question from a little amount of fMRI data measured while subjects are viewing images.
• Feature Integration Via Geometrical Supervised Multi-View Multi-Label Canonical Correlation For Incomplete Label Assignment.

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 46 - 50 1522-4880 2020 

  This paper presents feature integration via geometrical supervised multi-view multi-label canonical correlation analysis (GSM2CCA) for incomplete label assignment. The problem of incomplete labels is frequently encountered in the multi-label classification problem where the training labels are obtained via crowd-sourcing. In such a situation, consideration of only the label correlation, which is the basic approach, is not suitable for improvement of representation ability of features. For dealing with the incomplete label assignment, GSM2CCA constructs effective feature embedding space providing the discriminant ability by introducing both the multi-label correlation and feature similarity of the original feature space into its objective function. Since novel integrated features with high discriminant ability can be calculated by our GSM2CCA, performance improvement of multi-label classification with the incomplete label assignment is realized. The main contribution of this paper is the realization of the effective feature integration via the adoption of the combination use of label similarity and locality preserving projection of heterogeneous features for solving the problem of the incomplete label assignment. The effectiveness of GSM2CCA by applying GSM2CCA-based feature integration to heterogeneous features calculated from various convolutional neural network models is verified via experimental results.
• Unsupervised Domain Adaptation for Semantic Segmentation with Symmetric Adaptation Consistency.

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING 2263 - 2267 1520-6149 2020 

  Unsupervised domain adaptation, which leverages label information from other domains to solve tasks on a domain without any labels, can alleviate the problem of the scarcity of labels and expensive labeling costs faced by supervised semantic segmentation. In this paper, we utilize adversarial learning and semi-supervised learning simultaneously to solve the task of unsupervised domain adaptation in semantic segmentation. We propose a new approach that trains two segmentation models with the adversarial learning symmetrically and further introduces the consistency between the outputs of the two models into the semi-supervised learning to improve the accuracy of pseudo labels which significantly affect the final adaptation performance. We achieve state-of-the-art semantic segmentation performance on the GTA5-to-Cityscapes scenario, a widely used benchmark setting in unsupervised domain adaptation.
• Multi-View Bayesian Generative Model for Multi-Subject FMRI Data on Brain Decoding of Viewed Image Categories.

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING 1215 - 1219 1520-6149 2020 

  Brain decoding studies have demonstrated that viewed image categories can be estimated from human functional magnetic resonance imaging (fMRI) activity. However, there are still limitations with the estimation performance because of the characteristics of fMRI data and the employment of only one modality extracted from viewed images. In this paper, we propose a multi-view Bayesian generative model for multi-subject fMRI data to estimate viewed image categories from fMRI activity. The proposed method derives effective representations of fMRI activity by utilizing multi-subject fMRI data. In addition, we associate fMRI activity with multiple modalities, i:e:, visual features and semantic features extracted from viewed images. Experimental results show that the proposed method outperforms existing state-of-the-art methods of brain decoding.
• Interest Level Estimation Based on Feature Integration Considering Distribution of Partially Paired User's Behavior, Videos and Posters.

  Kyohei Kamikawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 944 - 945 2020
• Music Playlist Generation Based on Reinforcement Learning Using Acoustic Feature Map.

  Keigo Sakurai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 942 - 943 2020
• Estimation of Images Matched with Audio-Induced Brain Activity via Modified DGCCA.

  Yun Liang 0014, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 940 - 941 2020
• Estimation of User-Specific Visual Attention Considering Individual Tendency toward Gazed Objects.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 745 - 746 2020
• Estimation of Viewed Images Using Individual and Shared Brain Responses.

  Takaaki Higashi, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 716 - 717 2020
• Cross-domain Recommendation via Multi-layer Graph Analysis Using User-item Embedding.

  Taisei Hirakawa, Keisuke Maeda, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 714 - 715 2020
• Question Answering for Estimation of Seen Image Contents from Multi-subject fMRI Responses.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 712 - 713 2020
• Face Synthesis via User Manipulation of Disentangled Latent Representation.

  Nao Nakagawa, Ren Togo, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 692 - 693 2020
• Complexity Evaluation of Medical Image Data for Classification Problem Based on Spectral Clustering.

  Guang Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 667 - 669 2020
• Important Scene Prediction of Baseball Videos Using Twitter and Video Analysis Based on LSTM.

  Kaito Hirasawa, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  9th IEEE Global Conference on Consumer Electronics(GCCE) 636 - 637 2020
• Brain Decoding of Viewed Image Categories via Semi-Supervised Multi-View Bayesian Generative Model.

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE Trans. Signal Process. 68 5769 - 5781 1053-587X 2020 

  Brain decoding has shown that viewed image categories can be estimated from evoked functional magnetic resonance imaging (fMRI) activity. Recent studies attempted to estimate viewed image categories that were not used for training previously. Nevertheless, the estimation performance is limited since it is difficult to collect a large amount of fMRI data for training. This paper presents a method to accurately estimate viewed image categories not used for training via a semi-supervised multi-view Bayesian generative model. Our model focuses on the relationship between fMRI activity and multiple modalities, i.e., visual features extracted from viewed images and semantic features obtained from viewed image categories. Furthermore, in order to accurately estimate image categories not used for training, our semi-supervised framework incorporates visual and semantic features obtained from additional image categories in addition to image categories of training data. The estimation performance of the proposed model outperforms existing state-of-the-art models in the brain decoding field and achieves more than 95% identification accuracy. The results also have shown that the incorporation of additional image category information is remarkably effective when the number of training samples is small. Our semi-supervised framework is significant for the brain decoding field where brain activity patterns are insufficient but visual stimuli are sufficient.
• Few-Shot Personalized Saliency Prediction Based on Adaptive Image Selection Considering Object and Visual Attention.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Sensors 20 (8) 2170 - 2170 2020 

  A few-shot personalized saliency prediction based on adaptive image selection considering object and visual attention is presented in this paper. Since general methods predicting personalized saliency maps (PSMs) need a large number of training images, the establishment of a theory using a small number of training images is needed. To tackle this problem, although finding persons who have visual attention similar to that of a target person is effective, all persons have to commonly gaze at many images. Thus, it becomes difficult and unrealistic when considering their burden. On the other hand, this paper introduces a novel adaptive image selection (AIS) scheme that focuses on the relationship between human visual attention and objects in images. AIS focuses on both a diversity of objects in images and a variance of PSMs for the objects. Specifically, AIS selects images so that selected images have various kinds of objects to maintain their diversity. Moreover, AIS guarantees the high variance of PSMs for persons since it represents the regions that many persons commonly gaze at or do not gaze at. The proposed method enables selecting similar users from a small number of images by selecting images that have high diversities and variances. This is the technical contribution of this paper. Experimental results show the effectiveness of our personalized saliency prediction including the new image selection scheme.
• Tensor-Based Emotional Category Classification via Visual Attention-Based Heterogeneous CNN Feature Fusion.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Sensors 20 (7) 2146 - 2146 2020 

  The paper proposes a method of visual attention-based emotion classification through eye gaze analysis. Concretely, tensor-based emotional category classification via visual attention-based heterogeneous convolutional neural network (CNN) feature fusion is proposed. Based on the relationship between human emotions and changes in visual attention with time, the proposed method performs new gaze-based image representation that is suitable for reflecting the characteristics of the changes in visual attention with time. Furthermore, since emotions evoked in humans are closely related to objects in images, our method uses a CNN model to obtain CNN features that can represent their characteristics. For improving the representation ability to the emotional categories, we extract multiple CNN features from our novel gaze-based image representation and enable their fusion by constructing a novel tensor consisting of these CNN features. Thus, this tensor construction realizes the visual attention-based heterogeneous CNN feature fusion. This is the main contribution of this paper. Finally, by applying logistic tensor regression with general tensor discriminant analysis to the newly constructed tensor, the emotional category classification becomes feasible. Since experimental results show that the proposed method enables the emotional category classification with the F1-measure of approximately 0.6, and about 10% improvement can be realized compared to comparative methods including state-of-the-art methods, the effectiveness of the proposed method is verified.
• Chronic gastritis classification using gastric X-ray images with a semi-supervised learning method based on tri-training.

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  Medical Biol. Eng. Comput. 58 (6) 1239 - 1250 0140-0118 2020 

  High-quality annotations for medical images are always costly and scarce. Many applications of deep learning in the field of medical image analysis face the problem of insufficient annotated data. In this paper, we present a semi-supervised learning method for chronic gastritis classification using gastric X-ray images. The proposed semi-supervised learning method based on tri-training can leverage unannotated data to boost the performance that is achieved with a small amount of annotated data. We utilize a novel learning method named Between-Class learning (BC learning) that can considerably enhance the performance of our semi-supervised learning method. As a result, our method can effectively learn from unannotated data and achieve high diagnostic accuracy for chronic gastritis.
• Multi-Task Convolutional Neural Network Leading to High Performance and Interpretability via Attribute Estimation.

  Keisuke Maeda, Kazaha Horii, Takahiro Ogawa, Miki Haseyama

  IEICE Trans. Fundam. Electron. Commun. Comput. Sci. 103-A (12) 1609 - 1612 0916-8508 2020 

  A multi-task convolutional neural network leading to high performance and interpretability via attribute estimation is presented in this letter. Our method can provide interpretation of the classification results of CNNs by outputting attributes that explain elements of objects as a judgement reason of CNNs in the middle layer. Furthermore, the proposed network uses the estimated attributes for the following prediction of classes. Consequently, construction of a novel multi-task CNN with improvements in both of the interpretability and classification performance is realized.
• Inpainting via Sparse Representation Based on a Phaseless Quality Metric.

  Takahiro Ogawa, Keisuke Maeda, Miki Haseyama

  IEICE Trans. Fundam. Electron. Commun. Comput. Sci. 103-A (12) 1541 - 1551 0916-8508 2020 

  An inpainting method via sparse representation based on a new phaseless quality metric is presented in this paper. Since power spectra, phaseless features, of local regions within images enable more successful representation of their texture characteristics compared to their pixel values, a new quality metric based on these phaseless features is newly derived for image representation. Specifically, the proposed method enables spare representation of target signals, i.e., target patches, including missing intensities by monitoring errors converged by phase retrieval as the novel phaseless quality metric. This is the main contribution of our study. In this approach, the phase retrieval algorithm used in our method has the following two important roles: (1) derivation of the new quality metric that can be derived even for images including missing intensities and (2) conversion of phaseless features, i.e., power spectra, to pixel values, i.e., intensities. Therefore, the above novel approach solves the existing problem of not being able to use better features or better quality metrics for inpainting. Results of experiments showed that the proposed method using sparse representation based on the new phaseless quality metric outperforms previously reported methods that directly use pixel values for inpainting.
• Heterogeneous-Graph-Based Video Search Reranking Using Topic Relevance.

  Soh Yoshida, Mitsuji Muneyasu, Takahiro Ogawa, Miki Haseyama

  IEICE Trans. Fundam. Electron. Commun. Comput. Sci. 103-A (12) 1529 - 1540 0916-8508 2020 

  In this paper, we address the problem of analyzing topics, included in a social video group, to improve the retrieval performance of videos. Unlike previous methods that focused on an individual visual aspect of videos, the proposed method aims to leverage the "mutual reinforcement" of heterogeneous modalities such as tags and users associated with video on the Internet. To represent multiple types of relationships between each heterogeneous modality, the proposed method constructs three subgraphs: user-tag, video-video, and video-tag graphs. We combine the three types of graphs to obtain a heterogeneous graph. Then the extraction of latent features, i.e., topics, becomes feasible by applying graph-based soft clustering to the heterogeneous graph. By estimating the membership of each grouped cluster for each video, the proposed method defines a new video similarity measure. Since the understanding of video content is enhanced by exploiting latent features obtained from different types of data that complement each other, the performance of visual reranking is improved by the proposed method. Results of experiments on a video dataset that consists of YouTube-8M videos show the effectiveness of the proposed method, which achieves a 24.3% improvement in terms of the mean normalized discounted cumulative gain in a search ranking task compared with the baseline method.
• Deep convolutional neural network-based anomaly detection for organ classification in gastric X-ray examination.

  Ren Togo, Haruna Watanabe, Takahiro Ogawa, Miki Haseyama

  Comput. Biol. Medicine 123 103903 - 103903 0010-4825 2020 

  Aim: The aim of this study was to determine whether our deep convolutional neural network-based anomaly detection model can distinguish differences in esophagus images and stomach images obtained from gastric X-ray examinations.Methods: A total of 6012 subjects were analyzed as our study subjects. Since the number of esophagus X-ray images is much smaller than the number of gastric X-ray images taken in X-ray examinations, we took an anomaly detection approach to realize the task of organ classification. We constructed a deep autoencoding gaussian mixture model (DAGMM) with a convolutional autoencoder architecture. The trained model can produce an anomaly score for a given test X-ray image. For comparison, the original DAGMM, AnoGAN, and a One-Class Support Vector Machine (OCSVM) that were trained with features obtained by a pre-trained Inception-v3 network were used.Results: Sensitivity, specificity, and the calculated harmonic mean of the proposed method were 0.956, 0.980, and 0.968, respectively. Those of the original DAGMM were 0.932, 0.883, and 0.907, respectively. Those of AnoGAN were 0.835, 0.833, and 0.834, respectively, and those of OCSVM were 0.932, 0.935, and 0.934, respectively. Experimental results showed the effectiveness of the proposed method for an organ classification task.Conclusion: Our deep convolutional neural network-based anomaly detection model has shown the potential for clinical use in organ classification.
• Human-Centric Emotion Estimation Based on Correlation Maximization Considering Changes With Time in Visual Attention and Brain Activity.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE Access 8 203358 - 203368 2169-3536 2020 

  A human-centric emotion estimation method based on correlation maximization with consideration of changes with time in visual attention and brain activity when viewing images is proposed in this paper. Owing to the recent developments of many kinds of biological sensors, many researchers have focused on multimodal emotion estimation using both eye gaze data and brain activity data for improving the quality of emotion estimation. In this paper, a novel method that focuses on the following two points is introduced. First, in order to reduce the burden on users, we obtain brain activity data from users only in the training phase by using a projection matrix calculated by canonical correlation analysis (CCA) between gaze-based visual features and brain activity-based features. Next, for considering the changes with time in both visual attention and brain activity, we obtain novel features based on CCA-based projection in each time unit. In order to include these two points, the proposed method analyzes a fourth-order gaze and image tensor for which modes are pixel location, color channel and the changes with time in visual attention. Moreover, in each time unit, the proposed method performs CCA between gaze-based visual features and brain activity-based features to realize human-centric emotion estimation with a high level of accuracy. Experimental results show that accurate human emotion estimation is achieved by using our new human-centric image representation.
• Estimation of Interest Levels From Behavior Features via Tensor Completion Including Adaptive Similar User Selection.

  Keisuke Maeda, Tetsuya Kushima, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEEE Access 8 126109 - 126118 2169-3536 2020 

  A method for estimating interest levels from behavior features via tensor completion including adaptive similar user selection is presented in this paper. The proposed method focuses on a tensor that is suitable for data containing multiple contexts and constructs a third-order tensor in which three modes are "products", "users" and "user behaviors and interest levels" for these products. By complementing this tensor, unknown interest level estimation of a product for a target user becomes feasible. For further improving the estimation performance, the proposed method adaptively selects similar users for the target user by focusing on converged estimation errors between estimated interest levels and known interest levels in the tensor completion. Furthermore, the proposed method can adaptively estimate the unknown interest from the similar users. This is the main contribution of this paper. Therefore, the influence of users having different interests is reduced, and accurate interest level estimation can be realized. In order to verify the effectiveness of the proposed method, we show experimental results obtained by estimating interest levels of users holding books.
• Supervised Fractional-Order Embedding Geometrical Multi-View CCA (SFGMCCA) for Multiple Feature Integration.

  Keisuke Maeda, Yoshiki Ito, Takahiro Ogawa, Miki Haseyama

  IEEE Access 8 114340 - 114353 2169-3536 2020 

  Techniques for integrating different types of multiple features effectively have been actively studied in recent years. Multiset canonical correlation analysis (MCCA), which maximizes the sum of pairwise correlations of inter-view (i.e., between different features), is one of the powerful methods for integrating different types of multiple features, and various MCCA-based methods have been proposed. This work focuses on a supervised MCCA variant in order to construct a novel effective feature integration framework. In this paper, we newly propose supervised fractional-order embedding geometrical multi-view CCA (SFGMCCA). This method constructs not only the correlation structure but also two types of geometrical structures of intra-view (i.e., within each feature) and inter-view simultaneously, thereby realizing more precise feature integration. This method also supports the integration of small sample and high-dimensional data by using the fractional-order technique. We conducted experiments using four types of image datasets, i.e., MNIST, COIL-20, ETH-80 and CIFAR-10. Furthermore, we also performed an fMRI dataset containing brain signals to verify the robustness. As a result, it was confirmed that accuracy improvements using SFGMCCA were statistically significant at the significance level of 0.05 compared to those using conventional representative MCCA-based methods.
• Enhancing Cross-Modal Retrieval Based on Modality-Specific and Embedding Spaces.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Access 8 96777 - 96786 2169-3536 2020 

  A new approach that drastically improves cross-modal retrieval performance in vision and language (hereinafter referred to as & x201C;vision and language retrieval & x201D;) is proposed in this paper. Vision and language retrieval takes data of one modality as a query to retrieve relevant data of another modality, and it enables flexible retrieval across different modalities. Most of the existing methods learn optimal embeddings of visual and lingual information to a single common representation space. However, we argue that the forced embedding optimization results in loss of key information for sentences and images. In this paper, we propose an effective utilization of representation spaces in a simple but robust vision and language retrieval method. The proposed method makes use of multiple individual representation spaces through text-to-image and image-to-text models. Experimental results showed that the proposed approach enhances the performance of existing methods that embed visual and lingual information to a single common representation space.
• Context-Aware Network Analysis of Music Streaming Services for Popularity Estimation of Artists.

  Yui Matsumoto, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE Access 8 48673 - 48685 2169-3536 2020 [Refereed][Not invited]
   

  A novel trial for estimating popularity of artists in music streaming services (MSS) is presented in this paper. The main contribution of this paper is to improve extensibility for using multi-modal features to accurately analyze latent relationships between artists. In the proposed method, a novel framework to construct a network is derived by collaboratively using social metadata and multi-modal features via canonical correlation analysis. Different from conventional methods that do not use multi-modal features, the proposed method can construct a network that can capture social metadata and multi-modal features, i.e., a context-aware network. For effectively analyzing the context-aware network, a novel framework to realize popularity estimation of artists is developed based on network analysis. The proposed method enables effective utilization of the network structure by extracting node features via a node embedding algorithm. By constructing an estimator that can distinguish differences between the node features, the proposed method can archive accurate popularity estimation of artists. Experimental results using multiple real-world datasets that contain artists in various genres in Spotify, one of the largest MSS, are presented. Quantitative and qualitative evaluations show that our method is effective for both classifying and regressing the popularity.
• Retrieval of similar scenes based on multimodal distance metric learning in soccer videos

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  MMSports 2019 - Proceedings of the 2nd International Workshop on Multimedia Content Analysis in Sports, co-located with MM 2019 10 - 15 2019/10/15 [Refereed][Not invited]
   

  © 2019 Association for Computing Machinery. This paper presents a new method for retrieval of similar scenes based on multimodal distance metric learning in far-view soccer videos that broadly capture soccer fields and are not edited. We extract visual features and audio features from soccer video clips, and we extract text features from text data corresponding to these soccer video clips. In addition, distance metric learning based on Laplacian Regularized Metric Learning is performed to calculate the distances for each kind of features. Finally, by determining the final rank by integrating these distances, we realize successful multimodal retrieval of similar scenes from query scenes of soccer video clips. Experimental results show the effectiveness of our retrieval method.
• Semantic Shot Classification in Baseball Videos Based on Similarities of Visual Features

  K. Hirasawa, K. Maeda, T. Ogawa, M. Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 663 - 664 2019/10 [Refereed][Not invited]
  
• Region-based Distress Classification of Road Infrastructures via CNN Without Region Annotation

  N. Ogawa, K. Maeda, T. Ogawa, M. Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 764 - 765 2019/10 [Refereed][Not invited]
  
• Aesthetic style transfer through text-to-image synthesis and image-to-image translation

  Megumi Kotera, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 492 - 493 2019/10 [Refereed][Not invited]
  
• Voice-input multimedia information retrieval system based on text-to-image GAN

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 943 - 944 2019/10 [Refereed][Not invited]
  
• Estimation of drilling energy from tunnel cutting face image based on online learning

  Kentaro Yamamoto, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 77 (1) 794 - 795 2185-6591 2019/10 [Refereed][Not invited]
  
• Detection of distress region from subway tunnel images via U-net-based deep semantic segmentation

  An Wang, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 766 - 767 2019/10 [Refereed][Not invited]
  
• Estimation of Emotion Labels via Tensor-Based Spatiotemporal Visual Attention Analysis

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  Proceedings - International Conference on Image Processing, ICIP 2019-September 4105 - 4109 1522-4880 2019/09 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents emotion label estimation via tensor-based spatiotemporal visual attention analysis. It has been reported in the fields of psychology and neuroscience that human emotions are related to two elements, their visual attention change and objects included in a target image. Therefore, the proposed method focuses on the spatiotemporal change of visual attention of human gazing at objects in the target image and constructs two neural networks which enable the emotion label estimation considering both of the above two elements. Specifically, the proposed method newly constructs a fourth-order tensor, gaze and image tensor (GIT) whose modes correspond to the width, the height and the color channel of the target image and the time axis of visual attention which is used for representing the time change. Then the first network, which consists of general tensor discriminant analysis (GTDA) and extreme learning machine (ELM), estimates the emotion label from the fourth-order GIT with concerning their visual attention change. Furthermore, the second network, which consists of pre-trained convolutoinal neural network-based feature extraction, GTDA and ELM, enables the estimation from the second-order GIT including visual features obtained from objects focused at each time. Finally, the proposed method estimates emotion labels based on decision fusion of the outputs from the two networks. Experimental results show the effectiveness of the proposed method.
• Scene Retrieval for Video Summarization Based on Text-to-Image gan

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  Proceedings - International Conference on Image Processing, ICIP 2019-September 1825 - 1829 1522-4880 2019/09 [Refereed][Not invited]
   

  © 2019 IEEE. We present a new scene retrieval method based on text-to-image Generative Adversarial Network (GAN) and its application to query-based video summarization. Text-to-image GAN is a deep learning method that can generate images from their corresponding sentences. In this paper, we reveal a characteristic that deep learning-based visual features extracted from images generated by text-to-image GAN include semantic information sufficiently. By utilizing the generated images as queries, the proposed method achieves higher scene retrieval performance than those of the stateof-the-art methods. In addition, we introduce a novel architecture that can consider order relationship of the input sentences to our method for realizing a target video summarization. Specifically, the proposed method generates multiple images thorough text-to-image GAN from multiple sentences summarizing target videos. Their summarized video can be obtained by performing the retrieval of corresponding scenes from the target videos according to the generated images with considering the order relationship. Experimental results show the effectiveness of the proposed method in the retrieval and summarization performance.
• Gastritis Detection from Gastric X-Ray Images Via Fine-Tuning of Patch-Based Deep Convolutional Neural Network

  Misaki Kanai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  Proceedings - International Conference on Image Processing, ICIP 2019-September 1371 - 1375 1522-4880 2019/09 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method for gastritis detection from gastric X-ray images via fine-tuning techniques using a deep convolutional neural network (DCNN). DCNNs can learn parameters to capture high-dimensional features which express semantic contents of images by training on a large number of labeled images. However, lack of gastric X-ray images for training often occurs. To realize accurate detection with a small number of gastric X-ray images, the proposed method adopts fine-tuning techniques and newly introduces simple annotation of stomach regions to gastric X-ray images used for training. The proposed method fine-tunes a pre-trained DCNN with patches and three kinds of patch-level class labels considering not only the image-level ground truth ('gastritis'/'non-gastritis') but also the regions of a stomach since the outside of the stomach is not related to the image-level ground truth. In the test phase, by estimating the patch-level class labels with the fine-tuned DCNN, the proposed method enables the image-level class label estimation which excludes the effect of the unnecessary regions. Experimental results show the effectiveness of the proposed method.
• Neural Network Maximizing Ordinally Supervised Multi-View Canonical Correlation for Deterioration Level Estimation

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  Proceedings - International Conference on Image Processing, ICIP 2019-September 919 - 923 1522-4880 2019/09 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a neural network maximizing ordinally supervised multi-view canonical correlation for deterioration level estimation. The contributions of this paper are twofold. First, in order to calculate features representing deterioration levels on transmission towers, which is one of the infrastructures, a novel neural network handling multi-modal features is constructed from a small amount of training data. Specifically, in our method, effective transformation to features with high discriminant ability without using many hidden layers is realized by setting projection matrices maximizing correlation between multiple features into hidden layer's weights. Second, since there exists ordinal scale in deterioration levels, the proposed method newly derives ordinally supervised multi-view canonical correlation analysis (OsMVCCA). OsMVCCA enables estimation of the effective projection considering not only label information but also their ordinal scales. Experimental results show that the proposed method realizes accurate deterioration level estimation.
• Convolutional sparse coding-based deep random vector functional link network for distress classification of road structures

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  Computer-Aided Civil and Infrastructure Engineering 34 (8) 654 - 676 1093-9687 2019/08 [Refereed][Not invited]
   

  © 2019 Computer-Aided Civil and Infrastructure Engineering This paper presents a convolutional sparse coding (CSC)-based deep random vector functional link network (CSDRN) for distress classification of road structures. The main contribution of this paper is the introduction of CSC into a feature extraction scheme in the distress classification. CSC can extract visual features representing characteristics of target images because it can successfully estimate optimal convolutional dictionary filters and sparse features as visual features by training from a small number of distress images. The optimal dictionaries trained from distress images have basic components of visual characteristics such as edge and line information of distress images. Furthermore, sparse feature maps estimated on the basis of the dictionaries represent both strength of the basic components and location information of regions having their components, and these maps can represent distress images. That is, sparse feature maps can extract key components from distress images that have diverse visual characteristics. Therefore, CSC-based feature extraction is effective for training from a limited number of distress images that have diverse visual characteristics. The construction of a novel neural network, CSDRN, by the use of a combination of CSC-based feature extraction and the DRN classifier, which can also be trained from a small dataset, is shown in this paper. Accurate distress classification is realized via the CSDRN.
• Bilingual Lexicon Learning Using Tagged Images via Graph Trilateral Filter-based Feature Refinement

  Yui Matsumoto, Shota Hamano, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2019 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2019 2019/05 

  A novel method to realize bilingual lexicon learning (BLL) using tagged images is presented in this paper. Different from existing methods that require parallel corpora, the proposed method enables extraction of semantically similar words by utilizing not such corpora but tagged images on image sharing services. The main contribution of this paper is derivation of a novel framework to refine visual features of tagged images based on graph trilateral filter-based smoothing. This enables reduction of the influence of noisy tags that are irrelevant to contents of images. As a result, accurate BLL becomes feasible by nearest neighbor search using the refined visual features.
• Convolutional Sparse Coding-based Anomalous Event Detection in Surveillance Videos

  Masanao Matsumoto, Naoki Saito, Takahiro Ogawa, Miki Haseyama

  2019 IEEE International Conference on Consumer Electronics - Taiwan (ICCE-TW) 2019/05
• User-Specific Visual Attention Estimation Based on Visual Similarity and Spatial Information in Images

  Y. Moroto, K. Maeda, T. Ogawa, M. Haseyama

  IEEE International Conference on Consumer Electronics – Taiwan (ICCE-TW) 479 - 480 2019/05 [Refereed][Not invited]
  
• Estimating Viewed Image Categories from Human Brain Activity via Semi-supervised Fuzzy Discriminative Canonical Correlation Analysis

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings IEEE 2019-May 1105 - 1109 1520-6149 2019/05 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method to estimate viewed image categories from human brain activity via newly derived semi-supervised fuzzy discriminative canonical correlation analysis (Semi-FDCCA). The proposed method can estimate image categories from functional magnetic resonance imaging (fMRI) activity measured while subjects view images by making fMRI activity and visual features obtained from images comparable through Semi-FDCCA. To realize Semi-FDCCA, we first derive a new supervised CCA called FDCCA that can consider fuzzy class information based on image category similarities obtained from WordNet ontology. Second, we adopt SemiCCA that can utilize additional unpaired visual features in addition to pairs of fMRI activity and visual features in order to prevent overfitting to the limited pairs. Furthermore, Semi-FDCCA can be derived by combining FDCCA with SemiCCA. Experimental results show that Semi-FDCCA enables accurate estimation of viewed image categories.
• Extraction of regions related to cardiac sarcoidosis in polar map images

  Ren Togo, Takahiro Ogawa, Osamu Manabe, Kenji Hirata, Tohru Shiga, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 237 - 238 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method for extracting important regions for deep learning models in the identification of cardiac sarcoidosis using polar map images. Although deep learning-based detection methods have widely studied, they are still often called black boxes. Since high reliability for provided results from computer-aided diagnosis systems is important toward clinical applications, this problem should be solved. In this paper, we try to visualize important regions for deep learning-based models for improvement of understanding to clinicians. We monitor the variance of confidence of a model constructed with a deep learning-based feature and define it as a contribution value toward the estimated label. We visualize important regions for models based on the contribution value.
• Estimation of emotions evoked by images based on multiple gaze-based CNN features

  Taiga Matsui, Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 194 - 195 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method for estimating emotions evoked by watching images based on multiple visual features considering relationship with gaze information. The proposed method obtains multiple visual features from multiple middle layers of a Convolutional Neural Network. Then the proposed method newly derives their gaze-based visual features maximizing correlation with gaze information by using Discriminative Locality Preserving Canonical Correlation Analysis. The final estimation result is calculated by integrating multiple estimation results obtained from these gaze-based visual features. Consequently, successful emotion estimation becomes feasible by using such multiple estimation results which correspond to different semantic levels of target images.
• Estimation of users' interest levels using tensor completion with SemiCCA

  Tetsuya Kushima, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 239 - 240 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a new method for estimation of users' interest levels using tensor completion with SemiCCA. The proposed method extracts new features maximizing correlation between features calculated from partially paired users' behavior and contents with semi-supervised canonical correlation analysis (SemiCCA). By this approach, we can successfully use the contents that users have not viewed for the interest level estimation. Moreover, our method utilizes the tensor completion to estimate unknown interest levels. Consequently, in the proposed method, accurate estimation of interest levels using SemiCCA and the tensor completion is realized. Experimental results are shown to verify the effectiveness of the proposed method by using actual data.
• Classification of subcellular protein patterns in human cells with transfer learning

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 273 - 274 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. In this paper, we present a deep learning method for classifying subcellular protein patterns in human cells. Our method is mainly based on transfer learning and utilizes a newly proposed loss function named focal loss to deal with the problem of severe class imbalance existing in the task. The performance of our method is evaluated by a MacroF1 score of total 28 classes, and the final MacroF1 score of our method is 0.706.
• Estimation of visual attention via canonical correlation between visual and gaze-based features

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 229 - 230 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method for estimating visual attention via canonical correlation between visual and gaze-based features. The proposed method estimates user-specific visual attention by comparing a test image with training images including their corresponding individual eye gaze data in a common space. Specifically, canonical correlation analysis can derive projections which enable comparison between visual and gaze-based features in the common space. Therefore, given the new test image, our method projects its visual features to the common space and can estimate visual attention. Experimental results show the effectiveness of the proposed method.
• Chronic gastritis detection from gastric X-ray images via deep autoencoding Gaussian mixture models

  Masanao Matsumoto, Naoki Saito, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 231 - 232 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a detection method of chronic gastritis from gastric X-ray images. The conventional method cannot detect chronic gastritis accurately since the number of non-gastritis images is overwhelmingly larger than the number of gastritis images. To deal with this problem, the proposed method performs the detection of chronic gastritis by using Deep Autoencoding Gaussian Mixture Models (DAGMM) which is an anomaly detection approach. DAGMM enables construction of chronic gastritis detection model using only non-gastritis images. In addition, DAGMM is superior to conventional anomaly detection methods since the models of dimensionality reduction and density estimation can be learned simultaneously. Therefore, the proposed method realizes accurate detection of chronic gastritis by utilizing DAGMM.
• Fine-tuning of pre-trained DCNN for gastritis detection from gastric X-ray images

  Misaki Kanai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 196 - 197 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a detection method of gastritis from gastric X-ray images using fine-tuning techniques. With the development of deep convolutional neural networks (DCNNs), DCNN-based methods have achieved more accurate performance than conventional machine learning methods using hand-crafted features in the field of medical image analysis. However, lack of training images often occurs in clinical situations even though DCNNs require a large amount of training images to avoid overfitting. Therefore, the proposed method aims to consider the clinical situations that a limited amount of the training images are available. By fine-tuning a DCNN pre-trained with a large amount of annotated natural images, we avoid overfitting and realize accurate detection of the gastritis with a small amount of the training images.
• Bone metastatic tumor detection based on AnoGAN using CT images

  Haruna Watanabe, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 235 - 236 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. In this paper, we propose a method to detect bone metastatic tumors using computed tomography (CT) images. Bone metastatic tumors spread from primary cancer to other organs, and they can cause severe pain. Therefore, it is important to detect metastatic tumors earlier in addition to primary cancer. However, since metastatic tumors are very small, and they emerge from unpredictable regions in the body, collecting metastatic tumor images is difficult compared to primary cancer. In such a case, it can be considered that the idea of anomaly detection is suitable. The proposed method based on a generative adversarial network model trains with only non-metastatic bone tumor images and detects bone metastatic tumor in an unsupervised manner. Then the anomaly score is defined for each test CT image. Experimental results show the anomaly scores between non-metastatic bone tumor images and metastatic bone tumor images are clearly different. The anomaly detection approach may be effective for the detection of bone metastatic tumors in CT images.
• Semi-supervised discriminative CCA for estimating viewed image categories from fMRI data

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 233 - 234 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method that estimates viewed image categories from functional magnetic resonance imaging (fMRI) data via semi-supervised discriminative canonical correlation analysis (Semi-DCCA). We newly derive Semi-DCCA that enables direct comparison of fMRI data and visual features extracted from viewed images while taking into account the class information and additional visual features to avoid overfitting. The proposed method enables estimation of image categories from fMRI data measured when subjects view images by comparing fMRI data with visual features through Semi-DCCA. Experimental results show that Semi-DCCA can improve estimation performance of the viewed image categories.
• Video classification based on user preferences with soft-bag multiple instance learning

  Akira Toyoda, Takahiro Ogawa, Miki Haseyama

  2019 IEEE 1st Global Conference on Life Sciences and Technologies, LifeTech 2019 198 - 199 2019/03 [Refereed][Not invited]
   

  © 2019 IEEE. This paper presents a method to classify videos based on user preferences with soft-bag multiple instance learning (MIL). Our method classifies videos that a user has watched into two classes (preferred and not-preferred) with two-modal features extracted from the videos and brain signals measured while the user is watching the videos. Our method splits videos and brain signals into fixed-length segments and computes features used for classification from only a fixed-number of segments selected based on the idea of soft-bag MIL. By using the features computed from the selected segments, our method makes it possible to classify videos in the case that some videos that a user prefers contain some scenes the user does not prefer, and vice versa. Our main contribution allows methods classifying videos based on user preferences to treat such a case unlike conventional methods.
• The Friction Properties of Firebrat Scales

  Yuji Hirai, Naoto Okuda, Naoki Saito, Takahiro Ogawa, Ryuichiro Machida, Shuhei Nomura, Masahiro Ohara, Miki Haseyama, Masatsugu Shimomura

  BIOMIMETICS 4 (1) 2313-7673 2019/01 

  Friction is an important subject for sustainability due to problems that are associated with energy loss. In recent years, micro- and nanostructured surfaces have attracted much attention to reduce friction; however, suitable structures are still under consideration. Many functional surfaces are present in nature, such as the friction reduction surfaces of snake skins. In this study, we focused on firebrats, Thermobia domestica, which temporary live in narrow spaces, such as piled papers, so their body surface (integument) is frequently in contact with surrounding substrates. We speculate that, in addition to optical, cleaning effects, protection against desiccation and enemies, their body surface may be also adapted to reduce friction. To investigate the functional effects of the firebrat scales, firebrat surfaces were observed using a field-emission scanning electron microscope (FE-SEM) and a colloidal probe atomic force microscope (AFM). Results of surface observations by FE-SEM revealed that adult firebrats are entirely covered with scales, whose surfaces have microgroove structures. Scale groove wavelengths around the firebrat's head are almost uniform within a scale but they vary between scales. At the level of single scales, AFM friction force measurements revealed that the firebrat scale reduces friction by decreasing the contact area between scales and a colloidal probe. The heterogeneity of the scales' groove wavelengths suggests that it is difficult to fix the whole body on critical rough surfaces and may result in a "fail-safe" mechanism.
• Multimodal Retrieval of Similar Soccer Videos Based on Optimal Combination of Multiple Distance Measures.

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 665 - 666 2019
• Similarity Calculation Based on Pass Regions in Soccer Videos.

  Sho Takahashi, Marco Bertini, Alberto Del Bimbo, Miki Haseyama, Toru Hagiwara

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 515 - 516 2019
• Interest Estimation for Images Based on Eye Gaze-based Visual and Text Features.

  Masanao Matsumoto, Naoki Saito 0006, Takahiro Ogawa, Miki Haseyama

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 481 - 482 2019
• Performance Prediction Method of Examinees Based on Matrix Completion.

  Yutaka Yamada, Takahiro Ogawa, Miki Haseyama

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 229 - 230 2019 [Refereed][Not invited]
  
• Estimating Viewed Image Categories from fMRI Activity via Multi-view Bayesian Generative Model.

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 127 - 128 2019 [Refereed][Not invited]
  
• The Extraction of Individual Music Preference Based on Deep Time-series CCA.

  Ryosuke Sawata, Takahiro Ogawa, Miki Haseyama

  IEEE 8th Global Conference on Consumer Electronics(GCCE) 15 - 16 2019 [Refereed][Not invited]
  
• Estimation of User-Specific Visual Attention Based on Gaze Information of Similar Users.

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 477 - 478 2019 [Refereed][Not invited]
  
• Effectiveness Evaluation of Deep Features for Image Reconstruction from fMRI Signals.

  Saya Takada, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 479 - 480 2019 [Refereed][Not invited]
  
• Scene Retrieval Using Text-to-image GAN-based Visual Similarities and Image-to-text Model-based Textual Similarities.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Global Conference on Consumer Electronics (GCCE) 13 - 14 2019 [Refereed][Not invited]
  
• Query is GAN: Scene Retrieval with Attentional Text-To-Image Generative Adversarial Network

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 153183 - 153193 2169-3536 2019 [Refereed][Not invited]
   

  © 2013 IEEE. Scene retrieval from input descriptions has been one of the most important applications with the increasing number of videos on the Web. However, this is still a challenging task since semantic gaps between features of texts and videos exist. In this paper, we try to solve this problem by utilizing a text-To-image Generative Adversarial Network (GAN), which has become one of the most attractive research topics in recent years. The text-To-image GAN is a deep learning model that can generate images from their corresponding descriptions. We propose a new retrieval framework, 'Query is GAN', based on the text-To-image GAN that drastically improves scene retrieval performance by simple procedures. Our novel idea makes use of images generated by the text-To-image GAN as queries for the scene retrieval task. In addition, unlike many studies on text-To-image GANs that mainly focused on the generation of high-quality images, we reveal that the generated images have reasonable visual features suitable for the queries even though they are not visually pleasant. We show the effectiveness of the proposed framework through experimental evaluation in which scene retrieval is performed from real video datasets.
• Synthetic gastritis image generation via loss function-based conditional pggan

  Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 87448 - 87457 2169-3536 2019 [Refereed][Not invited]
   

  © 2013 IEEE. In this paper, a novel synthetic gastritis image generation method based on a generative adversarial network (GAN) model is presented. Sharing medical image data is a crucial issue for realizing diagnostic supporting systems. However, it is still difficult for researchers to obtain medical image data since the data include individual information. Recently proposed GAN models can learn the distribution of training images without seeing real image data, and individual information can be completely anonymized by generated images. If generated images can be used as training images in medical image classification, promoting medical image analysis will become feasible. In this paper, we targeted gastritis, which is a risk factor for gastric cancer and can be diagnosed by gastric X-ray images. Instead of collecting a large amount of gastric X-ray image data, an image generation approach was adopted in our method. We newly propose loss function-based conditional progressive growing generative adversarial network (LC-PGGAN), a gastritis image generation method that can be used for a gastritis classification problem. The LC-PGGAN gradually learns the characteristics of gastritis in gastric X-ray images by adding new layers during the training step. Moreover, the LC-PGGAN employs loss function-based conditional adversarial learning so that generated images can be used as the gastritis classification task. We show that images generated by the LC-PGGAN are effective for gastritis classification using gastric X-ray images and have clinical characteristics of the target symptom.
• Semi-Supervised Learning Based on Tri-Training for Gastritis Classification using Gastric X-ray Images.

  Zongyao Li, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE International Symposium on Circuits and Systems, ISCAS 2019, Sapporo, Japan, May 26-29, 2019 IEEE 1 - 5 0271-4302 2019 [Refereed][Not invited]
   

  This paper presents a method of semi-supervised learning based on tri-training for gastritis classification using gastric X-ray images. The proposed method is constructed based on the tri-training architecture, and the strategies of label smoothing regularization and random erasing augmentation are utilized in the method to enhance the performance. Although the task of gastritis classification is challenging, we report that the proposed semi-supervised learning method using only a small number of labeled data achieves 0.888 harmonic mean of sensitivity and specificity on test data composed of 615 patients.
• Synthetic Image Generation for Gastritis Detection Based on Auxiliary Classifier Generative Adversarial Network.

  Misaki Kanai, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE International Symposium on Circuits and Systems, ISCAS 2019, Sapporo, Japan, May 26-29, 2019 IEEE 1 - 5 0271-4302 2019 [Refereed][Not invited]
   

  With the development of convolutional neural networks (CNNs), CNN-based methods for medical image analysis have achieved more accurate performance than conventional machine learning methods using hand-crafted features. Although these methods utilize a large number of training images and realize high performance, lack of the training images often occurs in medical image analysis due to several reasons. This paper presents a novel image generation method to construct a dataset for gastritis detection from gastric X-ray images. The proposed method effectively utilizes two kinds of training images (gastritis and non-gastritis images) to generate images of each domain by introducing label conditioning into a generative model. Experimental results using real-world gastric X-ray images show the effectiveness of the proposed method.
• Text-to-Image GAN-Based Scene Retrieval and Re-Ranking Considering Word Importance.

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 169920 - 169930 2169-3536 2019 [Refereed][Not invited]
   

  In this paper, we propose a novel scene retrieval and re-ranking method based on a text-to-image Generative Adversarial Network (GAN). The proposed method generates an image from an input query sentence based on the text-to-image GAN and then retrieves a scene that is the most similar to the generated image. By utilizing the image generated from the input query sentence as a query, we can control semantic information of the query image at the text level. Furthermore, we introduce a novel interactive re-ranking scheme to our retrieval method. Specifically, users can consider the importance of each word within the first input query sentence. Then the proposed method re-generates the query image that reflects the word importance provided by users. By updating the generated query image based on the word importance, it becomes feasible for users to revise retrieval results through this re-ranking process. In experiments, we showed that our retrieval method including the re-ranking scheme outperforms recently proposed retrieval methods.
• Estimating Regions of Deterioration in Electron Microscope Images of Rubber Materials via a Transfer Learning-Based Anomaly Detection Model.

  Ren Togo, Naoki Saito 0006, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 162395 - 162404 2169-3536 2019 [Refereed][Not invited]
   

  A method for estimating regions of deterioration in electron microscope images of rubber materials is presented in this paper. Deterioration of rubber materials is caused by molecular cleavage, external force, and heat. An understanding of these characteristics is essential in the field of material science for the development of durable rubber materials. Rubber material deterioration can be observed by using on electron microscope but it requires much effort and specialized knowledge to find regions of deterioration. In this paper, we propose an automated deterioration region estimation method based on deep learning and anomaly detection techniques to support such material development. Our anomaly detection model, called Transfer Learning-based Deep Autoencoding Gaussian Mixture Model (TL-DAGMM), uses only normal regions for training since obtaining training data for regions of deterioration is difficult. TL-DAGMM makes use of extracted high representation features from a pre-trained deep learning model and can automatically learn the characteristics of normal rubber material regions. Regions of deterioration are estimated at the pixel level by calculated anomaly scores. Experiments on real rubber material electron microscope images demonstrated the effectiveness of our model.
• Multi-feature Fusion Based on Supervised Multi-view Multi-label Canonical Correlation Projection.

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2019, Brighton, United Kingdom, May 12-17, 2019 IEEE 3936 - 3940 1520-6149 2019 [Refereed][Not invited]
   

  This paper presents multi-feature fusion based on supervised multi view multi-label canonical correlation projection (sM2CP). The proposed method applies sM2CP-based feature fusion to multiple features obtained from various convolutional neural networks (CNNs) whose characteristics are different. Since new fused features with high representation ability can be obtained, performance improvement of multi-label classification is realized. Specifically, in order to tackle the multi-label problem, sM2CP introduces a label similarity information of label vectors into the objective function of supervised multi-view canonical correlation analysis. Thus, sM2CP can deal with complex label information such as multi-label annotation. The main contribution of this paper is the realization of feature fusion of multiple CNN features for the multi-label problem by introducing multi-label similarity information into the canonical correlation analysis-based feature fusion approach. Experimental results show the effectiveness of sM2CP, which enables effective fusion of multiple CNN features.
• Team Tactics Estimation in Soccer Videos Based on a Deep Extreme Learning Machine and Characteristics of the Tactics.

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 153238 - 153248 2169-3536 2019 [Refereed][Not invited]
   

  A novel method for estimating team tactics in soccer videos based on a Deep Extreme Learning Machine (DELM) and unique characteristics of tactics is presented in this paper. The proposed method estimates the tactics of each team from players formations and enables successful training from a limited amount of training data. Specifically, the estimation of tactics consists of two stages. First, by utilizing two DELMs corresponding to the two teams, the proposed method estimates the provisional tactics of each team. Second, the proposed method updates the team tactics based on unique characteristics of soccer tactics, the relationship between tactics of the two teams and information on ball possession. Consequently, since the proposed method estimates the team tactics that satisfy these characteristics, accurate estimation results can be obtained. In an experiment, the proposed method is applied to actual soccer videos to verify its effectiveness.
• Interest Level Estimation Based on Tensor Completion via Feature Integration for Partially Paired User's Behavior and Videos.

  Tetsuya Kushima, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 148576 - 148585 2169-3536 2019 [Refereed][Not invited]
   

  A novel method for interest level estimation based on tensor completion via feature integration for partially paired users' behavior and videos is presented in this paper. The proposed method defines a novel canonical correlation analysis (CCA) framework that is suitable for interest level estimation, which is a hybrid version of semi-supervised CCA (SemiCCA) and supervised locality preserving CCA (SLPCCA) called semi-supervised locality preserving CCA (S2LPCCA). For partially paired users' behavior and videos in actual shops and on the Internet, new integrated features that maximize the correlation between partially paired samples by the principal component analysis (PCA)-mixed CCA framework are calculated. Then videos that users have not watched can be used for the estimation of users' interest levels. Furthermore, local structures of partially paired samples in the same class are preserved for accurate estimation of interest levels. Tensor completion, which can be applied to three contexts, videos, users and "canonical features and interest levels," is used for estimation of interest levels. Consequently, the proposed method realizes accurate estimation of users' interest levels based on S2LPCCA and the tensor completion from partially paired training features of users' behavior and videos. Experimental results obtained by applying the proposed method to actual data show the effectiveness of the proposed method.
• Consensus Clustering of Tweet Networks via Semantic and Sentiment Similarity Estimation.

  Ryosuke Harakawa, Shoji Takimura, Takahiro Ogawa, Miki Haseyama, Masahiro Iwahashi

  IEEE Access 7 116207 - 116217 2169-3536 2019 [Refereed][Not invited]
   

  Although Twitter has become an important source of information, the number of accessible tweets is too large for users to easily find their desired information. To overcome this difficulty, a method for tweet clustering is proposed in this paper. Inspired by the reports that network representation is useful for multimedia content analysis including clustering, a network-based approach is employed. Specifically, a consensus clustering method for tweet networks that represent relationships among the tweets' semantics and sentiment are newly derived. The proposed method integrates multiple clustering results obtained by applying successful clustering methods to the tweet networks. By integrating complementary clustering results obtained based on semantic and sentiment features, the accurate clustering of tweets becomes feasible. The contribution of this work can be found in the utilization of the features, which differs from existing network-based consensus clustering methods that target only the network structure. Experimental results for a real-world Twitter dataset, which includes 65 553 tweets of 25 datasets, verify the effectiveness of the proposed method.
• Music Video Recommendation Based on Link Prediction Considering Local and Global Structures of a Network.

  Yui Matsumoto, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE Access 7 104155 - 104167 2169-3536 2019 [Refereed][Not invited]
   

  A novel method for music video recommendation is presented in this paper. The contributions of this paper are two-fold. (i) The proposed method constructs a network, which not only represents relationships between music videos and users but also captures multi-modal features of music videos. This enables collaborative use of multi-modal features such as audio, visual, and textual features, and multiple social metadata that can represent relationships between music videos and users on video hosting services. (ii) A novel scheme for link prediction considering local and global structures of the network (LP-LGSN) is newly derived by fusing multiple link prediction scores based on both local and global structures. By using the LP-LGSN to predict the degrees to which users desire music videos, the proposed method can recommend users' desired music videos. The experimental results for a real-world dataset constructed from YouTube-8M show the effectiveness of the proposed method.
• Scene Retrieval from Multiple Resolution Generated Images Based on Text-to-image GAN

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama

  2019 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS) 1 - 5 0271-4302 2019 [Refereed][Not invited]
   

  Text-to-image Generative Adversarial Network (GAN) is a deep learning model that generates an image from an input sentence. It is expressly attracting attentions because of its applicability of the generated images. However, many existing studies have still focused on generation of high-quality images, and there are few studies focusing on application of the generated images since text-to-image GANs still cannot produce visually pleasing images in the complicated tasks. In this paper, we apply a text-to-image GAN as a generator of query images for a scene retrieval task to show availability of the visually non-pleasant images. The proposed method utilizes a low-resolution generated image that focuses on a sentence and a high-resolution generated image that focuses on each word of the sentence to retrieve a desired scene. With this mechanism, the proposed method realizes a high-accuracy scene retrieval from a sentence input. Experimental results show the effectiveness of our method.
• Cardiac sarcoidosis classification with deep convolutional neural network-based features using polar maps.

  Ren Togo, Kenji Hirata, Osamu Manabe, Hiroshi Ohira, Ichizo Tsujino, Keiichi Magota, Takahiro Ogawa, Miki Haseyama, Tohru Shiga

  Computers in biology and medicine 104 81 - 86 0010-4825 2019/01 [Refereed][Not invited]
   

  AIMS: The aim of this study was to determine whether deep convolutional neural network (DCNN)-based features can represent the difference between cardiac sarcoidosis (CS) and non-CS using polar maps. METHODS: A total of 85 patients (33 CS patients and 52 non-CS patients) were analyzed as our study subjects. One radiologist reviewed PET/CT images and defined the left ventricle region for the construction of polar maps. We extracted high-level features from the polar maps through the Inception-v3 network and evaluated their effectiveness by applying them to a CS classification task. Then we introduced the ReliefF algorithm in our method. The standardized uptake value (SUV)-based classification method and the coefficient of variance (CoV)-based classification method were used as comparative methods. RESULTS: Sensitivity, specificity and the harmonic mean of sensitivity and specificity of our method with the ReliefF algorithm were 0.839, 0.870 and 0.854, respectively. Those of the SUVmax-based classification method were 0.468, 0.710 and 0.564, respectively, and those of the CoV-based classification method were 0.655, 0.750 and 0.699, respectively. CONCLUSION: The DCNN-based high-level features may be more effective than low-level features used in conventional quantitative analysis methods for CS classification.
• Distress classification of class-imbalanced inspection data via correlation-maximizing weighted extreme learning machine

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  Advanced Engineering Informatics 37 79 - 87 1474-0346 2018/08/01 [Refereed][Not invited]
   

  This paper presents distress classification of class-imbalanced inspection data via correlation-maximizing weighted extreme learning machine (CMWELM). For distress classification, it is necessary to extract semantic features that can effectively distinguish multiple kinds of distress from a small amount of class-imbalanced data. In recent machine learning techniques such as general deep learning methods, since effective feature transformation from visual features to semantic features can be realized by using multiple hidden layers, a large amount of training data are required. However, since the amount of training data of civil structures becomes small, it becomes difficult to perform successful transformation by using these multiple hidden layers. On the other hand, CMWELM consists of two hidden layers. The first hidden layer performs feature transformation, which can directly extract the semantic features from visual features, and the second hidden layer performs classification with solving the class-imbalanced problem. Specifically, in the first hidden layer, the feature transformation is realized by using projections obtained by maximizing the canonical correlation between visual and text features as weight parameters of the hidden layer without designing multiple hidden layers. Furthermore, the second hidden layer enables successful training of our classifier by using weighting factors concerning the class-imbalanced problem. Consequently, CMWELM realizes accurate distress classification from a small amount of class-imbalanced data.
• Strategy to develop convolutional neural network-based classifier for diagnosis of whole-body FDG PET images

  Keisuke Kawauchi, Kenji Hirata, Seiya Ichikawa, Osamu Manabe, Kentaro Kobayashi, Shiro Watanabe, Miki Haseyama, Takahiro Ogawa, Ren Togo, Tohru Shiga, Chietsugu Katoh

  Society of Nuclear Medicine and Molecular Imaging Annual Meeting (SNMMI) 59 0161-5505 2018/06 [Refereed][Not invited]
  
• Use of deep convolutional neural network-based features for detection of cardiac sarcoidosis from polar map

  Ren Togo, Kenji Hirata, Osamu Manabe, Hiroshi Ohira, Ichizo Tsujino, Takahiro Ogawa, Miki Haseyama, Tohru Shiga

  Society of Nuclear Medicine and Molecular Imaging Annual Meeting (SNMMI) 59 0161-5505 2018/06 [Refereed][Not invited]
  
• Estimating the quality of fractal compressed images using lacunarity

  Megumi Takezawa, Hirofumi Sanada, Takahiro Ogawa, Miki Haseyama

  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences E101A (6) 900 - 903 1745-1337 2018/06/01 [Refereed][Not invited]
   

  In this paper, we propose a highly accurate method for estimating the quality of images compressed using fractal image compression. Using an iterated function system, fractal image compression compresses images by exploiting their self-similarity, thereby achieving high levels of performance however, we cannot always use fractal image compression as a standard compression technique because some compressed images are of low quality. Generally, sufficient time is required for encoding and decoding an image before it can be determined whether the compressed image is of low quality or not. Therefore, in our previous study, we proposed a method to estimate the quality of images compressed using fractal image compression. Our previous method estimated the quality using image features of a given image without actually encoding and decoding the image, thereby providing an estimate rather quickly however, estimation accuracy was not entirely sufficient. Therefore, in this paper, we extend our previously proposed method for improving estimation accuracy. Our improved method adopts a new image feature, namely lacunarity. Results of simulation showed that the proposed method achieves higher levels of accuracy than those of our previous method.
• Graph-Based Video Search Reranking with Local and Global Consistency Analysis

  Soh Yoshida, Takahiro Ogawa, Miki Haseyama, Mitsuji Muneyasu

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E101D (5) 1430 - 1440 1745-1361 2018/05 [Refereed][Not invited]
   

  Video reranking is an effective way for improving the retrieval performance of text-based video search engines. This paper proposes a graph-based Web video search reranking method with local and global consistency analysis. Generally, the graph-based reranking approach constructs a graph whose nodes and edges respectively correspond to videos and their pairwise similarities. A lot of reranking methods are built based on a scheme which regularizes the smoothness of pairwise relevance scores between adjacent nodes with regard to a user's query. However, since the overall consistency is measured by aggregating only the local consistency over each pair, errors in score estimation increase when noisy samples are included within query-relevant videos' neighbors. To deal with the noisy samples, the proposed method leverages the global consistency of the graph structure, which is different from the conventional methods. Specifically, in order to detect this consistency, the propose method introduces a spectral clustering algorithm which can detect video groups, in which videos have strong semantic correlation, on the graph. Furthermore, a new regularization term, which smooths ranking scores within the same group, is introduced to the reranking framework. Since the score regularization is performed by both local and global aspects simultaneously, the accurate score estimation becomes feasible. Experimental results obtained by applying the proposed method to a real-world video collection show its effectiveness.
• 映像情報メディア年報2017シリーズ(第8回)メディア工学の研究動向

  長谷山 美紀, 河村 圭, 田良島 周平, 新井 啓之

  映像情報メディア学会誌 Vol.72 (No.2) pp.241-246  1342-6907 2018/03 [Not refereed][Not invited]
   

  本稿では、(1) 拡張現実・仮想現実の研究動向、(2) 深層学習の研究動向と実用展開、(3) IoT とメディア処理、等、メディア工学分野の研究動向を紹介する。(1)については、デプスセンサやヘッドマウントディスプレイ等のデバイスの登場と性能向上を踏まえつつ、これらを活用した研究開発の進展と動向を紹介する。(2) については、加速度的に進む画像映像メディアへの深層学習技術の適用について、具体的に画像分類や物体検出、さらには画像検索から教師データの検討、実用展開に至るまでを解説する。(3)では、(2) を受けてメディア工学の分野でAI 技術と相乗効果を生み出すIoT 技術について解説する。
• A Novel Framework for Estimating Viewer Interest by Unsupervised Multimodal Anomaly Detection

  Yuma Sasaka, Takahiro Ogawa, Miki Haseyama

  IEEE Access 6 8340 - 8350 2169-3536 2018/02/09 [Refereed][Not invited]
   

  A reliable method to estimate viewer interest is highly sought after for human-centered video information retrieval. A method that estimates viewer interest while users are watching Web videos is presented in this paper. The method uses a framework for anomaly detection based on collaborative use of facial expression and biological signals such as electroencephalogram (EEG) signals. To the best of our knowledge, there have been no studies that have taken into account two actual mechanisms of the behavior of users while they are watching Web videos. First, whereas most Web videos garner very little attention, a small number attract millions of views. Therefore, a framework for anomaly detection is newly applied to facial expression and EEG in order to model the imbalanced distribution of popularity. Second, since the number of Web videos that are labeled by users as interesting/not interesting is generally too small to estimate viewer interest by a supervised approach, the proposed method utilizes parametric techniques for anomaly detection, which estimates viewer interest in an unsupervised way. Unlike some related studies for estimating viewer interest, our method takes into account actual mechanisms of the behavior of users while they are watching Web videos by utilizing parametric techniques for anomaly detection. Then viewer interest can be estimated on the basis of an anomaly score calculated from our proposed method. Consequently, successful estimation of viewer interest based on a framework for anomaly detection, via collaborative use of facial expression and biological signals, becomes feasible.
• Accurate estimation of personalized video preference using multiple users' viewing behavior

  Yoshiki Ito, Takahiro Ogawa, Miki Haseyama

  IEICE Transactions on Information and Systems E101D (2) 481 - 490 1745-1361 2018/02/01 [Refereed][Not invited]
   

  A method for accurate estimation of personalized video preference using multiple users' viewing behavior is presented in this paper. The proposed method uses three kinds of features: a video, user's viewing behavior and evaluation scores for the video given by a target user. First, the proposed method applies Supervised Multiview Spectral Embedding (SMSE) to obtain lower-dimensional video features suitable for the following correlation analysis. Next, supervised Multi-View Canonical Correlation Analysis (sMVCCA) is applied to integrate the three kinds of features. Then we can get optimal projections to obtain new visual features, "canonical video features" reflecting the target user's individual preference for a video based on sMVCCA. Furthermore, in our method, we use not only the target user's viewing behavior but also other users' viewing behavior for obtaining the optimal canonical video features of the target user. This unique approach is the biggest contribution of this paper. Finally, by integrating these canonical video features, Support Vector Ordinal Regression with Implicit Constraints (SVORIM) is trained in our method. Consequently, the target user's preference for a video can be estimated by using the trained SVORIM. Experimental results show the effectiveness of our method.
• Sentiment-aware personalized tweet recommendation through multimodal FFM

  Ryosuke Harakawa, Daichi Takehara, Takahiro Ogawa, Miki Haseyama

  Multimedia Tools and Applications 77 (14) 18741 - 18759 1573-7721 2018 [Refereed][Not invited]
   

  For realizing quick and accurate access to desired information and effective advertisements or election campaigns, personalized tweet recommendation is highly demanded. Since multimedia contents including tweets are tools for users to convey their sentiment, users' interest in tweets is strongly influenced by sentiment factors. Therefore, successful personalized tweet recommendation can be realized if sentiment in tweets can be estimated. However, sentiment factors were not taken into account in previous works and the performance of previous methods may be limited. To overcome the limitation, a method for sentiment-aware personalized tweet recommendation through multimodal Field-aware Factorization Machines (FFM) is newly proposed in this paper. Successful personalized tweet recommendation becomes feasible through the following three contributions: (i) sentiment factors are newly introduced into personalized tweet recommendation, (ii) users' interest is modeled by deriving multimodal FFM that enables collaborative use of multiple factors in a tweet, i.e., publisher, topic and sentiment factors, and (iii) the effectiveness of using sentiment factors as well as publisher and topic factors is clarified from results of experiments using real-world datasets related to worldwide hot topics, "#trump", "#hillaryclinton" and "#ladygaga". In addition to showing the effectiveness of the proposed method, the applicability of the proposed method to other tasks such as advertisement and social analysis is discussed as a conclusion and future work of this paper.
• Favorite Video Estimation Based on Multiview Feature Integration via KMvLFDA.

  Array,Array,Miki Haseyama

  IEEE Access 6 63833 - 63842 2169-3536 2018 [Refereed][Not invited]
   

  This paper presents a novel method for favorite video estimation based on multiview feature integration via kernel multiview local fisher discriminant analysis (KMvLFDA). The proposed method first extracts electroencephalogram (EEG) features from users' EEG signals recorded while watching videos and multiple visual features from videos. Then, multiple EEG-based visual features are obtained by applying locality preserving canonical correlation analysis to EEG features and each visual feature. Next, KMvLFDA, which is newly derived in this paper, explores the complementary properties of different features and integrates the multiple EEG-based visual features. In addition, by using KMvLFDA, between-class scatter is maximized and within-class scatter is minimized in the integrated feature space. Consequently, it can be expected that the new features that are obtained by the above integration are more effective than each of the EEG-based visual features for the estimation of users' favorite videos. The main contribution of this paper is the new derivation of KMvLFDA. Successful estimation of users' favorite videos becomes feasible by using the new features obtained via KMvLFDA.
• Favorite Video Classification Based on Multimodal Bidirectional LSTM.

  Array,Array, Keisuke Maeda, Miki Haseyama

  IEEE Access 6 61401 - 61409 2169-3536 2018 [Refereed][Not invited]
   

  Video classification based on the user's preference (information of what a user likes: WUL) is important for realizing human-centered video retrieval. A better understanding of the rationale of WUL would greatly contribute to the support for successful video retrieval. However, a few studies have shown the relationship between information of what a user watches and WUL. A new method that classifies videos on the basis of WUL using video features and electroencephalogram (EEG) signals collaboratively with a multimodal bidirectional Long Short-Term Memory (Bi-LSTM) network is presented in this paper. To the best of our knowledge, there has been no study on WUL-based video classification using video features and EEG signals collaboratively with LSTM. First, we newly apply transfer learning to the WUL-based video classification since the number of labels (liked or not liked) attached to videos by users is small, and it is difficult to classify videos based on WUL. Furthermore, we conduct a user study for showing that the representation of psychophysiological signals calculated from Bi-LSTM is effective for the WUL-based video classification. Experimental results showed that our deep neural network feature representations can distinguish WUL for each subject.
• Selection of Significant Brain Regions Based on MvGTDA and TS-DLF for Emotion Estimation.

  Array,Array,Miki Haseyama

  IEEE Access 6 32481 - 32492 2169-3536 2018 [Refereed][Not invited]
   

  In this paper, we propose a novel method for estimating human emotion using functional brain images. The final goal of our study is contribution to affective brain computer interfaces (aBCIs), which use neuropsychological signals. In the proposed method, we newly derive multiview general tensor discriminant analysis (MvGTDA) in order to reveal significant brain regions and accurately estimate human emotion evoked by visual stimuli. This is because it is important to find activation of multiple brain regions for estimating emotional states. Since we regard a Brodmann area as a view and introduce -norm regularization for these views, MvGTDA can eliminate non-crucial Brodmann areas and select significant ones. Moreover, in general studies on functional brain images based on machine learning methodologies, there is an overfitting problem caused by a small sample size. Therefore, revealing significant Brodmann areas based on MvGTDA has another important role, i.e., solving the overfitting problem. By inputting estimation results respectively obtained from the significant areas and the MvGTDA-based feature, tensor-based supervised decision-level fusion (TS-DLF) integrates them and outputs the final estimation result of the users emotion. In experiments, we showed the effectiveness of our method by using actual functional brain images and we revealed the significant brain regions in emotional states.
• A Language-Independent Ontology Construction Method Using Tagged Images in Folksonomy.

  Array,Takahiro Ogawa, Miki Haseyama

  IEEE Access 6 2930 - 2942 2018 [Refereed][Not invited]
  
• Tracking topic evolution via salient keyword matching with consideration of semantic broadness for Web video discovery

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  Multimedia Tools and Applications 77 (16) 1 - 28 1573-7721 2017/12/06 [Refereed][Not invited]
   

  A method to track topic evolution via salient keyword matching with consideration of semantic broadness for Web video discovery is presented in this paper. The proposed method enables users to understand the evolution of topics over time for discovering Web videos in which they are interested. A framework that enables extraction and tracking of the hierarchical structure, which contains Web video groups with various degrees of semantic broadness, is newly derived as follows: Based on network analysis using multimodal features, i.e., features of video contents and metadata, our method extracts the hierarchical structure and salient keywords that represent contents of each Web video group. Moreover, salient keyword matching, which is newly developed by considering salient keyword distribution, semantic broadness of each Web video group and initial topic relevance, is applied to each hierarchical structure obtained in different time stamps. Unlike methods in previous works, by considering the semantic broadness as well as the salient keyword distribution, our method can overcome the problem of the desired semantic broadness of topics being different depending on each user. Also, the initial topic relevance enables correction of the gap from an initial topic at the start of tracking. Consequently, it becomes feasible to track the evolution of topics over time for finding Web videos in which the users are interested. Experimental results for real-world datasets containing YouTube videos verify the effectiveness of the proposed method.
• Extracting hierarchical structure of content groups from different social media platforms using multiple social metadata

  Daichi Takehara, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  MULTIMEDIA TOOLS AND APPLICATIONS 76 (19) 20249 - 20272 1380-7501 2017/10 [Refereed][Not invited]
   

  A novel scheme for retrieving users' desired contents, i.e., contents with topics in which users are interested, from multiple social media platforms is presented in this paper. In existing retrieval schemes, users first select a particular platform and then input a query into the search engine. If users do not specify suitable platforms for their information needs and do not input suitable queries corresponding to the desired contents, it becomes difficult for users to retrieve the desired contents. The proposed scheme extracts the hierarchical structure of content groups (sets of contents with similar topics) from different social media platforms, and it thus becomes feasible to retrieve desired contents even if users do not specify suitable platforms and do not input suitable queries. This paper has two contributions: (1) A new feature extraction method, Locality Preserving Canonical Correlation Analysis with multiple social metadata (LPCCA-MSM) that can detect content groups without the boundaries of different social media platforms is presented in this paper. LPCCA-MSM uses multiple social metadata as auxiliary information unlike conventional methods that only use content-based information such as textual or visual features. (2) The proposed novel retrieval scheme can realize hierarchical content structuralization from different social media platforms. The extracted hierarchical structure shows various abstraction levels of content groups and their hierarchical relationships, which can help users select topics related to the input query. To the best of our knowledge, an intensive study on such an application has not been conducted; therefore, this paper has strong novelty. To verify the effectiveness of the above contributions, extensive experiments for real-world datasets containing YouTube videos and Wikipedia articles were conducted.
• Wiener-Based Inpainting Quality Prediction

  Takahiro Ogawa, Akira Tanaka, Miki Haseyama

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E100D (10) 2614 - 2626 1745-1361 2017/10 [Refereed][Not invited]
   

  A Wiener-based inpainting quality prediction method is presented in this paper. The proposed method is the first method that can predict inpainting quality both before and after the intensities have become missing even if their inpainting methods are unknown. Thus, when the target image does not include any missing areas, the proposed method estimates the importance of intensities for all pixels, and then we can know which areas should not be removed. Interestingly, since this measure can be also derived in the same manner for its corrupted image already including missing areas, the expected difficulty in reconstruction of these missing pixels is predicted, i.e., we can know which missing areas can be successfully reconstructed. The proposed method focuses on expected errors derived from the Wiener filter, which enables least-squares reconstruction, to predict the inpainting quality. The greatest advantage of the proposed method is that the same inpainting quality prediction scheme can be used in the above two different situations, and their results have common trends. Experimental results show that the inpainting quality predicted by the proposed method can be successfully used as a universal quality measure.
• Visualizing Web Images Using Fisher Discriminant Locality Preserving Canonical Correlation Analysis

  Kohei Tateno, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E100D (9) 2005 - 2016 1745-1361 2017/09 [Refereed][Not invited]
   

  A novel dimensionality reduction method, Fisher Discriminant Locality Preserving Canonical Correlation Analysis (FDLPCCA), for visualizing Web images is presented in this paper. FDLP-CCA can integrate two modalities and discriminate target items in terms of their semantics by considering unique characteristics of the two modalities. In this paper, we focus onWeb images with text uploaded on Social Networking Services for these two modalities. Specifically, text features have high discriminate power in terms of semantics. On the other hand, visual features of images give their perceptual relationships. In order to consider both of the above unique characteristics of these two modalities, FDLPCCA estimates the correlation between the text and visual features with consideration of the cluster structure based on the text features and the local structures based on the visual features. Thus, FDLP-CCA can integrate the different modalities and provide separated manifolds to organize enhanced compactness within each natural cluster.
• Biomimetics Image Retrieval Platform

  Miki Haseyama, Takahiro Ogawa, Sho Takahashi, Shuhei Nomura, Masatsugu Shimomura

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E100D (8) 1563 - 1573 1745-1361 2017/08 [Refereed][Not invited]
   

  Biomimetics is a new research field that creates innovation through the collaboration of different existing research fields. However, the collaboration, i.e., the exchange of deep knowledge between different research fields, is difficult for several reasons such as differences in technical terms used in different fields. In order to overcome this problem, we have developed a new retrieval platform, "Biomimetics image retrieval platform," using a visualization-based image retrieval technique. A biological database contains a large volume of image data, and by taking advantage of these image data, we are able to overcome limitations of text-only information retrieval. By realizing such a retrieval platform that does not depend on technical terms, individual biological databases of various species can be integrated. This will allow not only the use of data for the study of various species by researchers in different biological fields but also access for a wide range of researchers in fields ranging from materials science, mechanical engineering and manufacturing. Therefore, our platform provides a new path bridging different fields and will contribute to the development of biomimetics since it can overcome the limitation of the traditional retrieval platform.
• Deterioration Level Estimation on Transmission Towers via Extreme Learning Machine based on Combination Use of Local Receptive Field and Principal Component Analysis

  K. Maeda, S. Takahashi, T. Ogawa, M. Haseyama

  International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC) 457 - 458 2017/07 [Refereed][Not invited]
  
• Effectiveness Evaluation of Imaging Direction for Estimation of Gastritis Regions on Gastric X-ray Images

  Ren Togo, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama

  International Technical Conference on Circuits, Systems, Computers, and Communications (ITC-CSCC) 459 - 460 2017/05 [Refereed][Not invited]
  
• Helicobacter Pylori infection detection from gastric X-ray images based on feature fusion and decision fusion

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama

  COMPUTERS IN BIOLOGY AND MEDICINE 84 69 - 78 0010-4825 2017/05 [Refereed][Not invited]
   

  In this paper, a fully automatic method for detection of Helicobacter pylori (H. pylori) infection is presented with the aim of constructing a computer-aided diagnosis (CAD) system. In order to realize a CAD system with good performance for detection of H. pylori infection, we focus on the following characteristic of stomach X-ray examination. The accuracy of X-ray examination differs depending on the symptom of H. pylori infection that is focused on and the position from which X-ray images are taken. Therefore, doctors have to comprehensively assess the symptoms and positions. In order to introduce the idea of doctors' assessment into the CAD system, we newly propose a method for detection of H. pylori infection based on the combined use of feature fusion and decision fusion. As a feature fusion scheme, we adopt Multiple Kernel Learning (MKL). Since MKL can combine several features with determination of their weights, it can represent the differences in symptoms. By constructing an MKL classifier for each position, we can obtain several detection results. Furthermore, we introduce confidence-based decision fusion, which can consider the relationship between the classifier's performance and the detection results. Consequently, accurate detection of H. pylori infection becomes possible by the proposed method. Experimental results obtained by applying the proposed method to real X-ray images show that our method has good performance, close to the results of detection by specialists, and indicate that the realization of a CAD system for determining the risk of H. pylori infection is possible.
• Tracking hierarchical structure of web video groups based on salient keyword matching including semantic broadness estimation

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings 1238 - 1242 2017/04/19 [Refereed][Not invited]
   

  This paper presents a novel method to track the hierarchical structure of Web video groups on the basis of salient keyword matching including semantic broadness estimation. To the best of our knowledge, this paper is the first work to perform extraction and tracking of the hierarchical structure simultaneously. Specifically, the proposed method first extracts the hierarchical structure of Web video groups and salient keywords of them on the basis of an improved scheme of our previously reported method. Moreover, to calculate similarities between Web video groups obtained in different time stamps, salient keyword matching is newly developed by considering both co-occurrences of the salient keywords and semantic broadness of each Web video group. Consequently, tracking of the hierarchical structure over time becomes feasible to easily understand popularity trends of many Web videos for realizing effective retrieval.
• Distress Classification of Class Imbalanced Data for Maintenance Inspection of Road Structures in Express Way

  K. Maeda, S. Takahashi, T. Ogawa, M. Haseyama

  International Conference on Civil and Building Engineering Informatics in conjunction with Conference on Computer Applications in Civil and Hydraulic Engineering (ICCBEI & CCACHE) 182 - 185 2017/04 [Refereed][Not invited]
  
• Human-Centered Video Feature Selection via mRMR-SCMMCCA for Preference Extraction

  Takahiro Ogawa, Yoshiaki Yamaguchi, Satoshi Asamizu, Miki Haseyama

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E100D (2) 409 - 412 1745-1361 2017/02 [Refereed][Not invited]
   

  This paper presents human-centered video feature selection via mRMR-SCMMCCA (minimum Redundancy and Maximum Relevance-Specific Correlation Maximization Multiset Canonical Correlation Analysis) algorithm for preference extraction. The proposed method derives SCMMCCA, which simultaneously maximizes two kinds of correlations, correlation between video features and users' viewing behavior features and correlation between video features and their corresponding rating scores. By monitoring the derived correlations, the selection of the optimal video features that represent users' individual preference becomes feasible.
• PERSONALIZED VIDEO PREFERENCE ESTIMATION BASED ON EARLY FUSION USING MULTIPLE USERS'VIEWING BEHAVIOR

  Yoshiki Ito, Takahiro Ogawa, Miki Haseyama

  2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 3006 - 3010 1520-6149 2017 [Refereed][Not invited]
   

  This paper presents a novel method for personalized video preference estimation based on early fusion using multiple users' viewing behavior. The proposed method adopts supervised Multi-View Canonical Correlation Analysis (sMVCCA) to estimate correlation between different types of features. Specifically, we estimate optimal projections maximizing the correlation between three features of video, target user's viewing behavior and evaluation scores for video. Then novel video features (canonical video features), which reflect the target user's individual preference, are obtained by the estimated projections. Furthermore, our method computes sMVCCA-based canonical video features by using multiple users' viewing behavior and a target user's evaluation scores. This non-conventional approach using the multiple users' viewing behavior for the preference estimation of the target user is the biggest contribution of our method, and it enables early fusion of the canonical video features. Consequently, successful video recommendation that reflects the users' individual preference can be expected via the evaluation score prediction from the integrated canonical video features. Experimental results show the effectiveness of our method.
• EXEMPLAR-BASED IMAGE COMPLETION VIA NEW QUALITY MEASURE BASED ON PHASELESS TEXTURE FEATURES

  Takahiro Ogawa, Miki Haseyama

  2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 1827 - 1831 1520-6149 2017 [Refereed][Not invited]
   

  This paper presents an exemplar-based image completion via a new quality measure based on phaseless texture features. The proposed method derives a new quality measure obtained by monitoring errors caused in power spectra, i.e., errors of phaseless texture features, converged through phase retrieval. Even if a target patch includes missing pixels, this measure enables selection of the best matched patch including the most similar texture features for realizing the exemplar-based image completion. Furthermore, since the phaseless texture features are robust to various changes such as spatial gaps and luminance changes, the new quality measure successfully provides the best matched patch from few training examples. Then, by solving an optimization problem that retrieves the phase of the target patch from the phaseless texture features of the best matched patch, its missing areas can be reconstructed. Consequently, accurate image completion using the new quality measure becomes feasible. Subjective and quantitative experimental results are shown to verify the effectiveness of our method using the new quality measure.
• EMOTION ESTIMATION VIA TENSOR-BASED SUPERVISED DECISION-LEVEL FUSION FROM MULTIPLE BRODMANN AREAS

  Kento Sugata, Takahiro Ogawa, Miki Haseyama

  2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 999 - 1003 1520-6149 2017 [Refereed][Not invited]
   

  This paper presents a novel method that estimates human emotion based on tensor-based supervised decision-level fusion (TS-DLF) from multiple Brodmann areas (BAs). From multiple brain data corresponding to these BAs captured by functional magnetic resonance imaging (fMRI), our method performs general tensor discriminant analysis (GTDA) to obtain features which can reflect the user's emotion. Furthermore, since the dimension of the obtained features becomes lower, this can avoid overfitting in the following training procedure of estimators. Next, by separately using the transformed BA data obtained after GTDA, we obtain multiple estimation results of the user's emotion based on logistic tensor regression (LTR). Then our method realizes the decision of the final result based on TS-DLF from the multiple estimation results. This approach, i.e., the integration of the multiple BAs' results for the whole-brain data, is the biggest contribution of this paper. TS-DLF successfully integrates the multiple estimation results with considering the performance of the LTR-based estimator constructed for each BA. Experimental results show that our method outperforms state-of-the-art approaches, and the effectiveness of our method can be confirmed.
• Extracting Hierarchical Structure of Web Video Groups Based on Sentiment-Aware Signed Network Analysis

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  IEEE ACCESS 5 16963 - 16973 2169-3536 2017 [Refereed][Not invited]
   

  Sentiment in multimedia contents has an influence on their topics, since multimedia contents are tools for social media users to convey their sentiment. Performance of applications such as retrieval and recommendation will be improved if sentiment in multimedia contents can be estimated; however, there have been few works in which such applications were realized by utilizing sentiment analysis. In this paper, a novel method for extracting the hierarchical structure of Web video groups based on sentiment-aware signed network analysis is presented to realize Web video retrieval. First, the proposed method estimates latent links between Web videos by using multimodalfeatures of contents and sentiment features obtained from texts attached to Web videos. Thus, our method enables construction of a signed network that reflects not only similarities but also positive and negative relations between topics of Web videos. Moreover, an algorithm to optimize a modularity-based measure, which can adaptively adjust the balance between positive and negative edges, was newly developed. This algorithm detects Web video groups with similar topics at multiple abstraction levels; thus, successful extraction of the hierarchical structure becomes feasible. By providing the hierarchical structure, users can obtain an overview of many Web videos and it becomes feasible to successfully retrieve the desired Web videos. Results of experiments using a new benchmark dataset, YouTube-8M, validate the contributions of this paper, i.e., 1) the first attempt to utilize sentiment analysis for Web video grouping and 2) a novel algorithm for analyzing a weighted signed network derived from sentiment and multimodal features.
• Retrieval of similar inspection records based on metric learning using experienced inspectors' evaluation

  Ryota Saito, Sho Takahashi, Takahiro Ogawa, Miki Hasayama

  2016 IEEE 5th Global Conference on Consumer Electronics, GCCE 2016 1 - 2 2016/12/27 [Refereed][Not invited]
   

  This paper presents a retrieval method of similar inspection records in road structures based on metric learning using experienced inspectors' evaluation. Inspection records of road structures include images and text-based information such as category of distress, damaged parts and degree of damage. The proposed method calculates distances from query inspection records, and rank lists of retrieval results are obtained for each feature. In this approach, the distance quantification are updated on the basis of experienced inspectors' evaluation. Finally, the proposed method obtains retrieval results by integrating the multiple rank lists. The experimental results show the effectiveness of the proposed method.
• Adaptive Subspace-Based Inverse Projections via Division Into Multiple Sub-Problems for Missing Image Data Restoration

  Takahiro Ogawa, Miki Haseyama

  IEEE TRANSACTIONS ON IMAGE PROCESSING 25 (12) 5971 - 5986 1057-7149 2016/12 [Refereed][Not invited]
   

  This paper presents adaptive subspace-based inverse projections via division into multiple sub-problems (ASIP-DIMSs) for missing image data restoration. In the proposed method, a target problem for estimating missing image data is divided into multiple sub-problems, and each sub-problem is iteratively solved with the constraints of other known image data. By projection into a subspace model of image patches, the solution of each sub-problem is calculated, where we call this procedure "subspace-based inverse projection" for simplicity. The proposed method can use higher dimensional subspaces for finding unique solutions in each sub-problem, and successful restoration becomes feasible, since a high level of image representation performance can be preserved. This is the biggest contribution of this paper. Furthermore, the proposed method generates several subspaces from known training examples and enables derivation of a new criterion in the above framework to adaptively select the optimal subspace for each target patch. In this way, the proposed method realizes missing image data restoration using ASIP-DIMS. Since our method can estimate any kind of missing image data, its potential in two image restoration tasks, image inpainting and super-resolution, based on several methods for multivariate analysis is also shown in this paper.
• A Web video retrieval method using hierarchical structure of Web video groups

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  MULTIMEDIA TOOLS AND APPLICATIONS 75 (24) 17059 - 17079 1380-7501 2016/12 [Refereed][Not invited]
   

  In this paper, we propose a Web video retrieval method that uses hierarchical structure of Web video groups. Existing retrieval systems require users to input suitable queries that identify the desired contents in order to accurately retrieve Web videos; however, the proposed method enables retrieval of the desired Web videos even if users cannot input the suitable queries. Specifically, we first select representative Web videos from a target video dataset by using link relationships between Web videos obtained via metadata "related videos" and heterogeneous video features. Furthermore, by using the representative Web videos, we construct a network whose nodes and edges respectively correspond to Web videos and links between these Web videos. Then Web video groups, i.e., Web video sets with similar topics are hierarchically extracted based on strongly connected components, edge betweenness and modularity. By exhibiting the obtained hierarchical structure of Web video groups, users can easily grasp the overview of many Web videos. Consequently, even if users cannot write suitable queries that identify the desired contents, it becomes feasible to accurately retrieve the desired Web videos by selecting Web video groups according to the hierarchical structure. Experimental results on actual Web videos verify the effectiveness of our method.
• Performance Improvement of Error-Resilient 3D DWT Video Transmission Using Invertible Codes

  Kotoku Omura, Shoichiro Yamasaki, Tomoko K. Matsushima, Hirokazu Tanaka, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E99A (12) 2256 - 2265 1745-1337 2016/12 [Refereed][Not invited]
   

  Many studies have applied the three-dimensional discrete wavelet transform (3D DWT) to video coding. It is known that corruptions of the lowest frequency sub-band (LL) coefficients of 3D DWT severely affect the visual quality of video. Recently, we proposed an error resilient 3D DWT video coding method (the conventional method) that employs dispersive grouping and an error concealment (EC). The EC scheme of our conventional method adopts a replacement technique of the lost LL coefficients. In this paper, we propose a new 3D DWT video transmission method in order to enhance error resilience. The proposed method adopts an error correction scheme using invertible codes to protect LL coefficients. We use half-rate Reed-Solomon (RS) codes as invertible codes. Additionally, to improve performance by using the effect of interleave, we adopt a new configuration scheme at the RS encoding stage. The evaluation by computer simulation compares the performance of the proposed method with that of other EC methods, and indicates the advantage of the proposed method.
• Classifying Insects from SEM Images Based on Optimal Classifier Selection and D-S Evidence Theory

  Takahiro Ogawa, Akihiro Takahashi, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E99A (11) 1971 - 1980 1745-1337 2016/11 [Refereed][Not invited]
   

  In this paper, an insect classification method using scanning electron microphotographs is presented. Images taken by a scanning electron microscope (SEM) have a unique problem for classification in that visual features differ from each other by magnifications. Therefore, direct use of conventional methods results in inaccurate classification results. In order to successfully classify these images, the proposed method generates an optimal training dataset for constructing a classifier for each magnification. Then our method classifies images using the classifiers constructed by the optimal training dataset. In addition, several images are generally taken by an SEM with different magnifications from the same insect. Therefore, more accurate classification can be expected by integrating the results from the same insect based on Dempster-Shafer evidence theory. In this way, accurate insect classification can be realized by our method. At the end of this paper, we show experimental results to confirm the effectiveness of the proposed method.
• Estimation of salient regions related to chronic gastritis using gastric X-ray images

  Ren Togo, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama

  COMPUTERS IN BIOLOGY AND MEDICINE 77 9 - 15 0010-4825 2016/10 [Refereed][Not invited]
   

  Since technical knowledge and a high degree of experience are necessary for diagnosis of chronic gastritis, computer-aided diagnosis (CAD) systems that analyze gastric X-ray images are desirable in the field of medicine. Therefore, a new method that estimates salient regions related to chronic gastritis/non-gastritis for supporting diagnosis is presented in this paper. In order to estimate salient regions related to chronic gastritis/non-gastritis, the proposed method monitors the distance between a target image feature and Support Vector Machine (SVM)-based hyperplane for its classification. Furthermore, our method realizes removal of the influence of regions outside the stomach by using positional relationships between the stomach and other organs. Consequently, since the proposed method successfully estimates salient regions of gastric X-ray images for which chronic gastritis and non-gastritis are unknown, visual support for inexperienced clinicians becomes feasible. (C) 2016 Elsevier Ltd. All rights reserved.
• Category Classification of Tourism Images in Image Sharing Services

  SAITO Naoki, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  電子情報通信学会論文誌D 情報・システム J99-D (9) 848 - 860 1880-4535 2016/09/01 [Refereed]
   

  In this paper, we propose a category classification method of tourism images in image sharing services. First, the proposed method performs classification of tourism images by applying a fuzzy K-nearest neighbor algorithm to the location feature vectors and newly estimates their confidence measures. Second, if the confidence measures of the classification results are lower, the proposed method performs decision level fusion of related visual and textual tag feature vectors to obtain final classification results. Consequently, the proposed method enables successful integration of multiple classification results obtained from the several kinds of features based on the above two stage classification scheme.
• NMF-Based Spectral Reflectance Estimation From Image Pairs Including Near-Infrared Components

  Takahiro Ogawa, Yuta Igarashi, Miki Haseyama

  IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 26 (5) 855 - 867 1051-8215 2016/05 [Refereed][Not invited]
   

  In this paper, a novel spectral reflectance estimation method from image pairs including near-infrared (NIR) components based on nonnegative matrix factorization (NMF) is presented. The proposed method enables estimation of spectral reflectance from only two kinds of input images: 1) an image including both visible light components and NIR components and 2) an image including only NIR components. These two images can be easily obtained using a general digital camera without an infrared-cut filter and one with a visible light-cut filter, respectively. Since RGB values of these images are obtained according to spectral sensitivity of the image sensor, the spectrum power distribution of the light source and the spectral reflectance, we have to solve the inverse problem for estimating the spectral reflectance. Therefore, our method approximates spectral reflectance by a linear combination of several bases obtained by applying NMF to a known spectral reflectance data set. Then estimation of the optimal solution to the above problem becomes feasible based on this approximation. In the proposed method, NMF is used for obtaining the bases used in this approximation from a characteristic that the spectral reflectance is a nonnegative component. Furthermore, the proposed method realizes simple approximation of the spectrum power distribution of the light source with direct and scattered light components. Therefore, estimation of spectral reflectance becomes feasible using the spectrum power distribution of the light source in our method. In the last part of this paper, we show some simulation results to verify the performance of the proposed method. The effectiveness of the proposed method is also shown using the method for several applications that are closely related to spectral reflectance estimation. Although our method is based on a simple scheme, it is the first method that realizes the estimation of the spectral reflectance and the spectrum power distribution of the light source from the above two kinds of images taken by general digital cameras and provides breakthroughs to several fundamental applications.
• A Most Resource-Consuming Disease Estimation Method from Electronic Claim Data Based on Labeled LDA

  Yasutaka Hatakeyama, Takahiro Ogawa, Hironori Ikeda, Miki Haseyama

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E99D (3) 763 - 768 1745-1361 2016/03 [Refereed][Not invited]
   

  In this paper, we propose a method to estimate the most resource-consuming disease from electronic claim data based on Labeled Latent Dirichlet Allocation (Labeled LDA). The proposed method models each electronic claim from its medical procedures as a mixture of resource-consuming diseases. Thus, the most resource-consuming disease can be automatically estimated by applying Labeled LDA to the electronic claim data. Although our method is composed of a simple scheme, this is the first trial for realizing estimation of the most resource-consuming disease.
• Distress Classification of Road Structures via Multiple Classifier-based Bayesian Network

  K. Maeda, S. Takahashi, T. Ogawa, M. Haseyama

  International Workshop on Advanced Image Technology (IWAIT) 1 - 4 2016 [Refereed][Not invited]
  
• 6. Visualization Methods for Encourage Experience of Sports in Soccer Videos

  Takahashi Sho, Haseyama Miki

  The Journal of The Institute of Image Information and Television Engineers 一般社団法人 映像情報メディア学会 70 (9) 722 - 724 1342-6907 2016 [Not refereed][Not invited]
  
• Multimodal Interest Level Estimation via Variational Bayesian Mixture of Robust CCA

  Yuma Sasaka, Takhiro Ogawa, Miki Haseyama

  MM'16: PROCEEDINGS OF THE 2016 ACM MULTIMEDIA CONFERENCE 387 - 391 2016 [Refereed][Not invited]
   

  This paper presents a method which estimates interest level while watching videos, based on collaborative use of facial expression and biological signals such as electroencephalogram (EEG) and electrocardiogram (ECG). To the best of our knowledge, no studies have been carried out on the collaborative use of facial expression and biological signals for estimating interest level. Since training data, which is used for estimating interest level, is generally small and imbalanced, Variational Bayesian Mixture of Robust Canonical Correlation Analysis (VBMRCCA) is newly applied to facial expression and biological signals, which are obtained from users while they are watching the videos. Unlike some related works, VBMRCCA is used to obtain the posterior distributions which represent the latent correlation between facial expression and biological signals in our method. Then, the users' interest level can be estimated by comparing the posterior distributions of the positive class data with those of the negative. Consequently, successful interest level estimation, via collaborative use of facial expression and biological signals, becomes feasible.
• Realization of Associative Image Search Development of Image Retrieval Platform for Enhancing Serendipity

  Miki Haseyama

  2016 IEEE 46TH INTERNATIONAL SYMPOSIUM ON MULTIPLE-VALUED LOGIC (ISMVL 2016) 56 - 59 0195-623X 2016 [Refereed][Not invited]
   

  This paper presents "Associative Image Search", a new image retrieval scheme and its specific engineering application, which enable value creation from big data. The main aim of the associative image search is the realization of information retrieval that enhances the potential for serendipities by providing users with new awareness. Thus, this paper presents the details of research for realizing associative image retrieval. Furthermore, as an example of its applications, a Biomimetics image retrieval platform is also introduced in this paper. By associatively and collaboratively using data accumulated in the fields of biology and material science, the Biomimetics image retrieval platform enables acceleration of their knowledge sharing in different research fields. From retrieval results actually obtained from this platform, there is discussion of the potential of serendipities such as new knowledge emergence
• A Virtual Vital Signs Sensor "MIRUWS" for Visualization of Healthy to Illness Transition (HIT)

  Shigenobu Minami, Miki Haseyama, Hirokazu Tanaka, Toru Takahashi, Tatsuya Komori

  2016 10TH INTERNATIONAL SYMPOSIUM ON MEDICAL INFORMATION AND COMMUNICATION TECHNOLOGY (ISMICT) 1 - 5 2326-828X 2016 [Refereed][Not invited]
   

  This paper proposes a virtual Vital Signs Sensor (VSS) for visualization of half illness (so called "Mibyou") and sudden illness. Since both half and sudden illnesses, which are located in between wellness and illness, are categorized in healthy(H) to illness(I) transition(T) status. So, we name this status as HIT. Since HIT happens in an ordinary life, VSS for HIT visualization should work for 24 hours continuously and unconsciously. The proposed virtual VSS is named as "MIRUWS" which means (M)multi-sensing, (I)integrated, (R)reliable, and (U)unconscious (V)virtual (V)vital (S) signs sensor. All these key words are meaningful and required for professional HIT (Pro-HIT) visualization which has real demand rather than purely personal one. MIRUWS is a virtual VSS in a cyber space targeting this Pro-HIT visualization, and is a projection of actual VSSs in a physical space. There are wide varieties of physical sensors such as patch, touch, proximity, and remote types. To handle these wide varieties of VSSs efficiently and consistently, MIRUWS plays as a common and unique virtual VSS in a cyber space. Unlikely to medical devices which need very high reliability rather than flexibility, MIRUWS is needed to satisfy both reliability and flexibility at the same time to cover wide range of Pro-HIT use-cases. To realize this, MIRUWS visualizes physical VSS's specifications throughout new API, which are dynamically determined by connected and released physical VSSs in a BAN. To certify MIRUWS performance objectively, conformance testing is desired. This paper also presents MIRUWS test environment example, which measures basic vital signs performances, heart rate variation (HRV) and pulse wave transition time (PWTT) using two target physical VSSs at once.
• An Accurate Mortality Prediction Method Based on Decision-level Fusion of Existing ICU Scoring Systems

  Yasutaka Hatakeyama, Takahiro Ogawa, Hirokazu Tanaka, Miki Haseyama

  PROCEEDINGS OF 2016 INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY AND ITS APPLICATIONS (ISITA 2016) 126 - 130 2016 [Refereed][Not invited]
   

  In this paper, we propose a mortality prediction method based on decision-level fusion (DLF) of existing intensive unit care (ICU) scoring systems. First, the proposed method obtains severity scores from the existing ICU scoring systems. Furthermore, we construct classifiers that categorize patients into survivors or non-survivors. Next, patient feature vectors are extracted based on the mortality rates that are estimated from the obtained severity scores by using a non-linear least squares method to obtain other types of classification results. In order to obtain the final severity score for each patient, we integrate the obtained multiple classification results based on DLF that can estimate the final severity scores. Finally, we performed the proposed method to actual ICU patient data and verified the effectiveness of the proposed method. Thus, the proposed method can realize accurate mortality prediction without any additional work by using the existing ICU scoring systems.
• GRAPH-BASED WEB VIDEO SEARCH RERANKING THROUGH CONSISTENCY ANALYSIS USING SPECTRAL CLUSTERING

  Soh Yoshida, Takahiro Ogawa, Miki Haseyaina

  2016 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA & EXPO (ICME) 1 - 6 1945-7871 2016 [Refereed][Not invited]
   

  This paper proposes a graph-based Web video search reranking method through consistency analysis using spectral clustering. Graph-based reranking is effective for refining text-based video search results. Generally, this approach constructs a graph where the vertices are videos and the edges reflect their pairwise similarities. A lot of reranking methods are built based on a scheme which regularizes the smoothness of pairwise ranking scores between adjacent nodes. However, since the overall consistency is measured by aggregating the individual consistency over each pair, errors in score estimation increase when noisy samples are included within their neighbors. To deal with the noisy samples, different from the conventional methods, the proposed method models the global consistency of the graph structure. Specifically, in order to detect this consistency, the propose method introduces a spectral clustering algorithm which can detect video groups, whose videos have strong semantic correlation, on the graph. Furthermore, a new regularization term, which smooths ranking scores within the same group, is introduced to the reranking framework. Since score regularization is performed by both local and global aspect simultaneously, the accurate score estimation becomes feasible. Experimental results obtained by applying the proposed method to a real-world video collection show its effectiveness.
• HIERARCHICAL CONTENT GROUP DETECTION FROM DIFFERENT SOCIAL MEDIA PLATFORMS USING WEB LINK STRUCTURE

  Daichi Takehara, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2016 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 479 - 483 1522-4880 2016 [Refereed][Not invited]
   

  This paper presents a method for hierarchical content group detection from different social media platforms, which can reveal hierarchical structure of content groups. In this paper, content groups are defined as sets of contents with similar topics. Based on the revealed hierarchical structure, our method enables users to efficiently find the desired contents from large amount of contents placed in diversified social media platforms. The main contributions of this paper are twofold. First, effective latent features for comparing the contents placed in different social media platforms can be extracted by the combination use of the correlation between features obtained from different social media platform and the Web link structure. Second, the hierarchical structure of the content groups, which captures their various abstraction levels, can be revealed by hierarchically detecting their content groups. Experimental results on the real-world dataset containing YouTube videos and Wikipedia articles show the effectiveness of our method.
• Distress Classification of Road Structures via Decision Level Fusion

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  2016 IEEE INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING (DSP) 589 - 593 2016 [Refereed][Not invited]
   

  A distress classification method of road structures via decision level fusion is presented in this paper. In order to classify various kinds of distresses accurately, the proposed method integrates multiple classification results with considering their performance, and this is the biggest contribution of this paper. By introducing this approach, it becomes feasible to adaptively integrate the multiple classification results based on the accuracy of each classifier for a target sample. Consequently, realization of the accurate distress classification can be expected. Experimental results show that our method outperforms existing methods.
• Improvement of video coding efficiency based on sparse contractive mapping approach

  Zaixing He, Takahiro Ogawa, Sho Takahashi, Miki Haseyama, Xinyue Zhao

  NEUROCOMPUTING 173 1898 - 1907 0925-2312 2016/01 [Refereed][Not invited]
   

  This paper presents a new method for improving video coding efficiency based on a sparse contractive mapping approach. The proposed method introduces a new sparse contractive mapping approach to replace the traditional intra prediction in the video coding standards such as H.264/AVC. Specifically, the intra- and its following inter-frame are respectively approximated by the sparse representation, satisfying contractive mapping. Then these two frames can be reconstructed from an arbitraryinitial image by utilizing a few representation coefficients. With this advantage, the proposed method reduces the total amount of bits by removing MBs in the target I frame, whose approximation performance is higher than the others in the encoder. Furthermore, by transmitting the representation coefficients of the removed MBs, these MBs can be accurately reconstructed in the decoder. Since the reconstruction performance is better than that of the conventional approach, the proposed method can remove more MBs from the target video sequences, and reduction of total amount of bits can be feasible. Therefore, the proposed method realizes the improvement of the video coding efficiency. Some experimental results are shown to verify the superior performance of the proposed method to that of H.264/AVC. The results also demonstrate that the bit-saving performance of the proposed method is comparable to that of H.2651 HEVC. (C) 2015 Elsevier B.V. All rights reserved.
• NOVEL FAVORITE MUSIC CLASSIFICATION USING EEG-BASED OPTIMAL AUDIO FEATURES SELECTED VIA KDLPCCA

  Ryosuke Sawata, Takahiro Ogawa, Miki Haseyama

  2016 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING PROCEEDINGS 759 - 763 1520-6149 2016 [Refereed][Not invited]
   

  This paper presents a novel method of favorite music classification using EEG-based optimal audio features. To select audio features related to user's music preference, our method utilizes a relationship between EEG features obtained from the user's EEG signals during listening to music and their corresponding audio features since EEG signals of human reflect his/her music preference. Specifically, cross-loadings, whose components denote the degree of the relationship, are calculated based on Kernel Discriminative Locality Preserving Canonical Correlation Analysis (KDLPCCA) which is newly derived in the proposed method. In contrast with standard CCA, KDLPCCA can consider (1) non-linear correlation, (2) class information and (3) local structures of input EEG and audio features, simultaneously. Therefore, KDLPCCA-based cross-loadings can reflect best correlation between the user's EEG and corresponding audio signals. Then an optimal set of audio features related to his/her music preference can be obtained by employing the cross-loadings as novel criteria for feature selection. Consequently, our method realizes favorite music classification successfully by using the EEG-based optimal audio features.
• Bregman pooling: Feature-space local pooling for image classification

  Alameen Najjar, Takahiro Ogawa, Miki Haseyama

  International Journal of Multimedia Information Retrieval 4 (4) 247 - 259 2192-662X 2015/12/01 [Refereed][Not invited]
   

  In this paper, we propose a novel feature-space local pooling method for the commonly adopted architecture of image classification. While existingmethods partition the feature space based on visual appearance to obtain pooling bins, learning more accurate space partitioning that takes semantics into account boosts performance even for a smaller number of bins. To this end, we propose partitioning the feature space over clusters of visual prototypes common to semantically similar images (i.e., images belonging to the same category). The clusters are obtained by Bregman co-clustering applied offline on a subset of training data. Therefore, being aware of the semantic context of the input image, our features have higher discriminative power than do those pooled from appearance-based partitioning. Testing on four datasets (Caltech-101, Caltech-256, 15 Scenes, and 17 Flowers) belonging to three different classification tasks showed that the proposed method outperforms methods in previous works on local pooling in he feature space for less feature dimensionality. Moreover, when implemented within a spatial pyramid, our method achieves comparable results on three of the datasets used.
• Perceptually Optimized Missing Texture Reconstruction via Neighboring Embedding

  Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E98A (8) 1709 - 1717 1745-1337 2015/08 [Refereed][Not invited]
   

  Perceptually optimized missing texture reconstruction via neighboring embedding (NE) is presented in this paper. The proposed method adopts the structural similarity (SSIM) index as a measure for representing texture reconstruction performance of missing areas. This provides a solution to the problem of previously reported methods not being able to perform perceptually optimized reconstruction. Furthermore, in the proposed method, a new scheme for selection of the known nearest neighbor patches for reconstruction of target patches including missing areas is introduced. Specifically, by monitoring the SSIM index observed by the proposed NE-based reconstruction algorithm, selection of known patches optimal for the reconstruction becomes feasible even if target patches include missing pixels. The above novel approaches enable successful reconstruction of missing areas. Experimental results show improvement of the proposed method over previously reported methods.
• 本格化するバイオミメティクス材料の開発〔解説‐2〕バイオミメティクス・画像検索システムの実現と活用

  HASEYAMA MIKI

  工業材料 63 (8) 29 - 33 0452-2834 2015/08/01 [Not refereed][Not invited]
  
• A Biological Classification Method Using Scanning Electron Mircroscope Images via Taxonomy-Based Decision Tree

  Piao Jun, Ogawa Takahiro, Haseyama Miki

  電子情報通信学会論文誌 D, 情報・システム 電子情報通信学会 98 (5) 823 - 834 2015/05/01 [Not refereed][Not invited]
   

  本論文では，走査型電子顕微鏡(Scanning Electron Microscop，SEM)で撮像された画像を用いた生物の分類法を提案する．提案手法では，生物学者が構築した分類体系に注目し，各ノードにその下位のノードへの分類を行う分類器を割り当てることで，決定木を構築する．これにより，構築された決定木を用いて生物の分類が可能となる．このように，生物の分類に有用な分類体系の構造を導入することで，画像特徴のみに注目する生物の分類法と比較して，高精度な分類が期待できる．また，提案手法では，更なる精度向上のため，以下の二つの処理も導入する．まず，分類体系において，画像特徴が類似する異なるノードに注目した決定木の変更を行い，誤分類を抑制する．次に，同種の生物の異なる撮像倍率の画像に対し，モーフィングを施すことで学習データの充足を行う．以上によって，提案手法では，SEMで撮像された生物の高精度な分類が可能となる．本論文の最後では，提案手法の有効性を確認するための実験結果を示す．
• ものづくりの発想を支援する―バイオミメティクス・画像検索基盤―

  HASEYAMA MIKI, HASEYAMA MIKI

  現代化学 (529) 31 - 34 0386-961X 2015/04/01 [Not refereed][Not invited]
  
• Biomimetics Image Retrieval Platform for Enhancing Serendipity

  HASEYAMA MIKI

  タクサ 日本動物分類学会 (38) 22 - 25 1342-2367 2015/02/28 [Not refereed][Not invited]
   

  Biomimetics is a new research area that creates innovation through the collaboration of different existing research fields. Since biomimetics brings together expert researchers with deep knowledge of various research fields, there is a need to facilitate the mutual exchange of that knowledge in order to create new research areas. However, this exchange is difficult due to several reasons, e.g., differences in technical terms between different fields. In order to overcome this problem, we started the development of a new data retrieval platform based on the theory of associative image retrieval. A biological database contains many image data, and by taking advantage of these image data, we are able to overcome limitations of text-only information retrieval. If the development of such a retrieval platform that does not depend on text data can be realized, individual biological databases of various species (insects, fish, etc.) will be integrated. This will allow not only the use for the study of the various species by researchers in different biological fields, but also access for a wide range of researchers in fields ranging from materials science, mechanical engineering and manufacturing.
• 「バイオミメティクス」日本企業の研究開発の現況:「生物・博物・情報」分野 バイオミメティクス・画像検索システム―生物の微細構造から材料開発の発想を支援する―

  長谷山美紀

  OHM 102 (1) 26 - 27 0386-5576 2015/01/12 [Not refereed][Not invited]
  
• Automatic Martian Dust Storm Detection from Multiple Wavelength Data Based on Decision Level Fusion

  Maeda Keisuke, Ogawa Takahiro, Haseyama Miki

  IMT Information and Media Technologies Editorial Board 10 (3) 473 - 477 1881-0896 2015 

  This paper presents automatic Martian dust storm detection from multiple wavelength data based on decision level fusion. In our proposed method, visual features are first extracted from multiple wavelength data, and optimal features are selected for Martian dust storm detection based on the minimal-Redundancy-Maximal-Relevance algorithm. Second, the selected visual features are used to train the Support Vector Machine classifiers that are constructed on each data. Furthermore, as a main contribution of this paper, the proposed method integrates the multiple detection results obtained from heterogeneous data based on decision level fusion, while considering each classifiers detection performance to obtain accurate final detection results. Consequently, the proposed method realizes successful Martian dust storm detection.
• NOVEL IMAGE CLASSIFICATION BASED ON INTEGRATION OF EEG AND VISUAL FEATURES VIA MSLPCCA

  Takuya Kawakami, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP) 957 - 961 1520-6149 2015 

  This paper presents a novel image classification method based on integration of EEG and visual features. In the proposed method, we obtain classification results by separately using EEG and visual features. Furthermore, we merge the above classification results based on a kernelized version of Supervised learning from multiple experts and obtain the final classification result. In order to generate feature vectors used for the final image classification, we apply Multiset supervised locality preserving canonical correlation analysis (MSLPCCA), which is newly derived in the proposed method, to EEG and visual features. Our method realizes successful multimodal classification of images by the object categories that they contain based on MSLPCCA-based feature integration.
• MISSING INTENSITY RESTORATION VIA ADAPTIVE SELECTION OF PERCEPTUALLY OPTIMIZED SUBSPACES

  Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP) 1628 - 1632 1520-6149 2015 

  A missing intensity restoration method via adaptive selection of perceptually optimized subspaces is presented in this paper. In order to realize adaptive and perceptually optimized restoration, the proposed method generates several subspaces of known textures optimized in terms of the structural similarity (SSIM) index. Furthermore, the SSIM-based missing intensity restoration is performed by a projection onto convex sets (POCS) algorithm whose constraints are the obtained subspace and known intensities within the target image. In this approach, a non-convex maximization problem for calculating the projection onto the subspace is reformulated as a quasi-convex problem, and the restoration of the missing intensities becomes feasible. Furthermore, the selection of the optimal subspace is realized by monitoring the SSIM index converged in the POCS algorithm, and the adaptive restoration becomes feasible. Experimental results show that our method outperforms existing methods.
• Heterogeneous Graph-based Video Search Reranking using Web Knowledge via Social Media Network

  Soh Yoshida, Takahiro Ogawa, Miki Haseyama

  MM'15: PROCEEDINGS OF THE 2015 ACM MULTIMEDIA CONFERENCE 871 - 874 2015 [Refereed][Not invited]
   

  Graph-based reranking is effective for refining text-based video search results by making use of the social network structure. Unlike previous works which only focus on an individual video graph, the proposed method leverages the mutual reinforcement of heterogeneous graphs, such as videos and their associated tags obtained by social influence mining. Specifically, propagation of information relevancy across different modalities is performed by exchanging information of inter- and intra-relations among heterogeneous graphs. The proposed method then formulates the video search reranking as an optimization problem from the perspective of Bayesian framework. Furthermore, in order to model the consistency over the modified video graph topology, a local learning regularization with a social community detection scheme is introduced to the framework. Since videos within the same social community have strong semantic correlation, the consistency score estimation becomes feasible. Experimental results obtained by applying the proposed method to a real-world video collection show its effectiveness.
• HELICOBACTER PYLORI INFECTION DETECTION FROM MULTIPLE X-RAY IMAGES BASED ON COMBINATION USE OF SUPPORT VECTOR MACHINE AND MULTIPLE KERNEL LEARNING

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 4728 - 4732 1522-4880 2015 [Refereed][Not invited]
   

  This paper presents a detection method of Helicobacter pylori (H. pylori) infection from multiple gastric X-ray images based on combination use of Support Vector Machine (SVM) and Multiple Kernel Learning (MKL). The proposed method firstly computes some types of visual features from multiple gastric X-ray images taken in several specific directions in order to represent the characteristics of X-ray images with H. pylori infection. Second, based on the minimal-Redundancy-Maximal-Relevance algorithm, we select the effective features for H. pylori infection detection from each type of visual feature and all visual features. The selected features are used to train the SVM classifier and the MKL classifier for each direction of gastric X-ray images. Finally, the proposed method integrates multiple detection results based on a late fusion scheme considering the detection performance of each classifier. Experimental results obtained by applying the proposed method to real X-ray images prove its effectiveness.
• AUTOMATIC DETECTION OF MARTIAN DUST STORMS FROM HETEROGENEOUS DATA BASED ON DECISION LEVEL FUSION

  Keisuke Maeda, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2246 - 2250 1522-4880 2015 [Refereed][Not invited]
   

  This paper presents automatic detection of Martian dust storms from heterogeneous data (raw data, reflectance data and background subtraction data of the reflectance data) based on decision level fusion. Specifically, the proposed method first extracts image features from these data and selects optimal features for dust storm detection based on the minimal-Redundancy-Maximal-Relevance algorithm. Second, the selected image features are used to train the Support Vector Machine classifier that is constructed on each data. Furthermore, as a main contribution of this paper, the proposed method combines the multiple detection results obtained from the heterogeneous data based on decision level fusion with considering each classifier's detection performance to obtain accurate final detection results. Consequently, the proposed method realizes automatic and accurate detection of Martian dust storms.
• EXTRACTION OF HIERARCHICAL STRUCTURE OF WEB COMMUNITIES INCLUDING SALIENT KEYWORD ESTIMATION FOR WEB VIDEO RETRIEVAL

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 1021 - 1025 1522-4880 2015 [Refereed][Not invited]
   

  In this paper, we present a method for extraction of hierarchical structure of Web communities including salient keyword estimation for Web video retrieval. The following two contributions of the proposed method enable retrieval of the desired Web videos even if users cannot input suitable queries that identify the desired contents. First, our method realizes the extraction of hierarchical structure of Web communities, i.e., Web video sets with similar topics by using heterogeneous features of Web videos and link relationships between Web videos obtained via metadata "related videos". Second, we can estimate salient keywords to identify the contents of each obtained Web community at a glance based on text attached to Web videos such as title, the heterogeneous features of Web videos and the link relationships between Web videos. Experimental results on actual Web videos verify that our method can realize accurate retrieval of the desired Web videos via the hierarchical structure of Web communities with their salient keywords.
• Human-centered Favorite Music Estimation: EEG-based Extraction of Audio Features Reflecting Individual Preference

  Ryosuke Sawata, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING (DSP) 818 - 822 2015 [Refereed][Not invited]
   

  This paper presents a human-centered method for favorite music estimation using EEG-based audio features. In order to estimate user's favorite musical pieces, our method utilizes his/her EEG signals for calculating new audio features suitable for representing the user's music preference. Specifically, projection, which transforms original audio features into the features reflecting the preference, is calculated by applying kernel Canonical Correlation Analysis (CCA) to the audio features and the EEG features which are extracted from the user's EEG signals during listening to favorite musical pieces. By using the obtained projection, the new EEG-based audio features can be derived since this projection provides the best correlation between the user's EEG signals and their corresponding audio signals. Thus, successful estimation of user's favorite musical pieces via a Support Vector Machine (SVM) classifier using the new audio features becomes feasible. Since our method does not need acquisition of EEG signals for obtaining new audio features from new musical pieces after calculating the projection, this indicates the high practicability of our method. Experimental results show that our method outperforms methods using original audio features or EEG features.
• Algorithm for Sparse Representation Minimizing Mean Square Error of Power Spectrograms

  Yuma Tanaka, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING (DSP) 618 - 622 2015 [Refereed][Not invited]
   

  Sparse representation is an idea to approximate a target signal by a linear combination of a small number of sample signals, and it is utilized in various research fields. In this paper, we evaluate the approximation error of signals by the mean square error of power spectrograms (P-MSE). Specifically, we propose a P-MSE minimization algorithm for sparse representation. Our method minimizes the P-MSE by an iterative approach. Specifically, in each iteration, we find the optimal sample signal and optimize the corresponding coefficients by a gradient-based method. In this approach, our method can utilize the result of the previous iteration for fast and stable convergence in the optimization of the coefficients. Based on this algorithm, the sparse representation which minimizes the P-MSE becomes feasible. Experimental results show the effectiveness of our method in terms of the P-MSE minimization.
• MISSING INTENSITY RESTORATION VIA ADAPTIVE SELECTION OF PERCEPTUALLY OPTIMIZED SUBSPACES

  Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP) 1628 - 1632 1520-6149 2015 [Refereed][Not invited]
   

  A missing intensity restoration method via adaptive selection of perceptually optimized subspaces is presented in this paper. In order to realize adaptive and perceptually optimized restoration, the proposed method generates several subspaces of known textures optimized in terms of the structural similarity (SSIM) index. Furthermore, the SSIM-based missing intensity restoration is performed by a projection onto convex sets (POCS) algorithm whose constraints are the obtained subspace and known intensities within the target image. In this approach, a non-convex maximization problem for calculating the projection onto the subspace is reformulated as a quasi-convex problem, and the restoration of the missing intensities becomes feasible. Furthermore, the selection of the optimal subspace is realized by monitoring the SSIM index converged in the POCS algorithm, and the adaptive restoration becomes feasible. Experimental results show that our method outperforms existing methods.
• NOVEL IMAGE CLASSIFICATION BASED ON INTEGRATION OF EEG AND VISUAL FEATURES VIA MSLPCCA

  Takuya Kawakami, Takahiro Ogawa, Miki Haseyama

  2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP) 957 - 961 1520-6149 2015 [Refereed][Not invited]
   

  This paper presents a novel image classification method based on integration of EEG and visual features. In the proposed method, we obtain classification results by separately using EEG and visual features. Furthermore, we merge the above classification results based on a kernelized version of Supervised learning from multiple experts and obtain the final classification result. In order to generate feature vectors used for the final image classification, we apply Multiset supervised locality preserving canonical correlation analysis (MSLPCCA), which is newly derived in the proposed method, to EEG and visual features. Our method realizes successful multimodal classification of images by the object categories that they contain based on MSLPCCA-based feature integration.
• Random combination for information extraction in compressed sensing and sparse representation-based pattern recognition

  Zaixing He, Xinyue Zhao, Shuyou Zhang, Takahiro Ogawa, Miki Haseyama

  NEUROCOMPUTING 145 160 - 173 0925-2312 2014/12 [Refereed][Not invited]
   

  In compressed sensing and sparse representation-based pattern recognition, random projection with a dense random transform matrix is widely used for information extraction. However, the complicated structure makes dense random matrices computationally expensive and difficult in hardware implementation. This paper considers the simplification of the random projection method. First, we propose a simple random method, random combination, for information extraction to address the issues of dense random methods. The theoretical analysis and the experimental results show that it can provide comparable performance to those of dense random methods. Second, we analyze another simple random method, random choosing, and give its applicable occasions. The comparative analysis and the experimental results show that it works well in dense cases but worse in sparse cases. Third, we propose a practical method for measuring the effectiveness of the feature transform matrix in sparse representation-based pattern recognition. A matrix satisfying the Representation Residual Restricted Isometry Property can provide good recognition results. (C) 2014 Elsevier B.V. All rights reserved.
• Adaptive missing texture reconstruction method based on kernel cross-modal factor analysis with a new evaluation criterion

  Takahiro Ogawa, Mild Haseyama

  SIGNAL PROCESSING 103 69 - 83 0165-1684 2014/10 [Refereed][Not invited]
   

  This paper presents an adaptive missing texture reconstruction method based on kernel cross-modal factor analysis (KCFA) with a new evaluation criterion. The proposed method estimates the latent relationship between two areas, which correspond to a missing area and its neighboring area, respectively, from known parts within the target image and realizes reconstruction of the missing textures. In order to obtain this relationship, KCFA is applied to each cluster containing similar known textures, and the optimal cluster is used for reconstructing each target missing area. Specifically, a new criterion obtained by monitoring errors caused in the latent space enables selection of the optimal cluster. Then each missing texture is adaptively estimated by the optimal cluster's latent relationship, which enables accurate reconstruction of similar textures. In our method, the above criterion is also used for estimating patch priority, which determines the reconstruction order of missing areas within the target image. Since patches, whose textures are accurately modeled by our KCFA-based method, can be selected by using the new criterion, it becomes feasible to perform successful reconstruction of the missing areas. Experimental results show improvements of our KCFA-based reconstruction method over previously reported methods. (C) 2013 Elsevier B.V. All rights reserved.
• Player Tracking in Far-View Soccer Videos Based on Composite Energy Function

  Kazuya Iwai, Sho Takahashi, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E97D (7) 1885 - 1892 1745-1361 2014/07 [Refereed][Not invited]
   

  In this paper, an accurate player tracking method in far-view soccer videos based on a composite energy function is presented. In far-view soccer videos, player tracking methods that perform processing based only on visual features cannot accurately track players since each player region becomes small, and video coding causes color bleeding between player regions and the soccer field. In order to solve this problem, the proposed method performs player tracking on the basis of the following three elements. First, we utilize visual features based on uniform colors and player shapes. Second, since soccer players play in such a way as to maintain a formation, which is a positional pattern of players, we use this characteristic for player tracking. Third, since the movement direction of each player tends to change smoothly in successive frames of soccer videos, we also focus on this characteristic. Then we adopt three energies: a potential energy based on visual features, an elastic energy based on formations and a movement direction-based energy. Finally, we define a composite energy function that consists of the above three energies and track players by minimizing this energy function. Consequently, the proposed method achieves accurate player tracking in far-view soccer videos.
• A new method for error degree estimation in numerical weather prediction via MKDA-based ordinal regression

  Takahiro Ogawa, Shintaro Takahashi, Sho Takahashi, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2014 115 - 115 1687-6180 2014/07 [Refereed][Not invited]
   

  This paper presents a new method for estimating error degrees in numerical weather prediction via multiple kernel discriminant analysis (MKDA)-based ordinal regression. The proposed method tries to estimate how large prediction errors will occur in each area from known observed data. Therefore, ordinal regression based on KDA is used for estimating the prediction error degrees. Furthermore, the following points are introduced into the proposed approach. Since several meteorological elements are related to each other based on atmospheric movements, the proposed method merges such heterogeneous features in the target and neighboring areas based on a multiple kernel algorithm. This approach is based on the characteristics of actual meteorological data. Then, MKDA-based ordinal regression for estimating the prediction error degree of a target meteorological element in each area becomes feasible. Since the amount of training data obtained from known observed data becomes very large in the training stage of MKDA, the proposed method performs simple sampling of those training data to reduce the number of samples. We effectively use the remaining training data for determining the parameters of MKDA to realize successful estimation of the prediction error degree.
• 歌謡番組における映像の構造に注目したシーン分割手法

  吉田壮, 小川貴弘, 長谷山美紀

  電子情報通信学会論文誌 D(Web) J97-D (7) WEB ONLY 1177-1188  1881-0225 2014/07/01 [Not refereed][Not invited]
  
• バイオミメティクスエンジニアリング 発想を支援するバイオミメティクス・データ検索と技術革新

  長谷山美紀

  化学工学 78 (6) 394 - 397 0375-9253 2014/06/05 [Not refereed][Not invited]
  
• A Cross-Modal Approach for Extracting Semantic Relationships Between Concepts Using Tagged Images

  Marie Katsurai, Takahiro Ogawa, Miki Haseyama

  IEEE TRANSACTIONS ON MULTIMEDIA 16 (4) 1059 - 1074 1520-9210 2014/06 [Refereed][Not invited]
   

  This paper presents a cross-modal approach for extracting semantic relationships between concepts using tagged images. In the proposed method, we first project both text and visual features of the tagged images to a latent space using canonical correlation analysis (CCA). Then, under the probabilistic interpretation of CCA, we calculate a representative distribution of the latent variables for each concept. Based on the representative distributions of the concepts, we derive two types of measures: the semantic relatedness between the concepts and the abstraction level of each concept. Because these measures are derived from a cross-modal scheme that enables the collaborative use of both text and visual features, the semantic relationships can successfully reflect semantic and visual contexts. Experiments conducted on tagged images collected from Flickr show that our measures are more coherent to human cognition than the conventional measures that use either text or visual features, or the WordNet-based measures. In particular, a new measure of semantic relatedness, which satisfies the triangle inequality, obtains the best results among different distance measures in our framework. The applicability of our measures to multimedia-related tasks such as concept clustering, image annotation and tag recommendation is also shown in the experiments.
• Trial Realization of Associative Search for Media Understanding

  HASEYAMA Miki

  Technical report of IEICE. PRMU 一般社団法人電子情報通信学会 113 (493) 73 - 77 0913-5685 2014/03/06 [Not refereed][Not invited]
   

  This paper presents a new associative search system for enhancing serendipity which collaboratively use several unstructured data such as images and videos and extracts their latent similarities. In order to realize such systems, the multimodal signal processing is essential. By using several different kinds of data such as images, videos and users' log data, the multimodal signal processing provides a solution to problems of not being able to improve the performance when using only single modalities. Therefore, in this paper, we introduce video retrieval using the multimodal signal processing and also show trial realization of new associative retrieval systems for effectively providing desired contents.
• 映像の構造に注目したMCMC法に基づくシーン分割法

  SONG Yan, 小川貴弘, 長谷山美紀

  電子情報通信学会論文誌 D J97-D (3) 560 - 573 1880-4535 2014/03 [Not refereed][Not invited]
  
• A Scene Segmentation Approach Based on the MCMC Method Using Video Structures

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edition) 一般社団法人電子情報通信学会 97 (3) 560 - 573 1880-4535 2014/03/01 [Not refereed][Not invited]
   

  以前,我々は,映像の類似したショットが隣接せずに出現する構造(映像の構造)に注目したシーン分割手法を提案した.しかしながら,映像の構造に注目したシーン分割手法の共通の問題として,映像の構造を取得する際に用いられるシーンの最長時間幅の設定が困難であるという点が存在した.そこで,本論文では,その改良手法として映像の構造に注目したMCMC法に基づくシーン分割手法を提案する.提案手法では,シーンの最長時間幅を変化させて映像の構造に基づき取得された全ての境界をシーン境界候補とし,それらの中からMCMC法に基づいて最適なシーン境界を推定する.これにより,従来の手法を適用する際に設定が困難であったシーンの最長時間幅を一意に決定することなく,MCMC法により最適なシーン境界を求めることが可能となる.本論文の最後では,実際にテレビで放送された映像に対して提案手法を適用し,その有効性を確認する.
• MFCC extraction in AAC domain for audio content analysis

  Ai Haojun, Miki Haseyama, Wang Kang

  INFORMATION SCIENCE AND MANAGEMENT ENGINEERING, VOLS 1-3 46 1413 - 1420 1746-4463 2014 

  We focus the attention on the extraction of Mel-frequency cepstral coefficients (MFCC) features from MDCT spectrum in AAC domain for audio content analysis. In particular, a MFCC extraction method is proposed, which is adaptive to the window switch in AAC encoding process, and independent of the audio sampling frequency. We discuss the fusion method of MFCC features from different window type in order to keep the balance of the frequency and temporal resolution. The audio scene segmentation and audio classification experimental results show that such approach based on compression domain can approach the performance of the system based on PCM audio, and the CPU overload decreased dramatically. It is meaningful to the real-time analysis of audio content.
• HELICOBACTER PYLORI INFECTION DETECTION FROM MULTIPLE X-RAY IMAGES BASED ON DECISION LEVEL FUSION

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama

  2014 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 2769 - 2773 1522-4880 2014 [Refereed][Not invited]
   

  This paper presents an automatic detection method of Helicobacter pylori (H. pylori) infection from multiple gastric X-ray images. As the biggest contribution of this paper, we combine multiple detection results based on a decision level fusion. In order to obtain multiple detection results, the proposed method first focuses on characteristics of gastric X-ray images with H. pylori infection and computes several visual features from multiple X-ray images taken in several specific directions. Second, we select effective features for H. pylori infection detection from all features based on the minimal-Redundancy-Maximal-Relevance algorithm, and the selected features are used to train the Support Vector Machine (SVM) classifiers that are constructed for each direction of gastric radiography. Therefore, the detection of H. pylori infection becomes feasible, and we can obtain multiple detection results from the SVM classifiers. Furthermore, we combine multiple detection results based on the decision level fusion scheme considering the detection performance of each SVM classifier. Experimental results obtained by applying the proposed method to real X-ray images prove the effectiveness of the proposed method.
• 2D SEMI-SUPERVISED CCA-BASED INPAINTING INCLUDING NEW PRIORITY ESTIMATION

  Takahiro Ogawa, Miki Haseyama

  2014 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 1837 - 1841 1522-4880 2014 [Refereed][Not invited]
   

  This paper presents an inpainting method based on 2D semi-supervised canonical correlation analysis (2D semi-CCA) including new priority estimation. The proposed method estimates relationship, i.e., the optimal correlation, between missing area and its neighboring area from known parts within the target image by using 2D CCA. In this approach, we newly introduce a semi-supervised scheme into the 2D CCA for deriving the 2D semi-CCA which corresponds to a hybrid version of 2D CCA and 2D principle component analysis (2D PCA). This enables successful relationship estimation even if sufficient number of training pairs cannot be provided. Then, by using the obtained relationship, accurate estimation of the missing intensities can be realized. Furthermore, in the proposed method, errors caused in the new variate space obtained by the 2D semi-CCA are effectively used for deriving patch priority determining inpainting order of missing areas. Experimental results show our inpainting method can outperform previously reported methods.
• NOVEL IMAGE CLASSIFICATION BASED ON DECISION-LEVEL FUSION OF EEG AND VISUAL FEATURES

  Takuya Kawakami, Takahiro Ogawa, Miki Haseyama

  2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 5874 - 5878 1520-6149 2014 [Refereed][Not invited]
   

  This paper presents a novel image classification based on decision-level fusion of EEG and visual features. In the proposed method, we extract the EEG features from EEG signals recorded while users stare at images, and the visual features are computed from these images. Then the classification of images is performed based on Support Vector Machine (SVM) by separately using the EEG and visual features. Furthermore, we merge the above classification results based on Supervised Learning from Multiple Experts to obtain the final classification result. This method focuses on the classification accuracy calculated from each classification result. Therefore, although classification accuracy based on EEG and visual features are different from each other, our method realizes effective integration of these classification results. In addition, we newly derive a kernelized version of the method in order to realize more accurate integration of the classification results. Consequently, our method realizes successful multimodal classification of images by the object categories that they contain.
• MISSING INTENSITY RESTORATION VIA PERCEPTUALLY OPTIMIZED SUBSPACE PROJECTION BASED ON ENTROPY COMPONENT ANALYSIS

  Takahiro Ogawa, Miki Haseyama

  2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 175 - 179 1520-6149 2014 [Refereed][Not invited]
   

  A missing intensity restoration method via perceptually optimized subspace projection based on entropy component analysis (ECA) is presented in this paper. The proposed method calculates the optimal subspace of known patches within a target image based on structural similarity (SSIM) index, and the optimal bases are determined based on ECA. Then missing intensity estimation whose results maximize the SSIM index is realized by using a projection onto convex sets (POCS) algorithm whose constraints are the obtained subspace and known intensities within the target image. In this approach, a non-convex maximization problem for calculating the projection onto the subspace is reformulated as a quasi-convex problem, and the restoration of the missing intensities becomes feasible. Experimental results show that our restoration method outperforms previously reported methods.
• Image inpainting based on sparse representations with a perceptual metric

  Takahiro Ogawa, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2013 179 - 179 1687-6180 2013/12 [Refereed][Not invited]
   

  This paper presents an image inpainting method based on sparse representations optimized with respect to a perceptual metric. In the proposed method, the structural similarity (SSIM) index is utilized as a criterion to optimize the representation performance of image data. Specifically, the proposed method enables the formulation of two important procedures in the sparse representation problem, 'estimation of sparse representation coefficients' and 'update of the dictionary', based on the SSIM index. Then, using the generated dictionary, approximation of target patches including missing areas via the SSIM-based sparse representation becomes feasible. Consequently, image inpainting for which procedures are totally derived from the SSIM index is realized. Experimental results show that the proposed method enables successful inpainting of missing areas.
• Exploring and visualizing tag relationships in photo sharing websites based on distributional representations

  Marie Katsurai, Miki Haseyama

  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings 3617 - 3621 1520-6149 2013/10/18 [Refereed][Not invited]
   

  This paper presents a method for exploring and visualizing tag relationships in photo sharing websites based on distributional representations of tags. First, we find a representative distribution of a tag, which is summarized by the mean and covariance, using features of tagged photos. This distributional representation can jointly consider the semantic meaning of tags and their abstraction levels. Then, based on the representative distributions, we derive two kinds of semantic measures on tag relationships. The extracted information is visualized in a graphical network to facilitate the understanding of tag usage. Experiments conducted using tagged photos collected from Flickr show that our tag network is more coherent to human cognition than other networks constructed by conventional methods. © 2013 IEEE.
• Vocal segment estimation in music pieces based on collaborative use of EEG and audio features

  Takuya Kawakami, Takahiro Ogawa, Miki Haseyama

  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings 1197 - 1201 1520-6149 2013/10/18 [Refereed][Not invited]
   

  This paper presents a novel estimation method of segments including vocals in music pieces based on collaborative use of features extracted from electroencephalogram (EEG) signals recorded while users are listening to music pieces and features extracted from these audio signals. From extracted EEG features and audio features, we estimate segments including vocals based on Support Vector Machine (SVM) by separately utilizing these two features. Furthermore, the final classification results are obtained by integrating these estimation results based on supervised learning from multiple experts. Therefore, our method realizes multimodal estimation of segments including vocals in music pieces. Experimental results show the improvement of our method over the methods utilizing only EEG or audio features. © 2013 IEEE.
• 胃X線検査における胃背景粘膜の自動解析による胃がんリスク評価の検討

  間部克裕, 長谷山美紀, 小川貴弘, 吉澤和哉, 大泉晴史, 中島滋美, 加藤元嗣

  日本消化器がん検診学会雑誌 51 (3) 96  1880-7666 2013/05/15 [Not refereed][Not invited]
  
• Player Tracking by Using Level-Set Method in Soccer Video

  TAKAHASHI Sho, LIM Wonkuk, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 96 (3) 695 - 703 1880-4535 2013/03 [Refereed][Not invited]
   

  本論文では,サッカー映像からレベルセット法を用いて選手を追跡する手法を提案する.提案手法では,サッカー映像を各フレームが時間軸方向に重なるように連結した三次元データとして扱う.このデータに対して,レベルセット法を適用することで抽出される三次元の領域は,複数フレームに渡って存在する同一選手を包含する.提案手法では,この三次元の領域をサッカー映像から抽出することで,選手の追跡を実現する.したがって,提案手法では,フレームごとに選手を検出する必要がないため,フレームを個別に処理する従来手法における選手の検出と追跡それぞれの誤差によって精度が低下する問題を解決可能である.また,我々は,ユニフォームの色成分をサッカー映像から色コリログラムを用いて推定し,これをレベルセット法を用いて追跡する選手の特徴として導入する.これにより,提案手法では,追跡対象の特徴を事前に与えることなく,選手の頑健な追跡が可能となる.本文の最後では,実際にテレビで放送されたサッカー映像に対する実験により,提案手法の有効性を確認する.
• Missing Texture Reconstruction Method Based on Error Reduction Algorithm Using Fourier Transform Magnitude Estimation Scheme

  Takahiro Ogawa, Miki Haseyama

  IEEE TRANSACTIONS ON IMAGE PROCESSING 22 (3) 1252 - 1257 1057-7149 2013/03 [Refereed][Not invited]
   

  A missing texture reconstruction method based on an error reduction (ER) algorithm, including a novel estimation scheme of Fourier transform magnitudes is presented in this brief. In our method, Fourier transform magnitude is estimated for a target patch including missing areas, and the missing intensities are estimated by retrieving its phase based on the ER algorithm. Specifically, by monitoring errors converged in the ER algorithm, known patches whose Fourier transform magnitudes are similar to that of the target patch are selected from the target image. In the second approach, the Fourier transform magnitude of the target patch is estimated from those of the selected known patches and their corresponding errors. Consequently, by using the ER algorithm, we can estimate both the Fourier transform magnitudes and phases to reconstruct the missing areas.
• Super-resolution for simultaneous realization of resolution enhancement and motion blur removal based on adaptive prior settings

  Takahiro Ogawa, Daisuke Izumi, Akane Yoshizaki, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2013 1 - 17 1687-6180 2013/02 [Refereed][Not invited]
   

  A super-resolution method for simultaneously realizing resolution enhancement and motion blur removal based on adaptive prior settings are presented in this article. In order to obtain high-resolution (HR) video sequences from motion-blurred low-resolution video sequences, both of the resolution enhancement and the motion blur removal have to be performed. However, if one is performed after the other, errors in the first process may cause performance deterioration of the subsequent process. Therefore, in the proposed method, a new problem, which simultaneously performs the resolution enhancement and the motion blur removal, is derived. Specifically, a maximum a posterior estimation problem which estimates original HR frames with motion blur kernels is introduced into our method. Furthermore, in order to obtain the posterior probability based on Bayes' rule, a prior probability of the original HR frame, whose distribution can adaptively be set for each area, is newly defined. By adaptively setting the distribution of the prior probability, preservation of the sharpness in edge regions and suppression of the ringing artifacts in smooth regions are realized. Consequently, based on these novel approaches, the proposed method can perform successful reconstruction of the HR frames. Experimental results show impressive improvements of the proposed method over previously reported methods.
• Randomized-Controlled Study on the Effects of Hand Bathing on Stroke Patients in Convalescence - Subjective Improvements in Hand Movement and Increased Positiveness in Talking

  Rika Yano, Miki Haseyama

  STROKE 44 (2) 0039-2499 2013/02
• Active grid-based method for visualizing pass regions in soccer videos

  Sho Takahashi, Miki Haseyama

  Electronic Proceedings of the 2013 IEEE International Conference on Multimedia and Expo Workshops, ICMEW 2013 1 - 6 2330-7927 2013 [Refereed][Not invited]
   

  This paper presents a method for visualizing pass regions that have high probabilities of the pass succeeding from broadcast soccer videos. In soccer matches, players discover pass regions based on player position geometry and player velocities. Therefore, by using player position geometry and player velocities, which are obtained from a broadcast soccer video, we can visualize pass regions. The proposed method is realized by the following two steps. 1) Generation of new three-dimensional data (volume data) for analyzing pass regions, which are not visible. 2) Visualization of pass regions. In the first step, volume data are generated from player position geometry and player velocities. By generating the volume data, which express the player position geometry and the player velocities, analysis of invisible pass regions is enabled. In the second step, by applying Active grid to the generated volume data, pass regions are visualized. Specifically, lattice points of the Active grid converge to the pass regions. Therefore, positions of the pass regions on the pitch can be visualized from densities of the lattice points. In the experiment, the proposed method is applied to actual TV programs to verify its effectiveness. © 2013 IEEE.
• An extraction method of hierarchical Web communities for Web video retrieval

  Ryosuke Harakawa, Yasutaka Hatakeyama, Takahiro Ogawa, Miki Haseyama

  2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings 4397 - 4401 1522-4880 2013 [Refereed][Not invited]
   

  This paper presents an extraction method of hierarchical Web communities for Web video retrieval. In the proposed method, Web communities containing Web videos whose topics are similar to each other are extracted by using hyperlinks between Web pages including Web videos and their video features. Furthermore, we focus on graph structure of hyperlinks between Web pages including Web videos which belong to the Web communities. Then, by using strongly connected components and betweenness centrality of the graph, hierarchical structure of the Web communities can be estimated. Consequently, users can easily find Web videos including related topics in each hierarchy, and desired Web videos can be effectively retrieved. © 2013 IEEE.
• Insect classification using Scanning Electron Microphotographs considering magnifications

  Akihiro Takahashi, Takahiro Ogawa, Miki Haseyama

  2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings 3269 - 3273 1522-4880 2013 [Refereed][Not invited]
   

  This paper presents a method of insect classification using images taken by Scanning Electron Microscope (SEM) considering magnifications. Generally, when images of the same insects are taken by SEM with different magnifications, visual features of these images are different from each other. Thus, the proposed method adopts a new scheme which groups images of different magnifications in such a way that the classification performance becomes the highest. Then a classifier is constructed for each group, and the insect classification becomes feasible based on a target image magnification. In addition, by integrating the classification results of several images obtained from the same sample, i.e., the same insect, performance improvement of the insect classification considering magnifications can be realized. Experimental results show the effectiveness of the proposed method. © 2013 IEEE.
• SPECTRAL REFLECTANCE ESTIMATION FROM VISIBLE LIGHT COMPONENTS AND NEAR-INFRARED COMPONENTS

  Yuta Igarashi, Takahiro Ogawa, Miki Haseyama

  2013 20TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP 2013) 2388 - 2392 1522-4880 2013 [Refereed][Not invited]
   

  This paper presents a novel method for estimating a spectral reflectance from two kinds of input images: an image including both visible light components and near-infrared (NIR) components, and an image including only NIR components. From these input images, we estimate the spectral reflectance based on the Non-negative Matrix Factorization algorithm using spectral sensitivities of a digital camera. The estimated spectral reflectance enables several important applications. In this paper, the e ff ectiveness of the proposed method is verified by using the estimated spectral reflectance in the two image processing applications.
• KCFA-BASED MISSING AREA RESTORATION INCLUDING NEW PRIORITY ESTIMATION

  Takahiro Ogawa, Miki Haseyama

  2013 20TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP 2013) 704 - 708 1522-4880 2013 [Refereed][Not invited]
   

  A kernel cross-modal factor analysis (KCFA) based missing area restoration method including a new priority estimation scheme is presented in this paper. The proposed method estimates latent relationship between missing areas and their neighboring areas by deriving projection matrices minimizing their errors in the latent space based on KCFA. This latent relationship represented by the derived projection matrices is optimal for accurately restoring missing areas within the target image. Furthermore, the proposed method adopts a new priority estimation scheme which determines the restoration order of missing areas. Specifically, this priority is estimated based on the criterion representing the restoration performance derived from KCFA, and it enables adaptive selection of missing areas successfully restored by our method. Consequently, it becomes feasible to accurately perform the restoration of missing areas by using the proposed KCFA-based method. Experimental results show subjective and quantitative improvements of the proposed method over previously reported restoration methods.
• Trial Realization of Human-Centered Multimedia Navigation for Video Retrieval

  Miki Haseyama, Takahiro Ogawa

  INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION 29 (2) 96 - 109 1044-7318 2013/01 [Refereed][Not invited]
   

  A trial realization of human-centered navigation for video retrieval is presented in this article. This system consists of the following functions: (a) multimodal analysis for collaborative use of multimedia data, (b) preference extraction for the system to adapt to users' individual demands, and (c) adaptive visualization for users to be guided to their desired contents. By using these functions, users can find their desired video contents more quickly and accurately than with the conventional retrieval schemes since our system can provide new pathways to the desired contents. Experimental results verify the effectiveness of the proposed system.
• Glare Detection for Night Wet Road Surfaces and Driver Visibility Improvement by Using Multiple Onboard Cameras

  公文宏明, 長谷山美紀

  映像情報メディア学会誌(Web) 67 (3) 1881-6908 2013
• Performance of Spatial and Temporal Error Concealment Method for 3D DWT Video Coding in Packet Loss Channel

  Hirokazu Tanaka, Sunmi Kim, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E95A (11) 2015 - 2022 0916-8508 2012/11 [Refereed][Not invited]
   

  A new spatial and temporal error concealment method for three-dimensional discrete wavelet transform (3D DWT) video coding is analyzed. 3D DWT video coding employing dispersive grouping (DG) and two-step error concealment is an efficient method in a packet loss channel [20], [21]. In the two-step error concealment method, the interpolations are only spatially applied however, higher efficiency of the interpolation can be expected by utilizing spatial and temporal similarities. In this paper, we propose an enhanced spatial and temporal error concealment method in order to achieve higher error concealment (EC) performance in packet loss networks. In the temporal error concealment method, structural similarity (SSIM) index is employed for inter group of pictures (GOP) EC and minimum mean square error (MMSE) is used for intra GOP EC. Experimental results show that the proposed method can obtain remarkable performance compared with the conventional methods.
• Super-Resolution Reconstruction for Spatio-Temporal Resolution Enhancement of Video Sequences

  Miki Haseyama, Daisuke Izumi, Makoto Takizawa

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E95D (9) 2355 - 2358 0916-8532 2012/09 [Refereed][Not invited]
   

  A method for spatio-temporal resolution enhancement of video sequences based on super-resolution reconstruction is proposed. A new observation model is defined for accurate resolution enhancement, which enables subpixel motion in intermediate frames to be obtained. A modified optimization formula for obtaining a high-resolution sequence is also adopted.
• Transmission and Vibration in Circuit Theory (7) Band-Pass Filters Having Impedance Transformation Ratio

  永井 信夫, 任 捷, 長谷山 美紀

  信号処理 [信号処理学会] 16 (5) 359 - 368 1342-6230 2012/09 [Not refereed][Not invited]
  
• Transmission and Vibration in Circuit Theory (6) Reconsideration of Left-Handed Circuit Viewing form Circuit Theory

  永井 信夫, 任 捷, 長谷山 美紀

  信号処理 [信号処理学会] 16 (4) 263 - 272 1342-6230 2012/07 [Not refereed][Not invited]
  
• A Novel Framework for Extracting Visual Feature-Based Keyword Relationships from an Image Database

  Marie Katsurai, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E95A (5) 927 - 937 1745-1337 2012/05 [Refereed][Not invited]
   

  In this paper, a novel framework for extracting visual feature-based keyword relationships from an image database is proposed. From the characteristic that a set of relevant keywords tends to have common visual features, the keyword relationships in a target image database are extracted by using the following two steps. First, the relationship between each keyword and its corresponding visual features is modeled by using a classifier. This step enables detection of visual features related to each keyword. In the second step, the keyword relationships are extracted from the obtained results. Specifically, in order to measure the relevance between two keywords, the proposed method removes visual features related to one keyword from training images and monitors the performance of the classifier obtained for the other keyword. This measurement is the biggest difference from other conventional methods that focus on only keyword co-occurrences or visual similarities. Results of experiments conducted using an image database showed the effectiveness of the proposed method.
• Transmission and Vibration in Circuit Theory (5) Generalization of Difference Equation and Left-handed Circuit

  永井 信夫, 任 捷, 長谷山 美紀

  信号処理 [信号処理学会] 16 (3) 187 - 196 1342-6230 2012/05 [Not refereed][Not invited]
  
• Transmission and Vibration in Circuit Theory (4) Characteristic Vibration and Resonance Created at Image Phase π/2

  永井 信夫, 任 捷, 長谷山 美紀

  信号処理 [信号処理学会] 16 (2) 101 - 110 1342-6230 2012/03 [Not refereed][Not invited]
  
• Haseyama Laboratory at Hokkaido University

  長谷山 美紀

  信号処理 [信号処理学会] 16 (2) 121 - 126 1342-6230 2012/03 [Not refereed][Not invited]
  
• An SVDD-Based Method for Taxonomy Estimation of Benthic Species from Microscopic Images

  HASEGAWA TAKASHI, OGAWA TAKAHIRO, WATANABE HIDEMI, HASEYAMA MIKI

  電子情報通信学会技術研究報告 111 (442(IE2011 105-132)) 73 - 78 0913-5685 2012/02 [Not refereed][Not invited]
  
• A cross-modal approach for extracting semantic relationships of concepts from an image database

  Marie Katsurai, Takahiro Ogawa, Miki Haseyama

  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings 2373 - 2376 1520-6149 2012 [Refereed][Not invited]
   

  This paper presents a cross-modal approach for extracting semantic relationships of concepts from an image database. First, canonical correlation analysis (CCA) is used to capture the cross-modal correlations between visual features and tag features in the database. Then, in order to measure inter-concept relationships and estimate semantic levels, the proposed method focuses on the distributions of images under the probabilistic interpretation of CCA. Results of experiments conducted by using an image database showed the improvement of the proposed method over existing methods. © 2012 IEEE.
• PERCEPTUALLY OPTIMIZED SUBSPACE ESTIMATION FOR MISSING TEXTURE RECONSTRUCTION

  Takahiro Ogawa, Miki Haseyama

  2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) 1141 - 1144 1520-6149 2012 [Refereed][Not invited]
   

  This paper presents a perceptually optimized subspace estimation method for missing texture reconstruction. The proposed method calculates the optimal subspace of known patches within a target image based on structural similarity (SSIM) index instead of calculating mean square error (MSE)-based eigenspace. Furthermore, from the obtained subspace, missing texture reconstruction whose results maximize the SSIM index is performed. In this approach, the non-convex maximization problem is reformulated as a quasi convex problem, and the reconstruction of the missing textures becomes feasible. Experimental results show that our method overcomes previously reported MSE-based reconstruction methods.
• Transmission and Vibration in Circuit Theory (3) Characteristic Vibration Related to Combinational Oscillation and Its Expansion

  永井 信夫, 任 捷, 長谷山 美紀

  信号処理 [信号処理学会] 16 (1) 13 - 22 1342-6230 2012/01 [Not refereed][Not invited]
  
• 8-3 A note on improvement of 3D pass region estimation method using player velocity in soccer videos

  TAKAHASHI Sho, HASEYAMA Miki

  PROCEEDINGS OF THE ITE WINTER ANNUAL CONVENTION 一般社団法人 映像情報メディア学会 2012 8 - 3-1 1343-4357 2012 

  This paper realizes an improvement of 3D pass region estimation method by using player velocity in soccer videos. In the previous method, since the pass region was estimated regardless of player velocity, the accuracy was limited. Therefore, by introducing the player velocity to the pass region estimation, we improve the performance of the previous method.
• Support Vector Data Dscription-Based Method for Finding New Benthic Species and Estimating Their Taxonomy Position from Microscopic Images.

  長谷川尭史, 小川貴弘, 渡邉日出海, 長谷山美紀

  映像情報メディア学会誌(Web) 66 (7) 1881-6908 2012
• Missing Image Data Reconstruction Based on Adaptive Inverse Projection via Sparse Representation

  Takahiro Ogawa, Miki Haseyama

  IEEE TRANSACTIONS ON MULTIMEDIA 13 (5) 974 - 992 1520-9210 2011/10 [Refereed][Not invited]
   

  In this paper, a missing image data reconstruction method based on an adaptive inverse projection via sparse representation is proposed. The proposed method utilizes sparse representation for obtaining low-dimensional subspaces that approximate target textures including missing areas. Then, by using the obtained low-dimensional subspaces, inverse projection for reconstructing missing areas can be derived to solve the problem of not being able to directly estimate missing intensities. Furthermore, in this approach, the proposed method monitors errors caused by the derived inverse projection, and the low-dimensional subspaces optimal for target textures are adaptively selected. Therefore, we can apply adaptive inverse projection via sparse representation to target missing textures, i.e., their adaptive reconstruction becomes feasible. The proposed method also introduces some schemes for color processing into the calculation of subspaces on the basis of sparse representation and attempts to avoid spurious color caused in the reconstruction results. Consequently, successful reconstruction of missing areas by the proposed method can be expected. Experimental results show impressive improvement of our reconstruction method over previously reported reconstruction methods.
• Cross Low-Dimension Pursuit for Sparse Signal Recovery from Incomplete Measurements Based on Permuted Block Diagonal Matrix

  Zaixing He, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E94A (9) 1793 - 1803 1745-1337 2011/09 [Refereed][Not invited]
   

  In this paper, a novel algorithm, Cross Low-dimension Pursuit, based on a new structured sparse matrix, Permuted Block Diagonal (PBD) matrix, is proposed in order to recover sparse signals from incomplete linear measurements. The main idea of the proposed method is using the PBD matrix to convert a high-dimension sparse recovery problem into two (or more) groups of highly low-dimension problems and crossly recover the entries of the original signal from them in an iterative way. By sampling a sufficiently sparse signal with a PBD matrix, the proposed algorithm can recover it efficiently. It has the following advantages over conventional algorithms: (1) low complexity, i.e., the algorithm has linear complexity, which is much lower than that of existing algorithms including greedy algorithms such as Orthogonal Matching Pursuit and (2) high recovery ability, i.e., the proposed algorithm can recover much less sparse signals than even l(1)-norm minimization algorithms. Moreover, we demonstrate both theoretically and empirically that the proposed algorithm can reliably recover a sparse signal from highly incomplete measurements.
• Transmission and vibration in circuit theory (2) Resonance and one-dimensional vibration for two atoms crystal

  永井 信夫, 任 捷, 長谷山 美紀

  Journal of signal processing 〔信号処理学会〕 15 (5) 331 - 340 1342-6230 2011/09 [Not refereed][Not invited]
  
• Adaptive Single Image Superresolution Approach Using Support Vector Data Description

  Takahiro Ogawa, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2011 1687-6180 2011 [Refereed][Not invited]
   

  An adaptive single image superresolution (SR) method using a support vector data description (SVDD) is presented. The proposed method represents the prior on high-resolution (HR) images by hyperspheres of the SVDD obtained from training examples and reconstructs HR images from low-resolution (LR) observations based on the following schemes. First, in order to perform accurate reconstruction of HR images containing various kinds of objects, training HR examples are previously clustered based on the distance from a center of a hypersphere obtained for each cluster. Furthermore, missing high-frequency components of the target image are estimated in order that the reconstructed HR image minimizes the above distances. In this approach, the minimized distance obtained for each cluster is utilized as a criterion to select the optimal hypersphere for estimating the high-frequency components. This approach provides a solution to the problem of conventional methods not being able to perform adaptive estimation of the high-frequency components. In addition, local patches in the target low-resolution (LR) image are utilized as the training HR examples from the characteristic of self-similarities between different resolution levels in general images, and our method can perform the SR without utilizing any other HR images.
• ADAPTIVE KPCA-BASED MISSING TEXTURE RECONSTRUCTION APPROACH INCLUDING CLASSIFICATION SCHEME VIA DIFFERENCE SUBSPACES

  Takahiro Ogawa, Miki Haseyama

  2011 18TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP) 1133 - 1136 1522-4880 2011 [Refereed][Not invited]
   

  This paper presents an adaptive kernel principal component analysis (KPCA) based missing texture reconstruction approach including a classification scheme via difference subspaces. The proposed method utilizes a KPCA-based nonlinear eigenspace, which is obtained from each kind of known texture within a target image, as a constraint for reconstructing missing textures with a constraint of known neighboring areas. Then since these two constraints are convex, we can estimate missing textures based on a projection onto convex sets (POCS) algorithm. Furthermore, in this approach, the proposed method derives a new criterion for selecting the optimal eigenspace by monitoring errors caused in the projection via a difference subspace of each kind of known texture. This provides a solution to conventional problems of not being able to perform accurate texture classification, and the adaptive reconstruction of missing textures can be realized by the proposed method. Experimental results show subjective and quantitative improvement of the proposed method over previously reported reconstruction methods.
• LINEAR TIME DECODING OF REAL-FIELD CODES OVER HIGH ERROR RATE CHANNELS

  Zaixing He, Takahiro Ogawa, Miki Haseyama

  2011 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING 3172 - 3175 1520-6149 2011 [Refereed][Not invited]
   

  This paper proposes a novel algorithm for decoding real-field codes over erroneous channels, where the encoded message is corrupted by sparse errors, i.e., impulsive noise. The main problem of decoding such a corrupted encoded message is to reconstruct the error vector; recently, a common way to reconstruct it is to find the sparsest solution to an underdetermined system that is constructed using a parity-check matrix. Unlike the conventional approaches reconstructing the high-dimensional error vector directly, the proposed method crossly recovers the elements of error vector from two (or several) groups of low-dimensional equations. Compared with the traditional algorithms, the proposed method can decode an encoded message with a much higher corruption rate. Furthermore, the complexity of our method is linear, which is much lower than those of the traditional methods. The experimental results verified the high error correction ability and speed of the proposed method.
• ADAPTIVE RECONSTRUCTION METHOD OF MISSING TEXTURES BASED ON PERCEPTUALLY OPTIMIZED ALGORITHM

  Takahiro Ogawa, Miki Haseyama

  2011 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING 1157 - 1160 1520-6149 2011 [Refereed][Not invited]
   

  This paper presents an adaptive reconstruction method of missing textures based on structural similarity (SSIM) index. The proposed method firstly performs SSIM-based selection of the optimal known local textures to adaptively obtain subspaces for reconstructing missing textures. Furthermore, from the selected known textures, the missing texture reconstruction maximizing the SSIM index is performed. In this approach, the non-convex maximization problem is reformulated as a quasi convex problem, and the adaptive reconstruction of the missing textures becomes feasible. Experimental results show impressive improvement of the proposed method over previously reported reconstruction methods.
• Adaptive example-based super-resolution using kernel PCA with a novel classification approach

  Takahiro Ogawa, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2011 1 - 29 1687-6180 2011 [Refereed][Not invited]
   

  An adaptive example-based super-resolution (SR) using kernel principal component analysis (PCA) with a novel classification approach is presented in this paper. In order to enable estimation of missing high-frequency components for each kind of texture in target low-resolution (LR) images, the proposed method performs clustering of high-resolution (HR) patches clipped from training HR images in advance. Based on two nonlinear eigenspaces, respectively, generated from HR patches and their corresponding low-frequency components in each cluster, an inverse map, which can estimate missing high-frequency components from only the known low-frequency components, is derived. Furthermore, by monitoring errors caused in the above estimation process, the proposed method enables adaptive selection of the optimal cluster for each target local patch, and this corresponds to the novel classification approach in our method. Then, by combining the above two approaches, the proposed method can adaptively estimate the missing high-frequency components, and successful reconstruction of the HR image is realized.
• Music recommendation according to human motion based on kernel CCA-based relationship

  Hiroyuki Ohkushi, Takahiro Ogawa, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2011 121 - 121 1687-6180 2011 [Refereed][Not invited]
   

  In this article, a method for recommendation of music pieces according to human motions based on their kernel canonical correlation analysis (CCA)-based relationship is proposed. In order to perform the recommendation between different types of multimedia data, i.e., recommendation of music pieces from human motions, the proposed method tries to estimate their relationship. Specifically, the correlation based on kernel CCA is calculated as the relationship in our method. Since human motions and music pieces have various time lengths, it is necessary to calculate the correlation between time series having different lengths. Therefore, new kernel functions for human motions and music pieces, which can provide similarities between data that have different time lengths, are introduced into the calculation of the kernel CCA-based correlation. This approach effectively provides a solution to the conventional problem of not being able to calculate the correlation from multimedia data that have various time lengths. Therefore, the proposed method can perform accurate recommendation of best matched music pieces according to a target human motion from the obtained correlation. Experimental results are shown to verify the performance of the proposed method.
• 足跡を用いた男女識別

  画像ラボ 22 (1) 17 - 21 2011 [Not refereed][Not invited]
  
• 固有空間BPLP法の補間精度に関する解析

  田中章, 小川貴弘, 長谷山美紀, 宮腰政明

  電子情報通信学会 論文誌(A) J94-A (2) 116 - 126 0913-5707 2011 [Not refereed][Not invited]
  
• Error-Resilient 3-D Wavelet Video Coding with Duplicated Lowest Sub-Band Coefficients and Two-Step Error Concealment Method

  Sunmi Kim, Hirokazu Tanaka, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E93A (11) 2173 - 2183 0916-8508 2010/11 [Refereed][Not invited]
   

  In this paper we propose a two step error concealment algorithm based on an error resilient three dimensional discrete wavelet transform (3 D DWT) video coding scheme. The proposed scheme consists of an error resilient encoder duplicating the lowest sub band bit streams for dispersive grouped frames and an error concealment decoder. The error concealment method of this decoder is decomposed of two steps the first step is replacement of erroneous coefficients in the lowest sub band by the duplicated coefficients and the second step is interpolation of the missing wavelet coefficients by minimum mean square error (MMSE) estimation. The proposed scheme can achieve robust transmission over unreliable channels. Experimental results provide performance comparisons in terms of peak signal to noise ratio (PSNR) and demonstrate increased performances compared to state of the art error concealment schemes.
• Recent Trends of Image and Video Semantic Analysis and Retrieval Interfaces(Developments of Vision Computing based on Probabilistic Information Processing)

  HASEYAMA Miki

  The Journal of the Institute of Electronics, Information, and Communication Engineers 一般社団法人電子情報通信学会 93 (9) 764 - 769 0913-5693 2010/09 [Not refereed][Not invited]
   

  画像・映像意味理解の研究動向とその検索への応用について紹介する.更に,画像及び映像が持つ固有の多義性とあいまい性から検索結果の可視化システムの必要性を議論し,その実現の試みについて紹介するとともに今後の展開について考える.
• Gender Identification Using Support Vector Machine Based on Shoe Print

  ASAMIZU Satoshi, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 93 (5) 642 - 646 1880-4535 2010/05/01 [Not refereed][Not invited]
   

  本論文では,施設や店舗などに出入りする人物の足跡を用いて男女識別する手法について検証する.取得した足跡から算出が可能な特徴量を用いてSVMにより男女を識別する.本手法を用いて被験者実験を行い,90%の識別率を実現した.
• ADAPTIVE RECONSTRUCTION METHOD OF MISSING TEXTURES BASED ON INVERSE PROJECTION VIA SPARSE REPRESENTATION

  Takahiro Ogawa, Miki Haseyama

  2010 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO (ICME 2010) 352 - 357 1945-7871 2010 [Refereed][Not invited]
   

  This paper presents an adaptive reconstruction method of missing textures based on an inverse projection via sparse representation. The proposed method approximates original and corrupted textures in lower-dimensional subspaces by using the sparse representation technique. Then, this approach effectively solves problems of not being able to directly estimate an inverse projection for reconstructing missing textures. Furthermore, even if target textures contain missing areas, the proposed method enables adaptive generation of the subspaces by monitoring errors caused in their known neighboring textures by the estimated inverse projection. Consequently, since the optimal inverse projection is adaptively estimated for each texture, successful reconstruction of the missing areas can be expected. Experimental results show impressive improvement of the proposed reconstruction technique over previously reported reconstruction techniques.
• THE SIMPLEST MEASUREMENT MATRIX FOR COMPRESSED SENSING OF NATURAL IMAGES

  Zaixing He, Takahiro Ogawa, Miki Haseyama

  2010 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING 4301 - 4304 1522-4880 2010 [Refereed][Not invited]
   

  There exist two main problems in currently existing measurement matrices for compressed sensing of natural images, the difficulty of hardware implementation and low sensing efficiency. In this paper, we present a novel simple and efficient measurement matrix, Binary Permuted Block Diagonal (BPBD) matrix. The BPBD matrix is binary and highly sparse (all but one or several "1"s in each column are "0"s). Therefore, it can simplify the compressed sensing procedure dramatically. The proposed measurement matrix has the following advantages, which cannot be entirely satisfied by existing measurement matrices. (1) It has easy hardware implementation because of the binary elements; (2) It has high sensing efficiency because of the highly sparse structure; (3) It is incoherent with different popular sparsity basis' like wavelet basis and gradient basis; (4) It provides fast and nearly optimal reconstructions. Moreover, the simulation results demonstrate the advantages of the proposed measurement matrix.
• Spatio-temporal resolution enhancement of video sequence based on super-resolution reconstruction.

  Miki Haseyama, Daisuke Izumi, Makoto Takizawa

  Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2010, 14-19 March 2010, Sheraton Dallas Hotel, Dallas, Texas, USA 870 - 873 2010 [Refereed][Not invited]
  
• A Genetic Algorithm for Generating Multiple Paths on Mesh Maps

  Jun Inagaki, Tomoaki Shirakawa, Tetsuo Shimono, Miki Haseyama

  2010 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC) 1 - 4 2010 [Refereed][Not invited]
   

  Path generation is an optimization problem mainly performed on grid square maps that combines generation of paths with minimization of their cost. Several methods that belong to the class of exhaustive searches are available; however, these methods are only able to obtain a single path as a solution for each iteration of the search. Hence, this paper proposes a new method using genetic algorithms for this problem with the goal of simultaneously searching for multiple candidate paths.
• Missing Texture Reconstruction Method Based on Perceptually Optimized Algorithm

  Takahiro Ogawa, Miki Haseyama

  EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING 2010 1687-6172 2010 [Refereed][Not invited]
   

  This paper presents a simple and effective missing texture reconstruction method based on a perceptually optimized algorithm. The proposed method utilizes the structural similarity (SSIM) index as a new visual quality measure for reconstructing missing areas. Furthermore, in order to adaptively reconstruct target images containing several kinds of textures, the following two novel approaches are introduced into the SSIM-based reconstruction algorithm. First, the proposed method performs SSIM-based selection of the optimal known local textures to adaptively obtain subspaces for reconstructing missing textures. Secondly, missing texture reconstruction that maximizes the SSIM index in the known neighboring areas is performed. In this approach, the nonconvex maximization problem is reformulated as a quasi convex problem, and adaptive reconstruction of the missing textures based on the perceptually optimized algorithm becomes feasible. Experimental results show impressive improvements of the proposed method over previously reported reconstruction methods.
• Erratum: Error-Resilient 3-D Wavelet Video Coding with Duplicated Lowest Sub-Band Coefficients and Two-Step Error Concealment Method [IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences E93.A (2010) , No. 11 pp.2173-2183]

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki

  IEICE Trans. Fundamentals The Institute of Electronics, Information and Communication Engineers 93 (12) 2763_e1 - 2763_e1 0916-8508 2010
• Interdisciplinary Research Project on Biological Science(Next-Generation Information Technology Based on Knowledge Discovery and Knowledge Federation: Hokkaido University Global COE Program and Research Projects on ICT in Hokkaido)

  WATANABE Hidemi, KANEKO Shun'ichi, HASEYAMA Miki

  The Journal of the Institute of Electronics, Information, and Communication Engineers 一般社団法人電子情報通信学会 92 (10) 822 - 827 0913-5693 2009/10 [Not refereed][Not invited]
  
• Adaptive Missing Texture Reconstruction Method Based on Kernel Canonical Correlation Analysis with a New Clustering Scheme

  Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (8) 1950 - 1960 0916-8508 2009/08 [Refereed][Not invited]
   

  In this paper, a method for adaptive reconstruction of missing textures based on kernel canonical correlation analysis (CCA) with a new clustering scheme is presented. The proposed method estimates the correlation between two areas, which respectively correspond to a missing area and its neighboring area, from known parts within the target image and realizes reconstruction of the missing texture. In order to obtain this correlation, the kernel CCA is applied to each cluster containing the same kind of textures, and the optimal result is selected for the target missing area. Specifically, a new approach monitoring errors caused in the above kernel CCA-based reconstruction process enables selection of the optimal result. This approach provides a solution to the problem in traditional methods of not being able to perform adaptive reconstruction of the target textures due to missing intensities. Consequently, all of the missing textures are successfully estimated by the optimal cluster's correlation, which provides accurate reconstruction of the same kinds of textures. In addition, the proposed method can obtain the correlation more accurately than our previous works, and more successful reconstruction performance can be expected. Experimental results show impressive improvement of the proposed reconstruction technique over previously reported reconstruction techniques.
• An ER Algorithm-Based Method for Removal of Adherent Water Drops from Images Obtained by a Rear View Camera Mounted on a Vehicle in Rainy Conditions

  Tomoki Hiramatsu, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (8) 1939 - 1949 0916-8508 2009/08 [Refereed][Not invited]
   

  In this paper, an ER (Error-Reduction) algorithm-based method for removal of adherent water drops from images obtained by a rear view camera mounted on a vehicle in rainy conditions is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two novel constraints for the removal of adherent water drops. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude of the previous frame in the obtained images as that of the target frame. Noting that images obtained by the rear view camera have the unique characteristics of objects moving like ripples because the rear view camera is generally composed of a fish-eye lens for a wide view angle, the proposed method assumes that the Fourier transform magnitudes of the target frame and the previous frame are the same in the polar coordinate system. The second constraint is the object-domain constraint that utilizes intensities in an area of the target frame to which water drops have adhered. Specifically, the proposed method models a deterioration process of intensities that are corrupted by the water drop adhering to the rear view camera lens. By utilizing these novel constraints, the proposed ER algorithm can remove adherent water drops from images obtained by the rear view camera. Experimental results that verify the performance of the proposed method are represented.
• Estimating Number of People Using Calibrated Monocular Camera Based on Geometrical Analysis of Surface Area

  Hiroyuki Arai, Isao Miyagawa, Hideki Koike, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (8) 1932 - 1938 0916-8508 2009/08 [Refereed][Not invited]
   

  We propose a novel technique for estimating the number of people in a video sequence; it has the advantages of being stable even in crowded situations and needing no ground-truth data. By analyzing the geometrical relationships between image pixels and their intersection volumes in the real world quantitatively, a foreground image directly indicates the number of people. Because foreground detection is possible even in crowded situations, the proposed method can be applied in such situations. Moreover, it can estimate the number of people in an a priori manner, so it needs no ground-truth data unlike existing feature-based estimation techniques. Experiments show the validity of the proposed method.
• An Accurate Scene Segmentation Method Based on Graph Analysis Using Object Matching and Audio Feature

  Makoto Yamamoto, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (8) 1883 - 1891 0916-8508 2009/08 [Refereed][Not invited]
   

  A method for accurate scene segmentation using two kinds of directed graph obtained by object matching and audio features is proposed. Generally, in audiovisual materials, such as broadcast programs and movies, there are repeated appearances of similar shots that include frames of the same background, object or place, and such shots are included in a single scene. Many scene segmentation methods based on this idea have been proposed; however, since they use color information as visual features, they cannot provide accurate scene segmentation results if the color features change in different shots for which frames include the same object due to camera operations such as zooming and panning. In order to solve this problem, scene segmentation by the proposed method is realized by using two novel approaches. In the first approach, object matching is performed between two frames that are each included in different shots. By using these matching. results, repeated appearances of shots for which frames include the same object can be successfully found and represented as a directed graph. The proposed method also generates another directed graph that represents the repeated appearances of shots with similar audio features in the second approach. By combined use of these two directed graphs, degradation of scene segmentation accuracy, which results from using only one kind of graph, can be avoided in the proposed method and thereby accurate scene segmentation can be realized. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• Maintaining the Stability of Solutions in Dynamic Fuzzy CSPs

  SUDO Yasuhiro, YANAGIDA Takuto, KURIHARA Masahito, HASEYAMA Miki

  Journal of Japan Society for Fuzzy Theory and Intelligent Informatics 日本知能情報ファジィ学会 21 (3) 372 - 380 1347-7986 2009/06/15 [Not refereed][Not invited]
   

  A Fuzzy Constraint Satisfaction Problem (FCSP) is an extension of the classical CSP, a powerful tool for modeling various problems based on constraints among variables. Meanwhile, a Dynamic CSP (DCSP) is a framework for modeling the transformation of problems, and the differences between past solutions and new solutions should be as small as possible. These schemes are the techniques to formulate real world problems as CSPs more easily. The CSP model that combines these (Dynamic Fuzzy CSP) has already existing researches. However, as CSPs are NP-hard problems in general, no efficient and complete algorithms for solving CSPs exist and the increase in the worst-case computation time is exponential in the size of the problems. In the work reported in this paper we tested a hybrid approximate method, called the SRS algorithm. Moreover, we introduce a post-filtering method, called the SRSD algorithm. We empirically show that SRS and SRSD algorithms keep the stability of solutions better than other algorithms. They are able to quickly get good-quality approximate and stabile solutions to large problems.
• Active Net-Based Non-interception Region Estimation in Soccer Videos

  TAKAHASHI Sho, KON Hirofumi, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 92 (4) 501 - 510 1880-4535 2009/04/01 [Not refereed][Not invited]
   

  本論文では,チームスポーツ映像からアクティブネットを用いてパス可能領域を推定する手法を提案する.チームスポーツ映像の一つであるサッカー映像の意味内容解析を行うために重要なサッカーの戦術は,選手の移動とボール運びによって表現されるため,ボール運びを実現するパスを分析することは重要である.一般にパスコースはボール保持者と味方チームの選手へとつながる緩やかな曲線で表される.提案手法では,新たなエネルギーの定義とパス可能領域を推定するための画像生成により,アクティブネットを用いて前述の曲線が存在する領域を抽出する.また,パス可能領域は守備の選手から離れるほど,パスが成功する可能性が高いという特徴をもつ.提案手法では,格子点の密度に着眼することで,パスが成功する可能性をパス可能領域の推定と同時に得る.更に,アクティブネットの収束結果は多少の選手位置の誤差を許容するため,選手の動きを用いた従来手法における,選手位置の誤差の影響を受けやすいという問題点を解決することが可能である.したがって,提案手法はカメラワークが存在し,高精度な選手位置の推定が困難であるテレビ映像に対しても,高精度にパス可能領域の推定が可能である.
• Kalman Filter-Based Error Concealment for Video Transmission

  Shigeki Takahashi, Takahiro Ogawa, Hirokazu Tanaka, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (3) 779 - 787 0916-8508 2009/03 [Refereed][Not invited]
   

  A novel error concealment method using a Kalman filter is presented ill this paper, In order to successfully utilize the Kalman filter, its state transition and observation models that are suitable for the video error concealment are newly defined as follows. The state transition model represents the video decoding process by a notion-compensated prediction. Furthermore, the new observation model that represents all image blurring process is defined. and calculation of the Kalman gain becomes possible. The problem of the traditional methods is solved by using the Kalman filter in the proposed method, and accurate reconstruction of corrupted video frames, is achieved. Consequently. an effective error concealment method using the Kalman filter is realized. Experimental results showed that the proposed method has better performance than that of traditional methods.
• An IFS-Based Image Enlargement Method Considering Edge Continuity

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 92 (3) 382 - 392 1880-4535 2009/03/01 [Not refereed][Not invited]
   

  本論文では,エッジの連続性を考慮した, Iterated Function System (IFS)に基づく画像拡大法を提案する.従来のIFS画像拡大法では,拡大後の画像において処理の最小単位となるブロックの境界で,本来存在しないはずの輝度値の変化が生じる.また,エッジの連続性を考慮しておらず,拡大後の画像におけるエッジが不連続となる問題が存在した.そこで提案手法では,まず,処理の最小単位となるブロックの重なりを許すことで, IFSによる高近似縮小写像を実現し,ブロック境界での輝度値の変化を抑制する.更に,連続性を保ったエッジの推定が可能であるラインプロセスを新たにIFS画像拡大法に導入する.このとき我々は,ラインプロセスを拡大後の画像のエッジを推定する手法に拡張することで,輝度値が未知である拡大後の画像に対しても,連続性を保ったエッジの推定を可能とする.また,提案手法では得られるエッジの存在を考慮した上でIFSに基づく画像拡大を行うことにより,従来法で発生していた拡大後の画像におけるエッジの不連続を解決し,高精細な拡大を実現する.本論文の最後では,提案手法の有効性を示すため比較実験を行い,その拡大性能を評価する.
• A Kalman Filter-Based Method for Restoration of Images Obtained by an In-Vehicle Camera in Foggy Conditions

  Tomoki Hiramatsu, Takahiro Ogawa, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E92A (2) 577 - 584 0916-8508 2009/02 [Refereed][Not invited]
   

  In this paper, a Kalman filter-based method for restoration of video images acquired by an in-vehicle camera in foggy conditions is proposed. In order to realize Kalman filter-based restoration, the proposed method clips local blocks from the target frame by using a sliding window and regards the intensities in each block as elements of the state variable of the Kalman filter. Furthermore, the proposed method designs the following two models for restoration of foggy images. The first one is an observation model, which represents a fog deterioration model. The proposed method automatically determines all parameters of the fog deterioration model from only the foggy images to design the observation model. The second one is a non-linear state transition model, which represents the target frame in the original video image from its previous frame based on motion vectors. By utilizing the observation and state transition models, the correlation between successive frames can be effectively utilized for restoration, and accurate restoration of images obtained in foggy conditions can be achieved. Experimental results show that the proposed method has better performance than that of the traditional method based on the fog deterioration model.
• "Common technologies" of information grand voyage project: Introduction to image and video processing technologies

  Miki Haseyama, Toru Hisamitsu

  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers 一般社団法人映像情報メディア学会 63 (1) 42 - 47 1342-6907 2009 [Not refereed][Not invited]
   

  2007年度に3年計画でスタートした情報大航海プロジェクトは,モデルサービスによる実証を通じて次世代の情報検索・解析技術を開発することを目的としている.同時に,実証を通じた制度的課題の洗い出しにより,市場創出に必要な環境整備を目指している.本稿では,開発された技術の中から,マルチメディア情報処理の中核を担う,画像・映像処理技術を紹介する.
• Improvement of video coding performance using adaptive interpolation filter without filter selection information

  Satoshi Kondor, Takaya Matsuno, Miki Haseyama

  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers 映像情報メディア学会 63 (11) 1592 - 1597 1342-6907 2009 [Not refereed][Not invited]
   

  We propose a method to improve performance of video coding using an adaptive interpolation filter technique. The adaptive interpolation technique was based on clustering using the k-means method and did not need filter selection information. To improve the performance of the clustering, we introduced autocorrelation coefficients of the pixel values and the directions of the motion vectors as the new features and the Mahalanobis distance as the distance scale in the k-means method. We also used vector quantization to reduce the number of interpolation filter coefficients. In the simulation, our proposed method was implemented in the MPEG-2 based video codec. The simulation results show that the proposed method can reduce the bit rate by up to nearly 7 % compared to conventional adaptive interpolation filter methods.
• Accurate Graph-Based Scene Segmentation Using Object Matching and Audio Feature

  Makoto Yamamoto, Miki Haseyama

  ISCE: 2009 IEEE 13TH INTERNATIONAL SYMPOSIUM ON CONSUMER ELECTRONICS, VOLS 1 AND 2 670 - 671 2158-3994 2009 [Refereed][Not invited]
   

  A method for accurate scene segmentation utilizing two kinds of directed graph obtained by object matching and by using audio features is proposed. Generally, in audiovisual materials, there are repeated appearances of shots that include frames of the same background, object or place. It should be assumed that such shots are included in a single scene. In the proposed method, by performing object matching between two frames that are each included in different shots, multiple shots for which frames include the same object can be successfully found and their repeated appearances are represented as a directed graph. The proposed method also generates another directed graph that represents the repeated appearances of shots with similar audio features. By the combined use of these two graphs, accurate scene segmentation can be realized. The effectiveness of the proposed method is verified by applying this method to news programs and another broadcast program.
• Semantic Image Retrieval Based on POCS Algorithm Using Kernel PCA And Its Performance Verification

  Takahiro Ogawa, Miki Haseyama

  ISCE: 2009 IEEE 13TH INTERNATIONAL SYMPOSIUM ON CONSUMER ELECTRONICS, VOLS 1 AND 2 342 - 343 2158-3994 2009 [Refereed][Not invited]
   

  This paper presents a projection onto convex sets (POCS)-based semantic image retrieval method and its performance verification. The main contributions of the proposed method are twofold: introduction of nonlinear eigenspace of visual and semantic features into the constraint of the POCS-based semantic image retrieval algorithm and adaptive selection of the annotated images utilized for this algorithm. Then, by combining these two approaches., the semantic features of the query image are successfully estimated, and accurate image retrieval can be expected. Finally, relationship between the performance of the proposed method and the kinds of the kernel functions utilized for the kernel PICA is shown in this paper.
• A new image retrieval interface and its practical use in "View Search Hokkaido"

  Miki Haseyama, Toshifumi Murata, Hisashi Ukawa

  ISCE: 2009 IEEE 13TH INTERNATIONAL SYMPOSIUM ON CONSUMER ELECTRONICS, VOLS 1 AND 2 624 - + 2158-3994 2009 [Refereed][Not invited]
   

  New image retrieval technology, which is used for a service demonstration project "View Search Hokkaido" in "Information Grand Voyage Project" conducted by Ministry of Economy, Trade and Industry, Japan, is presented in this paper. The new technology enables image retrieval based on low-level features without utilizing any tag-based scheme and realizes a 3D interface for image retrieval. This 3D interface can provide new user experience, which has not been provided by the conventional retrieval services. The effectiveness of the 3D interface has been verified in "View Search Hokkaido".
• 映像検索のための画像及び音響信号によるシーン間の相違度算出--正準相関分析とDynamic Time Warpingを用いたアプローチ

  道山 大悟, 長谷山 美紀

  画像ラボ 日本工業出版 19 (7) 1 - 4 0915-6755 2008/07 [Not refereed][Not invited]
  
• Innovation of Video Technology and Visual Information Media

  Haseyama Miki

  The journal of the Institute of Image Information and Television Engineers 一般社団法人映像情報メディア学会 62 (4) K12 - K12 1342-6907 2008/04/01 [Not refereed][Not invited]
  
• Media distribution and processing

  Kazuto Kamikura, Miki Haseyama, Kazuhito Murakami

  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers 一般社団法人映像情報メディア学会 62 (8) 1251 - 1254 1342-6907 2008 [Not refereed][Not invited]
  
• Estimating the number of people in a video sequence via geometrical model.

  Hiroyuki Arai, Isao Miyagawa, Hideki Koike, Miki Haseyama

  19th International Conference on Pattern Recognition (ICPR 2008)(ICPR) 1 - 4 2008
• KERNEL PCA-BASED SEMANTIC FEATURE ESTIMATION APPROACH FOR SIMILAR IMAGE RETRIEVAL

  Takahiro Ogawa, Miki Haseyama

  2008 IEEE International Conference on Image Processing, Proceedings 969 - 972 1522-4880 2008 [Refereed][Not invited]
   

  A kernel PCA-based semantic feature estimation approach for similar image retrieval is presented in this paper. Utilizing database images previously annotated by keywords, tire proposed method estimates unknown semantic features of a query image. First, our method performs semantic clustering of the database images and derives a new map from a nonlinear eigenspace of visual and semantic features in each c aster. This map accurately provides the semantic features for the images belonging to each cluster by using their visual features. Further, in order to select the optional cluster including the query image, the proposed method monitors errors of the visual features caused by the Semantic feature estimation process. Then, even if any semantics of the query image arc unknown, its semantic features are successfully estimated by tire optimal cluster. Experimental results verify the effectiveness of the proposed method for semantic image retrieval.
• カルマンフィルタと適応信号処理, 谷萩隆嗣著, (コロナ社, Tel.03-3941-3131, 2005年, A5判, 282頁, 定価4,515円(税込))

  長谷山 美紀

  日本音響学会誌 社団法人日本音響学会 63 (8) 0369-4232 2007/08/01 [Not refereed][Not invited]
  
• Audio-based shot classification for audiovisual indexing using PCA, MGD and fuzzy algorithm

  Naoki Nitanda, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E90A (8) 1542 - 1548 1745-1337 2007/08 [Refereed][Not invited]
   

  An audio-based shot classification method for audiovisual indexing is proposed in this paper. The proposed method mainly consists of two parts, an audio analysis part and a shot classification part. In the audio analysis part, the proposed method utilizes both principal component analysis (PCA) and Mahalanobis generalized distance (MGD). The effective features for the analysis can be automatically obtained by using PCA, and these features are analyzed based on MGD, which can take into account the correlations of the data set. Thus, accurate analysis results can be obtained by the combined use of PCA and MGD. In the shot classification part, the proposed method utilizes a fuzzy algorithm. By using the fuzzy algorithm, the mixing rate of the multiple audio sources can be roughly measured, and thereby accurate shot classification can be attained. Results of experiments performed by applying the proposed method to actual audiovisual materials are shown to verify the effectiveness of the proposed method.
• Players clustering based on graph theory tor tactics analysis purpose in soccer videos

  Hiroftuni Kon, Miki Haseyama

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E90A (8) 1528 - 1533 1745-1337 2007/08 [Refereed][Not invited]
   

  In this paper, a new method for clustering of players in order to analyze games in soccer videos is proposed. The proposed method classifies players who are closely related in terms of soccer tactics into one group. Considering soccer tactics, the players in one group are located near each other. For this reason, the Euclidean distance between the players is an effective measurement for the clustering of players. However, the distance is not sufficient to extract tactics-based groups. Therefore, we utilize a modified version of the community extraction method, which finds community structure by dividing a non-directed graph. The use of this method in addition to the distance enables accurate clustering of players.
• Visualization of relationship among academic papers in small database

  二反田 直己, 鎌倉 純一, 長谷山 美紀

  Journal of signal processing 〔信号処理学会〕 11 (2) 179 - 185 1342-6230 2007/03 [Not refereed][Not invited]
  
• Phase retrieval based on a snake for image reconstruction

  Keiko Kondo, Miki Haseyama, Hideo Kitajima

  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E90D (1) 283 - 287 0916-8532 2007/01 [Not refereed][Not invited]
   

  A new phase retrieval method using an active contour model (snake) for image reconstruction is proposed. The proposed method reconstructs a target image by retrieving the phase from the magnitude of its Fourier transform and the measured area of the image. In general, the measured area is different from the true area where the target image exists. Thus a snake, which can extract the shape of the target image, is utilized to renew the measured area. By processing this renewal iteratively, the area obtained by the snake converges to the true area and as a result the proposed method can accurately reconstruct a target image even when the measured area is different from the true area. Experimental results show the effectiveness of the proposed method.
• Steady-state properties of a CORDIC-based adaptive ARMA lattice filter

  Shin'ichi Shiraishi, Miki Haseyama, Hideo Kitajima

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E89A (12) 3724 - 3729 0916-8508 2006/12 [Not refereed][Not invited]
   

  This paper analyzes the steady-state properties of a CORDIC-based adaptive ARMA lattice filter. In our previous study, the convergence properties of the filter in the non-steady state were clarified; however, its behavior in the steady state was not discussed. Therefore, we develop a distinct analysis technique based on a Markov chain in order to investigate the steady-state properties of the filter. By using the proposed technique, the relationship between step size and coefficient estimation error is revealed.
• 基礎シリーズ 線形予測理論とラティスフィルタアルゴリズム(2)ARラティスフィルタの実現とその性質

  長谷山 美紀

  信号処理 〔信号処理学会〕 10 (5) 309 - 315 1342-6230 2006/09 [Not refereed][Not invited]
  
• Image Recognition Method Using Intensity Gradient Vectors

  HIRAMOTO Masao, OGAWA Takahiro, HASEYAMA Miki

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 89 (6) 1348 - 1358 1880-4535 2006/06/01 [Not refereed][Not invited]
   

  本論文では,撮像素子の多面素化や高画質化の流れを踏まえ,画像の回転・移動等の幾何学的変換にも対応できる大局的な画像識別方法を提案している.提案手法は,ベクトルを利用した投票方式を用いかものであり,画像を輝度こう配を表すベクトルと位置を示すベクトルで表現し,識別のための投票ベクトルと類似度を定義している.また提案手法では,同一画像であれば得票場所が原点に集中し,得票結果が幾何学的な変換に影響されないという特徴がある.原画像に対してガウシアン,メジアンのフィルタリング処理,JPEG圧縮処理を施した画像も含め,自然画像の識別実験を行ったところ,類似性において明確な差が現れ,画像に対して人工的な処理を加えても識別可能であることが分かった.更に提案手法の応用として,最多得票点を利用した画像の識別について検討したところ,識別能力が高く,1画像内に含まれる部分画像の識別も可能であることを示すことができた.
• 基礎シリーズ 線形予測理論とラティスフィルタアルゴリズム(1)線形予測とARモデル同定

  長谷山 美紀

  信号処理 〔信号処理学会〕 10 (3) 153 - 159 1342-6230 2006/05 [Not refereed][Not invited]
  
• A report on video retrieval for intelligent information access

  長谷山 美紀

  回路とシステム軽井沢ワークショップ論文集 [電子情報通信学会] 19 199 - 203 2006/04/24 [Not refereed][Not invited]
  
• Improvements on Watershed-Based Image Segmentation Using Color Edges and Parallel Region Merging

  ZHAO Yanjun, HASEYAMA Miki, KITAJIMA Hideo

  The IEICE transactions on information and systems (Japanese edetion) 一般社団法人電子情報通信学会 89 (4) 836 - 849 1880-4535 2006/04/01 [Not refereed][Not invited]
   

  画像内容に基づく画像処理では,画素ではなく,画像中の物体を構成する領域を対象とし,画像の圧縮,検索,認識等様々な処理を行う.そのため,領域を検出し,ラベルを付与する画像分割が必要になる.本論文では画像における輝度値の空間分布を地形とみなすWatershedアルゴリズムによる領域分割を検討する.Watershedアルゴリズムにより,閉じた単一の領域分割線が得られるが,分割結果には領域の未分割及び過剰分割という問題が発生する.本論文においてはこれらの問題を解決するために,原画像を直接処理することではなく,連続エッジと均質部分の両方が強調できる色エッジの強度画像を分割対象とする.更に,未分割と過剰分割とを防ぐために,色エッジの強度画像に対する分割結果を初期領域として,ノイズ抑制を考慮した並列実行の領域統合を行う.提案手法を各種の画像へ適用した結果によりその有効性を示す.
• A multi-objective service restoration method for power distribution systems

  Jun Inagaki, Jun Nakajima, Miki Haseyama

  2006 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-11, PROCEEDINGS 1784 - + 0271-4302 2006 [Refereed][Not invited]
   

  Service restoration problem in distribution systems is formulated as a multi-objective optimization problem which is demanded not only for minimizing the amount of unrestored total loads but also for minimizing the number of the switching operations. The solution of the multi-objective optimization problem is usually obtained with a set of Pareto optimal solutions. The Pareto optimal solutions for the service restoration problem are useful for users to obtain their desired restoration by comparing a Pareto optimal solution with the others. However, the conventional methods cannot obtain several Pareto optimal solutions in one trial. Therefore, this paper proposes a method for obtaining the Pareto optimal set for the service restoration problem with a genetic algorithm. The genetic algorithm produces many possible solutions in its search process. By utilizing this feature, the proposed method can obtain the Pareto optimal set.
• Restoration method of missing areas in still images using GMRF model

  T Ogawa, M Haseyama, H Kitajima

  2005 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), VOLS 1-6, CONFERENCE PROCEEDINGS 4931 - 4934 0271-4302 2005 [Refereed][Not invited]
   

  This paper proposes a GMRF-model based restoration method of missing areas in still images. The GMRF model used in the proposed method is realized by a new assumption that reasonably holds for an image source. This model can express important image features such as edges because of the use of the new assumption. Therefore, the proposed method restores the missing areas using the modified GMRF model and can correctly reconstruct the missing edges. Consequently, the proposed method achieves more accurate restoration than those of the traditional methods on both objective and subjective measures. Extensive experimental results demonstrate the improvement of the proposed method over the previous methods.
• GA-based applications for routing with an upper bound constraint

  J Inagaki, M Haseyama

  2005 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), VOLS 1-6, CONFERENCE PROCEEDINGS 2239 - 2242 0271-4302 2005 [Refereed][Not invited]
   

  This paper presents a method of searching for the shortest route via the most designated points among the routes whose lengths are less than the upper bound using a genetic algorithm (GA). If chromosomes whose route lengths exceed the upper bound are simply screened out in the GA process, the optimization probability gets worse. For the purpose of solving this problem, this paper proposes a new fitness function including an upper bound constraint which can be flexibly changed during the searching process. By using this function, the optimum is efficiently obtained and the optimization probability can be raised. Furthermore, the effectiveness of the proposed method is verified by the experiments applying it to the actual map data.
• Audio signal segmentation and classification for scene-cut detection

  N Nitanda, M Haseyama, H Kitajima

  2005 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), VOLS 1-6, CONFERENCE PROCEEDINGS 4030 - 4033 0271-4302 2005 [Refereed][Not invited]
   

  A scene is regarded as a basic unit of audiovisual material, and thereby the boundaries between two adjacent scenes, which are called scene-cuts, must be detected in advance for audiovisual indexing. This paper proposes a scene-cut detection method. Since scene-cuts are associated with a simultaneous change of visual and audio characteristics, both audio and visual analyses are required for the scene-cut detection. For the audio signal analysis, the proposed method utilizes an audio signal segmentation and classification method using fuzzy c-means clustering, which has been proposed by the authors. For the visual signal analysis, the proposed method utilizes some visual segmentation methods. By using these methods simultaneously, the proposed method can accurately detect the scene-cuts, and thereby it is highly valuable for the preprocessing of the audiovisual indexing. Experimental results performed by applying the proposed method to real audiovisual material are shown to verify its high performance.
• Quality improvement technique for JPEG images with fractal image coding

  M Takezawa, H Sanada, K Watanabe, M Haseyama

  2005 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), VOLS 1-6, CONFERENCE PROCEEDINGS 6320 - 6323 0271-4302 2005 [Refereed][Not invited]
   

  This paper proposes a quality improvement technique for JPEG images by using fractal image coding. JPEG coding is a commonly used standard method of compressing images. However, in its decoded images, quantization noise is sometimes visible in high frequency regions, such as the edges of objects. Hence, in order for the JPEG coding to become a more powerful image-coding method, the JPEG image quality must be improved. Therefore, our method solves this problem by adding the obtained codes by the fractal image coding to improve the image quality. Some simulation results verify that the proposed method achieved higher coding-performance than the traditional JPEG coding.
• A GA-based fast search algorithm for realizing efficiency motion compensation

  S Asamizu, M Haseyama

  8TH WORLD MULTI-CONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS, VOL VI, PROCEEDINGS 185 - 190 2004 [Refereed][Not invited]
   

  This paper proposes a fast search technique using a genetic algorithm (GA) [1] for realizing efficiency motion compensation. Previously proposed method[2] that the motion compensation based on table lookup refers to two or more frames with one motion table. Therefore, dispersion of the prediction error of each frame is not only dispersion of the image quality of each prediction image but also influences the search speed of the motion table designing. In this proposed method, the prediction error of each frame is obtained by using the block matching method[3] before the motion table is designed. The weight of each frame is put from the obtained prediction error to the fitness function of the GA, and the GA searches for the optimal motion table. Based on this strategy, the proposed method can search for a motion table more effectively.
• New JPEG coding algorithm including fractal image coding for color images

  M Takezawa, M Haseyama

  8TH WORLD MULTI-CONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS, VOL VI, PROCEEDINGS 333 - 336 2004 [Refereed][Not invited]
   

  This paper proposes an effective JPEG coding algorithm using fractal image coding for color images. In the JPEG images, quantization noise is sometimes visible in high frequency regions, such as the edges of objects. Hence, in order for the JPEG coding to become a more powerful image-coding method, the JPEG image quality must be improved. Therefore, in this paper, a quality improvement method for the color JPEG images is proposed by using the fractal image coding. Some simulation results verify that the proposed method can improve the image quality. in the high frequency regions and provide the decoded image with 0.4 dB higher quality than the traditional JPEG.
• A trainable retrieval system for cartoon character images

  M Haseyama, A Matsumura

  2003 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL III, PROCEEDINGS 673 - 676 2003 [Refereed][Not invited]
   

  This paper proposes a novel method to retrieve cartoon character images in a database or network. In this method, partial features of an image, defined as Regions and Aspects, are used as keys to identify cartoon character images. The similarities between a query cartoon character image and the images in the database are computed by using these features. Based on the similarities the cartoon images same or similar to the query image are identified and retrieved from the database. Moreover, our method adopts a training scheme to reflect the user's subjectivity. The training emphasizes the signficant Regions or Aspects by assigning more weight based on the user's preferences and actions, such as selecting a desired image or an area of an image. These processes make the retrieval more effective and accurate. Experiment results verify the effectiveness and retrieval accuracy of the method.
• 2-D functional AR model for image identification

  M Haseyama, Kondo, I

  2003 INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, VOL II, PROCEEDINGS 377 - 380 2003 [Refereed][Not invited]
   

  This paper proposes a 2-D Functional AR Model for image identification. The definition of the proposed model includes functions that can exploit the self-similarity nature in images to throughly extract image features. By introducing the functional scheme into the model, only a few number of parameters, which are called 2-D Functional AR parameters, can describe the image features simply and accurately. These characteristics make the model suitable for image identification applications. Some experiments of image identification are performed, and the results verify that the proposed model accurately represents the image feature, and the image can be correctly, identified. The calculation time is fast enough for practical use in image retrieval.
• E cient fixed-valued and random-valued impulse detection for accurate image restoration

  K Kondo, M Haseyama, H Kitajima

  ISPA 2003: PROCEEDINGS OF THE 3RD INTERNATIONAL SYMPOSIUM ON IMAGE AND SIGNAL PROCESSING AND ANALYSIS, PTS 1 AND 2 1009 - 1012 2003 [Refereed][Not invited]
   

  This paper proposes a novel impulse detection method for the restoration of images corrupted by impulse noise. Conventional impulse detection methods tend to work well for fixed-valued impulse noise but poorly for random-valued impulse noise. The proposed method can accurately detect not only fixed-valued but also random-valued impulse noise by using two different systems. The first system detects impulse noise by considering the di erences between the intensity of a target pixel and the output of a median filter The second system verifies whether the impulse detection results obtained by the first system are correct. By using these systems, the proposed method can accurately detect both types of impulse noise even in highly corrupted images. Furthermore, the proposed method can be e ectively used as a preprocessor for noise reduction filtering. Experiments are presented to demonstrate the e ectiveness of the proposed method.
• Two-dimensional analysis of magnetic microstructures in the DC-demagnetized state and magnetization fluctuations in the transition region using MFM images

  Takekuma, I, M Haseyama, K Sueoka, K Mukasa

  JOURNAL OF MAGNETISM AND MAGNETIC MATERIALS 239 (1-3) 359 - 362 0304-8853 2002/02 [Refereed][Not invited]
   

  This study analyzed the directional dependence of magnetic microstructures of media by focusing on the distribution of magnetic poles. Media, whose magnetic poles tend to be distributed along a specific direction in the DC-demagnetized state, have larger magnetization fluctuations in the transition region, larger medium noise and higher partial erasure (PE) probability than media with uniform magnetic pole distribution. (C) 2002 Elsevier Science B.V. All rights reserved.
• A new approach with IFS for image restoration.

  Miki Haseyama, Megumi Takezawa, Junichi Miura, Hideo Kitajima

  10th European Signal Processing Conference(EUSIPCO) 1 - 4 2000
• A practical method to reduce a number of reference signals for the ANC system

  M Akiho, M Haseyama, H Kitajima

  ICASSP '99: 1999 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, PROCEEDINGS VOLS I-VI 2387 - 2390 1520-6149 1999 [Refereed][Not invited]
   

  In this paper, we propose a practical method to reduce a number of reference signals for the active noise cancellation (ANC) system and the filter characteristics to generate the reduced number of reference signals, which maintain the original value of the coherence function. This method finds the number of independent noise sources and provides the filter characteristics based on SVD (singular value decomposition) of the power spectrum matrix of the reference signals. Then, we also use the multiple coherence function analysis to select dominant components in the reference signals. The method contributes greatly in reducing the number of reference signals for the ANC system that uses the large number of reference signals. We also discuss the characteristics of the filters that synthesis the new set of reference signals. And an experimental test is performed to confirm the theory.
• A genetic algorithm based image segmentation for image analysis

  M Haseyama, M Kumagai, H Kitajima

  ICASSP '99: 1999 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, PROCEEDINGS VOLS I-VI 3445 - 3448 1520-6149 1999 [Refereed][Not invited]
   

  In this paper a new genetic algorithm (GA) based image segmentation method is proposed for image analysis. This method using a mean square error (MSE) based criterion can segment an image into some regions, while estimating a suitable region representation. The criterion is defined as MSE caused by interpolating each region of an observed image with a parametric model. Since the criterion is expressed with not only the parameters of the model but also shape and location of the regions, the criterion can not be easily minimized by the usual optimization methods, the proposed method minimizes the criterion by a GA. The proposed method also includes a processor to eliminate fragile regions with the Markov random field (MRF) model. Though the thresholds of the existent methods negatively affect image segmentation results; since no thresholds are required in the proposed method, it segments images more accurately than the existent methods.
• A realization method of an ARMAX lattice filter

  M HASEYAMA, T HIROHKU, H KITAJIMA

  1995 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-3 365 - 368 0277-674X 1995 [Refereed][Not invited]
  
• FREQUENCY-WEIGHTING MODEL IDENTIFICATION WITH AN ADAPTIVE ARMA LATTICE FILTER

  M HASEYAMA, N NAGAI, N MIKI

  IEEE INTERNATIONAL CONFERENCE ON SYSTEMS ENGINEERING 543 - 546 1992 [Refereed][Not invited]
  

Books etc

• 原理がわかる信号処理

  長谷山, 美紀 
  共立出版 2021/08 (ISBN: 9784320086517) v, 115p
• バイオメティクス・エコミメティクスー持続可能な循環型社会へ導く技術革新のヒントー

  (Joint workpp.92-97)
  2021/01
• 人と共生する AI革命 活用事例から見る生活・産業・社会の未来展望

  (Joint workpp.21-29)
  2019/06
• 持続可能社会を拓くバイオミメティクス--生物学と工学が築く材料科学

  (Joint workpp.38-45)
  2018/03
• トコトンやさしいバイオミメティクスの本

  (Joint workpp.136-137)
  2016/03
• 生物の形や能力を利用する学問 バイオミメティクス

  (Joint workpp.124-132)
  2016/03
• 生物模倣技術と新材料・新製品開発への応用

  長谷山 美紀 (Joint work第5節)
  2014/07
• 著作画像の非破壊配信を目的とした新しい著作認証システムの実現

  長谷山, 美紀 
  [北海道大学大学院情報科学研究科] 2005 1冊

Conference Activities & Talks

• A Note on Estimation of Group Advantage for Group Tactics Analysis in Soccer Videos  [Not invited]

  TAKAHASHI SHO, HASEYAMA MIKI

  映像情報メディア学会技術報告  2015/12
• A Study on High Performance Parallel Computing for Active Net-based Pass Region Estimation in Soccer Videos  [Not invited]

  TAKAHASHI SHO, HASEYAMA MIKI

  映像情報メディア学会技術報告  2015/12
• 顔器官の動きに基づく関心の有無の推定に関する検討  [Not invited]

  SASAKA YUUMA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• サッカー映像における試合内容の理解を促すデータの提示に関する検討  [Not invited]

  TAKAHASHI SHO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 楽曲中の歌唱区間推定における簡易脳波計の導入に関する検討  [Not invited]

  SUGATA KENTO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 胃X線画像を用いたHelicobacter pylori感染の高精度識別に関する検討(2)―撮像方向が識別に有効な画像領域に与える影響に対する考察―  [Not invited]

  TOGO REN, ISHIHARA KENTA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 個々の道路構造物に関する点検項目の導入による道路構造物の変状推定の高精度化に関する検討  [Not invited]

  MAEDA KEISUKE, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• Web映像検索のためWebコミュニティの提示法に関する検討―Webコミュニティに含まれるトピックを考慮した代表キーフレーズ抽出の試み―  [Not invited]

  TAKEHARA DAICHI, HARAKAWA RYOSUKE, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 道路構造物の変状画像に対する類似検索の高精度化に関する検討  [Not invited]

  TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 画像共有サービスにおける投稿画像の分類結果の信頼度推定に関する検討  [Not invited]

  SAITO NAOKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015/11
• 情報科学が繋ぐ生物とものづくり―バイオミメティクス・画像検索システムの理論と実現―  [Not invited]

  HASEYAMA MIKI

  高分子学会予稿集(CD-ROM)  2015/05
• A Note on Image Retrieval Based on Evaluation of Retrieval Results-An Application to Similarity Evaluation of Surface Structure of Organisms Taken by Scanning Electron Microscope-  [Not invited]

  TAKAHASHI AKIHIRO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Preference Estimation Method based on Relationship between Videos and Viewing Behavior: Reduction of Performance Degradation for Few Training Pairs  [Not invited]

  YAMAGUCHI YOSHIAKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Helicobacter Pylori Infection Detection from Gastric X-ray Images-Performance Comparison between Feature Fusion and Decision Level Fusion-  [Not invited]

  ISHIHARA KENTA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• Super-Resolution for Microscope Images Focusing on Imaging Procedure  [Not invited]

  HIROKAWA MARIKO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• Improving Performance Classification of Biological Organisms using Scanning Electron Microscope Images-Introducing Training Data selection By Focusing on the Image Acquisition Method-  [Not invited]

  PIAO JUN, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Exhibition of Hierarchical Structure of Web Communities for Web Video Retrieval-An Approach for Representative Keyword Extraction of Web Communities-  [Not invited]

  HARAKAWA RYOSUKE, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Accurate Image Classification Method of Image Sharing Sites  [Not invited]

  SAITO NAOKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A note on Accurate Estimation of a Light Source and a Spectral Reflectance by Introduction of Superpixel Calculation Method  [Not invited]

  IGARASHI YUTA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Data Analysis for Maintenance Inspection of Infrastructures-Quantification of Relationship between Inspection Data Using Distress Images and Inspection Results-  [Not invited]

  MIZOROGI YUYA, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A note on Web video retrieval based on network structure analysis using video features and tag information-Retrieval performance improvement using local learning regularization-  [Not invited]

  YOSHIDA SO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Classification of Individual Favorite Musical Pieces Utilizing EEG Signals during Listening to Music-Performance Improvement via CCA Considering Class Information-  [Not invited]

  SAWATA RYOSUKE, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Image Classification Using EEG Features during Watching Images (4)-Verification of Using Multiple Users' EEG Signals-  [Not invited]

  KAWAKAMI TAKUYA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2015/02
• A Note on Network Analysis Based Detection of Important Player and Similar Scenes in Soccer Videos  [Not invited]

  TAKAHASHI SHO, HASEYAMA MIKI

  映像情報メディア学会技術報告  2014/12
• トピックモデルに基づく音響特徴と時間情報を考慮した流行曲識別に関する検討  [Not invited]

  木下奨平, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• 維持管理業務の効率化に向けた点検データの類似度算出における特徴選択に関する検討  [Not invited]

  三改木裕矢, 高橋翔, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• パワースペクトログラム誤差最小化に基づく消失音響区間の再構成手法に関する検討  [Not invited]

  田中佑磨, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• 複数の画像特徴を用いたベイジアンネットワークに基づく構造物の変状の推定の高精度化に関する検討  [Not invited]

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• Web映像検索のための意味内容を考慮したWebコミュニティの階層構造抽出に関する一検討  [Not invited]

  竹原大智, 原川良介, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• 画像共有サイトへの投稿画像およびそのメタデータに注目したマルチモーダル画像検索手法に関する検討  [Not invited]

  斉藤直輝, 小川貴弘, 浅水仁, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• トピックモデルを用いた複数特徴の統合による画像群の可視化に関する検討  [Not invited]

  館農浩平, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• 視聴動作に基づくユーザの関心対象を考慮した映像に対する関心の有無の識別に関する検討  [Not invited]

  山口由晃, 小川貴弘, 浅水仁, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014/10
• バイオミメティクス・データ検索基盤と新材料開発  [Not invited]

  長谷山美紀

  高分子学会予稿集(CD-ROM)  2014/09
• バイオミメティクス・データ検索による生物形態の類縁性の可視化とその産業応用  [Not invited]

  長谷山美紀

  日本動物学会大会予稿集  2014/08
• メディア理解のための発想支援型検索の試み  [Not invited]

  長谷山美紀

  電子情報通信学会技術研究報告  2014/03
• 個人の嗜好を表現可能なトピックモデルの提案とその楽曲推薦への応用に関する検討  [Not invited]

  中西亮太, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 色品質改善のための拡張ジョイントバイラテラルフィルタ  [Not invited]

  和田直史, 数井誠人, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 可視および近赤外領域に感度を有するディジタルカメラを用いた分光反射率推定に関する検討―光源スペクトル推定の導入―  [Not invited]

  五十嵐祐太, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 画像注視時の脳波特徴量を用いた画像分類に関する検討(2)―単一試行の脳波から算出された特徴量の有効性の検証―  [Not invited]

  川上拓也, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 生物の顕微鏡像を用いた科の推定に関する検討 分類体系を表す決定木の改良による高精度化  [Not invited]

  PIAO Jun, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• Web映像コミュニティの階層構造抽出の大規模データ適用に関する一検討  [Not invited]

  原川良介, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• Hidden Conditional Random Fieldsによる映像の構造解析に基づくシーン分割の高精度化に関する検討  [Not invited]

  吉田壮, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• タグ付き画像を用いた同一意味を表現するタグの選択に関する検討  [Not invited]

  SONG Yan, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 動画像の事例ベース超解像のランク最小化に基づく高精度化に関する検討  [Not invited]

  廣川真梨子, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 昆虫の走査型電子顕微鏡像による科の分類に関する検討―画像特徴を用いた階層型クラスタリングによる高精度化―  [Not invited]

  高橋昌弘, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• ネットワーク分析を用いたサッカー選手の特徴定義とその選手間類似度算出への応用  [Not invited]

  岩井和也, 高橋翔, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• 橋梁の点検画像を用いたベイジアンネットワークによる変状の推定に関する検討  [Not invited]

  小林克希, 高橋翔, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• カルマンフィルタを用いたエラーコンシールメントの高精度化に関する検討~物体の運動を考慮した状態遷移モデルの実現~  [Not invited]

  大貫修平, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2014/02
• MISSING INTENSITY RESTORATION VIA PERCEPTUALLY OPTIMIZED SUBSPACE PROJECTION BASED ON ENTROPY COMPONENT ANALYSIS  [Not invited]

  OGAWA TAKAHIRO, HASEYAMA MIKI

  Proc IEEE Int Conf Acoust Speech Signal Process  2014
• Exhibition Method of Hierarchical Structure of Web Communities Using Community Density for Web Video Retrieval  [Not invited]

  HARAKAWA RYOSUKE, OGAWA TAKAHIRO, HASEYAMA MIKI

  Dig Tech Pap IEEE Int Conf Consum Electron  2014
• NOVEL IMAGE CLASSIFICATION BASED ON DECISION-LEVEL FUSION OF EEG AND VISUAL FEATURES  [Not invited]

  KAWAKAMI TAKUYA, OGAWA TAKAHIRO, HASEYAMA MIKI

  Proc IEEE Int Conf Acoust Speech Signal Process  2014
• 胃X線二重造影像を用いたHelicobacter pylori感染の識別の高精度化に関する検討  [Not invited]

  石原賢太, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013/10
• 脳波特徴に基づく人間の楽曲聴取時における感情推定に関する検討  [Not invited]

  澤田充奨, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013/10
• 判別分析に基づいた順序回帰の導入による協調フィルタリングでのSparsity Problemの軽減に関する一検討  [Not invited]

  川島孝行, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013/10
• ビッグデータからの価値創出―「気づき」を生み出す画像検索―  [Not invited]

  長谷山美紀

  電子情報通信学会技術研究報告  2013/09
• Error Resilient 3D DWT Video Transmission  [Not invited]

  OMURA Kotoku, YAMASAKI Shoichiro, MATSUSHIMA Tomoko, TANAKA Hirokazu, HASEYAMA Miki

  電子情報通信学会技術研究報告. SIS, スマートインフォメディアシステム  2013/09 

  The coauthors proposed the method, which adopted duplication scheme of lowest frequency sub-band coefficients of three-dimensional discrete wavelet transform (3D DWT). In this paper, we propose a scheme of duplicating the coefficients of the neighborhood of LL along with the LL, and an interpolation scheme that replaces erroneous coefficients by value of zero. The simulation evaluations over burst error channels compare the performances of the proposed method with those of the conventional one, and show the performance improvement of the proposed method.
• 進化する生物模倣の世界 次世代バイオミメティクス最前線 データベースで未知の機能を探す  [Not invited]

  長谷山美紀

  JST News  2013/08
• Performance Evaluation of 3D DWT Video Transmission with Invertible Codes  [Not invited]

  OMURA Kotoku, YAMASAKI Shoichiro, MATSUSHIMA K. Tomoko, TANAKA Hirokazu, HASEYAMA Miki

  IEICE technical report. Information theory  2013/07 

  The objective of this study is to indicate efficacy of error resilience using invertible codes to three-dimensional discrete wavelet transform (3D DWT) video coding. There have been many studies of video coding to apply 3D DWT. It is known that lowest frequency sub-band coefficients (LL coefficients) of 3D DWT outputs are significant impact on the visual quality of video. This led us construct error resilient video transmission method using invertible codes with capability of error correction. Two of the authors proposed the method, which was adopted duplication scheme of LL coefficients. Experimental results show that the proposed method is obtained improvement of performance compared with the conventionally proposed method.
• 工学的気づきを生み出すバイオミメティクスデータ検索基盤  [Not invited]

  長谷山美紀

  高分子夏季大学講演予稿集  2013/07
• ボール保持チームに注目したサッカー映像分割の高精度化に関する検討  [Not invited]

  大貫修平, 高橋翔, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2013/02
• 視聴動作から推定される関心度を用いた映像の推薦の高精度化に関する検討  [Not invited]

  白石哲夫, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2013/02
• 映像検索結果の定量評価手法の応用に関する試み(1)―札幌市における実証実験を通した映像検索システムの評価―  [Not invited]

  山之内豊, 小川貴弘, 長谷山美紀

  電子情報通信学会技術研究報告  2013/02
• サッカー映像におけるパス可能領域の高精度な推定手法に関する検討~選手配置の時刻変化を導入した新たな評価関数の定義~  [Not invited]

  高橋翔, 長谷山美紀

  電子情報通信学会技術研究報告  2013/02
• A study of visual features for detecting deer from road image sequencest  [Not invited]

  KATSURAI MARIE, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Classical Music Classification Based on Acoustic Features  [Not invited]

  MATSUNO KEIICHI, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A note on shoe sole image retrieval for footprints identification in forensic  [Not invited]

  HATAKEYAMA YASUTAKA, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Using Bayesian Networks for Analyzing Snow-removal Work Data  [Not invited]

  TAKAHASHI SHINTARO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Soccer Player Tracking Using Elastic Model-Performance Improvement Based on New Potential Energy-  [Not invited]

  IWAI KAZUYA, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Trend Detection in Popular Music Database  [Not invited]

  NAKANISHI RYOTA, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Scene Segmentation Method Based on Video Structures in Music Programs-Verifying the Effect of Face Recognition on Scene Segmentation Accuracy-  [Not invited]

  YOSHIDA SO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Evaluating Image Search Result Visualization Methods Based on Feature Integration Theory-Verifying Effectiveness from Comparison of Different Visualization Methods-  [Not invited]

  KOBAYASHI KATSUKI, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Classification of Helicobacter Pylori Infection using Double Contrast X-ray Image of Stomach  [Not invited]

  YOSHIZAKI AKANE, KATSURAI MARIE, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• A Note on Accurate Estimation of Pitcher's Condition in Baseball Videos: Improvement of Features Based on Pitching Motions and Pitching Results  [Not invited]

  KUBO JUNKI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI

  電子情報通信学会技術研究報告  2013/02
• 選手の移動速度を考慮したサッカー映像における3次元パス可能領域の推定に関する検討  [Not invited]

  高橋翔, 長谷山美紀

  映像情報メディア学会冬季大会講演予稿集(CD-ROM)  2012/11
• サッカーのチーム戦術推定手法を用いた試合映像の分割に関する検討  [Not invited]

  大貫修平, 高橋翔, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 歌詞と音響特徴を用いたファジィ分類による音楽ジャンル帰属度の定義に関する検討  [Not invited]

  中西亮太, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 画像検索結果の提示方法の有効性評価に関する検討―特徴統合理論に基づく評価指標の実現―  [Not invited]

  小林克希, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 視聴動作から推定される関心度を用いた映像の推薦に関する検討  [Not invited]

  白石哲夫, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 順序回帰を用いた数値予報の誤差解析における高精度化に関する検討  [Not invited]

  高橋信太郎, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• クラシック楽曲信号から算出される特徴量の改良と時代分類の高精度化に関する考察  [Not invited]

  松野恵一, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 連想型映像検索における検索過程の分析に関する一考察  [Not invited]

  山之内豊, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 映像のシーン分割における複数特徴量の協調利用の有効性に関する評価  [Not invited]

  宋妍, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• 物体認識における識別器の高精度化に関する検討 SVMを用いた物体領域の自動選定手法の導入  [Not invited]

  吉田壮, 小川貴弘, 長谷山美紀

  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012/10
• A note on similarity calculation between shoe sole images for forensics  [Not invited]

  畠山 泰貴, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A Note on Music Recommendation Based on Human Motion and Its Application  [Not invited]

  大串 裕幸, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A note on accurate deer detection from road images using movement direction  [Not invited]

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• Color Correction for Images Including Visible and Infrared Components : A note on Color Transfer based on Color Components of Natural Images  [Not invited]

  和泉 大佑, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A Note on Accuracy Improvement of Error Analysis in Numerical Weather Prediction  [Not invited]

  高橋 信太郎, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (ITS)  [Not invited]

  何 再興, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 犯罪捜査のための靴裏画像の類似度算出に関する検討 (画像工学)  [Not invited]

  畠山 泰貴, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 人間の動きに基づく楽曲推薦手法の実データ適用に関する検討 (画像工学)  [Not invited]

  大串 裕幸, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• アマチュアサッカーチームの試合映像における重要場面の検出 : コーチングを目的として映像に付与されたコメントの利用 (画像工学)  [Not invited]

  高橋 翔, 嶌田 聡, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A note on the application of Web information to near-duplicate online video detection (画像工学)  [Not invited]

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 視聴動作から推定される関心度を用いたユーザクラスタリングに関する検討 (画像工学)  [Not invited]

  白石 哲夫, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 移動方向を考慮した道路映像中のシカ検出の高精度化に関する検討 (画像工学)  [Not invited]

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 可視光成分と近赤外光成分が重畳された画像の色補正 : 自然画像の色成分を用いた色変換の検討 (画像工学)  [Not invited]

  和泉 大佑, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (画像工学)  [Not invited]

  何 再興, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• 数値予報における予報誤差の解析の高精度化に関する検討 (画像工学)  [Not invited]

  高橋 信太郎, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A method for detecting important events in soccer videos of amateur teams : An approach using tactical comments added to videos for soccer coaching  [Not invited]

  高橋 翔, 嶌田 聡, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A note on the application of Web information to near-duplicate online video detection (ITS)  [Not invited]

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• A Note on User Clustering Using Measurement of Levels of Interest Estimated from Viewing Behavior  [Not invited]

  白石 哲夫, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告 : 信学技報  2012/02
• On Speeding Up of the Dynamic Coalescence Model  [Not invited]

  KAITA Takeshi, HASEYAMA Miki, KITAJIMA Hideo, TOMITA Shingo

  電子情報通信学会技術研究報告. ITS  2012/02 

  A cluster analysis method modified of Dynamic Coalescence Model(DCM) on reducing the calculation complexity is proposed. The number of component distributions of which a mixed distribution consists is generally unknown. The shape of each of the components is complicate in the structure. And, one of the components has complicate relationships to the others. DCM can find such a component as a corresponding single cluster of points. However, the DCM involves so many calculations that it is not considered practical. The calculation time of the proposed, modified DCM is examined to be short than that of the original DCM.
• Dynamic Coalescence Model の高速化に関する一考察  [Not invited]

  海田 健, 長谷山 美紀, 北島 秀夫, 富田 眞吾

  電子情報通信学会技術研究報告. ITS  2012/02
• 電子レセプトからの医療資源を最も投入した傷病の推定の高精度化に関する検討  [Not invited]

  池田 啓典, 畠山 泰貴, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告. ITS  2012/02
• Dynamic Coalescence Model の高速化に関する一考察  [Not invited]

  海田 健, 長谷山 美紀, 北島 秀夫, 富田 眞吾

  電子情報通信学会技術研究報告. IE, 画像工学  2012/02
• 電子レセプトからの医療資源を最も投入した傷病の推定の高精度化に関する検討  [Not invited]

  池田 啓典, 畠山 泰貴, 小川 貴弘, 長谷山 美紀

  電子情報通信学会技術研究報告. IE, 画像工学  2012/02
• A Note on Accurate estimation method of Most Resource-consuming Diseases in Electronic Receipt Data  [Not invited]

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2012/02 

  This paper proposes a method for estimating the most resource-consuming diseases from electronic receipt data. The proposed method constructs the model that represents medical treatments for diseases in a electronic receipt as a mixture distribution of medical treatments for a disease, which are represented as a probability distribution over medical practices, medical drugs and special treatment materials, by using Labeled LDA, one of probabilistic topic models. Using this model, the proposed method can realize estimation of points for each disease in a electronic receipt, and accurate estimation of the most resource-consuming diseases becomes feasible.
• On Speeding Up of the Dynamic Coalescence Model  [Not invited]

  KAITA Takeshi, HASEYAMA Miki, KITAJIMA Hideo, TOMITA Shingo

  IEICE technical report. Image engineering  2012/02 

  A cluster analysis method modified of Dynamic Coalescence Model(DCM) on reducing the calculation complexity is proposed. The number of component distributions of which a mixed distribution consists is generally unknown. The shape of each of the components is complicate in the structure. And, one of the components has complicate relationships to the others. DCM can find such a component as a corresponding single cluster of points. However, the DCM involves so many calculations that it is not considered practical. The calculation time of the proposed, modified DCM is examined to be short than that of the original DCM.
• A Note on Accurate estimation method of Most Resource-consuming Diseases in Electronic Receipt Data  [Not invited]

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki

  電子情報通信学会技術研究報告. ITS  2012/02 

  This paper proposes a method for estimating the most resource-consuming diseases from electronic receipt data. The proposed method constructs the model that represents medical treatments for diseases in a electronic receipt as a mixture distribution of medical treatments for a disease, which are represented as a probability distribution over medical practices, medical drugs and special treatment materials, by using Labeled LDA, one of probabilistic topic models. Using this model, the proposed method can realize estimation of points for each disease in a electronic receipt, and accurate estimation of the most resource-consuming diseases becomes feasible.
• A note on similarity calculation between shoe sole images for forensics  [Not invited]

  畠山 泰貴, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A Note on Music Recommendation Based on Human Motion and Its Application  [Not invited]

  大串 裕幸, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A method for detecting important events in soccer videos of amateur teams : An approach using tactical comments added to videos for soccer coaching  [Not invited]

  高橋 翔, 嶌田 聡, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A Note on the Application of Web Information to Near-Duplicate Online Video Detection (メディア工学 ヒューマンインフォメーション 映像表現&コンピュータグラフィックス)  [Not invited]

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A Note on User Clustering Using Measurement of Levels of Interest Estimated from Viewing Behavior  [Not invited]

  白石 哲夫, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A note on accurate deer detection from road images using movement direction  [Not invited]

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• Color Correction for Images Including Visible and Infrared Components : A note on Color Transfer based on Color Components of Natural Images  [Not invited]

  和泉 大佑, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• On Speeding Up of the Dynamic Coalescence Model  [Not invited]

  海田 健, 長谷山 美紀, 北島 秀夫

  映像情報メディア学会技術報告  2012/02
• A Note on Accuracy Improvement of Error Analysis in Numerical Weather Prediction  [Not invited]

  高橋 信太郎, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (メディア工学 ヒューマンインフォメーション 映像表現&コンピュータグラフィックス)  [Not invited]

  何 再興, 小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2012/02
• A note on detection of important events based on relationship between player action and sound on stadium in soccer videos  [Not invited]

  KUBO Junki, TAKAHASHI Sho, OGAWA Takahiro, HASEYAMA Miki

  Proceedings of the ... ITE annual convention  2011/08 

  This paper presents a detection method of important events based on relationship between player action and sound on stadium in soccer videos. Generally, since player action and sound on stadium have high correlation in the important events, we realize the detection by using their relationship based on canonical correlation analysis.
• A Note on Restoration of Latent Image from Blurred Image : Introduction of Adaptive Parameter Setting in Prior Probability Distribution  [Not invited]

  YOSHIZAKI Akane, IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  Proceedings of the ... ITE annual convention  2011/08 

  This paper presents a method for blurred image restoration. The proposed method adopts object matching for PSF estimation and adaptive parameter setting of a prior probability distribution for latent image estimation. Then, accurate blur removal becomes feasible without suffering from ringing artifacts.
• A note on visualization of user preference based on purchase data analysis  [Not invited]

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A traffic flow measurement method using simple image processing and its application to real data  [Not invited]

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers  [Not invited]

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A note on accurate scene segmentation based on the MCMC method using object matching  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description  [Not invited]

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• Improvement of Flow Estimation using Multiple Meteorological Data  [Not invited]

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• A note on improvement of classification performance based on SVDD including target object detection scheme  [Not invited]

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• A note on a fast Web community extraction method using features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A note on keyword hierachy construction using visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A note on visualization of user preference based on purchase data analysis  [Not invited]

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A traffic flow measurement method using simple image processing and its application to real data  [Not invited]

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers  [Not invited]

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A note on accurate scene segmentation based on the MCMC method using object matching  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• A note on improvement of classification performance based on SVDD including target object detection scheme  [Not invited]

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  IEICE technical report  2011/02 

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description  [Not invited]

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• Improvement of Flow Estimation using Multiple Meteorological Data  [Not invited]

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• A note on a fast Web community extraction method using features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, HASEYAMA Miki

  IEICE technical report  2011/02 

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A note on keyword hierachy construction using visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2011/02 

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A note on visualization of user preference based on purchase data analysis  [Not invited]

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A traffic flow measurement method using simple image processing and its application to real data  [Not invited]

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers  [Not invited]

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A note on accurate scene segmentation based on the MCMC method using object matching  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description  [Not invited]

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• Improvement of Flow Estimation using Multiple Meteorological Data  [Not invited]

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• A note on improvement of classification performance based on SVDD including target object detection scheme  [Not invited]

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  ITE Technical Report  2011/02 

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• A note on a fast Web community extraction method using features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, HASEYAMA Miki

  ITE Technical Report  2011/02 

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A note on keyword hierachy construction using visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2011/02 

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory  [Not invited]

  REN Jie, NAGAI Nobuo, HASEYAMA Miki

  IEICE technical report. Image engineering  2011/02 

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory  [Not invited]

  REN Jie, NAGAI Nobuo, HASEYAMA Miki

  IEICE technical report  2011/02 

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory  [Not invited]

  REN Jie, NAGAI Nobuo, HASEYAMA Miki

  ITE Technical Report  2011/02 

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• Analyses on Interpolation Accuracy of Eigenspace-BPLP  [Not invited]

  TANAKA Akira, OGAWA Takahiro, HASEYAMA Miki, MIYAKOSHI Masaaki

  The Transactions of the Institute of Electronics, Information and Communication Engineers. A  2011/02 

  欠損領域を有する数値データの補間技術として, 固有空間BPLP(Back Projection for Lost Pixels)法, 及び, その改良手法が提案されている.これらの手法は, 所与のデータから切り出したブロックデータの主成分構造を利用して欠損領域を推定する手法であり, 主要な固有空間の次元等を適切に選択することによって, 効果的に欠損部を補間することができる.一方, 重要なパラメータの一つである, 主要な固有空間の次元の選択の指針はこれまで与えられていなかった.本論文では, 主成分分析に用いる分散共分散行列と欠損ブロックに対応する分散共分散行列が等しいという理想的な条件下では, 固有空間BPLP法の改良手法の, 期待二乗誤差最小の意味での最適解が古典的なウィーナーフィルタであることを指摘するとともに, 固有空間として全空間を用いた解が, 上で述べた最適解であるウィーナーフィルタによる解に帰着することを示し, 結果として固有空間の最適な次元がブロックの次元そのものであることを示す.また, 主成分分析に用いる分散共分散行列と欠損ブロックに対応する分散共分散行列が完全に一致しない場合についても考察し, 上記理想的な条件下同様, 固有空間として全空間を採用することが最適となる十分条件を与える.また, 当該十分条件を満たさない場合についても, 数値実験によりウィーナーフィルタによる解の優位性を確認する.
• 足跡を用いた男女識別--足跡の特徴量からSVMにより男女を識別する試み  [Not invited]

  浅水 仁, 長谷山 美紀

  画像ラボ  2011/01
• Human-Centric Navigation System Video Vortex for Video Retrieval  [Not invited]

  2011 IEEE International Conference on Consumer Electronics  2011
• Missing Intensity Interpolation Using a Kernel PCA-Based POCS Algorithm and Its Applications  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEEE Transactions on Image Processing  2011
• Linear Time Decoding of Real-Field Codes over High Error Rate Channels  [Not invited]

  IEEE International Conference on AcousticsSpeech, and Signal Processing 2011  2011
• Adaptive Reconstruction Method of Missing Textures Based on Perceptually Optimized Algotithm  [Not invited]

  IEEE International Conference on AcousticsSpeech, and Signal Processing 2011  2011
• Adaptive KPCA-Based Missing Texture Reconstruction Approach Including Classification Scheme via Difference Subspaces  [Not invited]

  IEEE International Conference on Image Processing(ICIP 2011)  2011
• Cross Low-dimension Pursuit for Sparse Signal Recovery from Incomplete Measurements Based on Permuted Block Diagonal Matrix  [Not invited]

  IEICE Trans. Fundamentals of Electronics, Communications and Computer Sciences  2011
• AI-1-1 Cross-Media Theory and Its Application to Information Retrieval  [Not invited]

  Haseyama Miki

  Proceedings of the Society Conference of IEICE  2010/08
• A study on accurate scene segmentation based on the MCMC method utilizing video structures  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. CST  2010/06 

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. CST  2010/06 

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A note on keyword relationship extraction based on visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. CST  2010/06 

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• Music Recommendation According to Human Motion Based on Kernel CCA  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. CST  2010/06 

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• CAS2010-21 A study on accurate scene segmentation based on the MCMC method utilizinff video structures  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2010/06 

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• CAS2010-20 A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2010/06 

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• CAS2010-19 A note on keyword relationship extraction based on visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2010/06 

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• CAS2010-12 Music Recommendation According to Human Motion Based on Kernel CCA  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2010/06 

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• A study on accurate scene segmentation based on the MCMC method utilizing video structures  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2010/06 

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2010/06 

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A note on keyword relationship extraction based on visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2010/06 

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• Music Recommendation According to Human Motion Based on Kernel CCA  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2010/06 

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• Music Recommendation According to Human Motion Based on Kernel CCA  [Not invited]

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2010/06 

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution  [Not invited]

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2010/06 

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A study on accurate scene segmentation based on the MCMC method utilizing video structures  [Not invited]

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2010/06 

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A note on keyword relationship extraction based on visual features and its application to image annotation  [Not invited]

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2010/06 

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• 6C-3 A study on Gender Identification using Shoe Print with Support Vector Machine  [Not invited]

  ASAMIZU Satoshi, HASEYAMA Miki

  全国大会講演論文集  2010/03
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibiliy in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments  [Not invited]

  MURAYAMA Shota, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description (SVDD) to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• An accurate scene segmentation method based on video structure analysis using PLSA : A note on definition of new shot similarity by sequence alignment  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  ITE Technical Report  2010/02 

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping (DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis (PLSA) to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping  [Not invited]

  TOKUMOTO Ryota, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• A note on player and ball tracking in volleyball videos and its application to event estimation  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method  [Not invited]

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2010/02 

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of a articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibiliy in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments  [Not invited]

  MURAYAMA Shota, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description (SVDD) to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• An accurate scene segmentation method based on video structure analysis using PLSA : A note on definition of new shot similarity by sequence alignment  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping (DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis (PLSA) to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping  [Not invited]

  TOKUMOTO Ryota, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• A note on player and ball tracking in volleyball videos and its application to event estimation  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method  [Not invited]

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2010/02 

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of a articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos  [Not invited]

  TAKAHASHI Sho, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibility in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments  [Not invited]

  MURAYAMA Shota, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description(SVDD)to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• An accurate scene segmentation method based on video structure analysis using PLSA : Anote on definition of new shot similarity by sequence alignment  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  IEICE technical report  2010/02 

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping(DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis(PLSA)to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping  [Not invited]

  TOKUMOTO Ryota, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• A note on player and ball tracking in volleyball videos and its application to event estimation  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method  [Not invited]

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2010/02 

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of an articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• An fast Web community extraction using LSH-based Web video features and its application  [Not invited]

  Workshop on Picture Coding and Image Processing, PCSJ2010/IMPS2010  2010
• Players tracking approach using level-set method based on color components of players in soccer videos  [Not invited]

  Workshop on Picture Coding and Image Processing, PCSJ2010/IMPS2010  2010
• A Note on Improvement of Coding Efficiency of Distributed Video Coding Based on Frame Reconstruction : Reconstructon of Key and Wyner-Ziv Frames Based on Error Correction  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/10 

  This paper proposes a new method which improves coding efficiency of distributed video coding (DVC). Before the encoding process of key frames, the proposed method adds missing areas to the key frames and reduces the bit rate of the target video sequence. In the decoder, the proposed method interpolates the missing areas and reconstructs Wyner-Ziv (WZ) frames and successive key frames based on error correction results of the WZ frames. Then, since decoded WZ frames are accurately reconstructed, and interpolation errors of the missing ares are effectively reduced, the proposed method can achieve higher rate-distortion (RD) performance than the traditional DVC. Experimental results show the improvement of the proposed method over the traditional DVC.
• A Note on Visualization of Video Retrieval Based on Perceptual Grouping  [Not invited]

  TOKUMOTO Ryota, HASEYAMA Miki

  ITE Technical Report  2009/10 

  This paper proposes a visualization method of video retrieval based on perceptual grouping. The perceptual grouping is a perceptual phenomenon which a human perceives figures as unities in eyeshot. We have previously proposed a video retrieval system for providing the videos desired by the users if users do not have specific queries. However, in traditional method, the interface is not considered a past search history. Thus, users cannot distinguish precisely between a similar video and a video provided newly from past search results. And, it becomes difficult to confirm the change of the video search results in a series of flows along the search history. Therefore, the proposed method arranges the videos in the three-dimensional space. Furthermore, we visualize the video retrieval results based on the perceptual grouping. Consequently, the effective visualization of the video retrieval results is discussed.
• A note on accurate scene segmentation by video structure analysis : An approach using new features obtained by PLSA  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  ITE Technical Report  2009/10 

  A method for accurate scene segmentation based on statistical trend analysis of audiovisual features is proposed. In the proposed method, based on the assumption that audiovisual features in the same broadcast programs include similar trends, their trends are analyzed by using Probabilistic Latent Semantic Analysis (PLSA), the method for statistical trend analysis. Then the proposed method calculates new features representing these trends. By using these new features, degradation of scene segmentation accuracy, which results from the direct use of the features obtained from the audiovisual materials, can be avoided and thereby accurate scene segmentation becomes feasible. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A note distance between music pieces based on personal preferences and its application to music retrieval  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  ITE Technical Report  2009/10 

  This paper proposes a new calculation method of distance between music pieces based on personal preferences. In the proposed method, in order to utilize musical features for the distance calculation, feature vectors which represent melody, rhythm, and timbre, are firstly calculated from traning data sets. Then, by using the feature vectors, a metric having a general form of the Mahalanobis distance is defined and learned. Specifically, the metric is learned by minimizing a cost function designed so that the cost increases when the distance between dissimilar music pieces is shorter than the distance between similar music pieces. According to the distance based on the metric, the accurate retrieval of similar music can be expected.
• A Note on SVDD-Based Automatic Target Object Extraction for Image Retrieval  [Not invited]

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/08 

  This paper proposes an automatic extraction method of target objects from images based on the support vector data description (SVDD) for image retrieval. From training images including and not including target objects, the proposed method enables their automatic extraction. Specifically, probabilities which feature vectors of local areas belong to the object class is defined based on distances from center of the hypersphere, and positive areas including target objects and negative areas not including these objects can be obtained from the training images based on the probabilities. In this scheme, in order to obtain the hypersphere which can separate the positive areas and the negative ones more accurately, the feature vectors are selected adaptively. Furthermore, by iterating the calculation of the hypersphere and the extraction of the positive areas and the negative ones in various sizes, the detection of target objects in images can be realized without manually extracting areas of target objects. Therefore, image retrieval based on object extraction can be expected.
• A Note on Accurate Scene Segmentation Utilizing Speech Recognition : Introduction of Video Structure Analysis Using Object Matching  [Not invited]

  NARITA Yo, HASEYAMA Miki

  ITE Technical Report  2009/08 

  This paper proposes an improvement method for scene segmentation which uses new features defined from speech recognition result. In the proposed method, the new features are defined based on the appearance distribution of terms in fixed-length short segments of video materials. Furthermore, similarities between two segments are calculated from the obtained features as similarities of topics. Then, the changes of topics are detected from the time transition of the similarities, and the problem of over segmentations caused in the traditional methods can be alleviated by combining both their methods and our method. Finally, the proposed method realizes accurate scene segmentations based on the topics of video materials.
• Visualization interface for video scene retrieval  [Not invited]

  Haseyama Miki

  Technical report of IEICE. PRMU  2009/05 

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• Visualization interface for video scene retrieval  [Not invited]

  Haseyama Miki

  IEICE technical report. Image engineering  2009/05 

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• Visualization interface for video scene retrieval  [Not invited]

  Haseyama Miki

  IEICE technical report.  2009/05 

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method  [Not invited]

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/02 

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error (MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR (Peak Signal-to-Noise Ratio) and visual quality.
• A study on tracking player position and taking camera focusing area in soccer video  [Not invited]

  LIM Wonkuk, HASEYAMA Miki

  ITE Technical Report  2009/02 

  In this paper, we propose particle filters for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/02 

  In this paper, an ER (Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution (LR) image as that of a high-resolution (HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake  [Not invited]

  TAKIZAWA Makoto, HASEYAMA Miki

  ITE Technical Report  2009/02 

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/02 

  This paper presents a support vector data description (SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution (HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution (LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  ITE Technical Report  2009/02 

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2009/02 

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method  [Not invited]

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2009/01 

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error (MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR (Peak Signal-to-Noise Ratio) and visual quality.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  IEICE technical report  2009/01 

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2009/01 

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A study on tracking player position and taking camera focusing area in soccer video  [Not invited]

  LIM Wonkuk, HASEYAMA Miki

  IEICE technical report  2009/01 

  In this paper, we propose particle filters, for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2009/01 

  In this paper, an ER (Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution (LR) image as that of a high-resolution (HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake  [Not invited]

  TAKIZAWA Makoto, HASEYAMA Miki

  IEICE technical report  2009/01 

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2009/01 

  This paper presents a support vector data description (SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution (HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution (LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method  [Not invited]

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error(MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR(Peak Signal-to-Noise Ratio) and visual quality.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA  [Not invited]

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A study on tracking player position and taking camera focusing area in soccer video  [Not invited]

  LIM Wonkuk, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  In this paper, we propose particle filters, for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  In this paper, an ER(Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution(LR) image as that of a high-resolution(HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake  [Not invited]

  TAKIZAWA Makoto, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2009/01 

  This paper presents a support vector data description(SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution(HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution(LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A note on spatio-temporal resolution enhancement of video sequences based on super-resolution reconstruction : Introduction of clustering scheme based on feature point correspondence  [Not invited]

  TAKIZAWA Makoto, HASEYAMA Miki

  ITE Technical Report  2008/10 

  This paper proposes a new spatio-temporal resolution enhancement method of video sequences based on super-resolution reconstruction. The proposed method uses obtained homography matrix's parameters based on feature points correspondence for calculating the pixel correspondence between successive frames. In this procedure, the proposed method clusters the feature points correspondence and calculates the homography matrix's parameters from the feature points correspondence in the largest cluster. By using the obtained parameters, camera parameters are calculated accurately. Furthermore, from the camera parameters, we estimates the pixel correspondence between the successive frames and the interpolated one. With the obtained motion vector, we apply the super-resolution reconstruction. Consequently, we can obtain an accurate high-resolution and high frame rate video sequence. Simulation results are shown to confirm the high performance of the proposed method.
• A note on Similarity Evaluation of Images in Web Pages Based on Link Analysis  [Not invited]

  MURAYAMA Shota, HASEYAMA Miki

  ITE Technical Report  2008/10 

  This paper proposes a new similarity evaluation method for images in Web pages based on link analysis. The proposed method assumes that there are areas which contain common contents and have similar image features in the images, whose Web pages are connected by links. Based on the assumption, the proposed method defines similarity measure of the images. Consequently, with the similarity measure, the problem in the traditional methods of being affected by the image features of different contents can be solved, and an effective image retrieval is realized.
• A Note on Flow Estimation based on Particle Filter : Introduction of Helmholtz Decomposition Theorem into Observation Model  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  ITE Technical Report  2008/10 

  This paper proposes an effective flow estimation method based on a Helmholtz decomposition theorem with a particle filter. In a traditional method based on the Helmholtz decomposition theorem, since it does not consider translation and estimation errors included in previous flows, its performance tends to become degraded. Therefore, the proposed method extends the Helmholtz decomposition theorem with the translation. By using the extended theorem, the flow estimation considering the translation can be realized. Furthermore, the proposed method regards the flows as state variables in the particle filter. Then, the proposed method defines a process that the flows of the pixels in the current frame are provided from the ones of pixels in the previous frame as a state transition model. The proposed method defines the following two processes as an observation model. One is that intensities including additive noise in the current frame are generated from the flows of the pixels in the current frame. The other one is that the flows of the pixels in the current frame are decomposed into rotational and divergent components including the additive noise. By using the particle filter based on these models, the proposed method can estimate the flows of the pixels in the current frame in such a way that the estimation errors included in the previous flows do not affect the flow estimation scheme. Consequently, an accurate flow estimation can be achieved.
• Estimating Number of People from a Video Sequence via Geometrical Model  [Not invited]

  ARAI Hiroyuki, MIYAGAWA Isao, KOIKE Hideki, HASEYAMA Miki

  ITE Technical Report  2008/06 

  We propose a novel technique for estimating the number of people in a video sequence ; it has the advantages of being stable even in crowded situations and needing no ground-truth data. By analyzing the geometrical relationships between image pixels and their intersection volumes in the real world quantitatively, a foreground image can be directly indicate the number of people. Because foreground detection can be done even in crowded situations, the proposed method can be applied to such situations. Also it can estimate the number of people in an a-priori manner, so it needs no ground-truth data which is necessary for existing feature-based estimating techniques. Experiments show the validity of the proposed method.
• A Study on Frequency Domain Error Correction Scheme for MC-CDMA  [Not invited]

  TANAKA Hirokazu, JANG Kyung Woon, YAMASAKI Shochiro, HASEYAMA Miki

  IEICE technical report  2008/06 

  Recently, multi-code MC CDMA has a lot of attention in wireless multimedia communications because of its high bit rate performance. On the other hand, considering hardware implementation, it requires high power amplifier (HPA) with highly linear characteristics. HPA nonlinearity causes high out-of-band spurious emission and high peak to average power ration (PAPR). In order to solve these problems, employing bi-orthogonal modulation instead of multi-code modulation has been studied. As for the error correction coding of the MC-CDMA with bi-orthogonal modulation, the probability of the erroneous Hadamard code selection at receiver is critical for the BER performance compared to the sign bit error. In this paper, we propose an efficient error correcting scheme for the code selection error employing Reed-Solomon (RS) code. Additionally, an efficient decoding structure that can also correct the uncoded sign bit error using feed back RS decoded codeword is proposed. The basic performance of the proposed scheme was evaluated by both theoretical analysis and computer simulations.
• A Study on Frequency Domain Error Correction Scheme for MC-CDMA  [Not invited]

  TANAKA Hirokazu, JANG KyungWoon, YAMASAKI Shochiro, HASEYAMA Miki

  電子情報通信学会技術研究報告. SIS, スマートインフォメディアシステム  2008/06
• A note on player tracking by using level-set method based on color component in soccer video  [Not invited]

  LIM Wonkuk, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper proposes an accurate player tracking method in soccer video using a level-set method. The proposed method introduces features of baundary surfaces and those of color components, which compose the player's uniforms, into the evaluation function of the level-set method. Then, our method realizes the accurate tracking, that can distinguish the players in different teams.
• A note on improvement of human detection based on multiple detectors  [Not invited]

  KAGA Yosuke, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper presents a human detection method based on multiple detectors. In the proposed method, a human detection result is obtained by voting the results of a pedestrian detector and a face detector in 2-dimensional plane. Then, by integrating the multiple detection results of the different detectors into a single human detection result, the accurate human detection can be realized compared to each single detector. Experimental results performed by applying the proposed method to actual images are shown to verify its effectiveness.
• A note on adaptive interpolation filter design method based on clustering using image features  [Not invited]

  MATSUNO Takaya, KONDO Satoshi, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper presents an adaptive interpolation filter design method based on classification using image features. The proposed method classifies macroblocks into some clusters representing the interpolation directions based on their motion vectors. For each cluster, the optimal interpolation filters are designed in the least-squares sense. Introducing the classification of the macroblocks into the filter design scheme, our method considers the difference between the horizontal and vertical correlation and can design the optimal filters. Thus, by using these filters for motion compensated prediction, the total bit rate with the same PSNR value is reduced compared to the conventional methods.
• A note on video frame interpolation based on image morphing : An accurate frame segmentation used in adaptive correspondence estimation  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper presents an image morphing based interpolation method of video frames using an adaptive correspondence estimation. The proposed method estimates the pairs of points which represent the correspondence of image features between two successive frames. Based on these pairs of points, the interpolation frames are calculated by using the image morphing. In the estimation scheme of pairs of points, the proposed method segments each frame into background and foreground regions. Furthermore, in order to estimate the foreground and background regions more accurately, the verification of the segmentation results is performed by using the previous interpolation frame and its segmentation results. Then, in each region, the pairs of points which represent the correspondence of the image features between the two successive frames is estimated adaptively. Furthermore, in each region, the interpolation frames are calculated by using the appropriate scheme, and the intensities of the interpolation frames are obtained accurately around boundaries between foreground and background regions. Consequently, the proposed method can generate interpolation frames successfully.
• A note on Error Concealment Using Kalman Filter : Realization of State Transition Model Representing Video Decoding Process  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, TANAKA Hirokazu, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper proposes an accurate error concealment technique using a Kalman filter. In order to realize the error concealment, the proposed technique regards the intensities in each frames as elements of the state variable of the Kalman filter and designs the following state transition and observation models. First, the state transition model is designed by using motion vectors and represents the video decoding process. Further, the observation model is designed as a new degradation model and enables the Kalman gain to be calculated for the error concealment. Then, the compensation of the estimation results by the state transition becomes feasible. Consequently, by utilizing these new state transition and observation models, the accurate error concealment based on the rule of least squares can be realized. Experimental results show that the proposed technique has better performance than that of the traditional techniques.
• A note on player tracking by using level-set method based on color component in soccer video  [Not invited]

  LIM Wonkuk, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper proposes an accurate player tracking method in soccer video using a level-set method. The proposed method introduces features of baundary surfaces and those of color components, which compose the player's uniforms, into the evaluation function of the level-set method. Then, our method realizes the accurate tracking, that can distinguish the players in different teams.
• A note on improvement of human detection based on multiple detectors  [Not invited]

  KAGA Yosuke, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper presents a human detection method based on multiple detectors. In the proposed method, a human detection result is obtained by voting the results of a pedestrian detector and a face detector in 2-dimensional plane. Then, by integrating the multiple detection results of the different detectors into a single human detection result, the accurate human detection can be realized compared to each single detector. Experimental results performed by applying the proposed method to actual images are shown to verify its effectiveness.
• A note on adaptive interpolation filter design method based on clustering using image features  [Not invited]

  MATSUNO Takaya, KONDO Satoshi, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper presents an adaptive interpolation filter design method based on classification using image features. The proposed method classifies macroblocks into some clusters representing the interpolation directions based on their motion vectors. For each cluster, the optimal interpolation filters are designed in the least-squares sense. Introducing the classification of the macroblocks into the filter design scheme, our method considers the difference between the horizontal and vertical correlation and can design the optimal filters. Thus, by using these filters for motion compensated prediction, the total bit rate with the same PSNR value is reduced compared to the conventional methods.
• A note on video frame interpolation based on image morphing : An accurate frame segmentation used in adaptive correspondence estimation  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper presents an image morphing based interpolation method of video frames using an adaptive correspondence estimation. The proposed method estimates the pairs of points which represent the correspondence of image features between two successive frames. Based on these pairs of points, the interpolation frames are calculated by using the image morphing. In the estimation scheme of pairs of points, the proposed method segments each frame into background and foreground regions. Furthermore, in order to estimate the foreground and background regions more accurately, the verification of the segmentation results is performed by using the previous interpolation frame and its segmentation results. Then, in each region, the pairs of points which represent the correspondence of the image features between the two successive frames is estimated adaptively. Furthermore, in each region, the interpolation frames are calculated by using the appropriate scheme, and the intensities of the interpolation frames are obtained accurately around boundaries between foreground and background regions. Consequently, the proposed method can generate interpolation frames successfully.
• A note on Error Concealment Using Kalman Filter : Realization of State Transition Model Representing Video Decoding Process  [Not invited]

  TAKAHASHI Shigeki, OGAWA Takahiro, TANAKA Hirokazu, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper proposes an accurate error concealment technique using a Kalman filter. In order to realize the error concealment, the proposed technique regards the intensities in each frames as elements of the state variable of the Kalman filter and designs the following state transition and observation models. First, the state transition model is designed by using motion vectors and represents the video decoding process. Further, the observation model is designed as a new degradation model and enables the Kalman gain to be calculated for the error concealment. Then, the compensation of the estimation results by the state transition becomes feasible. Consequently, by utilizing these new state transition and observation models, the accurate error concealment based on the rule of least squares can be realized. Experimental results show that the proposed technique has better performance than that of the traditional techniques.
• A note on semantic feature estimation method of still images using kernel PCA : An approach for similar image retrieval based on semantic concepts  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper presents a semantic feature estimation method using kernel PCA in order to realize semantic image retrieval. The proposed method performs keyword-based clustering of database images and derives a new map by using a nonlinear eigenspace of visual and semantic features in each cluster. This map provides accurate semantic features for the images belonging to each cluster from their visual features. Therefore, the proposed method adaptively selects the optimal cluster including a query image, whose keywords are unknown, by monitoring the errors of the visual features caused in its semantic feature estimation process. Then, since the semantic features of the query image can be successfully estimated, semantic image retrieval for the query image can be easily achieved.
• A note on scene segmentation based on video structure : An approach by combination of audio and visual processing  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper proposes video scene segmentation method utilizing shot-based repetitive structures of video and audio signals. In the proposed method, two repetitive structures are respectively obtained by clustering shots based on their video and audio features. Specifically, these obtained structures are represented as directed graphs whose nodes are clusters containing similar shots and edges are formed based on time transition of shots. Thus, by calculating common cut edges in both of the two constructed graphs, video materials are effectively segmented into scenes from video and audio signals. Then, the proposed method can avoid over scene segmentation caused in the conventional method using only video signals.
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval  [Not invited]

  MICHIYAMA Daigo, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper proposes a method to calculate scene dissimilarities utilizing audio and visual signals for video retrieval. By utilizing Dynamic Time Warping, the proposed method calculates the scene dissimilarities by selecting the most similar features, which are selected from visual, audio, and their integrated features, and these selectings are perfomed between two shots in scenes. Furthermore, a new feature space which can match audio and visual features defined by utilizing canonical correlation analysis is introduced into these selectings. In conclusion, the proposed method can appropriately calculate the scene dissimilarities based on similar features between scenes. Experimental results performed by applying the proposed method to scene retrieval in actual video data are shown to verify its effectiveness.
• A note on music extraction based on similarity of melody line  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  IEICE technical report. Image engineering  2008/02 

  This paper verifies performance of a similar music extraction using the similarity measure computation method, which we have already proposed. This similarity measure is computed by comparing the melody lines of two different music pieces. In order to compare the melody lines between two music pieces, our method adopts Dynamic Time Warping (DTW). According to this similarity measure, the accurate extraction of similar music pieces can be expected. However, since this scheme had not been applied to a large music database yet, it was not sufficient to verify its performance for actual use. Therefore, this paper applies the proposed method to the larger database containing 284 music pieces and evaluates its extraction performance. Furthermore, some comparisons between our method and the traditional method are performed to evaluates its extraction performance.
• A note on semantic feature estimation method of still images using kernel PCA : An approach for similar image retrieval based on semantic concepts  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper presents a semantic feature estimation method using kernel PCA in order to realize semantic image retrieval. The proposed method performs keyword-based clustering of database images and derives a new map by using a nonlinear eigenspace of visual and semantic features in each cluster. This map provides accurate semantic features for the images belonging to each cluster from their visual features. Therefore, the proposed method adaptively selects the optimal cluster including a query image, whose keywords are unknown, by monitoring the errors of the visual features caused in its semantic feature estimation process. Then, since the semantic features of the query image can be successfully estimated, semantic image retrieval for the query image can be easily achieved.
• A note on scene segmentation based on video structure : An approach by combination of audio and visual processing  [Not invited]

  YAMAMOTO Makoto, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper proposes video scene segmentation method utilizing shot-based repetitive structures of video and audio signals. In the proposed method, two repetitive structures are respectively obtained by clustering shots based on their video and audio features. Specifically, these obtained structures are represented as directed graphs whose nodes are clusters containing similar shots and edges are formed based on time transition of shots. Thus, by calculating common cut edges in both of the two constructed graphs, video materials are effectively segmented into scenes from video and audio signals. Then, the proposed method can avoid over scene segmentation caused in the conventional method using only video signals.
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval  [Not invited]

  MICHIYAMA Daigo, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper proposes a method to calculate scene dissimilarities utilizing audio and visual signals for video retrieval. By utilizing Dynamic Time Warping, the proposed method calculates the scene dissimilarities by selecting the most similar features, which are selected from visual, audio, and their integrated features, and these selectings are perfomed between two shots in scenes. Furthermore, a new feature space which can match audio and visual features defined by utilizing canonical correlation analysis is introduced into these selectings. In conclusion, the proposed method can appropriately calculate the scene dissimilarities based on similar features between scenes. Experimental results performed by applying the proposed method to scene retrieval in actual video data are shown to verify its effectiveness.
• A note on music extraction based on similarity of melody line  [Not invited]

  KOBAYASHI Kazuya, HASEYAMA Miki

  IEICE technical report  2008/02 

  This paper verifies performance of a similar music extraction using the similarity measure computation method, which we have already proposed. This similarity measure is computed by comparing the melody lines of two different music pieces. In order to compare the melody lines between two music pieces, our method adopts Dynamic Time Warping (DTW). According to this similarity measure, the accurate extraction of similar music pieces can be expected. However, since this scheme had not been applied to a large music database yet, it was not sufficient to verify its performance for actual use. Therefore, this paper applies the proposed method to the larger database containing 284 music pieces and evaluates its extraction performance. Furthermore, some comparisons between our method and the traditional method are performed to evaluates its extraction performance.
• A note on player tracking by using level-set method based on color component in soccer video  [Not invited]

  林 原局, 長谷山 美紀

  ITE technical report  2008/02
• A note on improvement of human detection based on multiple detectors  [Not invited]

  加賀 陽介, 長谷山 美紀

  ITE technical report  2008/02
• A note on adaptive interpolation filter design method based on clustering using image features  [Not invited]

  松野 孝也, 近藤 敏志, 長谷山 美紀

  ITE technical report  2008/02
• A note on video frame interpolation based on image morphing: an accurate frame segmentation used in adaptive correspondence estimation  [Not invited]

  山本 貴史, 長谷山 美紀

  ITE technical report  2008/02
• A note on semantic feature estimation method of still images using kernel PCA: an approach for similar image retrieval based on semantic concepts  [Not invited]

  小川 貴弘, 長谷山 美紀

  ITE technical report  2008/02
• A note on scene segmentation based on video structure: an approach by combination of audio and visual processing  [Not invited]

  山本 誠, 長谷山 美紀

  ITE technical report  2008/02
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval  [Not invited]

  道山 大悟, 長谷山 美紀

  ITE technical report  2008/02
• A note on music extraction based on similarity of melody line  [Not invited]

  小林 和也, 長谷山 美紀

  ITE technical report  2008/02
• System for Providing Information about Emergency Rescue Procedures to Mobile Phone Users  [Not invited]

  Ogawa Takahiro, Sakuma Daisuke, Shiraishi Shin'ichi, Haseyama Miki

  The journal of the Institute of Image Information and Television Engineers  2007/12 

  We propose the use of a system that provides information about emergency rescue procedures for mobile phone users. To enable the users to easily understand the procedures, avatars demonstrate how they are used. Since the avatars are efficiently constructed on a subset of scalable vector graphics (SVG), they can be quickly transmitted to the mobile phone and with low computational loads. In addition to the avatars, complementary voice and text data that explain the procedures are also transmitted via a synchronized multimedia integration language (SMIL, pronounced "smile") format. Moreover, to verify the performance and the effectiveness of the proposed system, we implemented a dedicated SMIL player suitable for use with mobile phones.
• An Improved Third-order Discrimination System for Two-dimensional Point Distribution Using Its Topology and Similar Distribution Estimated Statistically  [Not invited]

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo

  Technical report of IEICE. PRMU  2007/11 

  We propose an improved third-order system for the discrimination of two-dimensional point distributions. This has training pattern creator to transpose a training pattern into its representative distribution that expresses relative location invariance feature of the pattern. Its feature extractor uses second-order feature that expresses topology of combinations of two points without redundancy. This transposes a training pattern into its feature vectors that express its similar patterns estimated statistically and its statistical feature. Discrimination accuracy by its discriminator of single-layered perceptron is high even if there is only one training pattern.
• Finding Representative of Cluster at Invariant Location by Discretized Clustering Method Dynamic Coalescence Model  [Not invited]

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo

  Technical report of IEICE. PRMU  2007/11 

  We propose a discreted clustering method that founds representative of cluster found by the clustering method Dynamic Coalescence Model (DCM) at invariant location. In our proposed method, points are expressed by particles, and they are moved by their acceleration calculated by its gravity force from other particles. So, representative of cluster is founded near the centre of gravity of the cluster. Availability of our method to projection, region partition, characters recognition, etc. is pointed out.
• An Improved Third-order Discrimination System for Two-dimensional Point Distribution Using Its Topology and Similar Distribution Estimated Statistically  [Not invited]

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo

  Technical report of IEICE. Multimedia and virtual environment  2007/11 

  We propose an improved third-order system for the discrimination of two-dimensional point distributions. This has training pattern creator to transpose a training pattern into its representative distribution that expresses relative location invariance feature of the pattern. Its feature extractor uses second-order feature that expresses topology of combinations of two points without redundancy. This transposes a training pattern into its feature vectors that express its similar patterns estimated statistically and its statistical feature. Discrimination accuracy by its discriminator of single-layered perceptron is high even if there is only one training pattern.
• Finding Representative of Cluster at Invariant Location by Discretized Clustering Method Dynamic Coalescence Model  [Not invited]

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo

  Technical report of IEICE. Multimedia and virtual environment  2007/11 

  We propose a discreted clustering method that founds representative of cluster found by the clustering method Dynamic Coalescence Model (DCM) at invariant location. In our proposed method, points are expressed by particles, and they are moved by their acceleration calculated by its gravity force from other particles. So, representative of cluster is founded near the centre of gravity of the cluster. Availability of our method to projection, region partition, characters recognition, etc. is pointed out.
• Development of shoeprint images acquisition system for a gender determination  [Not invited]

  浅水 仁, 長谷山 美紀

  ITE technical report  2007/10
• A note on spatial and temporal resolution enhancement of video sequences using super-resolution reconstruction  [Not invited]

  瀧澤 誠, 長谷山 美紀

  ITE technical report  2007/10
• A note on coding of coefficients in adaptive interpolation filter based on vector quantization  [Not invited]

  松野 孝也, 近藤 敏志, 長谷山 美紀

  ITE technical report  2007/10
• Removal of impulse noise with iterated function system: an approach for improvement using regional division of a block  [Not invited]

  西村 浩哉, 小川 貴弘, 長谷山 美紀

  ITE technical report  2007/10
• A note on adaptive resolution enhancement of still images using Kernel PCA  [Not invited]

  小川 貴弘, 長谷山 美紀

  ITE technical report  2007/10
• A note on detection of rotational and divergent structure in still image based on Helmoltz decomposition  [Not invited]

  覚幸 典弘, 小川 貴弘, 長谷山 美紀

  ITE technical report  2007/10
• A note on accurate restoration of in-vehicle camera images in foggy conditions : Application of fog deterioration model to Kalman filter  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2007/06 

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• A note on accurate restoration of in-vehicle camera images in foggy conditions : Application of fog deterioration model to Kalman filter  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2007/06 

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• A note on accurate restoration of in-vehicle camera images in foggy conditions : Application of fog deterioration model to Kalman filter  [Not invited]

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2007/06 

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• A Study on Precise Player Localization in a Basketball Video  [Not invited]

  SAKUMA Daisuke, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes a new method for precise player localization in basketball video images. In the proposed method, the positions of the players on a basketball court are estimated by a Mean-Shift method that tracks moving objects using color histograms. Further, the proposed method introduces a new scheme, which identifies the players who are affected by occlusions, into the Mean-Shift method. Then, our method searches these players in the subsequent frames by using the color histograms that are obtained from their regions in the previous frames. Consequently, the robust player tracking can be realized, so that our method can estimate the positions of the players precisely. In this paper, experimental results are shown to verify the high performance of the proposed method.
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes a new method for video frame interpolation. In the proposed method, the video frames are interpolated by image morphing. For accurate interpolation using the image morphing, it is effective to consider the characteristic of the objects in the frames. Thus, the proposed method introduces a new scheme that estimates the correspondence between consecutive frames based on the feature points extracted every frame. By using this scheme, the accurate video frame interpolation can be realized. Finally, simulation results are shown to demonstrate the validity of the proposed method.
• A Study on Precise Player Localization in a Basketball Video  [Not invited]

  SAKUMA Daisuke, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes a new method for precise player localization in basketball video images. In the proposed method, the positions of the players on a basketball court are estimated by a Mean-Shift method that tracks moving objects using color histograms. Further, the proposed method introduces a new scheme, which identifies the players who are affected by occlusions, into the Mean-Shift method. Then, our method searches these players in the subsequent frames by using the color histograms that are obtained from their regions in the previous frames. Consequently, the robust player tracking can be realized, so that our method can estimate the positions of the players precisely. In this paper, experimental results are shown to verify the high performance of the proposed method.
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes a new method for video frame interpolation. In the proposed method, the video frames are interpolated by image morphing. For accurate interpolation using the image morphing, it is effective to consider the characteristic of the objects in the frames. Thus, the proposed method introduces a new scheme that estimates the correspondence between consecutive frames based on the feature points extracted every frame. By using this scheme, the accurate video frame interpolation can be realized. Finally, simulation results are shown to demonstrate the validity of the proposed method.
• A Note on Improvement of HITS using Clusters based on Link Analysis  [Not invited]

  TAKEYOSHI Tomoya, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes an effective Web page scoring method based on HITS, which is traditionally used for Web page scoring using a link analysis. Recently, some Web page scoring methods that modify the HITS have been proposed. However, they cannot solve the problem that many Web pages including almost the same links as one page and the pages linked by them have high scores. Therefore, the proposed method previously classifies the target Web pages based on similarity of link structure. Then each cluster tends to contain those controversial pages. Thus, our method introduces weights, that assign lower values to the links of the Web pages included in the same clusters than the others, into the HITS and calculates the scores. Consequently, the proposed method can solve the problem of the traditional schemes, and satisfies search requests of various users enough. The experimental results show the high performance of the proposed scoring method.
• An Accurate IFS-based Image Enlargement Method using Edge Continuity  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS). The proposed method introduces a line process into the IFS-based enlargement procedures to solve edge discontinuity of an enlarged image in the traditional IFS schemes. The line process can detect the edges which retain continuity. However, the line process cannot apply to the enlarged image whose intensities are not known. Therefore, the proposed method estimates the enlarged image's edges from the target image's edges obtained by using the line process. Then, since the estimated edges retain the continuity, the proposed method can calculate the enlarged image in such a way that its edges retain the continuity. Consequently, an accurate image enlargement can be achieved.
• A Note on the similarity between music audio signals using the melody line and the rhythm  [Not invited]

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes a novel similarity between music audio signals using rhythmic contents. The proposed method estimates the rhythmic content utilizing the autocorrelation functions of the signal. Then, by applying DTW to the functions, the similarity between music audio signals can be appropriately computed. Furthermore, our method visualizes the similarity using both melody line and rhythm to try to easily retrieve the similar music audio signals.
• A Study on Group Tactics Analysis in Soccer Videos  [Not invited]

  KON Hirofumi, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper presents a method for group tactics analysis in soccer videos. For this analysis, the proposed method classifies players into some groups. As a result of this classification, the players in one group are closely related in terms of soccer tactics. In the proposed method, the relations between the players in the groups are analyzed by using player positions. Based on these relations, the proposed method estimates group advantage, which represents the degree of team's superiority or inferiority in each group. The group advantage helps viewers to understand game situations.
• A Study on Accurate Image Clustering Using Color Edges  [Not invited]

  WATANABE Takashi, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes a similar image clustering method. We have proposed a image clustering scheme using image structures and quadratic form distances between color histograms. However, since this scheme cannot effectively utilize the distribution of colors within the target images, misclassification of these images is inevitable. In order to utilize these features more effectively, the proposed method classifies the images by the K-means algorithm based on distances between color correlograms, which can be utilized for describing the relationship of colors to spatial distances. Furthermore, by reconstructing clusters based on colors at the edges and their neighborhoods, the accurate image clustering can be realized.
• A Note on Improvement of Similar Image Clustering Method using Neocognitron : Introduction of Novel Structure for Extracting Color Features  [Not invited]

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Image engineering  2007/02 

  This paper proposes a new image clustering method. We have proposed an image clustering method using a neocognitron which is one of the neural networks. In this paper, we use the edge features of the images as the input stimulus of the neocognitron. However, since this method notes only the edge features of the images, misclassifications are inevitable by changes of the target object's direction. Therefore, this paper proposes a new image clustering method based on the edge and color features. The color histogram that is one of the color features is robust to these problems. Therefore, the proposed method classifies the images based on the color histogram by using the self-organizing map. Further, the structure of the neocognitron is adaptively set from the obtained clustering result to perform the classification based on edges. Therefore, since both the color and edge features of the images can be utilized, our method accurately classifies the images. Some experimental results show the proposed method achieves high classification performance.
• A Note on Improvement of HITS using Clusters based on Link Analysis  [Not invited]

  TAKEYOSHI Tomoya, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes an effective Web page scoring method based on HITS, which is traditionally used for Web page scoring using a link analysis. Recently, some Web page scoring methods that modify the HITS have been proposed. However, they cannot solve the problem that many Web pages including almost the same links as one page and the pages linked by them have high scores. Therefore, the proposed method previously classifies the target Web pages based on similarity of link structure. Then each cluster tends to contain those controversial pages. Thus, our method introduces weights, that assign lower values to the links of the Web pages included in the same clusters than the others, into the HITS and calculates the scores. Consequently, the proposed method can solve the problem of the traditional schemes, and satisfies search requests of various users enough. The experimental results show the high performance of the proposed scoring method.
• An Accurate IFS-based Image Enlargement Method using Edge Continuity  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS). The proposed method introduces a line process into the IFS-based enlargement procedures to solve edge discontinuity of an enlarged image in the traditional IFS schemes. The line process can detect the edges which retain continuity. However, the line process cannot apply to the enlarged image whose intensities are not known. Therefore, the proposed method estimates the enlarged image's edges from the target image's edges obtained by using the line process. Then, since the estimated edges retain the continuity, the proposed method can calculate the enlarged image in such a way that its edges retain the continuity. Consequently, an accurate image enlargement can be achieved.
• A Note on the similarity between music audio signals using the melody line and the rhythm  [Not invited]

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes a novel similarity between music audio signals using rhythmic contents. The proposed method estimates the rhythmic content utilizing the autocorrelation functions of the signal. Then, by applying DTW to the functions, the similarity between music audio signals can be appropriately computed. Furthermore, our method visualizes the similarity using both melody line and rhythm to try to easily retrieve the similar music audio signals.
• A Study on Group Tactics Analysis in Soccer Videos  [Not invited]

  KON Hirofumi, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper presents a method for group tactics analysis in soccer videos. For this analysis, the proposed method classifies players into some groups. As a result of this classification, the players in one group are closely related in terms of soccer tactics. In the proposed method, the relations between the players in the groups are analyzed by using player positions. Based on these relations, the proposed method estimates group advantage, which represents the degree of team's superiority or inferiority in each group. The group advantage helps viewers to understand game situations.
• A Study on Accurate Image Clustering Using Color Edges  [Not invited]

  WATANABE Takashi, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes a similar image clustering method. We have proposed a image clustering scheme using image structures and quadratic form distances between color histograms. However, since this scheme cannot effectively utilize the distribution of colors within the target images, misclassification of these images is inevitable. In order to utilize these features more effectively, the proposed method classifies the images by the K-means algorithm based on distances between color correlograms, which can be utilized for describing the relationship of colors to spatial distances. Furthermore, by reconstructing clusters based on colors at the edges and their neighborhoods, the accurate image clustering can be realized.
• A Note on Improvement of Similar Image Clustering Method using Neocognitron : Introduction of Novel Structure for Extracting Color Features  [Not invited]

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report  2007/02 

  This paper proposes a new image clustering method. We have proposed an image clustering method using a neocognitron which is one of the neural networks. In this paper, we use the edge features of the images as the input stimulus of the neocognitron. However, since this method notes only the edge features of the images, misclassifications are inevitable by changes of the target object's direction. Therefore, this paper proposes a new image clustering method based on the edge and color features. The color histogram that is one of the color features is robust to these problems. Therefore, the proposed method classifies the images based on the color histogram by using the self-organizing map. Further, the structure of the neocognitron is adaptively set from the obtained clustering result to perform the classification based on edges. Therefore, since both the color and edge features of the images can be utilized, our method accurately classifies the images. Some experimental results show the proposed method achieves high classification performance.
• A study on precise player localization in a basketball video  [Not invited]

  佐久間 大輔, 長谷山 美紀

  ITE technical report  2007/02
• An accurate video frame interpolation using image morphing  [Not invited]

  山本 貴史, 長谷山 美紀

  ITE technical report  2007/02
• A note on improvement of HITS using clusters based on link analysis  [Not invited]

  武吉 朋也, 長谷山 美紀

  ITE technical report  2007/02
• An accurate IFS-based image enlargement method using edge continuity  [Not invited]

  覚幸 典弘, 小川 貴弘, 長谷山 美紀

  ITE technical report  2007/02
• A note on the similarity between music audio signals using the melody line and the rhythm  [Not invited]

  今野 聡司, 二反田 直己, 長谷山 美紀

  ITE technical report  2007/02
• A study on group tactics analysis in soccer videos  [Not invited]

  今 宏史, 長谷山 美紀

  ITE technical report  2007/02
• A study on accurate image clustering using color edges  [Not invited]

  渡辺 隆志, 長谷山 美紀

  ITE technical report  2007/02
• A note on improvement of similar image clustering method using neocognitron  [Not invited]

  大原 貴都, 小川 貴弘, 長谷山 美紀

  ITE technical report  2007/02
• A Study on Precise Player Localization in a Basketball Video  [Not invited]

  SAKUMA Daisuke, HASEYAMA Miki

  電子情報通信学会技術研究報告. SDM, シリコン材料・デバイス  2007/01
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  電子情報通信学会技術研究報告. SDM, シリコン材料・デバイス  2007/01
• A Note on Melody Detection in Musical Audio Signal using Phrases  [Not invited]

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki

  ITE technical report  2006/10
• PCA-based audio signal classification : An approach for accuracy improvement by using MGD  [Not invited]

  NITANDA Naoki, HASEYAMA Miki

  ITE technical report  2006/10
• A Note on POCS-based Restoration of Missing Areas in Still Images : Introduction of New Classification Scheme of Local Images  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  ITE technical report  2006/10
• A similar image clustering method using a neocognitron  [Not invited]

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki

  ITE technical report  2006/10
• A Study on Estimation of Pass Courses in Soccer Videos with Active Net  [Not invited]

  KON Hirofumi, HASEYAMA Miki

  ITE technical report  2006/10
• A note on estimation of moving direction for omnidirectional images by using optical flows  [Not invited]

  TAKAHASHI Hiroaki, HASEYAMA Miki

  ITE technical report  2006/10
• A Study on Image Clustering Using Color Histogram  [Not invited]

  WATANABE Takashi, HASEYAMA Miki

  ITE technical report  2006/10
• 動き補償予測誤差に基づく適応内挿フィルタ選択手法に関する考察  [Not invited]

  松野 孝也, 長谷山 美紀

  映像情報メディア学会技術報告  2006/10
• バスケットボール映像における選手位置推定に関する検討  [Not invited]

  佐久間 大輔, 長谷山 美紀

  映像情報メディア学会技術報告  2006/08
• 拘束条件の適応的選択を導入した静止画像中に存在する消失テクスチャの再構成に関する検討  [Not invited]

  小川 貴弘, 長谷山 美紀

  映像情報メディア学会技術報告  2006/08
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2006/06 

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• A note on video frame interpolation Application of image morphing technique  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report. Circuits and systems  2006/06 

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  IEICE technical report. Signal processing  2006/06 

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• A note on video frame interpolation Application of image morphing technique  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  IEICE technical report. Signal processing  2006/06 

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki

  Technical report of IEICE. VLD  2006/06 

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• A note on video frame interpolation Application of image morphing technique  [Not invited]

  YAMAMOTO Takashi, HASEYAMA Miki

  Technical report of IEICE. VLD  2006/06 

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• Restoration of Missing Intensities of Still Images Using GMRF Model  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  The IEICE transactions on information and systems (Japanese edetion)  2006/06 

  本論文では,GMRFモデルを用いた静止画像における失われた輝度値の復元手法を提案する.画像中に存在する輝度値消失領域を高精度に復元するためには,エッジ部を精度良く再構成する必要がある.そこで,我々はGMRFモデルを用いて,特に失われたエッジ部を高精変に復元する手法の実現を試みる.このとき,提案手法において用いられるGMRFモデルは,画像が複数の領域により構成され,各領域は弱定常な確率過程であるという仮定のもとで導出される.したがって,提案手法のモデルは画像中の異なる領域間でエッジの表現が可能となり,エッジ部において高精度な復元結果を与える.更に,・このモデルは輝度値消失領域に複数の領域が合まれ,それらが複数のエッジを構成する場合についても,その適用が可能であり,このような場合に従来手法が抱えていた問題である再構成されたエッジ部の過剰な平滑化が抑制される.本論文の最後では,提案手法の有効性を示すために実画像を用いた比較実験を行い,エッジ部における復元性能を評価する.
• A New Application of Fractal Coding to Video Images  [Not invited]

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki

  映像情報メディア学会技術報告  2006/02
• A Study on Quantification of Paper-Relations Using Co-occurrence and Its Visualization  [Not invited]

  KAMAKURA Junichi, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• Quality Improvement of Cell Animation Images Using a Region-based Algorithm  [Not invited]

  YAMAGUCHI Takumi, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• High-Quality Image Compression with Region-Based Filtering  [Not invited]

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  映像情報メディア学会技術報告  2006/02
• Detection of Person who Moves in Prohibited Direction by using Video Camera  [Not invited]

  ARAI Hiroyuki, YASUNO Takayuki, MIZUKAMI Midori, HSEYAMA Miki

  ITE technical report  2006/02
• A note on accurate reconstruction method of missing textures using ER algorithm : A study on subspaces used in constraints of ER algorithm  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• A Note on an Accurate IFS-based Image Enlargement Method using Line Process  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• A study on accurate audiovisual indexing using audio signal  [Not invited]

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• A Method of Soccer Video Analysis by Using Active Net  [Not invited]

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo

  映像情報メディア学会技術報告  2006/02
• A Study on Tracking Soccer Players using Active Contour Models  [Not invited]

  WATANABE Tomoki, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• Improvement of Target Tracking using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism  [Not invited]

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• A Note on Accurate Estimation of Chords in Musical Audio Signal : An Approach using Single Tone in Low-Frequency Band  [Not invited]

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02
• High-Quality Image Compression with Region-Based Filtering  [Not invited]

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  ITE technical report  2006/02 

  Recently, MPEG-2 is widely used in many fields such as DVD, digital broadcasting, etc. The video coding technique achieves efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that blocking and ringing artifacts appear in the decoded images at low bit rates. Therefore, we have proposed a method utilizes a two-dimensional FIR filter to reduce these artifacts. The method designs the filter to minimize the error between the original image and the coded image for each frames, and its coefficients are added into the bitstream. This paper proposes a method of region segmentation using the decoded images and filter design for each regions. By applying the filters to the decoded image, the proposed method improves the decoded video quality.
• A note on accurate reconstruction method of missing textures using ER algorithm : A study on subspaces used in constraints of ER algorithm  [Not invited]

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  IEICE technical report  2006/02 

  This paper proposes a reconstruction method of missing textures in still images using an ER(Error Reduction) algorithm which is one of phase retrieval methods. The proposed method defines constraints used in this algorithm in such a way that the target local image and its Fourier transform magnitude are respectively in two subspaces generated from some clipped local images and their Fourier transform magnitudes. Then, the Fourier transform magnitude and phase of the target local image are estimated so that the missing area can be reconstructed. Further, the proposed method adaptively selects the local images and Fourier transform magnitudes utilized to generate these subspaces for the target local image. Consequently, the proposed method can accurately reconstruct the missing textures by utilizing the ER algorithm including the novel constraints. Some experimental results are shown to verify the high performance of the proposed method.
• A Note on an Accurate IFS-based Image Enlargement Method using Line Process  [Not invited]

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo

  IEICE technical report  2006/02 

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS), which is traditionally used for image coding. The IFS can reconstruct an image of a different size from the coding target image's. Based on this property, some methods enlarging images have been proposed by using the IFS. However, their enlargement results suffer from block noise and edge discontinuity in the boundary between neighboring range blocks, which are units of the procedures in the IFS. The reasons for these problems are that they use non-overlapping range blocks and do not consider edge continuity in the boundary between the neighboring range blocks. Therefore, the proposed method allows selection of overlapping range blocks in order to avoid the block noise. Further, the proposed method introduces a line process, which is used for edge detection, into the enlargement procedures. The edges obtained by using the line process can retain the edge continuity. Therefore, the images enlarged by the proposed method retain the edge continuity. Consequently, an accurate image enlargement can be achieved.
• A New Application of Fractal Coding to Video Images  [Not invited]

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki

  ITE technical report  2006/02 

  Fractal image coding based on an iterated function system (IFS), one of the coding techniques for digital images, compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the decoded image quality by the fractal image coding. However, the relationship between the self-similarity of the original image and the quality of its decoded image has not been clarified so far. Therefore, in the previous study, we have clarified the relationship for still images. Then, in this paper, we clarify the relationship for video images and consider a new approach to video coding based on the fractal image coding.

MISC

• Efforts of Hokkaido university to develop data-driven interdisciplinary research

  長谷山美紀, 水谷祐輔, 田中晋吾  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• DEGRADATION DEGREE ESTIMATION OF DISTRESS IMAGE IN ROAD INFRASTRUCTURE USING MULTI-DATASET CONTRASTIVE LEARNING

  東孝明, 小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  AI・データサイエンス論文集(Web)  4-  (2)  2023
• A Note on Multi-label Image Classification in Animation Illustration Considering Hierarchical Relationships of Attributes

  LAM Ziwen, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A Note on Text Prompt Tuning in Cross-modal Image Retrieval for a Specific Database

  ZHANG Huaying, 柳凛太郎, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A Note on Improvement of Binauralization Performance Based on Multi-view Learning on 360° Videos

  吉田将規, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A Note on Improvement of Supervised Latent Variable Model with Graph-Encoded Class Information

  渡部航史, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A note on gaze-dependent image re-ranking for content-based image retrieval

  FENG Yuhu, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A Note on Accurate Shoot Prediction Considering Players’ Spatio-temporal Relations in Soccer Videos-Introduction of Complete Bipartite Graph Based on Players’ Team Information-

  五箇亮太, 諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• A Note on Interpretable Visual Question Answering Model with Visual Representation Based on Object Detection Model

  ZHU He, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  47-  (6(MMS2023 1-34/ME2023 21-54/AIT2023 1-34))  2023
• 知識蒸留を用いたFew-shot Learningに基づく画像の感情ラベル推定に関する検討—A Note on Visual Sentiment Prediction Based on Few-shot Learning Using Knowledge Distillation—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  叶 穎睿, 諸戸 祐哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  171  -175  2022/02
• Variational Autoencoderに基づく深層生成モデルを用いた潜在表現のDisentanglementに関する検討 : Disentanglement評価指標を含む正則化損失の導入—A Note on Disentanglement Using Deep Generative Model Based on Variational Autoencoder : Introduction of Regularization Losses Based on Metrics of Disentangled Representation—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  中川 真, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  97  -102  2022/02
• Captioning特徴を利用したグラフ畳み込みネットワークに基づくアニメイラストのマルチラベル画像分類に関する検討—A Note on Multi-label Image Recognition in Anime Illustration Based on Graph Convolutional Networks Using Captioning Features—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  蘭 子文, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  161  -165  2022/02
• ゴム材料開発のためのGenerative Adversarial Networkに基づく配合量および物性からの電子顕微鏡画像の生成に関する一検討—A Note on Electron Microscope Image Generation from Mix Proportion and Material Property via Generative Adversarial Network for Rubber Materials—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  柳 凛太郎, 藤後 廉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  187  -191  2022/02
• セマンティックセグメンテーションに対するマルチソースモデル適応に関する検討 : 複数のソースモデルからの不変な特徴表現の学習による適応精度の向上—A note on multi-source model adaptation for semantic segmentation : Improving adaptation performance by learning model-invariant representation from multiple source models—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  李 宗曜, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  37  -41  2022/02
• 地下鉄トンネル点検時の技術者から取得される生体信号と技術者の点検行動の関連性分析—Relevance Analysis between Bio-signals of Engineers Inspecting Subway Tunnels and Their Inspection Behaviors—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  平澤 魁人, 前田 圭介, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  365  -370  2022/02
• 高速道路の遮音壁画像を用いた物体検出手法による変状分類の高精度化に関する検討—A note on improvement of distress classification using noise barrier images on highway via object detection method—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  梁 鋆, 前田 圭介, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  359  -363  2022/02
• 道路構造物の維持管理効率化に向けた変状画像分類の高精度化に関する検討 : テキストデータに基づく類似事例の含有率の導入—A Note on Improvement of Accuracy in Classification of Destress Images for Efficient Inspection of Road Structures : Introduction of Ratio of Similar Cases Based on Text Data—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  平川 泰成, 小川 直輝, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  43  -48  2022/02
• 地下鉄トンネルの維持管理支援のためのマルチスケール解析を導入した深層学習に基づく変状検出に関する検討—A Note on Distress Detection Based on Deep Learning with Hierarchical Multi-Scale Attention Mechanism for Supporting Maintenance of Subway Tunnels—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  高田 紗弥, 前田 圭介, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  377  -381  2022/02
• A Note on Automatic Diagnosis of Helicobacter Pylori Infection Based on Self-Supervised Learning and Self-Knowledge Distillation: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  トウゴ レン, マベ カツヒロ, ニシダ シュンペイ, トモダ ヨシヒロ, オガワ タカヒロ, ハセヤマ ミキ  46-  (6)  49  -52  2022/02
• A Note on Realizing Adversarial Defense Based on Regularization of Multi-stage Squeeze-and-Excitation Features: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  マエダ ケイスケ, オガワ タカヒロ, ハセヤマ ミキ  46-  (6)  87  -90  2022/02
• ユーザの動作情報を用いたコンテンツの関心度推定に関する検討 : 複数ユーザを導入した特徴統合の有効性検証—A Note on Interest level Estimation Using Users' behavior Information : Validating the Effectiveness of Feature Integration with Multiple Users—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  上川 恭平, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  103  -107  2022/02
• 画像注視時の脳活動信号を用いた圧縮再構成ネットワークに基づく視覚認知内容の推定に関する検討—A Note on Perceived Visual Content Estimation Based on Compressed Reconstruction Network Using Brain Signals While Gazing on Images—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  東 孝明, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  349  -353  2022/02
• Attention mapに対する確信度を考慮可能な深層学習を用いた変状分類の高精度化に関する検討—A Note on Accurate Distress Classification Using Deep Learning Considering Confidence in Attention map—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  小川 直輝, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  371  -376  2022/02
• A Note on Transformer-based Scene Recognition in Soccer Videos Using Different Lengths of Clips: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  トウゴ レン, オガワ タカヒロ, ハセヤマ ミキ  46-  (6)  167  -170  2022/02
• 画像中の物体情報を考慮したユーザ類似度に基づく個人に特化した注視領域の推定に関する検討—A Note on Personalized Saliency Prediction Based on User Similarity Considering Object Information in Images—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  諸戸 祐哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  181  -186  2022/02
• 橋梁点検時の技術者の一人称および三人称視点映像を用いた点検動作の分類に関する検討—A Note on Inspection Action Classification Using First and Third Person Video of Engineers Inspecting Bridges—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  増田 毅, 前田 圭介, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  177  -180  2022/02
• ユーザの嗜好を考慮した強化学習と知識グラフに基づく楽曲プレイリスト生成に関する検討—A Note on User Preference-Aware Music Playlist Generation Based on Reinforcement Learning and Knowledge Graph—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  櫻井 慶悟, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  46-  (6)  109  -112  2022/02
• AIの実社会応用とインフラ分野の将来像

  長谷山美紀  高速道路と自動車  65-  (12)  2022
• DETERIORATION LEVEL CLASSIFICATION OF DISTRESS IMAGE CONSIDERING CORRELATION AMONG HETEROGENEOUS FEATURES AND ATTENTION MAP CONFIDENCE

  小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  AI・データサイエンス論文集(Web)  3-  (J2)  2022
• WINTER ROAD SURFACE CONDITION CLASSIFICATION USING DEEP LEARNING WITH FOCAL LOSS BASED ON TEXT AND IMAGE INFORMATION

  諸戸祐哉, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  AI・データサイエンス論文集(Web)  3-  (J2)  2022
• DISTRESS DETECTION BASED ON VISION TRANSFORMER USING EGOCENTRIC VIDEOS WHILE INSPECTING IN SUBWEY TUNNELS

  櫻井慶悟, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  AI・データサイエンス論文集(Web)  3-  (J2)  2022
• CLASSIFICATION OF INSPECTION ACTION OF ENGINEERS FOR EFFICIENT VIDEO PRESENTATION TO ASSIST ASSESSMENT OF INFRASTRUCTURE FACILITY DISTRESS

  上川恭平, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  AI・データサイエンス論文集(Web)  3-  (J2)  2022
• A note on multi-source model adaptation for semantic segmentation-Improving adaptation performance by learning model-invariant representation from multiple source models-

  LI Zongyao, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on User Preference-Aware Music Playlist Generation Based on Reinforcement Learning and Knowledge Graph

  櫻井慶悟, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Disentanglement Using Deep Generative Model Based on Variational Autoencoder-Introduction of Regularization Losses Based on Metrics of Disentangled Representation-

  中川真, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note of the Relationship between Popularity Bias and Embedding Representations of Latent Factor Models in Collaborative Filtering

  岡村洋希, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Robust Recommendation System for Domain-dependent Preference Based on Domain-shared Network

  山本一輝, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Image Retrieval Robust for Changing Camera Views Using Synthesized Images by pixelNeRF

  江良勇輝, 藤後廉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Study on Non-reference Image Quality Assessment Considering Phaseless Components of Neural Radiance Fields

  河合雅斗, 柳凛太郎, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Uncertainty-based Shoot Event Prediction from Scouting Videos of Soccer by Considering Spatio-Temporal Relations between Players

  五箇亮太, 諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Estimation of Viewed Images Using Brain Activity Data While Viewing Images-A Verification of Estimation Accuracy by Regression Model Used for fMRI Decoder-

  七田亮, 藤後廉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Estimation of Diabetic Retinopathy Grades Based on Unsupervised Domain Adaptation Using Fundus Images

  國枝翼, 藤後廉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2022-  2022
• A Note on Electron Microscope Image Generation from Mix Proportion and Material Property via Generative Adversarial Network for Rubber Materials

  柳凜太郎, 藤後廉, 前田圭介, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Improvement of Accuracy in Classification of Destress Images for Efficient Inspection of Road Structures- Introduction of Ratio of Similar Cases Based on Text Data-

  平川泰成, 小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Distress Detection Based on Deep Learning with Hierarchical Multi-Scale Attention Mechanism for Supporting Maintenance of Subway Tunnels

  高田紗弥, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Perceived Visual Content Estimation Based on Compressed Reconstruction Network Using Brain Signals While Gazing on Images

  東孝明, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Interest level Estimation Using Users’ behavior Information-Validating the Effectiveness of Feature Integration with Multiple Users-

  上川恭平, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Visual Sentiment Prediction Based on Few-shot Learning Using Knowledge Distillation

  YE Yingrui, 諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• Relevance Analysis between Bio-signals of Engineers Inspecting Subway Tunnels and Their Inspection Behaviors

  平澤魁人, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Personalized Saliency Prediction Based on User Similarity Considering Object Information in Images

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• A Note on Accurate Distress Classification Using Deep Learning Considering Confidence in Attention map

  小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  46-  (6(MMS2022 1-37/ME2022 26-62/AIT2022 1-37))  2022
• AIの実社会応用と次世代インフラ維持管理

  長谷山 美紀  学士会会報  2022-  (1)  78  -83  2022/01
• 招待講演 マルチメディアAI技術に基づく異分野融合研究と実社会応用—Invited Talk : Interdisciplinary Research and Real-world Applications Based on Multimedia AI technologies—情報センシング メディア工学 映像・信号の処理・解析・AI技術とその多分野応用

  小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (13)  73  -80  2021/06
• A Note on Distress Classification of High-resolution Subway Tunnel Images based on HRNet: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  トウゴ レン, オガワ タカヒロ, ハセヤマ ミキ  45-  (4)  13  -16  2021/02
• 地下鉄トンネルの維持管理支援を目的とした深層学習に基づく変状検出の高精度化に関する検討 : 壁面の施工方法に注目した精度検証—A Note on Improving Performance of Deep Learning-based Distress Detection for Supporting Maintenance of Subway Tunnels : Accuracy Verification Focusing on Tunnel Wall Characteristics—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  春山 知生, 前田 圭介, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  1  -6  2021/02
• 特別講演 実店舗の防犯カメラ映像データを用いた顧客の関心推定に関する検討 : 姿勢推定モデルから得られる特徴量を用いた異常検知モデルの精度検証—A Note on Customer's Interest Estimation Method Using Security Camera Video Data from the Real Store : Validation of the accuracy of an anomaly detection model using features obtained from a posture estimation model—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  本間 勇紀, 藤後 廉, 阿部 真育, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  193  -198  2021/02
• 地下鉄トンネル点検時の技術者から取得される視線およびモーションデータに基づく熟練度の推定に関する検討—A Note on Estimation of Expert-Novice Level Based on Eye Tracking and Motion Data from Engineers While Inspecting Subway Tunnel—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  赤松 祐亮, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  7  -12  2021/02
• 特別講演 ゴム材料開発のためのConditional StyleGANに基づく配合量からの電子顕微鏡画像の生成に関する一検討—A Note on Electron Microscope Image Generation from Mix Proportion via Conditional Style Generative Adversarial Network for Rubber Materials—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  柳 凜太郎, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  171  -175  2021/02
• 特別講演 路面画像を用いた深層学習に基づく路面状態の分類に関する検討—A Note on Discrimination of Road Surface Conditions Based on Deep Learning Using Road Images—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  諸戸 祐哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  165  -169  2021/02
• TCCに基づく自己教師学習による特徴表現を用いた映像中の人物動作の検出に関する検討 : 動作検出パラメータが与える影響に対する考察—A Note on Action Detection Using Feature Representation via Self-Supervised Learning based on TCC : Effectiveness evaluation of parameters for action detection—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  増田 毅, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  47  -51  2021/02
• 電子顕微鏡により撮像されたゴム材料からの異常検知に基づく劣化領域の推定に関する一検討 : 深層学習モデルにより抽出された特徴表現の有効性検証—A Note on Estimation of Deteriorated Regions Based on Anomaly Detection from Rubber Material Electron Microscope Images : Verification of Feature Representations Extracted from Deep Learning Models—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  松本 真直, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  43  -46  2021/02
• サッカー映像における情報の提示位置の推定に関する検討 : 視聴者の注視位置および選手位置に基づいたオブジェクト配置の決定—A Note on Position Estimation of Superimposing Information in Soccer Videos : Decision of Object Placement Based on User's Gaze Position and Players' Positions—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  鈴木 元樹, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  77  -81  2021/02
• A Note on Automatic Diagnosis of Helicobacter Pylori Infection Based on EfficientNet with Flooding Loss: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  トウゴ レン, マベ カツヒロ, ニシダ シュンペイ, トモダ ヨシヒロ, シミズ ヒカリ, オガワ タカヒロ, ハセヤマ ミキ  45-  (4)  23  -26  2021/02
• Twitterと映像を用いたMVAEに基づく野球映像の重要シーン予測に関する検討—A Note on Prediction of Important Scenes in Baseball Videos via Multimodal Variational Autoencoder Using Tweets and Videos—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  平澤 魁人, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  71  -75  2021/02
• Detecting axillary lymph node metastasis of breast cancer with FDG-PET/CT images based on attention mechanism: FDG-PET/CT画像を用いたattention mechanismに基づく乳癌の腋窩リンパ節転移の検出: マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  トウゴ レン, ヒラタ ケンジ, キタジマ カズヒロ, タケナカ ジュンキ, ミヨシ ヤスオ, クドウ コウスケ, オガワ タカヒロ, ハセヤマ ミキ  45-  (4)  33  -36  2021/02
• 特別講演 ユーザの嗜好を反映可能なインテリアコーディネート検索技術の構築 : コーディネートを表現可能な特徴の抽出と実データへの適用—A Note on Interior Coordination Retrieval Reflecting User's Preferences : Extraction of Features Representing Coordination and Application to Real Data—マルチメディアストレージ メディア工学 映像表現&コンピュータグラフィックス

  藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  45-  (4)  187  -191  2021/02
• A Note on Off-screen Sound Detection Based on Audio-visual Spatialization-Introduction of Audio-visual Features Based on Self-supervised Learning-

  吉田将規, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Note on Effect of Visual Representation Ability to Performance of World Models

  大羽賀駿也, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Verification of Vulnerability of Graph-based Recommender Models under Shilling Attacks

  小野寺望, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Note on Multi-label Image Recognition in Anime Illustration Based on Graph Convolutional Networks

  LAN Ziwen, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Note on Introducing Self-supervised Learning to Latent Variable Model for Low-dimensional Feature Extraction

  渡部航史, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• An Note on Use of Multiple Datasets in Visual Sentiment Prediction Using Few-shot Learning

  YE Yingrui, 諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Note on Accuracy Assessment Focusing on Manipulation Area by Text-Guided Image Manipulation

  渡邉優宇人, 藤後廉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2021-  2021
• A Note on Estimation of Semantic Content Based on a Question Answering Model Using Brain Activity Data while Viewing Images-Improvement of Estimation Performance Based on fine-tuning of the VQA model-

  高田紗弥, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  27  -31  2021
• A Note on Improvement of Image Sentiment Analysis Based on Introduction of Image Captioning

  LIANG Yun, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  65  -69  2021
• A Note on Cross-domain Recommendation Based on Multi-layer Graph Analysis with Visual Features

  平川泰成, 前田圭介, 小川貴弘, 浅水仁, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  59  -63  2021
• A Note on Improving Performance of Deep Learning-based Distress Detection for Supporting Maintenance of Subway Tunnels-Accuracy Verification Focusing on Tunnel Wall Characteristics-

  春山知生, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Customer’s Interest Estimation Method Using Security Camera Video Data from the Real Store-Validation of the accuracy of an anomaly detection model using features obtained from a posture estimation model-

  本間勇紀, 本間勇紀, 藤後廉, 阿部真育, 小川貴弘, 長谷山美紀, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Estimating of Deteriorated Refions Based on Anomaly Detection from Rubber Material Electron Microscope Images-Verification of Feature Representations Extracted from Deep Learning Models-

  松本真直, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Estimation of Expert-Novice Level Based on Eye Tracking and Motion Data from Engineers While Inspecting Subway Tunnel

  赤松祐亮, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Action Detection Using Feature Representation via Self-Supervised Learning based on TCC-Effectiveness evaluation of parameters for action detection-

  増田毅, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Discrimination of Road Surface Conditions Based on Deep Learning Using Road Images

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Position Estimation of Superimposing Information in Soccer Videos-Decision of Object Placement Based on User’s Gaze Position and Players’ Positions-

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Prediction of Important Scenes in Baseball Videos via Multimodal Variational Autoencoder Using Tweets and Videos

  平澤魁人, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Interior Coordination Retrieval Reflecting User’s Preferences-Extraction of Features Representing Coordination and Application to Real Data-

  藤後廉, 小川貴弘, 長谷山美紀, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• A Note on Accurate Distress Image Classification of Road Structures Using Attention Map Based on Text Data

  小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  17  -21  2021
• A Note on Electron Microscope Image Generation from Mix Proportion via Conditional Style Generative Adversarial Network for Rubber Materials

  柳凜太郎, 藤後廉, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (4(MMS2021 1-28/ME2021 1-28/AIT2021 1-28))  2021
• AIを中心とした医療デジタル技術基盤の構築へ向けた取り組み

  藤後廉, 小川貴弘, 長谷山美紀  日本消化器がん検診学会雑誌(Web)  59-  (Supplement 1)  2021
• Interdisciplinary Research and Real-world Applications Based on Multimedia AI technologies

  小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  45-  (13(IST2021 22-29/ME2021 53-60))  2021
• The Clinical Application of Artificial Intelligence for Detecting Atrophic Gastritis by Upper Gastrointestinal Barium X-ray Radiography

  竹内千尋, 藤後廉, 小川貴弘, 長谷山美紀, 山道信毅  臨床消化器内科  36-  (8)  1062  -1065  2021
• データサイエンスのあるべき姿とは : 高等教育・研究機関における取り組みを通じての意見交換会—特集 データサイエンス教育の潮流(その1)

  木島 正明, 竹村 彰通, 長谷山 美紀, 椿 広計, 蓮池 隆  オペレーションズ・リサーチ = Communications of the Operations Research Society of Japan : 経営の科学  65-  (10)  525  -536  2020/10  [Not refereed]
  
• メディア工学研究委員会—Technical Group on Media Engineering—特集 学会創立70周年記念特集 ; 映像情報メディア学会の研究会活動 : 過去から未来へ

  村上 和人, 長谷山 美紀, 田川 憲男, 田良島 周平, 石井 大祐  映像情報メディア学会誌 = The journal of the Institute of Image Information and Television Engineers  74-  (5)  779  -783  2020/09
• 画像内の物体に着目した画像検索に関する検討 : RetinaNetを用いた物体認識に基づく高精度化—A Note on Image Retrieval Focusing on Objects in Images : Improving Retrieval Performance Based on Object Recognition Using RetinaNet—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  柳 凛太郎, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  377  -381  2020/02
• 画像注視時の視線データを用いた関心度の推定に関する検討 : ラベルの逆量子化を導入したsMVCCAによる推定精度の高精度化—A Note on Eye Gaze Based User-specific Interest Estimation for Images : Estimation Performance Improvement Based on sMVCCA including Ordinal Label Dequantization—ITS : Intelligent Transport Systems Technology

  松本 真直, 斉藤 直輝, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  73  -77  2020/02
• 画像注視時の視線データを用いた関心度の推定に関する検討 : ラベルの逆量子化を導入したsMVCCAによる推定精度の高精度化—A Note on Eye Gaze Based User-specific Interest Estimation for Images : Estimation Performance Improvement Based on sMVCCA including Ordinal Label Dequantization—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  松本 真直, 斉藤 直輝, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  73  -77  2020/02
• 地下鉄トンネル点検時の生体信号に基づいた熟練および若手技術者の分類に関する検討—A Note on Classification of Experienced and Novice Inspectors Based on Bio-signals While Inspecting in Subway Tunnels—ITS : Intelligent Transport Systems Technology

  九島 哲哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  101  -105  2020/02
• 特別講演 社会インフラ維持管理効率化のための最先端AI技術の導入 : 点検データを用いた変状分類の精度向上に向けた取り組み—Introduction of Advanced AI Technology for Efficient Maintenance Inspection of Social Infrastructure : Performance Improvement of Distress Image Classification Utilizing Inspection Data—ITS : Intelligent Transport Systems Technology

  前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  359  -362  2020/02
• 道路構造物の変状評価における技術者の視線データと熟練度の分析に関する一考察—A Note on Analysis of Gaze Data and Skills of Inspectors in Distress Assessment of Road Structures—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  松井 太我, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  97  -100  2020/02
• 道路構造物の維持管理効率化に向けた変状画像分類における信頼性の向上に関する検討—A Note on Improvement of Reliability of Distress Image Classification for Efficient Maintenance of Road Structures—ITS : Intelligent Transport Systems Technology

  堀井 風葉, 前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  51  -56  2020/02
• 画像内の物体に着目した画像検索に関する検討 : RetinaNetを用いた物体認識に基づく高精度化—A Note on Image Retrieval Focusing on Objects in Images Improving Retrieval Performance Based on Object Recognition Using RetinaNet—ITS : Intelligent Transport Systems Technology

  柳 凜太郎, 藤後 廉, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  377  -381  2020/02
• 弱異常検知アルゴリズムに基づくCCTV映像を用いた河川利用者の異常行動の検出に関する検討—A Note on Abnormal Motion Detection of River Users on CCTV Videos Based on Weakly Supervised Anomaly Detection Algorithm—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  渡邊 はるな, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  367  -370  2020/02
• A note on retrieval of visually similar distress regions in subway tunnel images : Introduction of deep features extracted by semantic segmentation network: ITS : Intelligent Transport Systems Technology

  トウゴ レン, オガワ タカヒロ, ハセヤマ ミキ  119-  (421)  65  -68  2020/02
• 画像注視時のヒトの感情推定のための視線特徴の推定に関する検討—A Note on Estimation of Gaze Features for Human Emotion Estimation while Viewing Images—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  諸戸 祐哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  85  -89  2020/02
• ゴム材料の配合量を用いたAC-GANに基づく電子顕微鏡画像の生成に関する一検討—A Note on Generation of Electron Microscope Images via Auxiliary Classifier Generative Adversarial Network with Mix Proportions—ITS : Intelligent Transport Systems Technology

  金井 美岬, 藤後 廉, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  107  -111  2020/02
• ゴム材料の配合量を用いたAC-GANに基づく電子顕微鏡画像の生成に関する一検討—A Note on Generation of Electron Microscope Images via Auxiliary Classifier Generative Adversarial Network with Mix Proportions—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  金井 美岬, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  107  -111  2020/02
• 道路構造物の変状評価における技術者の視線データと熟練度の分析に関する一考察—A Note on Analysis of Gaze Data and Skills of Inspectors in Distress Assessment of Road Structures—ITS : Intelligent Transport Systems Technology

  松井 太我, 前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  97  -100  2020/02
• エッジの補完を導入したツイートクラスタリングの高精度化に関する検討—A Note on Tweet Clustering by Complementing Edges among Tweets—ITS : Intelligent Transport Systems Technology

  滝村 祥司, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  61  -64  2020/02
• 地下鉄トンネル維持管理支援を目的とした類似画像の検索に関する検討 : 技術者の評価を反映可能な距離計量学習の導入—A Note on Retrieval of Similar Images for Supporting Maintenance of Subway Tunnels : Introduction of Distance Metric Learning Reflecting Inspectors' Evaluation—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  松本 有衣, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  47  -50  2020/02
• 特別講演 社会インフラ維持管理効率化のための最先端AI技術の導入 : 点検データを用いた変状分類の精度向上に向けた取り組み—Introduction of Advanced AI Technology for Efficient Maintenance Inspection of Social Infrastructure : Performance Improvement of Distress Image Classification Utilizing Inspection Data—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  359  -362  2020/02
• 地下鉄トンネル維持管理支援を目的とした類似画像の検索に関する検討 : 技術者の評価を反映可能な距離計量学習の導入—A Note on Retrieval of Similar Images for Supporting Maintenance of Subway Tunnels : Introduction of Distance Metric Learning Reflecting Inspectors' Evaluation—ITS : Intelligent Transport Systems Technology

  松本 有衣, 前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  47  -50  2020/02
• 穿孔データを用いたオンライン学習に基づく岩盤の圧縮強度指数推定に関する検討—A Note on Estimation of rock compressive strength index from drilling data based on online learning—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  山本 健太郎, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  57  -60  2020/02
• 穿孔データを用いたオンライン学習に基づく岩盤の圧縮強度指数推定に関する検討—A Note on Estimation of rock compressive strength index from drilling data based on online learning—ITS : Intelligent Transport Systems Technology

  山本 健太郎, 藤後 廉, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  57  -60  2020/02
• サッカー映像視聴時の視線データを用いた情報の提示領域の算出に関する検討—Presentation Region Calculation of Information Based on Gaze Tracking Data in Soccer Videos—ITS : Intelligent Transport Systems Technology

  鈴木 元樹, 高橋 翔, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  91  -95  2020/02
• 地下鉄トンネル点検時の生体信号に基づいた熟練および若手技術者の分類に関する検討—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  九島 哲哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  101  -105  2020/02
• Bidirectional LSTMを用いたサッカー映像におけるイベント分類に関する検討—A Note on Event Classification in Soccer Videos Using Bidirectional LSTM—ITS : Intelligent Transport Systems Technology

  春山 知生, 高橋 翔, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  371  -375  2020/02
• 画像注視時のヒトの感情推定のための視線特徴の推定に関する検討—A Note on Estimation of Gaze Features for Human Emotion Estimation while Viewing Images—ITS : Intelligent Transport Systems Technology

  諸戸 祐哉, 前田 圭介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  119-  (421)  85  -89  2020/02
• 道路構造物の維持管理効率化に向けた変状画像分類における信頼性の向上に関する検討—A Note on Improvement of Reliability of Distress Image Classification for Efficient Maintenance of Road Structures—マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス

  堀井 風葉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  51  -56  2020/02
• A Note on Tweet Clustering by Complementing Edges among Tweets

  滝村 祥司, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  61  -64  2020/02
• A Note on Estimation of Image Categories Using Brain Activity While Viewing Images Based on MVBGM-MS

  赤松 祐亮, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  79  -83  2020/02
• サッカー映像視聴時の視線データを用いた情報の提示領域の算出に関する検討 (マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス)

  鈴木 元樹, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  91  -95  2020/02
• A Note on Event Classification in Soccer Videos Using Bidirectional LSTM

  春山 知生, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  44-  (6)  371  -375  2020/02
• A Note on Tweet Clustering by Complementing Edges among Tweets

  滝村祥司, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Event Classification in Soccer Videos Using Bidirectional LSTM

  春山知生, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• Introduction of Advanced AI Technology for Efficient Maintenance Inspection of Social Infrastructure-Performance Improvement of Distress Image Classification Utilizing Inspection Data-

  前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• Presentation Region Calculation of Information Based on Gaze Tracking Data in Soccer Videos

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Improvement of Reliability of Dstress Image Classification for Efficient Maintenace of Road Structures

  堀井風葉, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Retrieval of Similar Images for Supporting Maintenance of Subway Tunnels-Introduction of Distance Metric Learning Reflecting Inspectors’ Evaluation-

  松本有衣, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Estimation of rock compressive strength index from drilling data based on online learning

  山本健太郎, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Classification of Experienced and Novice Inspectors Based on Bio-signals While Inspecting in Subway Tunnels

  九島哲哉, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Abnormal Motion Detection of River Users on CCTV Videos Based on Weakly Supervised Anomaly Detection Algorithm

  渡邊はるな, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Eye Gaze Based User-specific Interest Estimation for Images-Estimation Performance Improvement Based on sMVCCA including Ordinal Label Dequantization-

  松本真直, 斉藤直輝, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• Efforts of Hokkaido University to Develop Human Resources and the Education of Mathematical and Data Science

  長谷山美紀, 大本亨, 湧田雄基, 阿部真育, 長谷山美紀, 大本亨  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Generation of Electron Microscope Images via Auxiliary Classifier Generative Adversarial Network with Mix Proportions

  金井美岬, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Analysis of Gaze Data and Skills of Inspectors in Distress Assessment of Road Structures

  松井太我, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Estimation of Image Categories Using Brain Activity While Viewing Images Based on MVBGM-MS

  赤松祐亮, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Image Retrieval Focusing on Objects in Images-Improving Retrieval Performance Based on Object Recognition Using RetinaNet-

  柳凛太郎, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• A Note on Estimation of Gaze Features for Human Emotion Estimation while Viewing Images

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (421(ITS2019 30-56))  2020
• 胃X線画像を用いたAIによるH.pylori感染識別と今後の展望

  藤後廉, 小川貴弘, 間部克裕, 加藤元嗣, 長谷山美紀  日本消化器がん検診学会雑誌(Web)  58-  (2)  2020
• SIMILAR INSPECTION DATA RETRIEVAL FOR ROAD STRUCTURE INSPECTION BASED ON CANONICAL CORRELATION BETWEEN EYE TRACKING DATA AND INSPECTION RECORDS

  前田圭介, 斉藤僚汰, 高橋翔, 小川貴弘, 長谷山美紀  土木学会論文集 F3(土木情報学)(Web)  76-  (1)  2020
• A Note on Music Playlist Generation Based on Reinforcement Learning Using Self-Organized Map

  櫻井慶悟, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Discrimination of Road Surface Conditions Based on Anomaly Detection Using Road Images

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Cross-domain Recommendation Based on Multilayer Graph Analysis-Performance Evaluation with Change in Dimension of Embedding Feature-

  平川泰成, 前田圭介, 小川貴弘, 浅水仁, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Detection of Human Actions Based on Temporal Cycle Consistency Learning

  増田毅, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on User Interest Estimation Based on Feature Integration Using m-SimGP

  上川恭平, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Image Sentiment Analysis based on Multi-level Deep Metric Net

  LIANG Yun, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Dataset Complexity Assessment Based on nAULS

  LI Guang, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Improvement of Distress Detection Performance in Subway Tunnel Images by Data Augmentation Based on RICAP

  春山知生, 前田圭介, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note of Estimation of Viewed Images Using fMRI Data While Viewing Images-Introduction of Shared Brain Responses of Multiple Subjects Based on Probabilistic Generative Model-

  東孝明, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• A Note on Disentangled Representation Learning via Pre-Trained Semantic Segmentation Model

  中川真, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2020-  2020
• Human-centered image classification via a neural network considering visual and biological features.

  Kazaha Horii, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  Multimedia Tools Appl.  79-  (7-8)  4395  -4415  2020  [Not refereed][Not invited]
   

  © 2019, Springer Science+Business Media, LLC, part of Springer Nature. In this paper, we propose a human-centered image classification via a neural network considering visual and biological features. The proposed method has two novelties. Firstly, we apply Group-Sparse Local Fisher Discriminant Analysis (GS-LFDA) to biological features. GS-LFDA realizes dimensionality reduction and noise elimination for biological features with consideration of local structures and class information. Secondly, we construct a Canonical Correlation Analysis (CCA)-based hidden layer via Discriminative Locality Preserving CCA (DLPCCA). DLPCCA transforms visual features into effective features by considering the relationships with biological information and class information. The CCA-based hidden layer enables transformation of visual features into effective features for image classification from a small number of training samples. Furthermore, once the projection can be obtained in the training phase, elimination of the need for biological data acquisition in the test phase is realized. This is another merit of our method.
• 胃バリウム検査におけるAIによるHelicobacter pylori診断 (Helicobacter Topics! : 「旬のHelicobacter」を知り,驚き,理解を深め,楽しむ)

  藤後 廉, 間部 克裕, 山道 信毅, 大泉 晴史, 小川 貴弘, 長谷山 美紀, 加藤 元嗣, 坂本 直哉  Helicobacter research : Journal of helicobacter research  23-  (2)  92,133  -137  2019/11  [Not refereed][Not invited]
  
• A Note on User-specific Visual Attention Estimation Based on Visual and Spatial Information in Images : A Study on Relationship between Estimation Performance and the Similarity of Visual Features

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (131)  13  -16  2019/07/19  [Not refereed][Not invited]
  
• A Note on Eye Gaze Based User-specific Interest Estimation for Images : Performance Improvement Based on Decision Level Fusion

  松本真直, 斉藤直輝, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (131)  17  -20  2019/07/19  [Not refereed][Not invited]
  
• A Note on Semantic Evaluation of Images Generated by Text-to-image Generative Adversarial Networks

  柳凜太郎, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  119-  (131)  21  -24  2019/07/19  [Not refereed][Not invited]
  
• Performance improvement of users' interest estimation using viewing behavior data obtained by sensors

  長谷山 美紀, 小川 貴弘, 髙橋 翔, 原川 良介  画像ラボ  30-  (7)  8  -12  2019/07  [Not refereed][Not invited]
  
• Detection of gastritis by a deep convolutional neural network from double-contrast upper gastrointestinal barium X-ray radiography

  Ren Togo, Nobutake Yamamichi, Katsuhiro Mabe, Yu Takahashi, Chihiro Takeuchi, Mototsugu Kato, Naoya Sakamoto, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  Journal of Gastroenterology  54-  (4)  321  -329  2019/04  [Not refereed][Not invited]
   

  © 2018, Japanese Society of Gastroenterology. Background: Deep learning has become a new trend of image recognition tasks in the field of medicine. We developed an automated gastritis detection system using double-contrast upper gastrointestinal barium X-ray radiography. Methods: A total of 6520 gastric X-ray images obtained from 815 subjects were analyzed. We designed a deep convolutional neural network (DCNN)-based gastritis detection scheme and evaluated the effectiveness of our method. The detection performance of our method was compared with that of ABC (D) stratification. Results: Sensitivity, specificity, and harmonic mean of sensitivity and specificity of our method were 0.962, 0.983, and 0.972, respectively, and those of ABC (D) stratification were 0.925, 0.998, and 0.960, respectively. Although there were 18 false negative cases in ABC (D) stratification, 14 of those 18 cases were correctly classified into the positive group by our method. Conclusions: Deep learning techniques may be effective for evaluation of gastritis/non-gastritis. Collaborative use of DCNN-based gastritis detection systems and ABC (D) stratification will provide more reliable gastric cancer risk information.
• A Note on Estimation of Deteriorated Regions Based on Anomaly Detection from Rubber Material Electron Microscope Images

  藤後 廉, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  265  -268  2019/02/19  [Not refereed][Not invited]
  
• Performance Verification of Stress Detection Using a High-performance Wristband with Multi-sensing

  伊藤 良起, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  269  -274  2019/02/19  [Not refereed][Not invited]
  
• A Note on Estimation of Rock Drilling Energy Using Tunnel Working Face Images

  山本 健太郎, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  291  -294  2019/02/19  [Not refereed][Not invited]
  
• A note on detection of distress regions in subway tunnels from laser data : Visualization of distress regions with fully convolutional networks

  豊田 陽, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  295  -299  2019/02/19  [Not refereed][Not invited]
  
• A Note on Twitter Followee Recommendation Based on Adaptive Integration of Multiple Classification Results

  滝村 祥司, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  307  -310  2019/02/19  [Not refereed][Not invited]
  
• A Note on Automatic Malignant Tumor Candidate Detection Based on a 3D Deep Residual Network with FDG-PET/CT Images (ITS)

  李 宗曜, 藤後 廉, 小川 貴弘, 平田 健司, 真鍋 治, 志賀 哲, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  311  -314  2019/02/19  [Not refereed][Not invited]
  
• A Note on Generation of Gastritis Images Based on Progressive Growing GAN for Gastritis Classification

  渡邊 はるな, 藤後 廉, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  319  -322  2019/02/19  [Not refereed][Not invited]
  
• A Note on Eye Gaze-based Estimation of Emotion Evoked by Watching Images : Performance Improvement via Feature Transformation Based on KDLPCCA

  松井 太我, 斉藤 直輝, 小川 貴弘, 浅水 仁, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  323  -327  2019/02/19  [Not refereed][Not invited]
  
• A note on interest level estimation for videos using users' behavior based on OpenPose

  九島 哲哉, 高橋 翔, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  351  -354  2019/02/19  [Not refereed][Not invited]
  
• A Note on Action Classification of River Users on CCTV Videos

  川本 舜, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  365  -369  2019/02/19  [Not refereed][Not invited]
  
• A Note on Anomaly Detection of River Users on CCTV Videos

  川本 舜, 原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (449)  371  -374  2019/02/19  [Not refereed][Not invited]
  
• A Note on Accurate Estimation of Deterioration Levels on Transmission Towers via Deep Learning Using Heterogeneous Features

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  361‐364  2019/02/12  [Not refereed][Not invited]
  
• A Note on Popularity Prediction of Artists on Music Streaming Services Based on Network Analysis Using Heterogeneous Features

  松本有衣, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  301‐306  2019/02/12  [Not refereed][Not invited]
  
• A Note on Retrieval of Similar Inspection Records Based on Distance Metric Learning for Supporting Maintenance of Subway Tunnels

  弦間奨, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  287‐290  2019/02/12  [Not refereed][Not invited]
  
• A Method for Predicting Importance of Attack Players based on Gaze Tracking Data in Soccer Videos

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  355‐359  2019/02/12  [Not refereed][Not invited]
  
• Leading edge of Education for Mathematical and Data Science in Hokkaido University―Concept of MDS Programs in Undergraduate and Graduate Schools―

  長谷山美紀, 長谷山美紀, 大本亨, 大本亨, 湧田雄基  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  3‐4  2019/02/12  [Not refereed][Not invited]
  
• A Note on Gastritis Detection from Gastric X‐ray Images via Transfer Learning Approach

  金井美岬, 藤後廉, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  315‐318  2019/02/12  [Not refereed][Not invited]
  
• A Note on Performance Improvement of Distress Detection in Subway Tunnels Using CNN

  WANG An, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  121‐124  2019/02/12  [Not refereed][Not invited]
  
• A Note on Attribute Estimation for Improving Interpretability of Convolutional Neural Network

  堀井風葉, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  275‐279  2019/02/12  [Not refereed][Not invited]
  
• A Note on Estimation of Characteristic Curve for Rubber Materials Using Electron Microscope Images and Mix Proportions

  斉藤直輝, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  261‐264  2019/02/12  [Not refereed][Not invited]
  
• A note on estimation of inspectors’ visual attention using distress images of subway tunnels―Trial introduction of deep learning‐based saliency prediction methods―

  斉藤僚汰, 前田圭介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  281‐285  2019/02/12  [Not refereed][Not invited]
  
• A Note on Performance Improvement of Distress Detection in Subway Tunnels Using CNN

  王 安, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  121  -124  2019/02  [Not refereed][Not invited]
  
• A Note on Estimation of Characteristic Curve for Rubber Materials Using Electron Microscope Images and Mix Proportions

  斉藤 直輝, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  261  -264  2019/02  [Not refereed][Not invited]
  
• A Note on Estimation of Deteriorated Regions Based on Anomaly Detection from Rubber Material Electron Microscope Images

  藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  265  -268  2019/02  [Not refereed][Not invited]
  
• Performance Verification of Stress Detection Using a High-performance Wristband with Multi-sensing

  伊藤 良起, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  269  -274  2019/02  [Not refereed][Not invited]
  
• A Note on Attribute Estimation for Improving Interpretability of Convolutional Neural Network

  堀井 風葉, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  275  -279  2019/02  [Not refereed][Not invited]
  
• A note on estimation of inspectors' visual attention using distress images of subway tunnels : Trial introduction of deep learning-based saliency prediction methods

  斉藤 僚汰, 前田 圭介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  281  -285  2019/02  [Not refereed][Not invited]
  
• A Note on Retrieval of Similar Inspection Records Based on Distance Metric Learning for Supporting Maintenance of Subway Tunnels

  弦間 奨, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  287  -290  2019/02  [Not refereed][Not invited]
  
• A Note on Estimation of Rock Drilling Energy Using Tunnel Working Face Images

  山本 健太郎, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  291  -294  2019/02  [Not refereed][Not invited]
  
• A note on detection of distress regions in subway tunnels from laser data : Visualization of distress regions with fully convolutional networks

  豊田 陽, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  295  -299  2019/02  [Not refereed][Not invited]
  
• A Note on Popularity Prediction of Artists on Music Streaming Services Based on Network Analysis Using Heterogeneous Features

  松本 有衣, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  301  -306  2019/02  [Not refereed][Not invited]
  
• A Note on Twitter Followee Recommendation Based on Adaptive Integration of Multiple Classification Results

  滝村 祥司, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  307  -310  2019/02  [Not refereed][Not invited]
  
• A Note on Automatic Malignant Tumor Candidate Detection Based on a 3D Deep Residual Network with FDG-PET/CT Images (マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス)

  李 宗曜, 藤後 廉, 小川 貴弘, 平田 健司, 真鍋 治, 志賀 哲, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  311  -314  2019/02  [Not refereed][Not invited]
  
• A Note on Gastritis Detection from Gastric X-ray Images via Transfer Learning Approach

  金井 美岬, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  315  -318  2019/02  [Not refereed][Not invited]
  
• A Note on Generation of Gastritis Images Based on Progressive Growing GAN for Gastritis Classification

  渡邊 はるな, 藤後 廉, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  319  -322  2019/02  [Not refereed][Not invited]
  
• A Note on Eye Gaze-based Estimation of Emotion Evoked by Watching Images : Performance Improvement via Feature Transformation Based on KDLPCCA

  松井 太我, 斉藤 直輝, 小川 貴弘, 浅水 仁, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  323  -327  2019/02  [Not refereed][Not invited]
  
• A note on interest level estimation for videos using users' behavior based on OpenPose

  九島 哲哉, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  351  -354  2019/02  [Not refereed][Not invited]
  
• A Method for Predicting Importance of Attack Players based on Gaze Tracking Data in Soccer Videos

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  IEEE 8th Global Conference on Consumer Electronics(GCCE)  43-  (5)  355  -359  2019/02  [Not refereed][Not invited]
  
• A Note on Accurate Estimation of Deterioration Levels on Transmission Towers via Deep Learning Using Heterogeneous Features

  前田 圭介, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  361  -364  2019/02  [Not refereed][Not invited]
  
• A Note on Action Classification of River Users on CCTV Videos

  川本 舜, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  365  -369  2019/02  [Not refereed][Not invited]
  
• A Note on Anomaly Detection of River Users on CCTV Videos

  川本 舜, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  43-  (5)  371  -374  2019/02  [Not refereed][Not invited]
  
• Efforts of Hokkaido University to Develop Data Science Human Resources-Construction of a foundation for creative human resource development: attempts to implement new functions of university education-

  長谷山美紀, 長谷山美紀, 大本亨, 大本亨, 湧田雄基  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  2019
• Efforts for Doctoral Human Resource Development at Hokkaido University-Smart Infrastructure Management Human Resources Development Consortium-

  長谷山美紀, 長谷山美紀, 大本亨, 大本亨, 湧田雄基  電子情報通信学会技術研究報告  118-  (449(ITS2018 59-85))  2019
• 人工知能が医療を変える!医療分野におけるAI研究開発最前線2019 II 領域別・画像診断におけるAI研究開発の最前線 7.その他 3)核医学におけるディープラーニングを用いた画像診断,画像処理

  平田健司, 藤後廉, 小川貴弘, 長谷山美紀, 志賀哲  Innervision  34-  (7)  2019
• ユーザの好みを反映した潜在特徴を用いたGCNに基づくTumblrポストの推薦

  大友一馬, 原川良介, 小川貴弘, 長谷山美紀, 岩橋政宏  信号処理シンポジウム講演論文集(CD-ROM)  34th-  2019
• 変状分類におけるGrad-CAM++に基づいたCNNの注目領域の可視化に関する検討

  小川直輝, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• サッカー映像視聴時の視線データを用いた周辺視に基づく攻撃選手の重要度算出に関する検討

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• 画像注視時の視線データを用いた関心の識別に関する検討-LDCCAに基づく特徴量の射影による高精度化-

  松本真直, 斉藤直輝, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• Sparse Bayesian Learningに基づく注視領域の時間変化を考慮したヒトの感情推定に関する検討

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• 複数の特徴量から得られる類似度の統合に基づくサッカー映像における類似場面検索の高精度化に関する検討

  春山知生, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• 複数変量の生成モデルに基づく画像注視時の脳活動データを用いた画像カテゴリの推定に関する検討

  赤松祐亮, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• Text-to-imageGANに基づくスタイル変換に関する検討-Image-to-textモデル導入による高精度化-

  古寺恵, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• Semantic Segmentationに基づく地下鉄トンネルにおける変状検出に関する検討

  WANG An, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• 脳活動データを用いた注視画像の再構成における用いる視覚野の領域に関する検討

  高田紗弥, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• Twitterを用いた異常検知に基づく野球映像の重要シーン検出に関する検討

  平澤魁人, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2019-  2019
• AI・IoT・ビッグデータ解析に基づいた次世代の社会インフラ維持管理技術

  長谷山美紀  電子情報通信学会大会講演論文集(CD-ROM)  2019-  2019
• 生体情報解析を導入した次世代AI技術の構築

  長谷山美紀, 小川貴弘  電子情報通信学会大会講演論文集(CD-ROM)  2019-  2019
• Novel Audio Feature Projection Using KDLPCCA-Based Correlation with EEG Features for Favorite Music Classification.

  Ryosuke Sawata, Takahiro Ogawa, Miki Haseyama  IEEE Trans. Affective Computing  10-  (3)  430  -444  2019  [Not refereed][Not invited]
   

  IEEE A novel audio feature projection using Kernel Discriminative Locality Preserving Canonical Correlation Analysis (KDLPCCA)-based correlation with electroencephalogram (EEG) features for favorite music classification is presented in this paper. The projected audio features reflect individual music preference adaptively since they are calculated by considering correlations with the user & #x0027;s EEG signals during listening to musical pieces that the user likes/dislikes via a novel CCA proposed in this paper. The novel CCA, called KDLPCCA, can consider not only a non-linear correlation but also local properties and discriminative information of each class sample, namely, music likes/dislikes. Specifically, local properties reflect intrinsic data structures of the original audio features, and discriminative information enhances the power of the final classification. Hence, the projected audio features have an optimal correlation with individual music preference reflected in the user & #x0027;s EEG signals, adaptively. If the KDLPCCA-based projection that can transform original audio features into novel audio features is calculated once, our method can extract projected audio features from a new musical piece without newly observing individual EEG signals. Our method therefore has a high level of practicability. Consequently, effective classification of user & #x0027;s favorite musical pieces via a Support Vector Machine (SVM) classifier using the new projected audio features becomes feasible. Experimental results show that our method for favorite music classification using projected audio features via the novel CCA outperforms methods using original audio features, EEG features and even audio features projected by other state-of-the-art CCAs.
• Classification of tourism categories based on heterogeneous features considering existence of reliable results

  Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  Proceedings of SPIE - The International Society for Optical Engineering  11049-  2019/01/01  [Not refereed][Not invited]
   

  © COPYRIGHT SPIE. This paper presents a classification method of tourism categories based on heterogeneous features considering existence of reliable results. The proposed method performs estimation of existence of reliable results based on one-versus-one scheme from three kinds of classification results obtained from tourism images, geotags and textual tags, separately. Then if the reliable result is included in the above results, this result is regarded as a final result. Otherwise, the final result is obtained by the multiple annotator logistic regression. The proposed method realizes accurate classification by estimating the existence of reliable results from more than two kinds of results.
• Field position estimation in soccer videos using convolutional neural network-based image features

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  Proceedings of SPIE - The International Society for Optical Engineering  11049-  2019/01/01  [Not refereed][Not invited]
   

  © COPYRIGHT SPIE. This paper presents a novel estimation method of field positions in soccer videos using Convolutional Neural Network (CNN)-based image features. CNN-based features have been well known to be ei€ective for tasks in machine learning. Therefore, the proposed method adopts CNN-based image features. By using these image features, the proposed method enables accurate estimation of soccer field positions than handcrafted features, i.e., Hough transform-based features. We show the ei€ectiveness of our method via experiment results using actual soccer videos.
• Gaze-based visual feature extraction via DLPCCA for visual sentiment estimation

  Taiga Matsui, Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  Proceedings of SPIE - The International Society for Optical Engineering  11049-  2019/01/01  [Not refereed][Not invited]
   

  © COPYRIGHT SPIE. This paper presents gaze-based visual feature extraction via Discriminative Locality Preserving Canonical Correlation Analysis (DLPCCA) for visual sentiment estimation. The proposed method calculates novel visual features reflecting users' visual sentiment by applying DLPCCA to gaze and original visual features. Consequently, accurate visual sentiment estimation becomes feasible by utilizing the novel visual features derived by the proposed method.
• Automatic Metastatic Bone Tumor Classification with DCNN-based Features Using Treatment-planning CT Images

  Haruna Watanabe, Ren Togo, Takahiro Ogawa, Miki Haseyama, Koichi Yasuda, Khin Khin Tha, Kohsuke Kudo, Hiroki Shirato  INTERNATIONAL FORUM ON MEDICAL IMAGING IN ASIA 2019  11050-  2019  [Not refereed][Not invited]
   

  In this paper, we propose a method to classify metastatic bone tumors using treatment-planning computed tomography images. The proposed method utilizes pre-trained deep convolutional neural network (DCNN) models as feature extractors and enables the metastatic bone tumor classification by using the obtained features. Performance of several state-of-the-art DCNN-based features was compared and evaluated in our experiment.
• Multi-classifier Decision: Integration of Multiple Brain Activity-based Classifications

  Takahiro Ogawa, Kento Sugata, Ren Togo, Miki Haseyama  ITE Transactions on Media Technology and Applications  7-  (1)  36  -44  2019/01/01  [Not refereed][Not invited]
   

  Copyright © 2019 by ITE Transactions on Media Technology and Applications (MTA) A novel method that integrates brain activity-based classifications obtained from multiple users is presented in this paper. The proposed method performs decision-level fusion (DLF) of the classifications using a kernelized version of extended supervised learning from multiple experts (KESLME), which is newly derived in this paper. In this approach, feature-level fusion of multiuser electroencephalogram (EEG) features is performed by multiset supervised locality preserving canonical correlation analysis (MSLPCCA). In the proposed method, the multiple classification results are obtained by classifiers separately constructed for the multiuser EEG features. Then DLF of these classification results becomes feasible based on KESLME, which can provide the final decision with consideration of the relationship between the MSLPCCA-based integrated EEG features and each classifier’s performance. In this way, a new multi-classifier decision technique, which depends only on users’ brain activities, is realized, and the performance in an image classification task becomes comparable to that of Inception-v3, one of the state-of-the-art deep convolutional neural networks.
• User-centric Visual Attention Estimation Based on Relationship between Image and Eye Gaze Data

  Yuya Moroto, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  388  -389  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a method for estimating user-centric visual attention based on the relationship between image and eye gaze data. The proposed method focuses on relationship between visual features calculated from images and saliency values calculated from eye gaze data. Specifically, our method calculates the saliency map of each training image by using individual eye gaze data obtained from only these images. Furthermore, from the pairs of visual features and the gaze-based saliency, the estimation of user-centric saliency from a new test image becomes feasible. Our contribution is the construction of a simple but successful estimation model which can train the relationship from limited amount of individual eye gaze data. Experimental results show the effectiveness of the proposed method.
• Missing Image Data Estimation via Iterative Convolutional Sparse Coding

  Masanao Matsumoto, Naoki Saito, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  331  -333  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. A novel method for missing image data estimation is presented in this paper. The proposed method realizes accurate estimation of missing image data by iterating dictionary learning and Convolutional Sparse Coding (CSC). Specifically, our method iterates estimation of missing image data via CSC by using a dictionary that is constructed from a target image, and reconstruction of the dictionary by using the obtained estimation results. As the main contribution of our paper, the proposed method enables the missing image data estimation by using more suitable dictionaries obtained by this iterative scheme. Experimental results show high missing image data estimation performance by the proposed method.
• Detection of Deformed PPG Waveforms Based on a Ratio of Their Average Accelerations

  Ken Kawakami, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  40  -44  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a new method to detect deformed Photoplethysmogram (PPG) waveforms for sufficient accuracy of signal processing. The PPG waveforms have been applied in many health indicators, such as blood pressure, blood viscosity and blood vessel elasticity. Usually, the measurements using a sensitive signals require user awareness so that all the PPG waveforms are kept accurate. Namely, accuracy of the calculated indicators becomes lower when the PPG waveform is deformed due to motion artifacts. In particular, detection methods of deformed PPG waveforms are important for incorporating the health indicators into general fitness trackers to find the correct waveform or to remove deformed PPG waveforms from the measurement. Therefore, we propose a new method which detects a badly formed PPG waveform by monitoring a ratio of average accelerations. Experimental results show the effectiveness of the method for detecting a deformed PPG waveform.
• Team Tactics Estimation in Soccer Videos via Deep Extreme Learning Machine Based on Players Formation

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  289  -290  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. A method of team tactics estimation in soccer videos is presented in this paper. Our method enables estimation of basic tactics in each team on the basis of the Deep-Extreme Learning Machine (DELM) by using features of players formation. In the soccer games, team tactics relate to each other team. Therefore, the proposed method obtains final estimation results by utilizing two DELMs of each team and their relationship. Since the proposed method takes into consideration the relevance of the estimated tactics in each team, we realize accurate tactics estibimation. Experimental results using actual soccer videos showed the effectiveness of our method.
• Blood Circulation Based on PPG Signals for Thermal Comfort Evaluation

  Ken Kawakami, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  198  -199  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a new index for monitoring transition of blood circulation from Photoplethysmogram (PPG) signals for thermal comfort evaluation of users. Heat dissipation reaction through the dilation of blood vessels is person's intrinsic ability to control the thermal comfort. When body temperature is higher than normal temperature, blood circulation changes according to the dilation of blood vessels in distal end of the extremities. Blood circulation is often evaluated by an index of peripheral resistance corresponding to changes in blood flow velocity such as systolic/diastolic ratio S/D of flow velocities, resistance index and the pulsatility index. Unfortunately, such an index cannot be utilized in daily life for healthcare with using fitness trackers since the blood flow velocity is measured by either an ultrasonic Doppler blood flowmeter (UDF) or a Laser-Doppler flowmeter. Therefore, we propose a new index which is easily acquirable from PPG signals. First, a couple of variables correlating to the blood flow velocity is calculated from a rate of volumetric strain of Photoplethysmogram signals. Then the new index can be obtained as difference of these variables. Experimental results show the effectiveness of this index by confirming high correlation with S/D of UDF.
• Estimation of Viewed Image Categories via CCA Using Human Brain Activity

  Yusuke Akamatsu, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  159  -160  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a method to estimate viewed image categories via canonical correlation analysis (CCA) using human brain activity measured by functional magnetic resonance imaging (fMRI). The proposed method enables estimation of image categories that a subject viewed by using only the subject's brain activity. Specifically, the proposed method calculates the projection matrices that enable direct comparison between human brain activity and images that subjects viewed through CCA. After projecting the human brain activity and the viewed images on the same latent space, k-Nearest Neighbor (k-NN) is performed to estimate the viewed image categories from only human brain activity. Through the projection matrices, the proposed method can increase training data for k-NN even if a large number of pairs of human brain activity and images cannot be prepared. Experimental results for ten subjects show the effectiveness of the proposed method.
• Estimation of Important Scenes in Soccer Videos Based on Collaborative Use of Audio-Visual CNN Features

  Tomoki Haruyama, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  710  -711  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a novel method for estimating important scenes in soccer videos based on collaborative use of audio-visual Convolutional Neural Network (CNN) features. In soccer games, since game situations influence not only players' movements but also audiences' cheers, analyses of their audio and visual sequences are useful for the estimation of important scenes. In our method, such scenes are estimated from audio and visual CNN features via support vector machine (SVM) in each feature. Furthermore, by applying weighted majority voting based on confidences defined from the SVM-based estimation results, accurate estimation of important scenes becomes feasible. Experimental results show the effectiveness of our method.
• Twitter Followee Recommendation Based on Multimodal FFM Considering Social Relations

  Shoji Takimura, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7th Global Conference on Consumer Electronics, GCCE 2018  196  -197  2018/12/12  [Not refereed][Not invited]
   

  © 2018 IEEE. A method for Twitter followee recommendation based on multimodal field-aware factorization machines considering social relations (MFFM-SR) is presented. MFFM-SR enables collaborative use of textual and visual features and social relations unlike conventional methods. Specifically, for distinguishing users' interest, visual features are extracted from images in their tweets and icons as well as textual features and social relations. Furthermore, to construct a model that accurately represents users' interest, MFFM-SR that enables calculation of high-level features via estimation of latent relationships among the obtained features and social relations is derived. By using the constructed model, successful followee recommendation becomes feasible.
• Binary sparse representation based on arbitrary quality metrics and its applications

  Takahiro Ogawa, Sho Takahashi, Naofumi Wada, Akira Tanaka, Miki Haseyama  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E101A-  (11)  1776  -1785  2018/11/01  [Not refereed][Not invited]
   

  © 2018 The Institute of Electronics, Information and Communication Engineers. Binary sparse representation based on arbitrary quality metrics and its applications are presented in this paper. The novelties of the proposed method are twofold. First, the proposed method newly derives sparse representation for which representation coefficients are binary values, and this enables selection of arbitrary image quality metrics. This new sparse representation can generate quality metric-independent subspaces with simplification of the calculation procedures. Second, visual saliency is used in the proposed method for pooling the quality values obtained for all of the parts within target images. This approach enables visually pleasant approximation of the target images more successfully. By introducing the above two novel approaches, successful image approximation considering human perception becomes feasible. Since the proposed method can provide lower-dimensional subspaces that are obtained by better image quality metrics, realization of several image reconstruction tasks can be expected. Experimental results showed high performance of the proposed method in terms of two image reconstruction tasks, image inpainting and super-resolution.
• Automatic detection of fish sounds based on multi-stage classification including logistic regression via adaptive feature weighting

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama, Tomonari Akamatsu  Journal of the Acoustical Society of America  144-  (5)  2709  -2718  2018/11/01  [Not refereed][Not invited]
   

  © 2018 Acoustical Society of America. This paper presents a method for automatic detection of fish sounds in an underwater environment. There exist two difficulties: (i) features and classifiers that provide good detection results differ depending on the underwater environment and (ii) there are cases where a large amount of training data that is necessary for supervised machine learning cannot be prepared. A method presented in this paper (the proposed hybrid method) overcomes these difficulties as follows. First, novel logistic regression (NLR) is derived via adaptive feature weighting by focusing on the accuracy of classification results by multiple classifiers, support vector machine (SVM), and k-nearest neighbors (k-NN). Although there are cases where SVM or k-NN cannot work well due to divergence of useful features, NLR can produce complementary results. Second, the proposed hybrid method performs multi-stage classification with consideration of the accuracy of SVM, k-NN, and NLR. The multi-stage acquisition of reliable results works adaptively according to the underwater environment to reduce performance degradation due to diversity of useful classifiers even if abundant training data cannot be prepared. Experiments on underwater recordings including sounds of Sciaenidae such as silver croakers (Pennahia argentata) and blue drums (Nibea mitsukurii) show the effectiveness of the proposed hybrid method.
• 画像注視時の注視領域の時間変化を考慮したテンソル解析に基づく感情推定に関する検討

  諸戸祐哉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.92  2018/10/27  [Not refereed][Not invited]
  
• 地下鉄トンネルの点検における視線データを用いた熟練度の分析に関する一考察

  斉藤僚汰, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.11  2018/10/27  [Not refereed][Not invited]
  
• 深層学習に基づく画像特徴量を利用した放射線治療用CT画像における転移性骨腫瘍の検出に関する検討

  渡邊はるな, 藤後廉, 小川貴弘, 長谷山美紀, 安田耕一, THA Khin Khin, 工藤與亮, 白土博樹  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.87  2018/10/27  [Not refereed][Not invited]
  
• AC‐GANに基づく胃炎識別のための画像生成に関する検討

  金井美岬, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.89  2018/10/27  [Not refereed][Not invited]
  
• Tri‐trainingに基づく胃X線画像を用いた胃炎の識別に関する検討

  LI Zongyao, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.88  2018/10/27  [Not refereed][Not invited]
  
• AttnGANを用いたシーン検索に関する検討―再検索の導入による高精度化―

  柳凜太郎, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  ROMBUNNO.10  2018/10/27  [Not refereed][Not invited]
  
• Simultaneous Realization of Multiple Music Video Applications Based on Heterogeneous Network Analysis Via Latent Link Estimation

  Yui Matsumoto, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  Proceedings - IEEE International Conference on Multimedia and Expo  2018-July-  1  -6  2018/10/08  [Not refereed][Not invited]
   

  © 2018 IEEE. To help users seek desired music videos and create attractive music videos, many methods that realize applications such as music video recommendation, captioning and generation have been proposed. In this paper, a novel method that realizes these applications simultaneously on the basis of heterogeneous network analysis via latent link estimation is proposed. To the best of our knowledge, this work is the first attempt to realize music video recommendation, captioning and generation simultaneously. The proposed method enables latent link estimation with consideration of multimodal information and multiple social metadata obtained from music videos via Laplacian multiset canonical correlation analysis. Thus, it becomes feasible to construct a heterogeneous network that enables direct comparison of audio, visual and textual information of music videos and user information on the same feature space. Furthermore, link prediction on the obtained heterogeneous network enables association with (i) user information and their desired audio information; (ii) audio information and textual information that describes contents of musical pieces; and (iii) audio information and visual information that represents contents of musical pieces visually. As a result, support for (i) music video recommendation; (ii) captioning; and (iii) generation becomes feasible, respectively. Experimental results for a real-world dataset constructed by using YouTube-8M show the effectiveness of the proposed method.
• Interest Level Estimation of Items via Matrix Completion Based on Adaptive User Matrix Construction

  Tetsuya Kushima, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  Proceedings - IEEE International Conference on Multimedia and Expo  2018-July-  1  -6  2018/10/08  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a novel method for interest level estimation of items via matrix completion based on adaptive user matrix construction. The proposed method introduces a new criterion for adaptively constructing a user matrix that consists of user behavior features and interest levels, which are evaluated by target users and similar users. In the estimation, the matrix completion via rank minimization using the truncated nuclear norm is applied to the constructed matrix. The proposed method enables both of the interest level estimation of the target users and the selection of the similar users suitable for the estimation by monitoring errors caused in the matrix completion algorithm. The caused errors indicate the minimum differences between the estimated interest levels and true ones, and they can be regarded as the criterion for both of the optimal estimation and the adaptive selection. Furthermore, the proposed method uses weight matrices for decreasing an influence of missing data on the estimation. Consequently, accurate estimation of the interest levels becomes feasible by using the adaptively constructed matrix. Experimental results obtained by applying the proposed method to users' behavior and interest data show the effectiveness of the proposed method.
• Semi-Supervised Multiple Feature Fusion for Video Preference Estimation

  Akira Toyoda, Takahiro Ogawa, Miki Haseyama  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings  2018-April-  891  -895  2018/09/10  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a new method to estimate user preferences for videos based on multiple feature fusion via semi-supervised Multiview Local Fisher Discriminant Analysis (sMvLFDA). The proposed method first extracts multiple visual features from videos and functional near-infrared spectroscopy (fNIRS) features from fNIRS signals recorded during watching videos. Next, we apply Locality Preserving Canonical Correlation Analysis (LPCCA) to each visual feature and fNIRS features and project each visual feature to the new feature spaces (fNIRS-based visual feature spaces). Consequently, since the correlation between each visual feature and fNIRS features which reflect user preferences is maximized, we can transform visual features into features which also reflect user preferences. In addition, we newly introduce sMvLFDA and fuse multiple fNIRS-based visual features via sMvLFDA. sMvLFDA fuses features while using labeled samples and unlabeled samples simultaneously to reduce overfitting to the labeled samples. Furthermore, sMvLFDA adequately uses complementary properties in multiple features. Therefore, it can be expected that the fused features are more effective for estimation of user preferences than each fNIRS-based visual feature. The main contribution of this paper is the new derivation of sMvLFDA. Consequently, by using the fused features, it becomes feasible to estimate user preferences for videos successfully.
• Sfemcca: Supervised Fractional-Order Embedding Multiview Canonical Correlation Analysis for Video Preference Estimation

  Yoshiki Ito, Takahiro Ogawa, Miki Haseyama  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings  2018-April-  3086  -3090  2018/09/10  [Not refereed][Not invited]
   

  © 2018 IEEE. In this paper, we present supervised fractional-order embedding multiview canonical correlation analysis (SFEMCCA). SFEMCCA is a CCA method realizing the following three points: (1) learning noisy data with small number of samples and large number of dimensions, (2) multiview learning that can integrate three or more kinds of features, and (3) supervised learning using labels corresponding to the samples. In real data, it is necessary to deal with high dimensional noisy data with limited number of samples, and there are many cases where three or more kinds of multimodal and supervised data are treated in order to calculate more accurate projections. Therefore, SFEMCCA, which takes the above advantages (1)-(3) into account, is effective for data obtained from real environments. From experimental results, it was confirmed that accuracy improvements using SFEMCCA were statistically significant compared to the several conventional methods of supervised multiview CCA.
• A Human-Centered Neural Network Model with Discriminative Locality Preserving Canonical Correlation Analysis for Image Classification

  Kazaha Horii, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2366  -2370  2018/08/29  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents a human-centered neural network model with discriminative locality preserving canonical correlation analysis (DLPCCA) for image classification. Although construction of multiple hidden layers adopted in recent deep learning methods is effective for extracting semantic features, a large amount of training images is required. In order to extract effective features for image classification successfully from a small amount of training images, the proposed method transforms visual features by using biological information obtained from image viewers as auxiliary information. The proposed method consists of two hidden layers. By constructing the first hidden layer, which can maximize canonical correlation between visual features and features based on biological information, the effective feature transformation can be realized. Specifically, the proposed method uses DLPCCA, which considers label information and preserves local structures. The second hidden layer constructed based on Extreme Learning Machine (ELM) enables classification. Consequently, since the first hidden layer performs the effective feature transformation, the proposed neural network model realizes accurate image classification from a quite small amount of training images.
• Anonymous Gastritis Image Generation via Adversarial Learning from Gastric X-Ray Images

  Ren Togo, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2082  -2086  2018/08/29  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents an anonymous gastritis image generation method based on a generative adversarial network approach. Since clinical individual data include highly confidential information, they must be handled carefully. Although data sharing is demanded to construct large-scale medical image datasets for deep learning-based recognition tasks, managing and annotating these data have been conducted manually. The proposed method enables the generation of anonymous images by an adversarial learning approach. Experimental results show that generated images by our method contribute to a gastritis recognition task. This will be helpful for constructing large-scale medical image datasets effectively.
• Anonymous Image Data Generation from Gastric X-Ray Images for Improving Gastritis Recognition Performance

  Ren Togo, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  2018 IEEE International Conference on Consumer Electronics-Taiwan, ICCE-TW 2018  2018/08/27  [Not refereed][Not invited]
   

  © 2018 IEEE. This paper presents an anonymous gastritis image generation method for improving gastritis recognition performance. We realize the generation of realistic gastritis images by considering label information. Experimental results showed that anonymous images generated by our method had a potential for a gastritis recognition task. Concretely, the recognition performance of a classifier constructed with the anonymous images outperformed the performance of the conventional image generation method-based classifier.
• Estimation of Deterioration Levels of Transmission Towers via Deep Learning Maximizing Canonical Correlation between Heterogeneous Features

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  IEEE Journal on Selected Topics in Signal Processing  12-  (4)  633  -644  2018/08/01  [Not refereed][Not invited]
   

  © 2007-2012 IEEE. This paper presents estimation of deterioration levels of transmission towers via deep learning maximizing the canonical correlation between heterogeneous features. In the proposed method, we newly construct a correlation-maximizing deep extreme learning machine (CMDELM) based on a local receptive field (LRF). For accurate deterioration level estimation, it is necessary to obtain semantic information that effectively represents deterioration levels. However, since the amount of training data for transmission towers is small, it is difficult to perform feature transformation by using many hidden layers such as general deep learning methods. In CMDELM-LRF, one hidden layer, which maximizes the canonical correlation between visual features and text features obtained from inspection text data, is newly inserted. Specifically, by using projections obtained by maximizing the canonical correlation as weight parameters of the hidden layer, feature transformation for extracting semantic information is realized without designing many hidden layers. This is the main contribution of this paper. Consequently, CMDELM-LRF realizes accurate deterioration level estimation from a small amount of training data.
• Inpainting Based on Low-dimensional Image Approximation and Its Applications

  小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  118-  (147)  1  -4  2018/07/20  [Not refereed][Not invited]
  
• Tourism category classification on image sharing services through estimation of existence of reliable results

  Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  ICMR 2018 - Proceedings of the 2018 ACM International Conference on Multimedia Retrieval  493  -496  2018/06/05  [Not refereed][Not invited]
   

  © 2018 ACM. A new tourism category classification method through estimation of existence of reliable classification results is presented in this paper. The proposed method obtains two kinds of classification results by applying a convolutional neural network to tourism images and applying a Fuzzy K-nearest neighbor algorithm to geotags attached to the tourism images. Then the proposed method estimates existence of reliable classification results in the above two results. If the reliable result is included, the result is selected as the final classification result. If any reliable result is not included, the final result is obtained by another approach based on a multiple annotator logistic regression model. Consequently, the proposed method enables accurate classification based on the new estimation scheme.
• Sentiment-aware personalized tweet recommendation through multimodal FFM

  Ryosuke Harakawa, Daichi Takehara, Takahiro Ogawa, Miki Haseyama  Multimedia Tools and Applications  77-  (14)  1  -19  2018/03/29  [Not refereed][Not invited]
   

  For realizing quick and accurate access to desired information and effective advertisements or election campaigns, personalized tweet recommendation is highly demanded. Since multimedia contents including tweets are tools for users to convey their sentiment, users’ interest in tweets is strongly influenced by sentiment factors. Therefore, successful personalized tweet recommendation can be realized if sentiment in tweets can be estimated. However, sentiment factors were not taken into account in previous works and the performance of previous methods may be limited. To overcome the limitation, a method for sentiment-aware personalized tweet recommendation through multimodal Field-aware Factorization Machines (FFM) is newly proposed in this paper. Successful personalized tweet recommendation becomes feasible through the following three contributions: (i) sentiment factors are newly introduced into personalized tweet recommendation, (ii) users’ interest is modeled by deriving multimodal FFM that enables collaborative use of multiple factors in a tweet, i.e., publisher, topic and sentiment factors, and (iii) the effectiveness of using sentiment factors as well as publisher and topic factors is clarified from results of experiments using real-world datasets related to worldwide hot topics, “#trump”, “#hillaryclinton” and “#ladygaga”. In addition to showing the effectiveness of the proposed method, the applicability of the proposed method to other tasks such as advertisement and social analysis is discussed as a conclusion and future work of this paper.
• Tag refinement based on multilingual tag hierarchies extracted from image folksonomy

  Shota Hamano, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  1327  -1331  2018/02/20  [Not refereed][Not invited]
   

  This paper presents a novel method for tag refinement using multilingual sources of tagged images in an image folksonomy. The proposed method enables accurate tag refinement by effectively leveraging multilingual sources of tags and considering the hierarchical structure of tags in the following way. First, synonymous tags across different languages are detected based on similarities between tagged images. In this stage, the proposed method utilizes visual similarities to effectively detect synonymous tags since the visual features extracted from images should be similar if they are assigned tags with the same meaning in different languages. Then hierarchical structure of the tags are extracted based on the similarity between the detected synonymous tags. The hierarchical structure provides hypernymous and hyponymous tags of the target tags, which are important for considering the relevance between tags and images. Consulting the hierarchical structure enables removal of irrelevant tags from the images and assignment of relevant tags to the images. The proposed method effectively utilizes tags in various languages in an image folksonomy. Experimental results show the effectiveness of introducing multilingual sources of tagged images for accuracy improvement in tag refinement.
• Image retrieval based on LRGA algorithm and relevance feedback for insect identification

  Susumu Gerund, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  3978  -3982  2018/02/20  [Not refereed][Not invited]
   

  This paper presents an image retrieval method based on local regression and global alignment (LRGA) algorithm and relevance feedback for insect identification. Based on LRGA algorithm, the proposed method enables estimation of ranking scores for image retrieval in such a way that the neighborhood structure of the database can be optimally preserved. This is the biggest contribution of this paper. Then our method measures relevance between the query image and all the images in the database and realizes retrieval of images based on the measured relevance. Furthermore, if positively labeled images obtained by a user are available, they are used as the query relevance information for the relevance feedback to improve the retrieval results. Experimental results show the effectiveness of our method.
• MvLFDA-based video preference estimation using complementary properties of features

  Akira Toyoda, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  635  -639  2018/02/20  [Not refereed][Not invited]
   

  This paper presents a new method to estimate users' video preferences using complementary properties of features via Multiview Local Fisher Discriminant Analysis (MvLFDA). The proposed method first extracts multiple visual features from video frames and electroencephalogram (EEG) features from users' EEG signals recorded during watching video. Then we calculate EEG-based visual features by applying Locality Preserving Canonical Correlation Analysis (LPCCA) to each visual feature and EEG features. The EEG-based visual features reflect users' preferences since the correlation between visual features and EEG features which reflect users' preferences is maximized. Next, MvLFDA, which is newly derived in this paper, integrates multiple EEG-based visual features. Since MvLFDA explores complementary properties of different features, it can be expected that the features obtained by integrating multiple EEG-based visual features are more effective for users' preference estimation than each EEG-based visual feature. The biggest contribution of this paper is the new derivation of MvLFDA. Then successful estimation of users' video preferences becomes feasible using features obtained by MvLFDA.
• Automatic estimation of deterioration level on transmission towers via deep extreme learning machine based on local receptive field

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  2379  -2383  2018/02/20  [Not refereed][Not invited]
   

  This paper presents an automatic estimation method of deterioration levels on transmission towers via Deep Extreme Learning Machine based on Local Receptive Field (DELM-LRF). Although Convolutional Neural Network (CNN) requires a large number of training images, it is difficult to prepare a sufficient number of training images of transmission towers. Thus, we generate a novel estimation method which enables training from a small number of training images. Specifically, we automatically extract image features based on Local Receptive Field (LRF) which combines convolution and pooling without using hand-craft features and estimate deterioration levels via Deep Extreme Learning Machine (DELM), which is a part of efficient deep learning methods. The derivation of DELM-LRF is the biggest contribution of this paper, and it can be trained from less training images compared to CNN. Experimental results show the effectiveness of DELM-LRF for the estimation of deterioration levels on transmission towers. Consequently, the proposed method makes it possible to approach challenging tasks with high expertise having difficulty in preparing enough images.
• Detection of gastric cancer risk from X-ray images via patch-based convolutional neural network

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  2055  -2059  2018/02/20  [Not refereed][Not invited]
   

  This paper presents a novel detection method of gastric cancer risk from X-ray images using the patch-based Convolutional Neural Network (CNN). Our method enables the training of the patch-based CNN which can accurately detect gastric cancer risk even though there is only the image-level ground truth. Furthermore, the proposed method can extract a feature vector that can represent the whole of symptoms associated with the presence or absence of the risk. Specifically, the proposed method selects the patches related to their true risk via the CNN, and it is the most innovative contribution of our method. Moreover, we extract the feature vector by applying the Bag-of-Feature representation to the output values from the CNN's intermediate layer obtained from the selected patches. Finally, the detection of gastric cancer risk is performed by inputting the extracted feature vector into Support Vector Machine. Experimental results confirm that the proposed method outperforms a previously reported method that combines the detection results obtained from X-ray images taken from multiple angles even though the proposed method only uses an X-ray image taken from a single angle, and we can achieve a higher performance than that of doctors.
• Automatic martian dust storm detection via decision level fusion basedondeep extreme learning machine

  Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  Proceedings - International Conference on Image Processing, ICIP  2017--  435  -439  2018/02/20  [Not refereed][Not invited]
   

  This paper presents an automatic Martian dust storm detection via decision level fusion (DLF) based on deep extreme learning machine (DELM). Since Martian images are taken in multi-wavelength bands, DLF techniques which output a final classification result by integrating multiple classification results are necessary. Furthermore, since the number of Martian images taken by satellites is different for each region, the number of the classification results to be integrated is different. Thus, we present a new DLF framework based on confidence values of the classification results. Specifically, we generate multiple extreme learning machines with kernel classifiers to obtain their classification results. Moreover, we monitor the classification results as confidence values and select the same number of the classification results with high confidence for each region. Finally, these selected results can be integrated by using a DLF based on DELM, which is a multilayered ELM. This integration framework is the biggest contribution of our method. Experimental results show the effectiveness of the DLF based on DELM.
• A Study on Active Net-based Estimation of Pass Region in Rugby Videos

  高橋 翔, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  117-  (432)  137  -142  2018/02/15  [Not refereed][Not invited]
  
• A Note on Image Retrieval Based on LRGA Algorithm for Insect Identification : Verification of Retrieval Accuracy Improvement Based on Data Augmentation

  弦間奨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  167  -171  2018/02/15  [Not refereed][Not invited]
  
• A Note on Pedestrian Tracking on Real CCTV Data of Rivers Using Pedestrian Detection via Background Subtraction and Deep Learning

  川本舜, 石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  179  -183  2018/02/15  [Not refereed][Not invited]
  
• Trial Introduction of Convolutional Sparse Coding for Accurate Distress Classification of Road Structures

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  189  -194  2018/02/15  [Not refereed][Not invited]
  
• A Note on Recurrent Neural Network-based Tactics Estimation in Soccer Videos

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  131  -135  2018/02/15  [Not refereed][Not invited]
  
• A Note on Accuracy Improvement of Category Classification for Tourism Images in Image Sharing Services Using Deep Learning

  斉藤直輝, 小川貴弘, 浅水仁, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  161  -165  2018/02/15  [Not refereed][Not invited]
  
• A note on video preference estimation using fNIRS signals : Fusion of multimodal features and introduction of distance metric learning

  豊田陽, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  173  -177  2018/02/15  [Not refereed][Not invited]
  
• A Note on Classification of Tunnel Working Face Images toward Estimation of Rock Elastic Wave Velocity

  濱野翔太, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  185  -188  2018/02/15  [Not refereed][Not invited]
  
• A note on accurate retrieval of similar inspection records based on canonical correlation between eye tracking data and inspection records

  斉藤僚汰, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  195  -200  2018/02/15  [Not refereed][Not invited]
  
• A Note on Estimation of Users' Emotion Evoked During Listening to Music : Performance Improvement Based on Deep Learning Methods

  段博瀟, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  201  -206  2018/02/15  [Not refereed][Not invited]
  
• Accuracy Improvement of Preference Estimation for Video Using SFEM-GS

  伊藤良起, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  315  -318  2018/02/15  [Not refereed][Not invited]
  
• A note on classification of brain metastases from MR images based on machine learning

  菅田健斗, 小川貴弘, 長谷山美紀, THA Khin Khin, THA Khin Khin, 工藤與亮, 工藤與亮, 白土博樹, 白土博樹  電子情報通信学会技術研究報告  117-  (431)  311  -314  2018/02/15  [Not refereed][Not invited]
  
• A Note on Use of Generative Adversarial Networks for Gastritis Classification from Gastric X-ray images

  藤後廉, 石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  299  -303  2018/02/15  [Not refereed][Not invited]
  
• Accurate Detection of Gastric Cancer Risk from X-ray Images via Deep Learning : Application to X-ray Images Taken from Multiple Angles

  石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  305  -309  2018/02/15  [Not refereed][Not invited]
  
• A note on the prevention of heatstroke using sensing data : Analysis of sensing data based on worker's physical condition

  佐坂勇磨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  117-  (431)  327  -331  2018/02/15  [Not refereed][Not invited]
  
• Preliminary study of automatic gastric cancer risk classification from photofluorography

  Ren Togo, Kenta Ishihara, Katsuhiro Mabe, Harufumi Oizumi, Takahiro Ogawa, Mototsugu Kato, Naoya Sakamoto, Shigemi Nakajima, Masahiro Asaka, Miki Haseyama  World Journal of Gastrointestinal Oncology  10-  (2)  62  -70  2018/02/15  [Not refereed][Not invited]
   

  AIM To perform automatic gastric cancer risk classification using photofluorography for realizing effective mass screening as a preliminary study. METHODS We used data for 2100 subjects including X-ray images, pepsinogen ? and ? levels, PG?/PG? ratio, Helicobacter pylori (H. pylori ) antibody, H. pylori eradication history and interview sheets. We performed two-stage classification with our system. In the first stage, H. pylori infection status classification was performed, and H. pylori -infected subjects were automatically detected. In the second stage, we performed atrophic level classification to validate the effectiveness of our system. RESULTS Sensitivity, specificity and Youden index (YI) of H. pylori infection status classification were 0.884, 0.895 and 0.779, respectively, in the first stage. In the second stage, sensitivity, specificity and YI of atrophic level classification for H. pylori -infected subjects were 0.777, 0.824 and 0.601, respectively. CONCLUSION Although further improvements of the system are needed, experimental results indicated the effectiveness of machine learning techniques for estimation of gastric cancer risk.
• A Study on Active Net-based Estimation of Pass Region in Rugby Videos

  高橋 翔, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  137  -142  2018/02  [Not refereed][Not invited]
  
• A Note on Recurrent Neural Network-based Tactics Estimation in Soccer Videos

  鈴木 元樹, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  131  -135  2018/02  [Not refereed][Not invited]
  
• A Note on Accuracy Improvement of Category Classification for Tourism Images in Image Sharing Services Using Deep Learning

  斉藤 直輝, 小川 貴弘, 浅水 仁, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  161  -165  2018/02  [Not refereed][Not invited]
  
• A Note on Image Retrieval Based on LRGA Algorithm for Insect Identification : Verification of Retrieval Accuracy Improvement Based on Data Augmentation

  弦間 奨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  167  -171  2018/02  [Not refereed][Not invited]
  
• A note on video preference estimation using fNIRS signals : Fusion of multimodal features and introduction of distance metric learning

  豊田 陽, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  173  -177  2018/02  [Not refereed][Not invited]
  
• A Note on Pedestrian Tracking on Real CCTV Data of Rivers Using Pedestrian Detection via Background Subtraction and Deep Learning

  川本 舜, 石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  179  -183  2018/02  [Not refereed][Not invited]
  
• A Note on Classification of Tunnel Working Face Images toward Estimation of Rock Elastic Wave Velocity

  濱野 翔太, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  185  -188  2018/02  [Not refereed][Not invited]
  
• Trial Introduction of Convolutional Sparse Coding for Accurate Distress Classification of Road Structures

  前田 圭介, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  189  -194  2018/02  [Not refereed][Not invited]
  
• A note on accurate retrieval of similar inspection records based on canonical correlation between eye tracking data and inspection records

  斉藤 僚汰, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  195  -200  2018/02  [Not refereed][Not invited]
  
• 楽曲聴取時のユーザに誘起される感情の推定に関する検討 : 深層学習の導入による高精度化 (マルチメディアストレージ ヒューマンインフォメーション メディア工学 映像表現&コンピュータグラフィックス)

  段 博瀟, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  201  -206  2018/02  [Not refereed][Not invited]
  
• A Note on Use of Generative Adversarial Networks for Gastritis Classification from Gastric X-ray images

  藤後 廉, 石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  299  -303  2018/02  [Not refereed][Not invited]
  
• Accurate Detection of Gastric Cancer Risk from X-ray Images via Deep Learning : Application to X-ray Images Taken from Multiple Angles

  石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  305  -309  2018/02  [Not refereed][Not invited]
  
• A note on classification of brain metastases from MR images based on machine learning

  菅田 健斗, 小川 貴弘, 長谷山 美紀, タ キンキン, 工藤 與亮, 白土 博樹  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  311  -314  2018/02  [Not refereed][Not invited]
  
• Accuracy Improvement of Preference Estimation for Video Using SFEM-GS

  伊藤 良起, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  315  -318  2018/02  [Not refereed][Not invited]
  
• A note on the prevention of heatstroke using sensing data : Analysis of sensing data based on worker's physical condition

  佐坂 勇磨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  42-  (4)  327  -331  2018/02  [Not refereed][Not invited]
  
• Leading edge of Education for Mathematical and Data Science in Hokkaido University-Implementation of a Custom-made Practical Education Program-

  長谷山美紀, 大本亨, 高橋翔, 長谷山美紀, 大本亨  電子情報通信学会技術研究報告  117-  (431(ITS2017 61-83))  2018
• Leading edge of Education for Mathematical and Data Science in Hokkaido University-Implementation of General Education Program and Specialized Education Program-

  長谷山美紀, 大本亨, 高橋翔, 長谷山美紀, 大本亨  電子情報通信学会技術研究報告  117-  (431(ITS2017 61-83))  2018
• Trends of Education and Research for Mathematical and Data Science-Implementation in Education and Research Center for Mathematical and Data Science, Hokkaido University-

  長谷山美紀, 大本亨, 高橋翔, 長谷山美紀, 大本亨  電子情報通信学会技術研究報告  117-  (431(ITS2017 61-83))  2018
• バイオミメティクス画像検索:情報科学が繋ぐ博物学とナノテクノロジー

  長谷山美紀  CSJ Current Review  (28)  2018
• フラクタル画像符号化における復号画質とフラクタル存在量の関係に関する考察

  竹沢恵, 真田博文, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 複数の分類器から得られる確信度に注目したサッカー映像における重要シーンの検出に関する検討

  春山知生, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• SemiCCAに基づくfMRIデータを用いた注視画像のカテゴリ推定の高精度化に関する検討

  赤松祐亮, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 複数の識別器から得られる予測値の統合に基づくフォロイー推薦の高精度化に関する検討

  滝村祥司, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• FCNとCNNを用いた地下鉄トンネルにおける変状検出に関する検討

  WANG An, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 視線データを用いた画像注視時の感情推定に関する検討-正準相関分析に基づく画像特徴変換手法の導入-

  松井太我, 斉藤直輝, 小川貴弘, 浅水仁, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 深層学習に基づく特徴量を用いたサッカー映像に撮像されたフィールド位置の推定に関する検討

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• テンソル補完に基づいたユーザの動作からの関心度推定に関する検討

  九島哲哉, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 地下鉄トンネルにおける変状評価支援のための類似点検データ検索の初期検討

  弦間奨, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 河川CCTV映像を用いた利用者の行動分類に関する初期検討

  川本舜, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• トンネル切羽画像を用いた岩盤の弾性波速度の推定に関する検討

  山本健太郎, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 画像共有サービスにおける画像の人気度予測に関する検討-回帰問題のための特徴量変換手法の導入-

  豊田陽, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• 反復型のMulti-channel Convolutional Sparse Codingを用いた欠損した画素値の推定に関する検討

  松本真直, 斉藤直輝, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2018-  2018
• A RETRIEVAL METHOD OF SIMILAR INSPECTION RECORDS BASED ON EXPERIENCED INSPECTORS' EVALUATION

  SAITO Ryota, TAKAHASHI Sho, OGAWA Takahiro, HASEYAMA Miki  土木学会論文集F3（土木情報学）  74-  (1)  67  -77  2018  [Not refereed][Not invited]
  
• Image Retrieval from Vague Description Based on AttnGAN

  Rintaro Yanagi, Ren Togo, Takahiro Ogawa, Miki Haseyama  2018 IEEE 7TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE 2018)  198  -199  2018  [Not refereed][Not invited]
   

  Image retrieval plays an important role in the information society. Many studies have been conducted to improve accuracy of the image retrieval. However, there exists a major limitation in their input methods. For example, if users only have a vague description that does not include detailed information such as its name and do not have an appropriate input image, it is difficult to retrieve their desired images. To solve this problem, we propose a novel image retrieval method that enables retrieval of a desired image from a vague description. In the proposed method, we generate a query image from a vague description through an Attentional Generative Adversarial Network. By using the generated query image, the proposed method enables users to retrieve images even if they do not have a clear retrieval description as an input. Experimental results show the effectiveness of our method.
• A Language-Independent Ontology Construction Method Using Tagged Images in Folksonomy

  Shota Hamano, Takahiro Ogawa, Miki Haseyama  IEEE Access  6-  2930  -2942  2017/12/21  [Not refereed][Not invited]
   

  This paper presents a language-independent ontology (LION) construction method that uses tagged images in an image folksonomy. Existing multilingual frameworks that construct an ontology deal with concepts translated on the basis of parallel corpora, which are not always available however, the proposed method enables LION construction without parallel corpora by using visual features extracted from tagged images as the alternative. In the proposed method, visual similarities in tagged images are leveraged to aggregate synonymous concepts across languages. The aggregated concepts take on intrinsic semantics of themselves, while they also hold distinct characteristics in different languages. Then relationships between concepts are extracted on the basis of visual and textual features. The proposed method constructs a LION whose nodes and edges correspond to the aggregated concepts and relationships between them, respectively. The LION enables successful image retrieval across languages since each of the aggregated concepts can be referred to in different languages. Consequently, the proposed method removes the language barriers by providing an easy way to access a broader range of tagged images for users in the folksonomy, regardless of the language they use.
• Aesthetic quality assessment of images via Supervised Locality Preserving CCA

  Misaki Kanai, Ren Togo, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  Aesthetic quality assessment plays an important role in how people organize large image collections. Many studies on aesthetic quality assessment are based on design of hand-crafted features without considering whether attributes conveyed by images can actually affect image aesthetics. This paper presents an aesthetic quality assessment method which uses new visual features. The proposed method utilizes Supervised Locality Preserving Canonical Correlation Analysis (SLPCCA) to derive the new features which maximize correlation between attributes and visual features. Finally, by applying ridge regression to the SLPCCA-based features, successful aesthetic quality assessment is realized.
• Construction of network using heterogeneous social metadata for music video recommendation

  Yui Matsumoto, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  A novel method to construct a network based on heterogeneous features obtained from music videos and social metadata for music video recommendation is presented in this paper. The proposed method enables construction of the network that can accurately associate users with music videos corresponding to their preference by the collaborative use of audio and textual features obtained from music videos and social metadata 'related videos', 'tags', and 'keywords' through sub-sampled canonical correlation analysis. By performing link prediction on the obtained network, our method enables users to obtain desired music videos that are not linked to each other in the network but corresponding to users' preference, that is, music video recommendation becomes feasible. Experimental results for real-world datasets show the effectiveness of our method.
• Interest level estimation based on matrix completion via rank minimization

  Tetsuya Kushima, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  This paper presents a novel method for interest level estimation based on matrix completion via rank minimization. The proposed method estimates interest levels of target objects from human behavior features which are extracted during selecting these objects. Specifically, by adopting matrix completion via rank minimization, unknown interest levels can be estimated. Furthermore, the proposed method can also estimate unknown interest levels with some missing behavior features which are not correctly extracted by sensors. Experimental results show the effectiveness of the proposed method.
• Personalized tweet recommendation based on field-aware factorization machines with adaptive field organization

  Shoji Takimura, Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  A novel method for personalized tweet recommendation based on Field-aware Factorization Machines (FFMs) with adaptive field organization is presented in this paper. The proposed method realizes accurate recommendation of tweets in which users are interested by the following two contributions. First, sentiment factors such as opinions, thoughts and feelings included in tweets are newly introduced into FFMs in addition to their publisher and topic factors. Second, the proposed method newly enables adaptive organization of fields via canonical correlation analysis for multiple features extracted from each tweet. Experimental results for real-world datasets confirm the performance improvement of personalized tweet recommendation through the two contributions.
• Image classification for trend prediction based on integration of fNIRS and visual features

  Kazaha Horii, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  This paper presents a novel method of image classification for trend prediction based on integration of visual and fNIRS features. It is expected that classification of images in the same object category in terms of generation enables trend prediction. However, since images in the same object category have similar visual features, a limit of accuracy exists for image classification by using only visual features. To overcome this problem, we utilize fNIRS features which represent brain activity in addition to visual features. Specifically, we apply Discriminative Locality Preserving Canonical Correlation Analysis (DLPCCA) to fNIRS and visual features for utilizing them collaboratively. The main contribution of this paper is the improvement of classification performance of images in the same object category for trend prediction by using the visual features projected to the DLPCCA-based space.
• Video preference estimation using fNIRS signals

  Akira Toyoda, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  This paper presents a new method for video preference estimation using functional near-infrared spectroscopy signals (fNIRS signals). The proposed method first computes fNIRS features from fNIRS signals recorded while users are watching videos and multiple visual features from these videos. Next, by applying Locality Preserving Canonical Correlation Analysis to fNIRS features and each visual feature, we can obtain multiple new visual features. In addition, Multiview Local Fisher Discriminant Analysis fuses multiple new visual features and optimizes within and between class scatter in the fused feature space while using complementary properties in these features. Consequently, we can realize video preference estimation by using the fused features.
• Preference estimation for video recommendation using DCNN features and viewing behavior

  Yoshiki Ito, Takahiro Ogawa, Miki Haseyama  2017 IEEE 6th Global Conference on Consumer Electronics, GCCE 2017  2017--  1  -2  2017/12/19  [Not refereed][Not invited]
   

  This paper presents a personalized preference estimation method for video recommendation. Our method not only uses deep convolutional neural network (DCNN)-based video features but also transforms them based on user's viewing behavior in order to improve accuracy of preference estimation for a video. Specifically, we adopt supervised multi-view canonical correlation analysis (sMVCCA) in order to calculate 'canonical video features', which have a maximal correlation between the following three kinds of features: a video, user's viewing behavior and user's evaluation scores for the video. By using the canonical video features, our method can estimate the user's personalized preference for a video more accurately than using only the DCNN-based video features. Experimental results show the effectiveness of our method.
• Tracking topic evolution via salient keyword matching with consideration of semantic broadness for Web video discovery

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  Multimedia Tools and Applications  77-  (16)  1  -28  2017/12/06  [Not refereed][Not invited]
   

  A method to track topic evolution via salient keyword matching with consideration of semantic broadness for Web video discovery is presented in this paper. The proposed method enables users to understand the evolution of topics over time for discovering Web videos in which they are interested. A framework that enables extraction and tracking of the hierarchical structure, which contains Web video groups with various degrees of semantic broadness, is newly derived as follows: Based on network analysis using multimodal features, i.e., features of video contents and metadata, our method extracts the hierarchical structure and salient keywords that represent contents of each Web video group. Moreover, salient keyword matching, which is newly developed by considering salient keyword distribution, semantic broadness of each Web video group and initial topic relevance, is applied to each hierarchical structure obtained in different time stamps. Unlike methods in previous works, by considering the semantic broadness as well as the salient keyword distribution, our method can overcome the problem of the desired semantic broadness of topics being different depending on each user. Also, the initial topic relevance enables correction of the gap from an initial topic at the start of tracking. Consequently, it becomes feasible to track the evolution of topics over time for finding Web videos in which the users are interested. Experimental results for real-world datasets containing YouTube videos verify the effectiveness of the proposed method.
• Next Generation of Information Retrieval for Value Creation from Big Data-Introduction of AI and IoT Technologies into Associative Search-

  長谷山美紀  電子情報通信学会技術研究報告  117-  (329(CS2017 62-77))  101‐106  -106  2017/11/23  [Not refereed][Not invited]
  
• A Study on Detection of Similar Scenes based on Player Positions in Soccer Video

  髙橋 翔, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (41)  5  -8  2017/11  [Not refereed][Not invited]
  
• 異種特徴の関連性を活用したネットワーク解析に基づく音楽動画コンテンツの推薦に関する一検討―ユーザの好みを反映可能とする特徴変換による高精度化―

  松本有衣, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.112  2017/10/28  [Not refereed][Not invited]
  
• 画像特徴量とfNIRS特徴量の関連性に注目した画像分類の高精度化に関する検討

  堀井風葉, 前田圭介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.114  2017/10/28  [Not refereed][Not invited]
  
• fNIRS信号を利用した好みの映像識別に関する検討―特徴量選択の導入による高精度化―

  豊田陽, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.76  2017/10/28  [Not refereed][Not invited]
  
• フラクタル画像符号化における復号画質と空隙性の関係に関する考察

  竹沢恵, 真田博文, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.77  2017/10/28  [Not refereed][Not invited]
  
• 敵対的学習により生成された画像が与える審美的印象に関する検討

  金井美岬, 藤後廉, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.115  2017/10/28  [Not refereed][Not invited]
  
• 背景差分法と深層学習を用いた河川CCTV映像における歩行者検出の高精度化に関する検討

  川本舜, 石原賢太, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.75  2017/10/28  [Not refereed][Not invited]
  
• 適応的フィールド構成を導入したFFMに基づくツイート推薦の高精度化に関する検討

  滝村祥司, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.111  2017/10/28  [Not refereed][Not invited]
  
• 音響特徴と心拍特徴を用いた楽曲聴取時のユーザに誘起される感情の推定の高精度化に関する検討

  DUAN Boxiao, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.74  2017/10/28  [Not refereed][Not invited]
  
• ランク最小化に基づく行列補完を用いた関心度推定の高精度化に関する検討

  九島哲哉, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2017-  ROMBUNNO.113  2017/10/28  [Not refereed][Not invited]
  
• 正準相関最大化を導入した深層学習に基づく送電鉄塔の劣化レベル分類に関する検討

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  41-  (29)  11  -14  2017/09  [Not refereed][Not invited]
  
• Deep Extreme Learning Machineに基づくサッカー映像に撮像されたフィールド位置の推定に関する検討

  鈴木元樹, 高橋翔, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  41-  (29)  25  -28  2017/09  [Not refereed][Not invited]
  
• 変状評価支援のための類似点検データ検索の高精度化に関する検討―熟練技術者の判定に関する推定に基づいた学習データ生成法の導入―

  斉藤僚汰, 高橋翔, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  41-  (29)  49  -52  2017/09  [Not refereed][Not invited]
  
• 複数の条件で撮像された走査型電子顕微鏡画像を用いたゴム材料の物性値推定に関する検討

  斉藤直輝, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  41-  (29)  31  -34  2017/09  [Not refereed][Not invited]
  
• Distress Classification of Road Structures via Adaptive Bayesian Network Model Selection

  K. Maeda, S. Takahashi, T. Ogawa, M. Haseyama  JOURNAL OF COMPUTING IN CIVIL ENGINEERING  31-  (5)  2017/09  [Not refereed][Not invited]
   

  This paper presents an accurate distress classification method via adaptive Bayesian network model selection for maintenance inspection of road structures. The main contribution of this paper is adaptive selection of two Bayesian network models concerning classification performance. The proposed method trains a tag-based Bayesian network model based on inspection items and estimates its classification performance. Furthermore, for distresses that degrade the classification performance of the tag-based Bayesian network model, the proposed method trains another multifeature Bayesian network model based on inspection items and distress images. Consequently, the proposed method can adaptively select optimal Bayesian network models according to the estimated performance of the tag-based Bayesian network model. In actual maintenance inspection, distresses are generally classified either from inspection items alone or from both inspection items and visual information of distress images-i.e., distress classification has two patterns. Therefore the adaptive model selection approach is suitable for this classification scheme. Experimental results show that the proposed method outperforms several comparative methods and is suitable for actual maintenance inspection due to its low computation costs. (C) 2017 American Society of Civil Engineers.
• Estimation of Regions Related to Helicobacter Pylori Infection from Gastric X-ray Images

  Ren Togo, Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  7  -8  2017/07  [Not refereed][Not invited]
  
• Tracking hierarchical structure of web video groups based on salient keyword matching including semantic broadness estimation

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings  1238  -1242  2017/04/19  [Not refereed][Not invited]
   

  This paper presents a novel method to track the hierarchical structure of Web video groups on the basis of salient keyword matching including semantic broadness estimation. To the best of our knowledge, this paper is the first work to perform extraction and tracking of the hierarchical structure simultaneously. Specifically, the proposed method first extracts the hierarchical structure of Web video groups and salient keywords of them on the basis of an improved scheme of our previously reported method. Moreover, to calculate similarities between Web video groups obtained in different time stamps, salient keyword matching is newly developed by considering both co-occurrences of the salient keywords and semantic broadness of each Web video group. Consequently, tracking of the hierarchical structure over time becomes feasible to easily understand popularity trends of many Web videos for realizing effective retrieval.
• A note on quantification of relationship between users and musical pieces using graph structure analysis(2)Verification of using link prediction

  木下奨平, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  31  -34  2017/02/20  [Not refereed][Not invited]
  
• A Note on Selection of Representative Images for Deterioration Diagnosis of Steel Tower

  藤後廉, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  47  -50  2017/02/20  [Not refereed][Not invited]
  
• A Note on Accurate Distress Classification for Maintenance Inspection of Road Structures via Deep Learning

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  51  -54  2017/02/20  [Not refereed][Not invited]
  
• A Note on Property Estimation Using Electron Microscope Images and Mix Proportions of Rubbers

  斉藤直輝, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  41  -45  2017/02/20  [Not refereed][Not invited]
  
• A Study on Object Extraction in Steel Tower Videos Shot by Inspection of Transmission Line Tower

  館農浩平, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  25  -30  2017/02/20  [Not refereed][Not invited]
  
• A note on the estimation of viewer interest level using facial expression : Applying a framework for anomaly detection

  佐坂勇磨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  77  -80  2017/02/20  [Not refereed][Not invited]
  
• A Note on Retrieval Using Hierarchical Structure of Web Video Groups : Evaluation of Effectiveness through a Visualization Interface

  竹原大智, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  63  -66  2017/02/20  [Not refereed][Not invited]
  
• A Note on Deformation Detection in Subway Tunnel Using Convolutional Neural Network

  石原賢太, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  81  -86  2017/02/20  [Not refereed][Not invited]
  
• A Note on Auditory Target Detection via Collaborative Use of Auditory and fMRI Features

  田中佑磨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  59  -62  2017/02/20  [Not refereed][Not invited]
  
• A Note on Accurate Extraction of Concept Subsumption Relationships Using Tagged Images

  濱野翔太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  67  -70  2017/02/20  [Not refereed][Not invited]
  
• A note on estimating human emotion evoked by visual stimuli using fNIRS signals

  菅田健斗, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  87  -90  2017/02/20  [Not refereed][Not invited]
  
• A note on estimation of users' emotion evoked during listening to music : Performance improvement based on fusion of multiple estimation results

  段博瀟, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (463)  349  -353  2017/02/20  [Not refereed][Not invited]
  
• A Study on Object Extraction in Steel Tower Videos Shot by Inspection of Transmission Line Tower

  館農 浩平, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  25  -30  2017/02  [Not refereed][Not invited]
  
• A note on quantification of relationship between users and musical pieces using graph structure analysis(2)Verification of using link prediction

  木下 奨平, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  31  -34  2017/02  [Not refereed][Not invited]
  
• A Note on Property Estimation Using Electron Microscope Images and Mix Proportions of Rubbers

  斉藤 直輝, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  41  -45  2017/02  [Not refereed][Not invited]
  
• A Note on Selection of Representative Images for Deterioration Diagnosis of Steel Tower

  藤後 廉, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  47  -50  2017/02  [Not refereed][Not invited]
  
• A Note on Accurate Distress Classification for Maintenance Inspection of Road Structures via Deep Learning

  前田 圭介, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  51  -54  2017/02  [Not refereed][Not invited]
  
• A Note on Auditory Target Detection via Collaborative Use of Auditory and fMRI Features

  田中 佑磨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  59  -62  2017/02  [Not refereed][Not invited]
  
• A Note on Retrieval Using Hierarchical Structure of Web Video Groups : Evaluation of Effectiveness through a Visualization Interface

  竹原 大智, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  63  -66  2017/02  [Not refereed][Not invited]
  
• A Note on Accurate Extraction of Concept Subsumption Relationships Using Tagged Images

  濱野 翔太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  67  -70  2017/02  [Not refereed][Not invited]
  
• A note on the estimation of viewer interest level using facial expression : Applying a framework for anomaly detection

  佐坂 勇磨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  77  -80  2017/02  [Not refereed][Not invited]
  
• A Note on Deformation Detection in Subway Tunnel Using Convolutional Neural Network

  石原 賢太, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  81  -86  2017/02  [Not refereed][Not invited]
  
• A note on estimating human emotion evoked by visual stimuli using fNIRS signals

  菅田 健斗, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  87  -90  2017/02  [Not refereed][Not invited]
  
• A note on estimation of users' emotion evoked during listening to music : Performance improvement based on fusion of multiple estimation results

  段 博瀟, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  41-  (5)  349  -353  2017/02  [Not refereed][Not invited]
  
• 農業ICT-IoT・ビッグデータ・AI活用で農業を成長産業へ-3 農業ビッグデータ解析基盤の構築-発想支援型検索による害虫同定支援の試み-

  長谷山美紀  情報処理  58-  (9)  2017
• A method of important player extraction based on link analysis in soccer videos

  Sho Takahashi, Miki Haseyama  ITE Transactions on Media Technology and Applications  5-  (2)  42  -48  2017  [Not refereed][Not invited]
   

  In this paper, a method for extraction of important players in soccer videos based on link analysis is proposed. In a soccer match, players perform shoot tackles, assistance, and covering. Furthermore, the soccer tactics are defined the formation of players based on various relationships between players. The proposed method extracts the important players, in order to obtain information for understanding the soccer matches for various audiences. Specifically, our method notes that relationship between players, who cooperate with each other by the pass and the covering, is similar to relationship between web pages which are connected by links. First, the proposed method obtains player networks based on relationship between players in each team. The relationships are defined based on player positions and the possibility of the pass or the covering between players. Finally, in the proposed method, by applying the link analysis to the obtained player networks, important players are extracted. By realizing this approach, important players are extracted from the player networks based on the possibility of the pass or the covering between players. In the last of this paper, the above link analysis-based method was applied to actual soccer matches to show the reasonability of our method.
• バイオミメティクス画像検索エンジン―発想支援型画像検索の応用―

  長谷山美紀  可視化情報学会誌  37-  (144)  24‐29,1(4)  -29,図巻頭1p  2017/01/01  [Not refereed][Not invited]
  
• 特定集中治療室用の重症度推定に関する一考察~疾病による推定精度の違い~

  畠山泰貴, 小川貴弘, 田中宏和, 長谷山美紀  情報理論とその応用シンポジウム予稿集(CD-ROM)  39th-  ROMBUNNO.9.3.2  2016/12/06  [Not refereed][Not invited]
  
• A Study on Detection of Similar Scenes based on Pass Regions in Soccer Videos

  高橋翔, 長谷山美紀  映像情報メディア学会技術報告  40-  (43(ME2016 125-132))  9‐12  -12  2016/11/28  [Not refereed][Not invited]
   

  This paper proposes a pass region-based method for detecting similar scenes in soccer videos. In the soccer games, since the pass is very useful for analyses of various soccer tactics, the visualization of the pass regions is a very important task. Generally, since the tactics are defined by the formation of players, in the case of same tactic in two soccer scenes, the pass course is considered to be similar. Therefore, in this paper, we propose a detection method of similar scenes by utilizing the pass regions in soccer videos.
• 画像注視時における脳機能画像のテンソル解析に基づく感情推定に関する検討

  菅田健斗, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.109  2016/11/05  [Not refereed][Not invited]
  
• 深層学習によって得られる画像特徴量を用いた道路構造物の点検データ検索の高精度化に関する検討

  斉藤僚汰, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.114  2016/11/05  [Not refereed][Not invited]
  
• Signedネットワーク解析に基づくWeb映像集合の階層構造抽出に関する一検討

  原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.103  2016/11/05  [Not refereed][Not invited]
  
• 音響特徴と心拍特徴を用いた感情推定に基づく楽曲分類に関する検討―Decision‐Level Fusionによる高精度化―

  DUAN Boxiao, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.101  2016/11/05  [Not refereed][Not invited]
  
• 画像注視時のユーザの脳波信号を利用した好みの画像の識別に関する検討

  豊田陽, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.108  2016/11/05  [Not refereed][Not invited]
  
• 深層学習を用いた胃X線画像におけるHelicobacter pylori感染識別に関する初期検討

  石原賢太, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.115  2016/11/05  [Not refereed][Not invited]
  
• タグ付き画像を用いたオントロジー構築におけるコンセプト間の包摂関係の定量化に関する検討

  濱野翔太, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.102  2016/11/05  [Not refereed][Not invited]
  
• 昆虫の同定を目的とした類似画像検索の高精度化に関する検討

  弦間奨, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.100  2016/11/05  [Not refereed][Not invited]
  
• ユーザの視聴動作を用いた映像の評価値推定の高精度化に関する検討

  伊藤良起, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2016-  ROMBUNNO.107  2016/11/05  [Not refereed][Not invited]
  
• IoTヘルスケア時代の仮想生体情報センサ′′MIRUWS′′

  南重信, 長谷山美紀, 田中宏和, 高橋透, 小森達也  電子情報通信学会大会講演論文集(CD-ROM)  2016-  ROMBUNNO.BS‐6‐2  2016/09/06  [Not refereed][Not invited]
  
• Web映像検索を目的としたリランキングの高精度化に関する検討

  吉田壮, 小川貴弘, 長谷山美紀, 棟安実治  電子情報通信学会大会講演論文集(CD-ROM)  2016-  ROMBUNNO.A‐15‐5  2016/09/06  [Not refereed][Not invited]
  
• 映像情報を用いた物体追跡技術の最前線とスポーツ科学への応用動向 6章 サッカー映像における試合内容の理解を促すデータの可視化

  高橋翔, 長谷山美紀  映像情報メディア学会誌  70-  (5)  722‐724  2016/09/01  [Not refereed][Not invited]
  
• 発想支援型検索と異分野連携がもたらす可能性

  長谷山美紀  化学経済  63-  (11)  42‐46  2016/09/01  [Not refereed][Not invited]
  
• Category Classification of Tourism Images in Image Sharing Services

  SAITO Naoki, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  電子情報通信学会論文誌 D(Web)  99-  (9)  848  -860  2016/09/01  [Not refereed][Not invited]
   

  本論文では，画像共有サービスに投稿される画像の観光名所に関するカテゴリー（観光カテゴリー）の分類手法を提案する．提案手法では，画像共有サービスにおいて画像とともに投稿される様々なデータの中で，位置座標が観光カテゴリーの分類に最も効果的であることに注目する．更に，このデータに基づいた分類において誤りが発生する場合を，分類結果から導出される確信度に基づいて判定する．誤分類と判定された場合，提案手法では，画像及びタグの特徴に基づいた分類結果を求め，それらを各々の分類精度に注目して統合することで，最終的な分類結果を高精度に推定可能とする．このとき，位置座標による分類結果に基づいて分類候補の観光カテゴリーを限定することで，多クラス分類問題におけるクラス数を減少させ，より正確な分類を可能とする．以上の提案手法によって，異なる種類のデータの分類精度が互いに大きく異なる場合に従来手法では困難であった高精度な最終分類結果の取得が可能となる．
• マダラシミ鱗粉の摩擦特性

  平井悠司, 奥田直人, 斉藤直樹, 小川貴弘, 町田龍一郎, 野村周平, 大原昌宏, 長谷山美紀, 下村政嗣  高分子学会予稿集(CD-ROM)  65-  (2)  ROMBUNNO.2U01  2016/08/24  [Not refereed][Not invited]
  
• 音響聴取時の脳活動信号を用いた目的音検出に関する検討

  田中佑磨, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  40-  (28(ME2016 88-115))  81‐82  2016/08/24  [Not refereed][Not invited]
  
• 生物の走査型電子顕微鏡像の撮像部位の分類における特徴選択方法に関する検討

  斉藤直輝, 小川貴弘, 平井悠司, 長谷山美紀  映像情報メディア学会技術報告  40-  (28(ME2016 88-115))  19‐20  2016/08/24  [Not refereed][Not invited]
  
• 材料開発のための新しい情報科学―バイオミメティクス画像検索基盤の実装―

  長谷山美紀  高分子学会予稿集(CD-ROM)  65-  (2)  ROMBUNNO.1U12  2016/08/24  [Not refereed][Not invited]
  
• 個人の画像注視時の脳機能画像を利用した感情推定に関する検討―複数のブロードマン領野から得られる推定結果の統合による高精度化―

  菅田健斗, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  40-  (28(ME2016 88-115))  71‐72  2016/08/24  [Not refereed][Not invited]
  
• 火星に発生するdust stormの自動検出に関する検討―不均衡データを考慮した識別器の構築―

  前田圭介, 小川貴弘, 長谷山美紀  映像情報メディア学会技術報告  40-  (28(ME2016 88-115))  47‐48  2016/08/24  [Not refereed][Not invited]
  
• 高速道路の維持管理における点検記録データ検索に対する技術者の評価を用いた検索精度向上に関する検討

  高橋翔, 小川貴弘, 長谷山美紀  映像情報メディア学会年次大会講演予稿集(CD-ROM)  2016-  ROMBUNNO.22B‐1  2016/08/17  [Not refereed][Not invited]
  
• Multimedia Signal Processing and the Next Generation of Information Retrieval

  長谷山美紀  電子情報通信学会技術研究報告  116-  (167(HIP2016 31-41))  33‐36  2016/07/22  [Not refereed][Not invited]
  
• A Note on Helicobacter Pylori Infection Detection from Gastric X-ray Images(3)A Study on Relationship between Imaging Environment in Medical Facility and Detection Performance

  石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  116-  (107)  57  -62  2016/06/20  [Not refereed][Not invited]
  
• A Note on Helicobacter Pylori Infection Detection from Gastric X-ray Images(3)A Study on Relationship between Imaging Environment in Medical Facility and Detection Performance

  石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (18)  57  -62  2016/06/13  [Not refereed][Not invited]
  
• 生物から学ぶトライボロジー 生物表面の摩擦―自己組織化によるバイオミメティクス―

  下村政嗣, 平井悠司, 奥田直人, 町田龍一郎, 町田龍一郎, 野村周平, 大原昌宏, 長谷山美紀  トライボロジスト  61-  (4)  215‐221  2016/04/15  [Not refereed][Not invited]
  
• ボディエリアネットワークと今後の技術課題―ETSI SmartBAN標準化最新動向から―

  田中宏和, 南重信, 長谷山美紀, 新實朋子, 高橋透  電子情報通信学会大会講演論文集(CD-ROM)  2016-  (1)  ROMBUNNO.B‐20‐12  -659  2016/03/01  [Not refereed][Not invited]
  
• Quantitative Evaluation of Power Spectrum-based Sparse Representation of Audio Signals

  田中佑磨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458)  217  -220  2016/02/22  [Not refereed][Not invited]
  
• A note on quantification of relationship between users and musical pieces using graph structure analysis

  木下奨平, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  209‐212  2016/02/15  [Not refereed][Not invited]
  
• A Note on Accurate Detection of Helicobacter Pylori Infection from Gastric X-ray Images (3)

  藤後廉, 石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  333‐336  2016/02/15  [Not refereed][Not invited]
  
• A Note on Accurate Distress Classification for Maintenance Inspection of Road Structures-Integration of Multiple Classification Results based on Tag Data and Distress Images-

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  181‐184  2016/02/15  [Not refereed][Not invited]
  
• A Note on Hierarchical Topic Detection from Heterogeneous Social Media Contents

  竹原大智, 原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  193‐196  2016/02/15  [Not refereed][Not invited]
  
• A note on dimensionality reduction for Web image visualization

  館農浩平, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  337‐340  2016/02/15  [Not refereed][Not invited]
  
• A Note on Photographed Part Classification from Scanning Electron Microphotographs of Biological Organisms-Classification Performance Improvement of Decision Level Fusion Based on Kernel Method-

  斉藤直輝, 小川貴弘, 平井悠司, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  51‐54  2016/02/15  [Not refereed][Not invited]
  
• A Note on Detecting Time Series Variation of Hierarchical Structure of Web Video Groups

  原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  201‐204  2016/02/15  [Not refereed][Not invited]
  
• A Note on Data Analysis for Supporting Distress Evaluation in Expressway Maintenance-Retrieval of Similar Inspection Records Using Distress Images-

  三改木裕矢, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  169‐172  2016/02/15  [Not refereed][Not invited]
  
• A Note on Analysis of Gaze Data and Skill of Inspector in Embankment Inspection

  高橋翔, 三改木裕矢, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  177‐180  2016/02/15  [Not refereed][Not invited]
  
• Improving Performance of Graph-Based Web Video Search Reranking

  吉田壮, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  197‐200  2016/02/15  [Not refereed][Not invited]
  
• A Trial of Gaze Data Acquisition in Embankment Inspection of River Management

  三改木裕矢, 高橋翔, 小川貴弘, 秋山泰祐, 巖倉啓子, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  173‐176  2016/02/15  [Not refereed][Not invited]
  
• A Note on H. Pylori Infection Detection from Gastric X-ray images (2)-Introduction of Effective Image Feature Based on Serum Hp Antibody Level and Serum PG Levels-

  石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  329‐332  2016/02/15  [Not refereed][Not invited]
  
• A Note on fMRI-based Extraction of Active Brain Areas Related to Music Listening

  澤田充奨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  115-  (458(ITS2015 56-83))  213‐216  2016/02/15  [Not refereed][Not invited]
  
• A Note on Photographed Part Classification from Scanning Electron Microphotographs of Biological Organisms : Classification Performance Improvement of Decision Level Fusion Based on Kernel Method

  斉藤 直輝, 小川 貴弘, 平井 悠司, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  51  -54  2016/02  [Not refereed][Not invited]
  
• A Note on Data Analysis for Supporting Distress Evaluation in Expressway Maintenance : Retrieval of Similar Inspection Records Using Distress Images

  三改木 裕矢, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  169  -172  2016/02  [Not refereed][Not invited]
  
• A Trial of Gaze Data Acquisition in Embankment Inspection of River Management

  三改木 裕矢, 高橋 翔, 小川 貴弘, 秋山 泰祐, 巖倉 啓子, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  173  -176  2016/02  [Not refereed][Not invited]
  
• A Note on Analysis of Gaze Data and Skill of Inspector in Embankment Inspection

  高橋 翔, 三改木 裕矢, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  177  -180  2016/02  [Not refereed][Not invited]
  
• A Note on Accurate Distress Classification for Maintenance Inspection of Road Structures : Integration of Multiple Classification Results based on Tag Data and Distress Images

  前田 圭介, 高橋 翔, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  181  -184  2016/02  [Not refereed][Not invited]
  
• A Note on Hierarchical Topic Detection from Heterogeneous Social Media Contents

  竹原 大智, 原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  193  -196  2016/02  [Not refereed][Not invited]
  
• Improving Performance of Graph-Based Web Video Search Reranking

  吉田 壮, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  197  -200  2016/02  [Not refereed][Not invited]
  
• A Note on Detecting Time Series Variation of Hierarchical Structure of Web Video Groups

  原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  201  -204  2016/02  [Not refereed][Not invited]
  
• A note on quantification of relationship between users and musical pieces using graph structure analysis

  木下 奨平, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  209  -212  2016/02  [Not refereed][Not invited]
  
• A Note on fMRI-based Extraction of Active Brain Areas Related to Music Listening

  澤田 充奨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  213  -216  2016/02  [Not refereed][Not invited]
  
• Quantitative Evaluation of Power Spectrum-based Sparse Representation of Audio Signals

  田中 佑磨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  217  -220  2016/02  [Not refereed][Not invited]
  
• A Note on H. Pylori Infection Detection from Gastric X-ray images(2)Introduction of Effective Image Feature Based on Serum Hp Antibody Level and Serum PG Levels

  石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  329  -332  2016/02  [Not refereed][Not invited]
  
• A Note on Accurate Detection of Helicobacter Pylori Infection from Gastric X-ray Images(3)Estimation of Regions Causing Classification Performance Degradation

  藤後 廉, 石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  333  -336  2016/02  [Not refereed][Not invited]
  
• A note on dimensionality reduction for Web image visualization

  館農 浩平, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  40-  (6)  337  -340  2016/02  [Not refereed][Not invited]
  
• A Touism Category Classification Method Based on Estimation of Reliable Decision

  Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  In this paper, we propose a tourism category classification method based on estimation of reliable decision. The proposed method performs tourism category classification using location, visual, and textual tag features obtained from tourism images in image sharing services. As the biggest contribution of this paper, the proposed method performs successful classification based on two classification results obtained from a fuzzy K-nearest neighbor algorithm using the location features and a decision level fusion approach using the visual and textual tag features. The proposed method enables estimation of reliable decision from above two classifiers.
• Different Language Association Using Mutual Information Between Tags and Visual Features

  Shota Hamano, Takahiro Ogawa, Mild Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  This paper presents a method for associating tags in one language with the tags representing the same meaning in another language. Since recent image search and sharing services highly rely on annotations like tags with images for obtaining the desired images, the proposed method utilizes the visual features extracted from images with tags. In the proposed method, mutual information between tags and visual features are calculated. Tag similarity is then calculated based on the mutual information. Mutual information takes into consideration the relevance between tags and visual features. Therefore, the similarity based on the mutual information represents tag-to-tag relationships more effectively than direct use of the visual features. Experimental results show the effectiveness of the proposed method in associating English tags with Japanese tags representing the same meanings.
• Novel Video Feature-based Favorite Video Estimation Using Users' Viewing Behavior and Evaluation

  Yoshiki Ito, Takahiro Ogawa, Miki Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  This paper presents novel video feature-based favorite video estimation method. In the proposed method, we use three features, videos, users' viewing behavior and users' evaluation scores for these videos. In order to calculate the novel video features, Multiset Canonical Correlations Analysis (MCCA) is applied to these features to integrate the different types of features. Specifically, MCCA maximizes the sum of three kinds of correlations between three pairs of these features. Then the novel video features that represent the users' individual preference can be obtained by using the projection maximizing the three correlations. Finally, Supported Vector Ordinal Regression (SVOR) is trained by using the novel video features to estimate favorite videos. Experimental results show the effectiveness of our method.
• Image Retrieval for Identification of Insects Based on Saliency Map and Distance Metric Learning

  Susumu Genma, Takahiro Ogawa, Miki Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  This paper presents an image retrieval method for insect identification based on saliency map and distance metric learning. First, the proposed method extracts regions of insects from target images by using saliency map and calculates visual features from the extracted insect regions. Next, in order to realize accurate retrieval of insects based on the calculated features, distance metric learning is newly adopted. Consequently, through users' evaluation in the retrieval, optimal distance can be obtained for the calculated visual features to obtain successful retrieval results, and the identification of insects becomes feasible. Experimental results show the effectiveness of our method.
• Estimating Human Emotion Evoked by Visual Stimuli Using fMRI Data

  Kento Sugata, Takahiro Ogawa, Miki Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  This paper presents a method that estimates human emotion evoked by visual stimuli using functional magnetic resonance imaging (fMRI) data. First, in our method, preprocessing and masking procedures are applied to the fMRI data. These procedures provide the multiple brain data corresponding to Brodmann areas (BA). In most cases, the dimensionality of fMRI data and the BA data is larger than the number of observations, and this results in overfilling. Thus, in order to reduce the dimensionality, we apply general tensor discriminant analysis (GTDA), which can take into account the information related to the users' emotion. Then multiple estimation results of the users' emotion are obtained from support vector machine by separately using the multiple BA data obtained after the dimensionality reduction via GTDA. Furthermore, our method obtains the final estimation result from effective supervised decision-level fusion of the above estimation results.
• Decision Level Fusion-based Team Tactics Estimation in Soccer Videos

  Genki Suzuki, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  2016 IEEE 5TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS  1  -2  2016  [Not refereed][Not invited]
   

  A decision-level fusion (DLF)-based team tactics estimation method in soccer videos is newly presented. In our method, tactics estimation based on audio-visual and formation features is newly adopted since the tactics of the soccer game are closely related to the audio-visual sequences and player positions. Therefore, by using these features, we classify the tactics via Support. Vector Machine (SVM). Furthermore, by applying DIA' to the SVM-based classification results, the two modalities are integrated to obtain more accurate tactics estimation results. Some results of experiments verify the superiority of our method.
• An Accurate Mortality Prediction Method Based on Decision-level Fusion of Existing ICU Scoring Systems

  Yasutaka Hatakeyama, Takahiro Ogawa, Hirokazu Tanaka, Miki Haseyama  PROCEEDINGS OF 2016 INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY AND ITS APPLICATIONS (ISITA 2016)  126  -130  2016  [Not refereed][Not invited]
   

  In this paper, we propose a mortality prediction method based on decision-level fusion (DLF) of existing intensive unit care (ICU) scoring systems. First, the proposed method obtains severity scores from the existing ICU scoring systems. Furthermore, we construct classifiers that categorize patients into survivors or non-survivors. Next, patient feature vectors are extracted based on the mortality rates that are estimated from the obtained severity scores by using a non-linear least squares method to obtain other types of classification results. In order to obtain the final severity score for each patient, we integrate the obtained multiple classification results based on DLF that can estimate the final severity scores. Finally, we performed the proposed method to actual ICU patient data and verified the effectiveness of the proposed method. Thus, the proposed method can realize accurate mortality prediction without any additional work by using the existing ICU scoring systems.
• DLF-based speech segment detection and its application to audio noise removal for video conferences

  Kazuto Sasaki, Takahiro Ogawa, Sho Takahashi, Miki Haseyama  ITE Transactions on Media Technology and Applications  4-  (1)  68  -77  2016  [Not refereed][Not invited]
   

  A new decision-level fusion (DLF)-based speech segment detection method and its application to audio noise removal for video conferences are presented in this paper. The proposed method calculates visual and audio features from video sequences and audio signals, respectively, obtained in video conferences. Features extracted from mouth regions of participants and attribution degrees of speech class are used as visual and audio features, respectively, and Support Vector Machine (SVM)-based classification is performed by using each kind of feature. The SVM classifier performs two-class classification of speech and non-speech segments to realize speech segment detection. From the detection results obtained from the visual and audio features, DLF based on Supervised Learning from Multiple Experts is performed to successfully obtain the final detection results with focus on the accuracy of each detection result. Then, from audio signals in the non-speech segments detected by our method, we can extract noise information to realize accurate audio noise removal in the speech segments.
• Accurate and efficient extraction of hierarchical structure ofweb communities for web video retrieval

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  4-  (1)  49  -59  2016  [Not refereed][Not invited]
   

  This paper presents an accurate and efficient method for extracting hierarchical structure of Web communities, i.e., Web video sets with similar topics for Web video retrieval. First, efficient canonical correlation analysis (CCA), named sub-sampled CCA, is derived to obtain link relationships that represent similarities between latent features of Web videos. Moreover, the obtained link relationships enable application of an algorithm based on recursive modularity optimization to extract hierarchical structure of Web communities. Different from previously reported methods, our method can extract the hierarchical structure for the whole target dataset since the algorithm enables recursive reduction of its processing targets. This means it becomes unnecessary to perform screening of Web videos, and we can avoid performance degradation caused by discarding relevant Web videos in the screening, which occurred in previously reported methods. Consequently, our method enables extraction of the hierarchical structure with high accuracy as well as low computational cost.
• Classification of gastric cancer risk from X-ray images based on efficient image features related to serum Hp antibody level and serum PG levels

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  4-  (4)  337  -348  2016  [Not refereed][Not invited]
   

  In this paper, a fully automatic gastric cancer risk classification method with the aim of constructing a computer-aided diagnosis (CAD) system is presented. Two-stage classification is used in the proposed method for determining gastric cancer risk. In the first stage, the proposed method detects H. pylori-infected patients, i.e., detection of patients who have gastric cancer risk, and the proposed method classifies the level of gastric cancer risk, i.e., high or low, from H. pylori-infected patients in the second stage. In each stage, we derive new image features that are closely related to values of blood examination via kernel canonical correlation analysis. The introduction of these new image features provides classification improvement in each stage, and it is the main contribution of this paper. Consequently, accurate classification becomes feasible by the proposed method. Experimental results obtained by applying the proposed method to real X-ray images show that our method outperforms several comparative methods.
• Biological surface friction - Biomimetic approach based on self-organization

  Masatsugu Shimomura, Yuji Hirai, Naoto Okuda, Ryuichiro Machida, Syûhei Nomura, Masahiro Ôhara, Miki Haseyama  Toraibarojisuto/Journal of Japanese Society of Tribologists  61-  215  -221  2016/01/01  [Not refereed][Not invited]
  
• Web映像検索のためのグラフマイニングを用いたリランキング手法

  吉田壮, 小川貴弘, 長谷山美紀, 棟安実治  信号処理シンポジウム講演論文集(CD-ROM)  31st-  ROMBUNNO.A6‐5  2016  [Not refereed][Not invited]
  
• 深層学習を用いた胃X線画像におけるHelicobacter pylori感染識別手法の高精度化に関する検討

  石原賢太, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  31st-  ROMBUNNO.B4‐1  2016  [Not refereed][Not invited]
  
• 胃X線画像を用いたHelicobacter Pylori感染の高精度識別に関する検討(4)

  藤後廉, 石原賢太, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  31st-  ROMBUNNO.B4‐2  2016  [Not refereed][Not invited]
  
• A Study on High Performance Parallel Computing for Active Net-based Pass Region Estimation in Soccer Videos

  高橋翔, 長谷山美紀  映像情報メディア学会技術報告  39-  (49(ME2015 118-124))  1‐6  -6  2015/12/03  [Not refereed][Not invited]
   

  This paper proposes a high performance parallel computing for Active net-based pass region estimation in soccer videos. In the soccer games, since the pass is very useful for analyses of various soccer tactics, the visualization of the pass regions is a very important task. Therefore, the pass region estimation method based on Active net is proposed. However, computation time of the pass region estimation must be reduced. Thus, in this paper, we propose a high performance parallel computing for Active net-based pass region estimation. By utilizing CUDA^ as the computation environment, the high performance parallel computing is realized.
• A Note on Estimation of Group Advantage for Group Tactics Analysis in Soccer Videos

  高橋翔, 長谷山美紀  映像情報メディア学会技術報告  39-  (49(ME2015 118-124))  7‐12  -12  2015/12/03  [Not refereed][Not invited]
   

  This paper reports a effectiveness of the estimation method of group advantage in soccer videos. The group advantage represents the degree of team's superiority or inferiority in each group. For this analysis, the proposed method classifies players into some groups. As a result of this classification, the players in one group are closely related in terms of soccer tactics. In this method, the relationship between the players in the groups are analyzed by using player positions. Then, this method estimates group advantages by utilizing the relationship of the players. In the last of this paper, by utilizing the actual soccer videos, we evaluate the estimation method of the group advantage.
• 画像共有サービスにおける投稿画像の分類結果の信頼度推定に関する検討

  斉藤直輝, 小川貴弘, 浅水仁, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.128  2015/11/07  [Not refereed][Not invited]
  
• 道路構造物の変状画像に対する類似検索の高精度化に関する検討

  高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.134  2015/11/07  [Not refereed][Not invited]
  
• Web映像検索のためWebコミュニティの提示法に関する検討―Webコミュニティに含まれるトピックを考慮した代表キーフレーズ抽出の試み―

  竹原大智, 原川良介, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.122  2015/11/07  [Not refereed][Not invited]
  
• 個々の道路構造物に関する点検項目の導入による道路構造物の変状推定の高精度化に関する検討

  前田圭介, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.133  2015/11/07  [Not refereed][Not invited]
  
• 胃X線画像を用いたHelicobacter pylori感染の高精度識別に関する検討(2)―撮像方向が識別に有効な画像領域に与える影響に対する考察―

  藤後廉, 石原賢太, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.123  2015/11/07  [Not refereed][Not invited]
  
• 楽曲中の歌唱区間推定における簡易脳波計の導入に関する検討

  菅田健斗, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.121  2015/11/07  [Not refereed][Not invited]
  
• サッカー映像における試合内容の理解を促すデータの提示に関する検討

  高橋翔, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.135  2015/11/07  [Not refereed][Not invited]
  
• 顔器官の動きに基づく関心の有無の推定に関する検討

  佐坂勇磨, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2015-  ROMBUNNO.120  2015/11/07  [Not refereed][Not invited]
  
• Toward the Future of Search Technology(Message from the 2014 ITE Awards Winners and New Fellow Members,Special Column)

  Haseyama Miki  The journal of the Institute of Image Information and Television Engineers  69-  (7)  642  -642  2015/09/01
• 情報科学が繋ぐ生物とものづくり―バイオミメティクス・画像検索システムの理論と実現―

  長谷山美紀  高分子学会予稿集(CD-ROM)  64-  (1)  ROMBUNNO.3B15IL  2015/05/12  [Not refereed][Not invited]
  
• A Biological Classification Method Using Scanning Electron Mircroscope Images via Taxonomy-Based Decision Tree

  Piao Jun, Ogawa Takahiro, Haseyama Miki  電子情報通信学会論文誌 D(Web)  98-  (5)  823  -834  2015/05/01  [Not refereed][Not invited]
   

  本論文では，走査型電子顕微鏡(Scanning Electron Microscop，SEM)で撮像された画像を用いた生物の分類法を提案する．提案手法では，生物学者が構築した分類体系に注目し，各ノードにその下位のノードへの分類を行う分類器を割り当てることで，決定木を構築する．これにより，構築された決定木を用いて生物の分類が可能となる．このように，生物の分類に有用な分類体系の構造を導入することで，画像特徴のみに注目する生物の分類法と比較して，高精度な分類が期待できる．また，提案手法では，更なる精度向上のため，以下の二つの処理も導入する．まず，分類体系において，画像特徴が類似する異なるノードに注目した決定木の変更を行い，誤分類を抑制する．次に，同種の生物の異なる撮像倍率の画像に対し，モーフィングを施すことで学習データの充足を行う．以上によって，提案手法では，SEMで撮像された生物の高精度な分類が可能となる．本論文の最後では，提案手法の有効性を確認するための実験結果を示す．
• A Note on Exhibition of Hierarchical Structure of Web Communities for Web Video Retrieval : An Approach for Representative Keyword Extraction of Web Communities

  原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  89  -94  2015/02/16  [Not refereed][Not invited]
  
• A note on Web video retrieval based on network structure analysis using video features and tag information : Retrieval performance improvement using local learning regularization

  吉田 壮, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  77  -82  2015/02/16  [Not refereed][Not invited]
  
• A Note on Image Retrieval Based on Evaluation of Retrieval Results : An Application to Similarity Evaluation of Surface Structure of Organisms Taken by Scanning Electron Microscope

  高橋 昌弘, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  83  -88  2015/02/16  [Not refereed][Not invited]
  
• Super-Resolution for Microscope Images Focusing on Imaging Procedure

  廣川 真梨子, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  167  -172  2015/02/16  [Not refereed][Not invited]
  
• Improving Performance Classification of Biological Organisms using Scanning Electron Microscope Images : Introducing Training Data selection By Focusing on the Image Acquisition Method

  朴 君, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  173  -178  2015/02/16  [Not refereed][Not invited]
  
• A Note on Classification of Individual Favorite Musical Pieces Utilizing EEG Signals during Listening to Music : Performance Improvement via CCA Considering Class Information

  澤田 充奨, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  179  -184  2015/02/16  [Not refereed][Not invited]
  
• A Note on Image Classification Using EEG Features during Watching Images(4)Verification of Using Multiple Users' EEG Signals

  川上 拓也, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  185  -190  2015/02/16  [Not refereed][Not invited]
  
• A Note on Helicobacter Pylori Infection Detection from Gastric X-ray Images : Performance Comparison between Feature Fusion and Decision Level Fusion

  石原 賢太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  191  -196  2015/02/16  [Not refereed][Not invited]
  
• A note on Accurate Estimation of a Light Source and a Spectral Reflectance by Introduction of Superpixel Calculation Method

  五十嵐 祐太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  39-  (7)  247  -252  2015/02/16  [Not refereed][Not invited]
  
• A Note on Image Classification Using EEG Features during Watching Images (4)-Verification of Using Multiple Users' EEG Signals-

  川上拓也, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  185  -190  2015/02/16  [Not refereed][Not invited]
  
• A Note on Classification of Individual Favorite Musical Pieces Utilizing EEG Signals during Listening to Music-Performance Improvement via CCA Considering Class Information-

  澤田充奨, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  179  -184  2015/02/16  [Not refereed][Not invited]
  
• A note on Web video retrieval based on network structure analysis using video features and tag information-Retrieval performance improvement using local learning regularization-

  吉田壮, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  77  -82  2015/02/16  [Not refereed][Not invited]
  
• A Note on Data Analysis for Maintenance Inspection of Infrastructures-Quantification of Relationship between Inspection Data Using Distress Images and Inspection Results-

  三改木裕矢, 高橋翔, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  257  -262  2015/02/16  [Not refereed][Not invited]
  
• A note on Accurate Estimation of a Light Source and a Spectral Reflectance by Introduction of Superpixel Calculation Method

  五十嵐祐太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  247  -252  2015/02/16  [Not refereed][Not invited]
  
• A Note on Accurate Image Classification Method of Image Sharing Sites

  斉藤直輝, 小川貴弘, 浅水仁, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  71  -75  2015/02/16  [Not refereed][Not invited]
  
• A Note on Exhibition of Hierarchical Structure of Web Communities for Web Video Retrieval-An Approach for Representative Keyword Extraction of Web Communities-

  原川良介, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  89  -94  2015/02/16  [Not refereed][Not invited]
  
• Improving Performance Classification of Biological Organisms using Scanning Electron Microscope Images-Introducing Training Data selection By Focusing on the Image Acquisition Method-

  PIAO Jun, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  173  -178  2015/02/16  [Not refereed][Not invited]
  
• Super-Resolution for Microscope Images Focusing on Imaging Procedure

  廣川真梨子, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  167  -172  2015/02/16  [Not refereed][Not invited]
  
• A Note on Helicobacter Pylori Infection Detection from Gastric X-ray Images-Performance Comparison between Feature Fusion and Decision Level Fusion-

  石原賢太, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  191  -196  2015/02/16  [Not refereed][Not invited]
  
• A Preference Estimation Method based on Relationship between Videos and Viewing Behavior: Reduction of Performance Degradation for Few Training Pairs

  山口由晃, 小川貴弘, 浅水仁, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  253  -256  2015/02/16  [Not refereed][Not invited]
  
• A Note on Image Retrieval Based on Evaluation of Retrieval Results-An Application to Similarity Evaluation of Surface Structure of Organisms Taken by Scanning Electron Microscope-

  高橋昌弘, 小川貴弘, 長谷山美紀  電子情報通信学会技術研究報告  114-  (460(IE2014 67-86))  83  -88  2015/02/16  [Not refereed][Not invited]
  
• Video Genre Estimation from Relationship between Motion and Facial Features using SLPCCA

  Yuma Sasaka, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  250  -251  2015  [Not refereed][Not invited]
   

  In this paper, we propose an efficient video genre estimation method based on the relationship between facial features and motion features. In the proposed method, we utilize supervised locality preserving canonical correlation analysis (SLPCCA), which is derived in the proposed method, to maximize the correlation between facial features and motion features. Moreover, by using SLPCCA, we can consider not only the correlation but also class information. Finally, by applying Support Vector Machine (SVM) to the SLPCCA-based feature vectors, we realize a successful video genre estimation. Experimental results show the effectiveness of our method.
• Web Image Visualization Considering Image Content Based on Visual and Tag Features

  Kouhei Tateno, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  254  -255  2015  [Not refereed][Not invited]
   

  This paper presents a Web image visualization considering image content based on visual and tag features. In this paper, we focus on tagged images on social media websites. Since these tags represent the image content according to the subjectivity of the user, using these tags is efficient for the image visualization. Thus, by using visual and tag features, the proposed method can take account of the semantic contents. Specifically, the proposed method applies Locality Preserving Canonical Correlation Analysis (LPCCA) to these two features to obtain the dimensionality reduction results, i.e., the visualization result.
• Image Classification Based on Integration of EEG and Visual Features Via LFDA-MSLPCCA

  Kento Sugata, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  513  -514  2015  [Not refereed][Not invited]
   

  This paper presents a novel image classification based on the integration of EEG and visual features. In the proposed method, we first obtain classification results by separately using EEG and visual features. Then we merge the above classification results based on kernelized version of Supervised Learning from Multiple Experts (KSLME) via Multiset Supervised Locality Preserving Canonical Correlation Analysis (MSLPCCA) to obtain final classification results. It should be noted that when the number of samples is fewer than the dimension of a sample data used in MSLPCCA, we have to reduce the dimension. Therefore, we propose MSLPCCA based on Local Fisher Discriminant Analysis (LFDA) which can take class information into account. Then the integration of all of the classifications results becomes feasible by MSLPCCA based on LFDA.
• Helicobacter Pylori Infection Detection From Gastric X-ray Images Using KLFDA-based Decision Fusion

  Kenta Ishihara, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  204  -205  2015  [Not refereed][Not invited]
   

  This paper presents the performance improvement of Helicobacter pylori (H.pylori) infection detection using Kernel Local Fisher Discriminant Analysis (KLFDA)-based decision fusion. As the biggest contribution of this paper, the proposed method extracts more discriminative features based on KLFDA for the decision fusion. Since the decision fusion employed in this paper can consider not only the detection results but also the visual features, by calculating more discriminative features via KLFDA, more accurate decision fusion becomes feasible. Furthermore, experimental results show the effectiveness of the proposed method.
• A new approach with IFS for image restoration

  Miki Haseyama, Megumi Takezawa, Junichi Miura, Hideo Kitajima  European Signal Processing Conference  2015-March-  2015/01/01  [Not refereed][Not invited]
   

  © 2000 EUSIPCO.This paper proposes a new image-restoration method based on Iterated Function System (IFS). The proposed method can restore images contaminated by impulsive noise according to self-similarity represented by the IFS parameters. Since the IFS is usually used for image coding, it has never been applied to image restoration; and it cannot be utilized for the image restoration as it is. In order to adapt the IFS for image restoration, this paper reforms the conventional criterion for the computation of the IFS parameters to suit for image restoration, and as preprocessing we apply an e-filter in which a median filter is embedded to contaminated images prior to computing the IFS parameters. Some simulation results are presented to demonstrate the effectiveness of this method.
• Automatic martian dust storm detection from multiple wavelength data based on decision level fusion

  Keisuke Maeda, Takahiro Ogawa, Miki Haseyama  IPSJ Transactions on Computer Vision and Applications  7-  79  -83  2015  [Not refereed][Not invited]
   

  This paper presents automatic Martian dust storm detection from multiple wavelength data based on decision level fusion. In our proposed method, visual features are first extracted from multiple wavelength data, and optimal features are selected for Martian dust storm detection based on the minimal-Redundancy-Maximal-Relevance algorithm. Second, the selected visual features are used to train the Support Vector Machine classifiers that are constructed on each data. Furthermore, as a main contribution of this paper, the proposed method integrates the multiple detection results obtained from heterogeneous data based on decision level fusion, while considering each classifier's detection performance to obtain accurate final detection results. Consequently, the proposed method realizes successful Martian dust storm detection.
• Estimation of video authenticity through collaborative use of available video signals

  Michael Penkov, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  3-  (3)  214  -225  2015  [Not refereed][Not invited]
   

  This paper proposes a method for identifying the video that retains the most information from its parent video. Since the parent video is often unavailable, the proposed method estimates its content through the collaborative use of the available video signals that are edited copies of the parent video. By reducing the difference between the video signals of the edited videos, the proposed method then enables the use of conventional no-reference video quality assessment algorithms. Since editing a video requires recompressing it, and since quality assessment algorithms can detect signs of recompression, the proposed method can identify the edited video that retains the most information from the parent video. The effectiveness of the proposed method is verified by subjective experiments over artifical and real-world data sets that include a total of over 400 videos.
• LDA-based Music Recommendation with CF-based Similar User Selection

  Shohei Kinoshita, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  215  -216  2015  [Not refereed][Not invited]
   

  This paper presents a Latent Dirichlet Allocation (LDA)-based music recommendation method with collaborative filtering (CF)-based similar user selection. By applying LDA to music, we can estimate latent topics of music. However, we have to effectively reduce the size of the target dataset applied to LDA in order to recommend music from a large dataset. Hence, we use CF techniques, which recommend items using evaluation information of users who have similar tastes to a target user. Therefore, the proposed method limits the size of the dataset by using information of similar users and enables the recommendation of music considering latent topics of music. By using the idea of CF, our method can use LDA for music recommendation. Experimental results show the effectiveness of our method.
• Missing Texture Reconstruction via Power Spectrum-based Sparse Representation

  Yuma Tanaka, Takahiro Ogawa, Miki Haseyama  2015 IEEE 4TH GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  221  -222  2015  [Not refereed][Not invited]
   

  This paper presents a method for missing texture reconstruction via power spectrum-based sparse representation. We reconstruct missing areas based on minimizing the mean square error between power spectra (P-MSE). In our method, missing areas are reconstructed by embedding some known patches. Mathematically, we obtain the optimal linear combination of measurement patches by P-MSE minimization. The optimization can be solved as a combinatorial problem based on sparse representaion. In this way, the optimal approximation which minimizes the P-MSE is obtained and we embed it in the missing area. Experimental results show effectiveness of our method for reconstructing texture images.
• Extended Joint Bilateral Filter for the Reduction of Color Bleeding in Compressed Image and Video

  Naofumi Wada, Masato Kazui, Miki Haseyama  ITE Transactions on Media Technology and Applications  3-  (1)  95  -106  2015  [Not refereed][Not invited]
   

  In this paper, we propose a simple and effective technique, named extended joint bilateral filter, for reducing color bleeding which is defined as a smearing of the color between areas of strongly contrasting chroma in lossy-compressed images. The color bleeding comes from not only quantization but also sub-sampling in the YCbCr 4:2:0 color format. In order to deal with the two problems simultaneously, we extend the joint bilateral filter formulation by referring all of Y, Cb and Cr components based on a correlation between luma and chroma. As a result, our method achieves both noise reduction and sharpness enhancement for chroma without color edge blurring. Experimental results show that our method is effective in terms of both objective quality and subjective quality.
• 胃X線画像を用いた胃がん発症リスクの推定―リスク推定に用いる分類手法の比較による初期検討―

  石原賢太, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  30th-  ROMBUNNO.C4‐5  2015  [Not refereed][Not invited]
  
• グラフ構造に基づくリランキングを利用したWeb映像検索に関する検討~初期検索結果の誤りを考慮した最適化に基づく高精度化~

  吉田壮, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  30th-  ROMBUNNO.A2‐3  2015  [Not refereed][Not invited]
  
• 胃X線画像を用いたHelicobacter pylori感染の高精度識別に関する検討(1)―識別精度向上に有効な画像領域の抽出手法―

  藤後廉, 石原賢太, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  30th-  ROMBUNNO.C4‐4  2015  [Not refereed][Not invited]
  
• 画像注視時の脳波特徴量を用いた画像分類に関する検討(5)―KESLMEを用いた教師あり分類結果統合に基づく高精度化―

  菅田健斗, 小川貴弘, 長谷山美紀  信号処理シンポジウム講演論文集(CD-ROM)  30th-  ROMBUNNO.A5‐4  2015  [Not refereed][Not invited]
  
• A Note on Network Analysis Based Detection of Important Player and Similar Scenes in Soccer Videos

  TAKAHASHI Sho, HASEYAMA Miki  ITE Technical Report  38-  (51)  1  -4  2014/12/05  [Not refereed][Not invited]
   

  This paper proposes a link analysis-based method for detecting important players and similar scenes in soccer videos. We define important players as follows: 1) the attacking player who have great relevancy to a score, 2) the defending player on the opposing team, and 3) players who assist the above players. Since soccer tactic analysis focuses not only on player skill but also relationships between players, this paper expresses the relationships between players as a network, which is constructed from player positions in the soccer video. The proposed method analyses the constructed network to detect important players and similar scenes.
• 視聴動作に基づくユーザの関心対象を考慮した映像に対する関心の有無の識別に関する検討

  YAMAGUCHI YOSHIAKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.165  2014/10/25  [Not refereed][Not invited]
  
• トピックモデルを用いた複数特徴の統合による画像群の可視化に関する検討

  TATENO KOHEI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.136  2014/10/25  [Not refereed][Not invited]
  
• 画像共有サイトへの投稿画像およびそのメタデータに注目したマルチモーダル画像検索手法に関する検討

  SAITO NAOKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.141  2014/10/25  [Not refereed][Not invited]
  
• Web映像検索のための意味内容を考慮したWebコミュニティの階層構造抽出に関する一検討

  TAKEHARA HIROTOMO, HARAKAWA RYOSUKE, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.164  2014/10/25  [Not refereed][Not invited]
  
• 複数の画像特徴を用いたベイジアンネットワークに基づく構造物の変状の推定の高精度化に関する検討

  MAEDA KEISUKE, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.140  2014/10/25  [Not refereed][Not invited]
  
• パワースペクトログラム誤差最小化に基づく消失音響区間の再構成手法に関する検討

  田中佑磨, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.135  2014/10/25  [Not refereed][Not invited]
  
• 維持管理業務の効率化に向けた点検データの類似度算出における特徴選択に関する検討

  三改木裕矢, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.139  2014/10/25  [Not refereed][Not invited]
  
• トピックモデルに基づく音響特徴と時間情報を考慮した流行曲識別に関する検討

  木下奨平, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2014-  ROMBUNNO.134  2014/10/25  [Not refereed][Not invited]
  
• A Note on Improving Video Authenticity Degree Estimation through Automatic Speech Recognition (画像工学)

  PENKOV Michael, OGAWA Takahiro, HASEYAMA Miki  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (434)  341  -346  2014/02/17
• A Note on Image Classification Using EEG Features during Watching Images(2)Verification of Using EEG Features Extracted from Single-Trial EEG for Classification

  川上 拓也, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  199  -204  2014/02/17  [Not refereed][Not invited]
  
• A Note on Selection of Tags Representing the Same Meaning Using Tagged Images

  宋 妍, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  211  -216  2014/02/17  [Not refereed][Not invited]
  
• A Note on Spectral Reflectance Estimation Using a Digital Camera Including Visible and Near-infrared Spectral Sensitivity : Introduction of Spectrum Estimation of a Light Source

  五十嵐 祐太, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  223  -228  2014/02/17  [Not refereed][Not invited]
  
• A Note on Insect Classification by using Scanning Electron Microphotographs : Performance Improvement Based on Hierarchical Clustering by using visual feature

  高橋 昌弘, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  241  -246  2014/02/17  [Not refereed][Not invited]
  
• A Note on Biological Family Estimation Using Scanning Electron Microscope Images of Organism : Performance Improvement Based on Modification of Decision Tree Constructed from Taxonomy

  朴 君, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  247  -252  2014/02/17  [Not refereed][Not invited]
  
• A note on hierarchical structure extraction of Web video communities and its application to large-scale datasets

  原川 良介, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  275  -280  2014/02/17  [Not refereed][Not invited]
  
• A Note on Improving Video Scene Segmentation Performance Using Structure Analysis Based on Hidden Conditional Random Fields

  吉田 壮, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  285  -290  2014/02/17  [Not refereed][Not invited]
  
• A Note on Probabilistic Topic Modeling with User Preferences and Its Application to Music Recommendation

  中西 亮太, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  313  -318  2014/02/17  [Not refereed][Not invited]
  
• A note on Performance Improvement of Error Concealment Using Kalman Filter : Realization of State Transition Model Considering Object Motion

  大貫 修平, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  325  -330  2014/02/17  [Not refereed][Not invited]
  
• A note on performance improvement of example-based video super-resolution based on rank minimization

  廣川 真梨子, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (433)  347  -350  2014/02/17  [Not refereed][Not invited]
  
• A Note on Estimating Deformation Based on Bayesian Networks Using Images from Bridge Inspection

  KOBAYASHI KATSUKI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  113-  (434(IE2013 95-132))  217  -221  2014/02/10  [Not refereed][Not invited]
  
• Definition of Soccer Player's Features Using Network Analysis and Its Application to Similarity Calculation between Players

  IWAI KAZUYA, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  113-  (434(IE2013 95-132))  319  -324  2014/02/10  [Not refereed][Not invited]
  
• 色品質改善のための拡張ジョイントバイラテラルフィルタ—Extended Joint Bilateral Filtering for Color Quality Enhancement—画像工学

  和田 直史, 数井 誠人, 長谷山 美紀  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (434)  1  -6  2014/02
• Extended Joint Bilateral Filtering for Color Quality Enhancement

  和田 直史, 数井 誠人, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  1  -6  2014/02
• A Note on Probabilistic Topic Modeling with User Preferences and Its Application to Music Recommendation

  中西 亮太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  313  -318  2014/02
• A note on Performance Improvement of Error Concealment Using Kalman Filter : Realization of State Transition Model Considering Object Motion

  大貫 修平, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  325  -330  2014/02
• A note on performance improvement of example-based video super-resolution based on rank minimization

  廣川 真梨子, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  347  -350  2014/02
• A Note on Improving Video Authenticity Degree Estimation through Automatic Speech Recognition (メディア工学 映像表現&コンピュータグラフィックス ヒューマンインフォメーション)

  PENKOV Michael, OGAWA Takahiro, HASEYAMA Miki  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  341  -346  2014/02
• A Note on Spectral Reflectance Estimation Using a Digital Camera Including Visible and Near-infrared Spectral Sensitivity : Introduction of Spectrum Estimation of a Light Source

  五十嵐 祐太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  223  -228  2014/02
• A Note on Insect Classification by using Scanning Electron Microphotographs : Performance Improvement Based on Hierarchical Clustering by using visual feature

  高橋 昌弘, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  241  -246  2014/02
• A Note on Image Classification Using EEG Features during Watching Images(2)Verification of Using EEG Features Extracted from Single-Trial EEG for Classification

  川上 拓也, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  199  -204  2014/02
• A Note on Selection of Tags Representing the Same Meaning Using Tagged Images

  宋 妍, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  211  -216  2014/02
• A Note on Biological Family Estimation Using Scanning Electron Microscope Images of Organism : Performance Improvement Based on Modification of Decision Tree Constructed from Taxonomy

  朴 君, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  247  -252  2014/02
• A Note on Improving Video Scene Segmentation Performance Using Structure Analysis Based on Hidden Conditional Random Fields

  吉田 壮, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  285  -290  2014/02
• A note on hierarchical structure extraction of Web video communities and its application to large-scale datasets

  原川 良介, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  38-  (7)  275  -280  2014/02
• Extended Joint Bilateral Filtering for Color Quality Enhancement

  和田直史, 数井誠人, 長谷山美紀  電子情報通信学会技術研究報告  113-  (434(IE2013 95-132))  2014
• バイオミメティクス・データ検索による生物形態の類縁性の可視化とその産業応用

  長谷山美紀  日本動物学会大会予稿集  85th-  2014
• バイオミメティクス・データ検索基盤と新材料開発

  長谷山美紀  高分子学会予稿集(CD-ROM)  63-  (2)  2014
• Exhibition Method of Hierarchical Structure of Web Communities Using Community Density for Web Video Retrieval

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  18TH IEEE INTERNATIONAL SYMPOSIUM ON CONSUMER ELECTRONICS (ISCE 2014)  2014  [Not refereed][Not invited]
   

  In this paper, we present an exhibition method of hierarchical structure of Web communities using community density for Web video retrieval. In the proposed method, Web communities that are Web video sets with similar topics is hierarchically extracted by using hyperlinks between Web videos and their video features, and density of each Web community is estimated. Then we exhibit the hierarchical structure of Web communities according to the obtained density to users. Since users can grasp detail degree of topics contained in each Web community, the desired Web community can be easily selected for Web video retrieval.
• Entomomimetics and Pest Management

  Takahiko Hariyama, Yoshihiro Uozu, Hiromi Mukai, Yumi Yamahama, Mantaro Hironaka, Yasuharu Takaku, Daisuke Ishii, Masahiro Ohara, Shuhei Nomura, Miki Haseyama, Shigeo Hara, Tateo Shimozawa, Masatsugu Shimomura  JAPANESE JOURNAL OF APPLIED ENTOMOLOGY AND ZOOLOGY  58-  (2)  79  -91  2014  [Not refereed][Not invited]
  
• An efficient extraction method of hierarchical structure of web communities for web video retrieval

  Ryosuke Harakawa, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  2-  (3)  287  -297  2014  [Not refereed][Not invited]
   

  In this paper, we propose a method that enables efficient extraction of hierarchical structure of Web communities containing Web videos that have similar topics in order to retrieve users' desired Web videos. Specifically, the proposed method first calculates Web video features by applying canonical correlation analysis to a small number of Web video samples obtained on the basis of a clustering scheme. Furthermore, we construct a "community graph" of which each node consists of multiple Web videos and each edge corresponds to hyperlinks of Web pages including these videos. Then, based on strongly connected components, edge betweenness and modularity of the community graph, hierarchical structure of Web communities is estimated. In this way, our method can efficiently extract the hierarchical structure of Web communities, and users' desired Web videos can be retrieved by selecting Web communities according to their hierarchical structure.
• Popular Music Estimation Based on Topic Model Using Time Information and Audio Features

  Shohei Kinoshita, Takahiro Ogawa, Miki Haseyama  2014 IEEE 3RD GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  102  -103  2014  [Not refereed][Not invited]
   

  This paper presents popular music estimation based on a topic model using time information and audio features. The proposed method calculates latent topic distribution using Latent Dirichlet Allocation to obtain more accurate music features. In this approach, we also use release date information of each music as time information for concerning the relationship between music trends and each age. Then, by using the obtained latent topic distribution features, the estimation of the popular music becomes feasible based on a Support Vector Machine classifier. Experimental results show the effectiveness of our method.
• Missing Audio Segment Reconstruction Based on Sparse Representation with Power Spectrogram

  Yuma Tanaka, Takahiro Ogawa, Miki Haseyama  2014 IEEE 3RD GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  86  -87  2014  [Not refereed][Not invited]
   

  This paper presents a method for reconstructing missing audio segments based on sparse representation with power spectrogram. In the proposed method, an error of power spectrograms is utilized as a quality measure representing reconstruction performance. Then the proposed method estimates missing segments based on sparse representation optimized with respect to the error of power spectrograms. This error minimization problem can be solved with a greedy algorithm by limitting the solution to only sparse one. By using our method, perceptually optimized reconstruction becomes feasible since missing segments are estimated by using the quality measure which represens auditory properties. Experimental results obtained by applying the proposed method to actual music signals from RWC Music Database show its effectiveness.
• Effective Multiple Feature Fusion Using Topic Model for Social Image Visualization

  Kouhei Tateno, Takahiro Ogawa, Miki Haseyama  2014 IEEE 3RD GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  182  -183  2014  [Not refereed][Not invited]
   

  This paper presents a multiple feature fusion method using topic model for social image visualization. Images in social media are represented from several aspects such as their visual information and tags. The proposed method extracts low-level features from social images and their tags and calculates their integrated high-level features. Specifically, the proposed method applies multilayer multimodal probabilistic Latent Semantic Analysis (mm-pLSA) to the low-level visual and tag features to obtain the high-level features. Then, by applying dimensionality reduction techniques to the obtained features, successful visualization becomes feasible.
• Bayesian Network-based Distress Estimation Using Image Features in Road Structure Assessment

  Keisuke Maeda, Sho Takahashi, Takahiro Ogawa, Miki Haseyama  2014 IEEE 3RD GLOBAL CONFERENCE ON CONSUMER ELECTRONICS (GCCE)  169  -170  2014  [Not refereed][Not invited]
   

  This paper presents a Bayesian network-based method for estimating a distress of road structures from inspection data. The distress is represented by a damage of road structures and its degree. In the previous work, the distress was estimated by utilizing Bayesian network based on categories of road structures, details of road structures and damaged parts. However, inspection data include not only the above items but also images of the distress. Therefore, by introducing the use of the images to the previous work, improvement of the distress estimation accuracy can be expected. The proposed method calculates Bayesian network from inspection items and their corresponding images to perform the distress estimation. Experimental results show the effectiveness of the proposed method.
• Welcome to the Special Section on ITE Awards Selection

  Miki Haseyama  ITE Transactions on Media Technology and Applications  2-  (1)  51  -51  2014/01/01  [Not refereed][Not invited]
  
• A compressed sensing-based low-density parity-check real-number code

  Zaixing He, Takahiro Ogawa, Miki Haseyama, Xinyue Zhao, Shuyou Zhang  Radioengineering  22-  (3)  851  -860  2013/10/22  [Not refereed][Not invited]
   

  In this paper, we propose a novel low-density parity-check real-number code, based on compressed sensing. A real-valued message is encoded by a coding matrix (with more rows than columns) and transmitted over an erroneous channel, where sparse errors (impulsive noise) corrupt the codeword. In the decoding procedure, we apply a structured sparse (low-density) parity-check matrix, the Permuted Block Diagonal matrix, to the corrupted output, and the errors can be corrected by solving a compressed sensing problem. A compressed sensing algorithm, Cross Low-dimensional Pursuit, is used to decode the code by solving this compressed sensing problem. The proposed code has high error correction performance and decoding efficiency. The comparative experimental results demonstrate both advantages of our code. We also apply our code to cryptography.
• 脳波特徴に基づく人間の楽曲聴取時における感情推定に関する検討

  澤田充奨, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.171  2013/10/19  [Not refereed][Not invited]
  
• 判別分析に基づいた順序回帰の導入による協調フィルタリングでのSparsity Problemの軽減に関する一検討

  川島孝行, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.172  2013/10/19  [Not refereed][Not invited]
  
• 胃X線二重造影像を用いたHelicobacter pylori感染の識別の高精度化に関する検討

  石原賢太, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.165  2013/10/19  [Not refereed][Not invited]
  
• 重み付きロバストPCAを用いた動画像フレーム補間の高精度化に関する検討

  廣川真梨子, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.174  2013/10/19  [Not refereed][Not invited]
  
• 画像検索結果提示法におけるユーザの探索時間と特徴量との関係に関する一考察

  三改木裕矢, 小林克希, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.160  2013/10/19  [Not refereed][Not invited]
  
• ベイジアンネットワークを用いた構造物の点検データからの変状の推定

  小林克希, 高橋翔, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2013-  ROMBUNNO.173  2013/10/19  [Not refereed][Not invited]
  
• Error Resilient 3D DWT Video Transmission

  OMURA Kotoku, YAMASAKI Shoichiro, MATSUSHIMA Tomoko, TANAKA Hirokazu, HASEYAMA Miki  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (202)  1  -6  2013/09/12  

  The coauthors proposed the method, which adopted duplication scheme of lowest frequency sub-band coefficients of three-dimensional discrete wavelet transform (3D DWT). In this paper, we propose a scheme of duplicating the coefficients of the neighborhood of LL along with the LL, and an interpolation scheme that replaces erroneous coefficients by value of zero. The simulation evaluations over burst error channels compare the performances of the proposed method with those of the conventional one, and show the performance improvement of the proposed method.
• A Study on an Accurate Method for Estimating Pass Region from Soccer Video : Definition of a New Evaluation Function Based on Changes in Player Position Geometry over Time

  高橋 翔, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  153  -158  2013/02/18  [Not refereed][Not invited]
  
• A Note on Using Bayesian Networks for Analyzing Snow-removal Work Data

  高橋 信太郎, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  39  -43  2013/02/18  [Not refereed][Not invited]
  
• A Note on Accurate Recommendation of Videos Using User Interest Extracted from Viewing Behavior

  白石 哲夫, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  95  -100  2013/02/18  [Not refereed][Not invited]
  
• A Scene Segmentation Method Based on Video Structures in Music Programs : Verifying the Effect of Face Recognition on Scene Segmentation Accuracy

  吉田 壮, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  101  -106  2013/02/18  [Not refereed][Not invited]
  
• A Note on Trend Detection in Popular Music Database

  中西 亮太, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  107  -110  2013/02/18  [Not refereed][Not invited]
  
• A Note on Classical Music Classification Based on Acoustic Features

  松野 恵一, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  111  -115  2013/02/18  [Not refereed][Not invited]
  
• A Note on the Quantitative Evaluation of Video Retrieval Results (1) An Evaluation of a Video Retrieval System through Subjective Public Experiments in Sapporo

  山之内 豊, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  159  -163  2013/02/18  [Not refereed][Not invited]
  
• A Note on Evaluating Image Search Result Visualization Methods Based on Feature Integration Theory : Verifying Effectiveness from Comparison of Different Visualization Methods

  小林 克希, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  112-  (433)  165  -169  2013/02/18  [Not refereed][Not invited]
  
• A Note on Using Bayesian Networks for Analyzing Snow-removal Work Data

  高橋 信太郎, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  39  -43  2013/02/11  [Not refereed][Not invited]
  
• A Note on Accurate Recommendation of Videos Using User Interest Extracted from Viewing Behavior

  白石 哲夫, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  95  -100  2013/02/11  [Not refereed][Not invited]
  
• A Scene Segmentation Method Based on Video Structures in Music Programs : Verifying the Effect of Face Recognition on Scene Segmentation Accuracy

  吉田 壮, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  101  -106  2013/02/11  [Not refereed][Not invited]
  
• A Note on Trend Detection in Popular Music Database

  中西 亮太, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  107  -110  2013/02/11  [Not refereed][Not invited]
  
• A Note on Classical Music Classification Based on Acoustic Features

  松野 恵一, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  111  -115  2013/02/11  [Not refereed][Not invited]
  
• A Note on the Quantitative Evaluation of Video Retrieval Results(1)An Evaluation of a Video Retrieval System through Subjective Public Experiments in Sapporo

  山之内 豊, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  159  -163  2013/02/11  [Not refereed][Not invited]
  
• A Note on Evaluating Image Search Result Visualization Methods Based on Feature Integration Theory : Verifying Effectiveness from Comparison of Different Visualization Methods

  小林 克希, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  165  -169  2013/02/11  [Not refereed][Not invited]
  
• A Note on Accurate Estimation of Pitcher's Condition in Baseball Videos: Improvement of Features Based on Pitching Motions and Pitching Results

  KUBO JUNKI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  141  -146  2013/02/11  [Not refereed][Not invited]
  
• A Note on Classification of Helicobacter Pylori Infection using Double Contrast X-ray Image of Stomach

  YOSHIZAKI AKANE, KATSURAI MARIE, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  25  -28  2013/02/11  [Not refereed][Not invited]
  
• A Note on Soccer Player Tracking Using Elastic Model-Performance Improvement Based on New Potential Energy-

  IWAI KAZUYA, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  135  -139  2013/02/11  [Not refereed][Not invited]
  
• A Note on Accurate Soccer Video Segmentation Based on the Team Possessing the Ball

  ONUKI SHUHEI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  147  -151  2013/02/11  [Not refereed][Not invited]
  
• A study of visual features for detecting deer from road image sequences

  桂井 麻里衣, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  125  -128  2013/02
• A note on shoe sole image retrieval for footprints identification in forensic

  畠山 泰貴, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  171  -175  2013/02
• A Study on an Accurate Method for Estimating Pass Region from Soccer Video : Definition of a New Evaluation Function Based on Changes in Player Position Geometry over Time

  高橋 翔, 長谷山 美紀  映像情報メディア学会技術報告 = ITE technical report  37-  (8)  153  -158  2013/02  [Not refereed][Not invited]
  
• A note on shoe sole image retrieval for footprints identification in forensic

  畠山泰貴, 長谷山美紀  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  2013
• A study of visual features for detecting deer from road image sequencest

  桂井麻里衣, 長谷山美紀  電子情報通信学会技術研究報告  112-  (434(IE2012 110-137))  2013
• 進化する生物模倣の世界 次世代バイオミメティクス最前線 データベースで未知の機能を探す

  長谷山美紀  JST News  2013-  (Aug)  2013
• Performance Evaluation of 3D DWT Video Transmission with Invertible Codes

  大村光徳, 山嵜彰一郎, 松嶋智子, 田中宏和, 長谷山美紀  電子情報通信学会技術研究報告  113-  (153(IT2013 11-31))  2013
• 工学的気づきを生み出すバイオミメティクスデータ検索基盤

  長谷山美紀  高分子夏季大学講演予稿集  58th-  2013
• Value Creation from Big Data-Image Retrieval for Enhancing Serendipity-

  長谷山美紀  電子情報通信学会技術研究報告  113-  (218(EMCJ2013 49-59))  2013
• ビデオ伝送における誤り耐性強化の研究

  大村光徳, 山嵜彰一郎, 松嶋智子, 田中宏和, 長谷山美紀  職業大フォーラム 職業能力開発研究発表講演会講演論文集  21st-  2013
• 変状評価の支援を目的とした点検データの可視化に関する検討

  TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  日本道路会議論文集(CD-ROM)  30th-  ROMBUNNO.2053  2013  [Not refereed][Not invited]
  
• A method for improving SVM-Based image classification performance based on a target object detection scheme

  Soh Yoshida, Hiroshi Okada, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  1-  (3)  237  -243  2013  [Not refereed][Not invited]
   

  This paper presents a new method to improve performance of SVM-based classification, which contains a target object detection scheme. The proposed method tries to detect target objects from training images and improve the performance of the image classification by calculating the hyperplane from the detection results. Specifically, the proposed method calculates a Support Vector Machine (SVM) hyperplane, and detects rectangular areas surrounding the target objects based on the distances between their feature vectors and the separating hyperplane in the feature space. Then modification of feature vectors becomes feasible by removing features that exist only in background areas. Furthermore, a new hyperplane is calculated by using the modified feature vectors. Since the removed features are not part of the target object, they are not relevant to the learning process. Therefore, their removal can improve the performance of the image classification. Experimental results obtained by applying the proposed methods to several existing SVM-based classification method show its effectiveness.
• A review of broadcasting technology: From production to distribution

  Nobuyuki Yagi, Norifumi Egami, Naoki Shimidzu, Miki Haseyama  ITE Transactions on Media Technology and Applications  1-  (1)  10  -19  2013/01/01  [Not refereed][Not invited]
   

  Copyright © 2013 by ITE Transactions on Media Technology and Applications (MTA).This paper presents a survey of research trends in broadcasting technology from production to distribution and from fundamentals to applications. It briefly reviews work being done on UHDTV, 3DTV, IPTV, hybrid broadcasting, transmission technology, audio-visual coding, image sensing devices, display devices, media accessibility, content production, metadata, and content management.
• Active grid-Based pass region estimation from multiple frames of broadcast soccer videos

  Sho Takahashi, Miki Haseyama  ITE Transactions on Media Technology and Applications  1-  (3)  220  -225  2013  [Not refereed][Not invited]
   

  An Active grid-based method for estimating pass regions from broadcast soccer videos is presented in this paper. It is assumed that the pass region has a high probability of the pass succeeding. In soccer matches, players discover pass regions based on previous and current player positions. In conventional methods, pass regions are estimated by applying Active Net to only a single frame of a soccer video. In the proposed method, Active grid is applied to three-dimensional data by which frames of the soccer video are connected with the temporal dimension. The proposed method then realizes robust estimation of pass regions based on multiple frames of player positions. The proposed method was applied to actual TV programs to verify its effectiveness.
• Novel evaluation criterion for visualization of image search results based on feature integration theory

  Katsuki Kobayashi, Takahiro Ogawa, Miki Haseyama  ITE Transactions on Media Technology and Applications  1-  (4)  333  -342  2013  [Not refereed][Not invited]
   

  This paper presents a new evaluation criterion for visualization of image search results based on the feature integration theory. This criterion is derived by combining two elements, visual saliency on visualization and grouping degree of similar images. Visual saliency, which is calculated from the feature integration theory, on visualization of image search results enables representation of users' attention, which is closely related to the effectiveness of finding images. Furthermore, since users perceive similar images that are close to each other as one group, grouping degree of similar images enables evaluation of the effectiveness when users find images similar to a desired image. Therefore, by combining visual saliency on visualization and grouping degree of similar images, we can derive the novel criterion and evaluate the effectiveness of visualization of image search results.
• A review of video retrieval based on image and video semantic understanding

  Miki Haseyama, Takahiro Ogawa, Nobuyuki Yagi  ITE Transactions on Media Technology and Applications  1-  (1)  2  -9  2013  [Not refereed][Not invited]
   

  Research trends in new video retrieval based on image and video semantic understanding are presented in this paper. First, recent studies related to image and video semantic analysis are introduced to understand leading-edge multimedia retrieval technologies. Several works related to visualization interfaces for multimedia retrieval are also presented. Finally, trends in state-of-the-art studies and the future outlook are described.
• Quality estimation method for fractal compressed images

  Megumi Takezawa, Hirofumi Sanada, Miki Haseyama  ITE Transactions on Media Technology and Applications  1-  (2)  178  -183  2013  [Not refereed][Not invited]
   

  A method for estimating the quality of images compressed by fractal image compression is presented in this paper. Fractal image compression based on an iterated function system is one of the compression techniques for digital images. It utilizes the self-similarity of images and achieves high image-compression performance. However, fractal image compression is currently not being in widespread use because it does not necessarily provide high-quality compressed images. We cannot determine whether a given image is unsuitable for fractal image compression without encoding it. Therefore, in this paper, we propose a new criterion for estimating the suitability of fractal image compression for a given image. By using the proposed criterion, we can estimate the quality of the compressed image in a short time without actually encoding the image.
• Glare detection for night wet road surfaces and driver visibility improvement by using multiple onboard cameras

  Hiroaki Kumon, Miki Haseyama  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  67-  (3)  J95  -J103  2013  [Not refereed][Not invited]
   

  Driver's visibility diminishes severely in rainy conditions at night. The glare on wet road surfaces causes this poor visibility. In this paper, we propose an image processing method that detects the glare on wet road surfaces by Using multiple onboard cameras. We also propose an image generating method that reduces the effect of glare. By Using our method, glare-reduced images can be obtained and by displaying these images on the monitor equipped in the vehicle, the number of traffic accidents can be reduced. The proposed method is based on the characteristic that the road plane has a common planar surface in multiple camera images. By Using this characteristic and a homography matrix, the glare can be detected. Experimental results obtained by applying the proposed method to real image data show its high performance.
• Error Resilient 3D DWT Video Transmission Using Invertible Codes

  Kotoku Omura, Shoichiro Yamasaki, Tomoko K. Matsushima, Hirokazu Tanaka, Miki Haseyama  2013 INTERNATIONAL SYMPOSIUM ON INTELLIGENT SIGNAL PROCESSING AND COMMUNICATIONS SYSTEMS (ISPACS)  180  -185  2013  [Not refereed][Not invited]
   

  There have been many studies that apply three-dimensional discrete wavelet transform (3D DWT) to video coding. It is known that lowest frequency sub-band coefficients of 3D DWT outputs affect the visual quality of video. In order to maintain image quality of video that is transmitted over channel with error such as a wireless, it is necessary to development of error resilient transmission scheme for protecting the lowest sub-band coefficients. Coauthors of the study proposed the method (conventional method), which adopted duplication scheme of lowest frequency subband coefficients of 3D DWT. The objective of this study is to indicate improvement of error resilience by replacing the duplication scheme of conventional method with the error correction scheme using invertible codes. In addition, conventional study is assumed the packet loss channel model that packets are lost or received without bit errors. However, in wireless communications of the actual, the packet is often received with errors, and errors occur in bursty. The simulation evaluations compare the performances of the proposed method with those of the conventional one over burst error channel.
• A rating prediction method for e-commerce application using ordinal regression based on LDA with multi-modal features

  Takayuki Kawashima, Takahiro Ogawa, Miki Haseyama  2013 IEEE 2nd Global Conference on Consumer Electronics, GCCE 2013  260  -261  2013  [Not refereed][Not invited]
   

  This paper presents a new method for rating prediction in e-commerce, which uses ordinal regression based on linear discriminant analysis (LDA) with multi-modal features. In order to realize accurate recommendation in e-commerce, the proposed method estimates each user's rating for target items. Note that we define the rating as 'the degree of preference for each item by a user.' For estimating the target user's preference of each item from the past ratings of other items, the proposed method performs training from pairs of 'ratings of items' and their feature vectors using ordinal regression based on LDA. Furthermore, in this approach, new features are obtained by applying canonical correlation analysis (CCA) to textual and visual features extracted from review's texts and images on the Web, respectively. Therefore, higher performance of the rating prediction can be realized by our method than that when using single kind of features. Experimental results obtained by applying the proposed method to an actual movie data set, which has been provided by SNAP, show the effectiveness of the proposed method. © 2013 IEEE.
• Adaptive parameter setting for pass region estimation in soccer videos and its performance verification

  Sho Takahashi, Miki Haseyama  2013 IEEE 2nd Global Conference on Consumer Electronics, GCCE 2013  271  -272  2013  [Not refereed][Not invited]
   

  This paper proposes an accurate pass region estimation method by introducing adaptive parameter settings. Our previous paper proposed a pass region estimation method by utilizing average values of ball and player velocities. However, such velocities vary according to player density and skill. Therefore, in order to realize a more accurate pass region estimation, the proposed method obtains parameters, which are ball and player velocities, from player positions in a target soccer video. By introducing the above parameter settings to pass region estimation, more realistic pass region can be obtained. Consequently, the accurate method of pass region estimation is realized. © 2013 IEEE.
• 半教師あり学習による走査型電子顕微鏡で撮像された昆虫の識別に関する検討

  TAKAHASHI MASAHIRO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.146  2012/10/20  [Not refereed][Not invited]
  
• モーフィングを用いた新たな学習データの生成による昆虫の自動分類の高精度化

  BOKU KUN, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.125  2012/10/20  [Not refereed][Not invited]
  
• 弾性モデルを用いたサッカー映像における選手追跡の高精度化に関する検討

  IWAI KAZUYA, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.149  2012/10/20  [Not refereed][Not invited]
  
• 映像の特徴を用いたWebコミュニティ抽出の高精度化に関する検討―最短リンク経路長の導入による試み―

  HARAKAWA RYOSUKE, HATAKEYAMA YASUTAKA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.150  2012/10/20  [Not refereed][Not invited]
  
• 近赤外光除去フィルタを用いずに取得した画像からの可視光成分の抽出に関する検討

  IGARASHI YUTA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.162  2012/10/20  [Not refereed][Not invited]
  
• 楽曲中の歌唱区間推定における脳波特徴量の導入に関する検討

  KAWAKAMI TAKUYA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.157  2012/10/20  [Not refereed][Not invited]
  
• サポートベクトル回帰を用いたぶれによる劣化画像の復元に関する一検討

  YOSHIZAKI AKANE, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.145  2012/10/20  [Not refereed][Not invited]
  
• 物体認識における識別器の高精度化に関する検討 SVMを用いた物体領域の自動選定手法の導入

  YOSHIDA SO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.122  2012/10/20  [Not refereed][Not invited]
  
• 映像のシーン分割における複数特徴量の協調利用の有効性に関する評価

  SO KEN, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.123  2012/10/20  [Not refereed][Not invited]
  
• 連想型映像検索における検索過程の分析に関する一考察

  YAMANOUCHI YUTAKA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.153  2012/10/20  [Not refereed][Not invited]
  
• クラシック楽曲信号から算出される特徴量の改良と時代分類の高精度化に関する考察

  MATSUNO KEIICHI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.155  2012/10/20  [Not refereed][Not invited]
  
• 順序回帰を用いた数値予報の誤差解析における高精度化に関する検討

  TAKAHASHI SHINTARO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.147  2012/10/20  [Not refereed][Not invited]
  
• 視聴動作から推定される関心度を用いた映像の推薦に関する検討

  SHIRAISHI TETSUO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.151  2012/10/20  [Not refereed][Not invited]
  
• 画像検索結果の提示方法の有効性評価に関する検討―特徴統合理論に基づく評価指標の実現―

  KOBAYASHI KATSUKI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.152  2012/10/20  [Not refereed][Not invited]
  
• 歌詞と音響特徴を用いたファジィ分類による音楽ジャンル帰属度の定義に関する検討

  NAKANISHI RYOTA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.156  2012/10/20  [Not refereed][Not invited]
  
• サッカーのチーム戦術推定手法を用いた試合映像の分割に関する検討

  ONUKI SHUHEI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2012-  ROMBUNNO.148  2012/10/20  [Not refereed][Not invited]
  
• 高速道路の構造物の点検・診断業務の効率化に向けた検討

  紫桃孝一郎, 三石晃, 黒田尚士, 石川雄章, 湧田雄基, 長谷山美紀, 小川貴弘  土木学会年次学術講演会講演概要集(CD-ROM)  67th-  ROMBUNNO.VI-237  2012/08/01  [Not refereed][Not invited]
  
• On Speeding Up of the Dynamic Coalescence Model

  KAITA Takeshi, HASEYAMA Miki, KITAJIMA Hideo, TOMITA Shingo  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  11  -14  2012/02/20  [Not refereed][Not invited]
   

  A cluster analysis method modified of Dynamic Coalescence Model(DCM) on reducing the calculation complexity is proposed. The number of component distributions of which a mixed distribution consists is generally unknown. The shape of each of the components is complicate in the structure. And, one of the components has complicate relationships to the others. DCM can find such a component as a corresponding single cluster of points. However, the DCM involves so many calculations that it is not considered practical. The calculation time of the proposed, modified DCM is examined to be short than that of the original DCM.
• A Note on Accurate estimation method of Most Resource-consuming Diseases in Electronic Receipt Data

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  313  -317  2012/02/20  [Not refereed][Not invited]
   

  This paper proposes a method for estimating the most resource-consuming diseases from electronic receipt data. The proposed method constructs the model that represents medical treatments for diseases in a electronic receipt as a mixture distribution of medical treatments for a disease, which are represented as a probability distribution over medical practices, medical drugs and special treatment materials, by using Labeled LDA, one of probabilistic topic models. Using this model, the proposed method can realize estimation of points for each disease in a electronic receipt, and accurate estimation of the most resource-consuming diseases becomes feasible.
• On Speeding Up of the Dynamic Coalescence Model

  KAITA Takeshi, HASEYAMA Miki, KITAJIMA Hideo, TOMITA Shingo  IEICE technical report. Image engineering  111-  (442)  11  -14  2012/02/20  [Not refereed][Not invited]
   

  A cluster analysis method modified of Dynamic Coalescence Model(DCM) on reducing the calculation complexity is proposed. The number of component distributions of which a mixed distribution consists is generally unknown. The shape of each of the components is complicate in the structure. And, one of the components has complicate relationships to the others. DCM can find such a component as a corresponding single cluster of points. However, the DCM involves so many calculations that it is not considered practical. The calculation time of the proposed, modified DCM is examined to be short than that of the original DCM.
• A Note on Accurate estimation method of Most Resource-consuming Diseases in Electronic Receipt Data

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report. Image engineering  111-  (442)  313  -317  2012/02/20  [Not refereed][Not invited]
   

  This paper proposes a method for estimating the most resource-consuming diseases from electronic receipt data. The proposed method constructs the model that represents medical treatments for diseases in a electronic receipt as a mixture distribution of medical treatments for a disease, which are represented as a probability distribution over medical practices, medical drugs and special treatment materials, by using Labeled LDA, one of probabilistic topic models. Using this model, the proposed method can realize estimation of points for each disease in a electronic receipt, and accurate estimation of the most resource-consuming diseases becomes feasible.
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (画像工学)

  何 再興, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  1  -6  2012/02/20  [Not refereed][Not invited]
  
• A Note on Accuracy Improvement of Error Analysis in Numerical Weather Prediction

  高橋 信太郎, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  7  -10  2012/02/20  [Not refereed][Not invited]
  
• Color Correction for Images Including Visible and Infrared Components : A note on Color Transfer based on Color Components of Natural Images

  和泉 大佑, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  21  -25  2012/02/20  [Not refereed][Not invited]
  
• A note on accurate deer detection from road images using movement direction

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  67  -71  2012/02/20  [Not refereed][Not invited]
  
• A Note on User Clustering Using Measurement of Levels of Interest Estimated from Viewing Behavior

  白石 哲夫, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  189  -193  2012/02/20  [Not refereed][Not invited]
  
• A note on the application of Web information to near-duplicate online video detection (画像工学)

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  201  -205  2012/02/20  [Not refereed][Not invited]
  
• A method for detecting important events in soccer videos of amateur teams : An approach using tactical comments added to videos for soccer coaching

  高橋 翔, 嶌田 聡, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  275  -280  2012/02/20  [Not refereed][Not invited]
  
• A Note on Music Recommendation Based on Human Motion and Its Application

  大串 裕幸, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  281  -285  2012/02/20  [Not refereed][Not invited]
  
• A note on similarity calculation between shoe sole images for forensics

  畠山 泰貴, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (442)  303  -306  2012/02/20  [Not refereed][Not invited]
  
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (ITS)

  何 再興, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  1  -6  2012/02/20  [Not refereed][Not invited]
  
• A Note on Accuracy Improvement of Error Analysis in Numerical Weather Prediction

  高橋 信太郎, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  7  -10  2012/02/20  [Not refereed][Not invited]
  
• Color Correction for Images Including Visible and Infrared Components : A note on Color Transfer based on Color Components of Natural Images

  和泉 大佑, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  21  -25  2012/02/20  [Not refereed][Not invited]
  
• A note on accurate deer detection from road images using movement direction

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  67  -71  2012/02/20  [Not refereed][Not invited]
  
• A Note on User Clustering Using Measurement of Levels of Interest Estimated from Viewing Behavior

  白石 哲夫, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  189  -193  2012/02/20  [Not refereed][Not invited]
  
• A note on the application of Web information to near-duplicate online video detection (ITS)

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  201  -205  2012/02/20  [Not refereed][Not invited]
  
• A method for detecting important events in soccer videos of amateur teams : An approach using tactical comments added to videos for soccer coaching

  高橋 翔, 嶌田 聡, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  275  -280  2012/02/20  [Not refereed][Not invited]
  
• A Note on Music Recommendation Based on Human Motion and Its Application

  大串 裕幸, 小川 貴弘, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  281  -285  2012/02/20  [Not refereed][Not invited]
  
• A note on similarity calculation between shoe sole images for forensics

  畠山 泰貴, 長谷山 美紀  電子情報通信学会技術研究報告 : 信学技報  111-  (441)  303  -306  2012/02/20  [Not refereed][Not invited]
  
• A Note on Estimation of Pitcher's Condition Based on Pitching Motion in Baseball Video and Scorebook

  KUBO JUNKI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  111-  (442(IE2011 105-132))  287  -292  2012/02/13  [Not refereed][Not invited]
  
• A Note on Accurate Restoration of Latent Image from Blurred Image-Ringing Artifact Suppression using Prior Probability including Direction-based Weighting Parameters-

  YOSHIZAKI AKANE, IZUMI DAISUKE, OGAWA TAKAHIRO, HASEYAMA MIKI  電子情報通信学会技術研究報告  111-  (442(IE2011 105-132))  27  -32  2012/02/13  [Not refereed][Not invited]
  
• A Note on Accurate estimation method of Most Resource-consuming Diseases in Electronic Receipt Data

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  電子情報通信学会技術研究報告. IE, 画像工学  111-  (442)  313  -317  2012/02/13  [Not refereed][Not invited]
  
• Dimensionality Reduction of Sparse Visual Features via Recoverable Projection for Large-Scale Image Retrieval (メディア工学 ヒューマンインフォメーション 映像表現&コンピュータグラフィックス)

  何 再興, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  1  -6  2012/02  [Not refereed][Not invited]
  
• A Note on Accuracy Improvement of Error Analysis in Numerical Weather Prediction

  高橋 信太郎, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  7  -10  2012/02  [Not refereed][Not invited]
  
• On Speeding Up of the Dynamic Coalescence Model

  海田 健, 長谷山 美紀, 北島 秀夫  映像情報メディア学会技術報告  36-  (9)  11  -14  2012/02  [Not refereed][Not invited]
  
• Color Correction for Images Including Visible and Infrared Components : A note on Color Transfer based on Color Components of Natural Images

  和泉 大佑, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  21  -25  2012/02  [Not refereed][Not invited]
  
• A note on accurate deer detection from road images using movement direction

  桂井 麻里衣, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  67  -71  2012/02  [Not refereed][Not invited]
  
• A Note on User Clustering Using Measurement of Levels of Interest Estimated from Viewing Behavior

  白石 哲夫, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  189  -193  2012/02  [Not refereed][Not invited]
  
• A Note on the Application of Web Information to Near-Duplicate Online Video Detection (メディア工学 ヒューマンインフォメーション 映像表現&コンピュータグラフィックス)

  ペンコフ マイケル, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  201  -205  2012/02  [Not refereed][Not invited]
  
• A method for detecting important events in soccer videos of amateur teams : An approach using tactical comments added to videos for soccer coaching

  高橋 翔, 嶌田 聡, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  275  -280  2012/02  [Not refereed][Not invited]
  
• A Note on Music Recommendation Based on Human Motion and Its Application

  大串 裕幸, 小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  281  -285  2012/02  [Not refereed][Not invited]
  
• A note on similarity calculation between shoe sole images for forensics

  畠山 泰貴, 長谷山 美紀  映像情報メディア学会技術報告  36-  (9)  303  -306  2012/02  [Not refereed][Not invited]
  
• 選手の移動速度を考慮したサッカー映像における3次元パス可能領域の推定に関する検討

  高橋翔, 長谷山美紀  映像情報メディア学会冬季大会講演予稿集(CD-ROM)  2012-  2012
• An image copyright-protectionapproach without signature insertion

  Miki Haseyama, Isao Kondo  Proceedings of the 6th IASTED International Conference on Software Engineering and Applications, SEA 2002  275  -279  2012/01/01  [Not refereed][Not invited]
   

  Various watermarking algorithms have provided successful solutions for copyright protection and authentication of multimedia data. Watermarking techniques usually embed a watermark into an original host therefore the original content is altered. When authors/owners of art works desire to distribute their products in the original quality while require the copyrights to be properly protected, the watermark embedding techniques cannot satisfy such a requirement. To respond to such requirements, this paper proposes a new image copyright-protection system, which does not alter the original images at all while provides a robust and reliable authentication scheme to the protected images. The proposed system consists of a signature generation subsystem and a signature extraction subsystem. Neither of the subsystems requires any signatures to be embedded in the original images. The experimental results show that our system can survive attacks to the images and can correctly extract the signature for authentication.
• Support vector data dscription-based method for finding new benthic species and estimating their taxonomy position from microscopic images

  Takashi Hasegawa, Takahiro Ogawa, Hidemi Watanabe, Miki Haseyama  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  66-  (7)  J240  -J250  2012  [Not refereed][Not invited]
   

  This paper presents a support vector data description (SVDD)-based method for finding new benthic species from microscopic images and its application to taxonomy position estimation. First, the proposed method generates hyperspheres that represent taxonomic species taxa of known species and enables automatic species classification. Furthermore, weight estimation of visual features based on multiple kernel learning (MKL) is used in this approach to realize automatic weighting of categorical traits that are traditionally determined by taxonomists. Next, based on the traditional taxonomic classification scheme, the proposed method merges the hyperspheres of similar species and generates new hyperspheres that represent ultra-species taxa in higher hierarchies. Then, from the obtained results, a new decision tree, whose nodes are hyperspheres of species taxa and ultra-species taxa, is constructed. By using this decision tree, new benthic species can be found from target samples, and their taxonomic positions can also be estimated.
• Image Inpainting by Block-Based Linear Regression with Optimal Block Selection

  Akira Tanaka, Takahiro Ogawa, Miki Haseyama  2012 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC)  1  -4  2012  [Not refereed][Not invited]
   

  Estimation of missing entries in a multivariate data is one of classical problems in the field of statistical science. One of most popular approaches for this problem is linear regression based on the EM algorithm. When we consider to apply this approach to block-based image inpainting problems, we have additional information, that is, a target lost pixel could be included in multiple blocks, which implies that we have multiple candidates of estimates for the pixel. In such cases, we have to choose a good estimate among the multiple candidates. In this paper, we propose a novel image inpainting method incorporating optimal block selection in terms of the expected squared errors among multiple candidates of the estimate for the target pixel. Results of numerical examples are also shown to verify the efficacy of the proposed method.
• サッカー映像における選手位置抽出結果を用いたチーム戦術の推定に関する検討

  ONUKI SHUHEI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.154  2011/10/22  [Not refereed][Not invited]
  
• 楽曲特徴量に基づくクラシック楽曲の時代分類に関する考察

  MATSUNO KEIICHI, OKUSHI HIROYUKI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.148  2011/10/22  [Not refereed][Not invited]
  
• 数値予報における予報誤差の解析に関する検討

  TAKAHASHI SHINTARO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.151  2011/10/22  [Not refereed][Not invited]
  
• 画像検索結果の提示方法の有効性に関する一検討

  KOBAYASHI KATSUKI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.145  2011/10/22  [Not refereed][Not invited]
  
• 視聴行動解析に基づく関心度の推定に関する検討

  SHIRAISHI TETSUO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.157  2011/10/22  [Not refereed][Not invited]
  
• 相互部分空間法を用いた音楽番組のインデキシング高精度化に関する検討

  YOSHIDA SO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2011-  ROMBUNNO.144  2011/10/22  [Not refereed][Not invited]
  
• 8-4 A note on detection of important events based on relationship between player action and sound on stadium in soccer videos

  KUBO Junki, TAKAHASHI Sho, OGAWA Takahiro, HASEYAMA Miki  Proceedings of the ... ITE annual convention  2011-  (2011)  8  -4-1-"8-4-2"  2011/08/24  [Not refereed][Not invited]
   

  This paper presents a detection method of important events based on relationship between player action and sound on stadium in soccer videos. Generally, since player action and sound on stadium have high correlation in the important events, we realize the detection by using their relationship based on canonical correlation analysis.
• 8-5 A Note on Color Component Estimation of Team Uniforms in Soccer Videos based on Color Co-occurrence Histograms

  TAKAHASHI Sho, HASEYAMA Miki  Proceedings of the ... ITE annual convention  2011-  (2011)  8  -5-1-"8-5-2"  2011/08/24  [Not refereed][Not invited]
   

  This paper proposes a color estimation method of team uniforms in soccer videos based on color co-occurrence. The proposed method calculates color correlograms and selects sets of color components, whose co-occurrences are higher than the other sets. This enables the color component estimation of the team uniforms, which contain multiple colors.
• 10-11 A Note on Restoration of Latent Image from Blurred Image : Introduction of Adaptive Parameter Setting in Prior Probability Distribution

  YOSHIZAKI Akane, IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  Proceedings of the ... ITE annual convention  2011-  (2011)  10  -11-1-"10-11-2"  2011/08/24  [Not refereed][Not invited]
   

  This paper presents a method for blurred image restoration. The proposed method adopts object matching for PSF estimation and adaptive parameter setting of a prior probability distribution for latent image estimation. Then, accurate blur removal becomes feasible without suffering from ringing artifacts.
• 選手動作と歓声の関係性に注目したサッカー映像における重要場面の検出に関する検討

  KUBO JUNKI, TAKAHASHI SHO, OGAWA TAKAHIRO, HASEYAMA MIKI  映像情報メディア学会年次大会講演予稿集(CD-ROM)  2011-  ROMBUNNO.8-4  2011/08/01  [Not refereed][Not invited]
  
• ぶれにより劣化した画像の復元に関する一考察―事前確率分布の適応的なパラメータ設定の導入―

  YOSHIZAKI AKANE, IZUMI DAISUKE, OGAWA TAKAHIRO, HASEYAMA MIKI  映像情報メディア学会年次大会講演予稿集(CD-ROM)  2011-  ROMBUNNO.10-11  2011/08/01  [Not refereed][Not invited]
  
• A note on keyword hierachy construction using visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  69  -72  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A note on a fast Web community extraction method using features of video materials

  HATAKEYAMA Yasutaka, HASEYAMA Miki  ITE technical report  35-  (9)  73  -76  2011/02/14  [Not refereed][Not invited]
   

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings

  TAKAHASHI Sho, HASEYAMA Miki  ITE technical report  35-  (9)  77  -82  2011/02/14  [Not refereed][Not invited]
   

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  83  -87  2011/02/14  [Not refereed][Not invited]
   

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A note on improvement of classification performance based on SVDD including target object detection scheme

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  113  -117  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Improvement of Flow Estimation using Multiple Meteorological Data

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  119  -124  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki  ITE technical report  35-  (9)  125  -129  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• A note on accurate scene segmentation based on the MCMC method using object matching

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  131  -135  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  159  -164  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  171  -174  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A traffic flow measurement method using simple image processing and its application to real data

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  193  -198  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A note on visualization of user preference based on purchase data analysis

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  35-  (9)  199  -202  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory

  REN Jie, NAGAI Nobuo, HASEYAMA Miki  ITE technical report  35-  (9)  203  -208  2011/02/14  [Not refereed][Not invited]
   

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• A note on keyword hierachy construction using visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  69  -72  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A note on a fast Web community extraction method using features of video materials

  HATAKEYAMA Yasutaka, HASEYAMA Miki  IEICE technical report  110-  (420)  73  -76  2011/02/14  [Not refereed][Not invited]
   

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings

  TAKAHASHI Sho, HASEYAMA Miki  IEICE technical report  110-  (420)  77  -82  2011/02/14  [Not refereed][Not invited]
   

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  83  -87  2011/02/14  [Not refereed][Not invited]
   

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A note on improvement of classification performance based on SVDD including target object detection scheme

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  113  -117  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Improvement of Flow Estimation using Multiple Meteorological Data

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  119  -124  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki  IEICE technical report  110-  (420)  125  -129  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• A note on accurate scene segmentation based on the MCMC method using object matching

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  131  -135  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  159  -164  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  171  -174  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A traffic flow measurement method using simple image processing and its application to real data

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  193  -198  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A note on visualization of user preference based on purchase data analysis

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (420)  199  -202  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory

  REN Jie, NAGAI Nobuo, HASEYAMA Miki  IEICE technical report  110-  (420)  203  -208  2011/02/14  [Not refereed][Not invited]
   

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• A note on keyword hierachy construction using visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  69  -72  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a keyword hierarchy construction method and its application to automatic image annotation. This method takes advantage of the characteristic that keywords with high-level semantic meaning tend to high visial diversity across the corresponding image set. Based on this characteristic, the keyword hierarchy is constructed by using the following approach: First, a visual feature clustering method is applied to a set of images that are annotated with a target keyword. Then, a novel criterion that represents visual diversity of the keyword is introduced. Specifically, this criterion is calculated using both intra-cluster and inter-cluster visual similarities. Finally, the keyword hierarchy can be constructed by sorting the keyword criterions. Based on the obtained keyword hierarchy, low-level keywords are first estimated from visual features, and then high-level keywords are provided by using semantic relationships of the low-level keywords. This will improve performance of image annotation.
• A note on a fast Web community extraction method using features of video materials

  HATAKEYAMA Yasutaka, HASEYAMA Miki  IEICE technical report  110-  (421)  73  -76  2011/02/14  [Not refereed][Not invited]
   

  In this paper, we proposed a fast Web community extraction method based on Web video features using Locality Sensitive Hashing. First, the proposed method applies Locality Sensitive Hashing to the low level Web video features such as visual, audio, and textual features, and enables fast calculation of similarity between Web pages containing Web videos. Furthermore, on the basis of the obtained similarities and hyperlinks between Web videos, Web communities containing similar Web videos are extracted. Therefore, the proposed method enables fast Web community extraction, and we can apply the proposed method to huge dataset.
• A Study on Accurate Pass Region Estimation in Soccer Videos : Introduction of Adaptive Parameter Settings

  TAKAHASHI Sho, HASEYAMA Miki  IEICE technical report  110-  (421)  77  -82  2011/02/14  [Not refereed][Not invited]
   

  In this paper, an accurate method of pass region estimation method is proposed by introduceing an adaptive parameter settings. We have proposed a pass region estimation method by utilizing average values of ball's velocity and player's velocity. However, velocities of a ball and players are different according to player's density and player's skill. Therefore, in the proposed method, parameters, which are a velocity of a ball and velocities of players, are set by using obtained player's positions from a target soccer video. Consicuently, the accurate method of pass region estimation is realized.
• Music Recommendation Based on Relationship between Human Motions and Music Pieces : A note on its automation and accuracy improvement

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  83  -87  2011/02/14  [Not refereed][Not invited]
   

  This paper prepares an automatic and accurate music recommendation method based on the relation-ship between human motions and music pieces extracted from video sequences. In the proposed method, it is necessary to extract human regions from video sequences when modelling this relationship based on kernel Canonical Correlation Analysis (CCA). Therefore, the proposed method realizes automatic music recommendation from human motions by automating its extraction according to motion features extracted from video sequences. The experimental results show the effectiveness of our method of human region detection and its application to music recommendation.
• A note on improvement of classification performance based on SVDD including target object detection scheme

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  113  -117  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a new method for improving classification performance based on the SVDD including a target object detection scheme. In the proposed method, regions including the target objects are automatically selected based on the distance between their visual feature vectors and the center of the hypersphere calculated by the SVDD. Then a generation of new positive examples is realized. Thus, even though from training images with various position, direction, scale, and shape of target objects, it can be expected to use only local blocks including the target objects as the new positive examples. Furthermore, by using also local blocks including the selected regions, it can be realized to increase a variation of the positive examples including the target objects. Therefore, the automatic selection of the regions including the target objects and the generation of the new training images based on the obtained regions become possible, and a highly accurate classifier is realized. Experimental results are shown to verify the performance of the proposed method.
• Improvement of Flow Estimation using Multiple Meteorological Data

  LI Dezhi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  119  -124  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method, which improves the accuracy of flow estimation by using multiple meteorological data. Specifically, the proposed method extends a 2-Dimensional flow estimation algorithm to 3-Dimensional one in order to estimate flows in 3-Dimensional space. In addition, a new constraint is introduced into the above extended algorithm by using the direction of wind velocity and vertical updraft, and the accuracy of flow estimation is improved. Consequently, the accurate flow estimation can be achieved.
• Automated Classification Method of Benthic Invertebrate Samples from Microscopic Images : A Note on Improvement of Classification Performance Based on Support Vector Data Description

  HASEGAWA Takashi, OGAWA Takahiro, WATANABE Hidemi, HASEYAMA Miki  IEICE technical report  110-  (421)  125  -129  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for automatic extraction and classification of organisms from microscopic images based on Support Vector Data Description (SVDD). In the proposed method, the species are classified from microscopic images including some species of benthoses based on SVDD. Since it is difficult to classify many species at once, a hierachical classification scheme is introduced into the above procedure. Then, based on the non-conventional scheme, the proposed method realizes high performance of the classification a highly accurate classification by removal species that classified high accuracy one by one and narrow down species.
• A note on accurate scene segmentation based on the MCMC method using object matching

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  131  -135  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes an accurate scene segmentation method based on the Markov Chain Monte Carlo (MCMC) algorithm using object matching. In the proposed approach, the SIFT features which are robust to rotation, scale transition and illumination change of the video subjects are calculated. These features can be used to find correspondences between the various video subjects. The proposed method uses these correspondences to calculate a similarity measure between different shots, then extracts the scene boundary candidates using this similarity measure. The real scene boundaries can then be estimated frome the scene boundary candidates. Compared to existing MCMC techniques, the proposed makes more effective scene segmentation by focusing on the structure of the video. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. Finally, the paper presents a consideration of the features used by the proposed and the conventional scene segmentation methods.
• A Note on Super Resolution Method of Blurred Video Sequences : Accurate Estimation of Posterior Probability Including PSF

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  159  -164  2011/02/14  [Not refereed][Not invited]
   

  This paper presents a resolution enhancement method of motion blur video sequences. From a blurred low resolution video sequence, the proposed method simultaneously estimates a high resolution frame and a point spread function (PSF) by maximizing their posterior probability, i.e., Maximum a Posterior estimation scheme. In the PSF estimation, a region including many edge pixels are automatically selected within the target frame based on the intensity gradient. By utilizing the region suitable for the PSF estimation, its accuracy is improved, and accurate motion blur removal and the resolution enhancement are also realized.
• A Note on Disease Estimation from Electronic Receipt Data Using Multiple Kinds of Classifiers

  IKEDA Hironori, HATAKEYAMA Yasutaka, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  171  -174  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a method for estimating the diseases that need the largest medical resources from electronic receipt data. The proposed method uses Diagnosis Procedure Combination (DPC) data in which the diseases that need the largest medical resources are specified as training sets. From the training sets, the proposed method extracts pairs of feature vectors calculated from medical care information and labels of the diseases, which are used to train a classifier. Then the use of the classifier enables estimation of the diseases from electronic receipt data. This paper considers realization of more accurate estimation technique of the deseases that need the largest medical resources by using multiple kinds of classifiers.
• A traffic flow measurement method using simple image processing and its application to real data

  NISHINO Yasunori, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  193  -198  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a novel method for vehicle speed measurement and discrimination of vehicle types using the results of vehicle detection from traffic surveillance video. The proposed method adapts to the current lighting conditions by automatically switching between vehicle body detection and headlight detection. This automatic switching approach allows the proposed method to be applied in rapidly changing lighting conditions (e.g. twilight), thereby increasing the robustness of the method than using luminance histogram.
• A note on visualization of user preference based on purchase data analysis

  KOHTANI Hirotsugu, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (421)  199  -202  2011/02/14  [Not refereed][Not invited]
   

  This paper proposes a visualization method of user sets' preferences by analyzing purchase data. In this method, feature vectors are calculated by using pairs of products purchased at the same time. Furthermore, by comparing user sets which are classified in advance to the other user sets, differences of the preferences between different user sets are represented. In addition, we analyze purchase data based on Bag-of-Words. Therefore, even if we cannot obtain enough a user's purchase data, we are able to extract the user's preference by utilizing purchase data of user set which the user is belong to. Furthermore, we apply multidimensional scaling to feature vectors which represent the preferences of user sets. Consequently, the visualization of the user set's preference is realized effectively.
• A Note on Analysis of One Dimensional Lattice Vibration by Circuit Theory

  REN Jie, NAGAI Nobuo, HASEYAMA Miki  IEICE technical report  110-  (421)  203  -208  2011/02/14  [Not refereed][Not invited]
   

  One-dimensional lattice vibrations are represented by the equations of motion, which are represented by difference equations. By using circuit theory, equivalent LC cascade circuits are obtained by the difference equations. Here L means inductance and C means capacitance. This paper obtains eigen-frequencies of one-dimensional lattice vibration through resonances of the equivalent circuit by using transmission circuit theory.
• Missing Intensity Interpolation Using a Kernel PCA-Based POCS Algorithm and its Applications

  Takahiro Ogawa, Miki Haseyama  IEEE TRANSACTIONS ON IMAGE PROCESSING  20-  (2)  417  -432  2011/02  [Not refereed][Not invited]
   

  A missing intensity interpolation method using a kernel principal component analysis (PCA)-based projection onto convex sets (POCS) algorithm and its applications are presented in this paper. In order to interpolate missing intensities within a target image, the proposed method reconstructs local textures containing the missing pixels by using the POCS algorithm. In this reconstruction process, a nonlinear eigenspace is constructed from each kind of texture, and the optimal subspace for the target local texture is introduced into the constraint of the POCS algorithm. In the proposed method, the optimal subspace can be selected by monitoring errors converged in the reconstruction process. This approach provides a solution to the problem in conventional methods of not being able to effectively perform adaptive reconstruction of the target textures due to missing intensities, and successful interpolation of the missing intensities by the proposed method can be realized. Furthermore, since our method can restore any images including arbitrary-shaped missing areas, its potential in two image reconstruction tasks, image enlargement and missing area restoration, is also shown in this paper.
• 色の共起頻度を用いたサッカー映像におけるユニフォームの色成分の推定に関する検討

  高橋翔, 長谷山美紀  映像情報メディア学会年次大会講演予稿集(CD-ROM)  2011-  2011
• 画像・映像処理の基礎と検索の最先端

  長谷山美紀  映像情報メディア学会冬季大会講演予稿集(CD-ROM)  2011-  2011
• Human-Centric Navigation System Video Vortex for Video Retrieval

  Miki Haseyama, Takahiro Ogawa  IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS (ICCE 2011)  167  -168  2011  [Not refereed][Not invited]
   

  This paper presents a human-centric navigation system "Video Vortex" for video retrieval. This system is realized by the following functions: (i) Collaborative use of multimedia data for their multi-modal analysis, (ii) Preference extraction for the system to adapt to users' individual demands, and (iii) Adaptive visualization for users to be guided to their desired contents. These functions effectively provide a solution to the conventional problems of not being able to satisfy various users' demands, and successful video retrieval becomes feasible.
• 足跡を用いた男女識別--足跡の特徴量からSVMにより男女を識別する試み

  浅水 仁, 長谷山 美紀  画像ラボ  22-  (1)  17  -21  2011/01  [Not refereed][Not invited]
  
• AI-1-1 Cross-Media Theory and Its Application to Information Retrieval

  Haseyama Miki  Proceedings of the Society Conference of IEICE  2010-  "SS  -68"-"SS-69"  2010/08/31  [Not refereed][Not invited]
  
• Music Recommendation According to Human Motion Based on Kernel CCA

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (88)  63  -68  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• A note on keyword relationship extraction based on visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (88)  105  -108  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (88)  109  -114  2010/06/14  [Not refereed][Not invited]
   

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A study on accurate scene segmentation based on the MCMC method utilizing video structures

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (88)  115  -120  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• Music Recommendation According to Human Motion Based on Kernel CCA

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (86)  63  -68  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• A note on keyword relationship extraction based on visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (86)  105  -108  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (86)  109  -114  2010/06/14  [Not refereed][Not invited]
   

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A study on accurate scene segmentation based on the MCMC method utilizing video structures

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (86)  115  -120  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• Music Recommendation According to Human Motion Based on Kernel CCA

  OHKUSHI Hiroyuki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (89)  63  -68  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a method for recommendation of suitable music for human motion based on kernel canonical correlation analysis (CCA). The kernel CCA is used to find the relationship between different data sets, human motion data and music data. In this approach, the proposed method newly uses similarity of human motions, which is robust to temporal expantions of the motion data, as the kernel function. Therefore, we can successfully model the relationship between the human motion and the music to recommend suitable music for human motions. Experimental results are shown to verify the performance of the proposed method.
• A note on keyword relationship extraction based on visual features and its application to image annotation

  KATSURAI Marie, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (89)  105  -108  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a novel keyword relationship extraction method using visual features, and its application to image annotation. The proposed method assumes that semantically relevant keywords share common visual features, and the keyword relationships are constructed using visual features in the following approaches. First, visual features relevant to each keyword are detected by using the logistic regression. Then, by removing these features from images, the strength of keyword correlations are calculated. Thus these approach can extract visual feature based keyword relationships from a target database, and the proposed method realize an accurate image annotation by utilizing both visual features and the obtained keyword relationships.
• A note on simultaneous restoration and resolution enhancement of a blurred video sequence : Estimation of prior probability distributions for super-resolution

  IZUMI Daisuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (89)  109  -114  2010/06/14  [Not refereed][Not invited]
   

  This paper presents a method for simultaneous restoration and resolution enhancement of blurred video sequences. In the proposed method, prior probability distributions, which are necessary for MAP estimation scheme of high resolution video sequences, are estimated by utilizing intensity gradient in each region. Therefore, the proposed method avoids the reduce of intensity gradients and preserves sharpness in edge regions. On the other hand, the proposed method avoids the increase of intensity gradients in smooth regions from the motion blurred frame in order to reduce ringing artifacts. Then the proposed method simultaneously performs accurate restoration and resolution enhancement.
• A study on accurate scene segmentation based on the MCMC method utilizing video structures

  SONG Yan, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  110-  (89)  115  -120  2010/06/14  [Not refereed][Not invited]
   

  This paper proposes a video scene segmentation method based on Markov Chain Monte Carlo(MCMC) method utilizing video structures. In the proposed method, the scene boundary candidates are extracted from shot boundaries based on structures of video sequences, and then ture scene boundaries are selected from these candidates by the image feature based MCMC method. The proposed method enables accurate scene cut detection which effectively utilizes the structure of video sequences. In this scheme, the prior probability distribution of the number of scenes in the target video sequence must be obtained for the use of the conventional method by MCMC method. Thus, the proposed method estimates the parameters of the prior probability distribution from training video sequences by the multiple regression analysis. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A study on Gender Identification using Shoe Print with Support Vector Machine

  ASAMIZU Satoshi, HASEYAMA Miki  全国大会講演論文集  72-  (0)  5  -6  2010/03/08  [Not refereed][Not invited]
  
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  34-  (6)  1  -4  2010/02/15  [Not refereed][Not invited]
   

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of a articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• A note on player and ball tracking in volleyball videos and its application to event estimation

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  34-  (6)  17  -21  2010/02/15  [Not refereed][Not invited]
   

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping

  TOKUMOTO Ryota, HASEYAMA Miki  ITE technical report  34-  (6)  119  -122  2010/02/15  [Not refereed][Not invited]
   

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• An accurate scene segmentation method based on video structure analysis using PLSA : A note on definition of new shot similarity by sequence alignment

  YAMAMOTO Makoto, HASEYAMA Miki  ITE technical report  34-  (6)  123  -128  2010/02/15  [Not refereed][Not invited]
   

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping (DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis (PLSA) to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  ITE technical report  34-  (6)  129  -132  2010/02/15  [Not refereed][Not invited]
   

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model

  KOBAYASHI Kazuya, HASEYAMA Miki  ITE technical report  34-  (6)  171  -174  2010/02/15  [Not refereed][Not invited]
   

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description (SVDD) to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments

  MURAYAMA Shota, HASEYAMA Miki  ITE technical report  34-  (6)  175  -178  2010/02/15  [Not refereed][Not invited]
   

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos

  TAKAHASHI Sho, HASEYAMA Miki  ITE technical report  34-  (6)  185  -190  2010/02/15  [Not refereed][Not invited]
   

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibiliy in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  109-  (414)  1  -4  2010/02/08  [Not refereed][Not invited]
   

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of an articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• A note on player and ball tracking in volleyball videos and its application to event estimation

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  109-  (414)  17  -21  2010/02/08  [Not refereed][Not invited]
   

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping

  TOKUMOTO Ryota, HASEYAMA Miki  IEICE technical report  109-  (414)  119  -122  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• An accurate scene segmentation method based on video structure analysis using PLSA : A note on definition of new shot similarity by sequence alignment

  YAMAMOTO Makoto, HASEYAMA Miki  IEICE technical report  109-  (414)  123  -128  2010/02/08  [Not refereed][Not invited]
   

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping(DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis(PLSA)to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report  109-  (414)  129  -132  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model

  KOBAYASHI Kazuya, HASEYAMA Miki  IEICE technical report  109-  (414)  171  -174  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description(SVDD)to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments

  MURAYAMA Shota, HASEYAMA Miki  IEICE technical report  109-  (414)  175  -178  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos

  TAKAHASHI Sho, HASEYAMA Miki  IEICE technical report  109-  (414)  185  -190  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibility in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• A Note on Upgrade of 3D Model Based Pose Estimation Using Parametric Eigenspace Method

  KADONO Kohsuke, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  109-  (415)  1  -4  2010/02/08  [Not refereed][Not invited]
   

  This paper presents a pose estimation method using the parametric eigenspace method. The proposed method obtains an eigenspace from a set of silhouette images which are generated from the 3D model of a articulated object. Next, a manifold is constructed by interpolating the silhouette images projected to the eigenspace. Furthermore, the proposed method projects a silhouette of a target object to the eigenspace and estimates optimal parameters by calculating a distance between the projected silhouette and the manifold. In this procedure, we represent the target articulated object as a tree structure. Then, by calculating the parameters from a root node, we realize the reduction of pose estimation.
• A note on player and ball tracking in volleyball videos and its application to event estimation

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  109-  (415)  17  -22  2010/02/08  [Not refereed][Not invited]
   

  This paper presents a novel ball tracking method in volleyball videos and its application to the estimation of events such as receive and toss. By utilizing characteristics that the ball moves parabolically in the volleyball videos, the proposed method calculates conceivable ball trajectories based on tracked player positions. Then, the proposed method realizes not only the ball tracking but also the detection of players who contact the ball, which is very important for the event estimation. Furthermore, we discuss the application of the proposed method to the event estimation based on volleyball rules. The effectiveness of the proposed method is verified in the experiments.
• A Study on Quantitative Evaluation of Associative Video Retrieval Based on Perceptual Grouping

  TOKUMOTO Ryota, HASEYAMA Miki  IEICE technical report  109-  (415)  119  -122  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a quantitative evaluation of associative video retrieval results based on perceptual grouping. We have previously proposed an associative video retrieval system for providing the videos desired by users even if users don't have specific queries. However, the traditional method does not include a quantitative evaluation. Therefore, the proposed method defines the model to evaluate the associative video retrieval results. We regard group of videos as perceptual grouping through the model, its quantitative values are calculated. Furthermore, we define a value to evaluate an effectiveness of the associate video retrieval results using the quantitative values. Consequently, we achieve the effective quantitative evaluation of the associative video retrieval results by comparing the evaluation values.
• An accurate scene segmentation method based on video structure analysis using PLSA : A note on definition of new shot similarity by sequence alignment

  YAMAMOTO Makoto, HASEYAMA Miki  IEICE technical report  109-  (415)  123  -128  2010/02/08  [Not refereed][Not invited]
   

  A method for accurate scene segmentation based on video structure analysis using sequences of continuous shots in audiovisual materials is proposed. In the proposed method, for efficient video structure analysis, similarities between the sequences of continuous shots are calculated by utilizing Dynamic Time Warping (DTW), one of the methods for sequence alignment. In this calculation, by applying Probabilistic Latent Semantic Analysis (PLSA) to audiovisual features extracted from the audiovisual materials, new features of the shots are obtained, and the proposed method newly introduces these features in definition of the costs used in DTW. Consequently, the proposed method can realize efficient video structure analysis based on the similarities between the sequences of continuous shots, and thereby accurate scene segmentation becomes feasible.
• A note on an accurate video retrieval method based on Web community extraction : Introduction of clustering method using similarity between features of video materials

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report  109-  (415)  129  -132  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes an accurate retrieval method based on Web community extraction introducing clustering of video materials using their similarities. In the proposed method, on the basis of visual and audio features of video materials and textual features obtained from Web pages including them, link relationship are weighted by similarities between their features. Next, by applying link analysis to the weighted link relationships, Web communities containing video materials whose topics are similar to each other can be extracted. Furthermore, the video materials which belong to a same Web community are clustered based on similarities between their features. Then, in the proposed method, users can effectively retrieve video materials which are similar to desired video materials using their features. Consequently, the proposed method enables users to effectively retrieve the video materials whose topics and features are similar to each other.
• A note on music recommendation based on personal preferences : Introduction of similarity index for preference model

  KOBAYASHI Kazuya, HASEYAMA Miki  IEICE technical report  109-  (415)  171  -174  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a new music recommendation method based on preference similarity between users. In the proposed method, by applying Support Vector Data Description (SVDD) to musical feature vectors which are calculated from training data, we calculate hyperspheres discriminating preferred music pieces. Furthermore, we newly define a degree of preference similarity between users by using the hyperspheres, and make the combination of similar users in preference based on this degree. Finally, recommended music pieces are outputted based on two probabilities calculated by each center of the hypersphere for the pair of users. Therefore, based on the preference similarity between users, recommendation of music pieces can be effectively realized.
• A note on traffic flow measurement for traffic surveillance video : Reduction of performance degradation in various environments

  MURAYAMA Shota, HASEYAMA Miki  IEICE technical report  109-  (415)  175  -178  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes a novel vehicle detection method for traffic surveillance video, that reduces the performance degradation caused by the change in environmental conditions. The proposed method detects the vehicles which are driving between traffic lanes based on the vehicle body or headlights. In order to detect the vehicles without suffering from performance degradation by the change, some parameters for detecting vehicles should be appropriately set. Then we estimate the parameters by multi-regression analysis for robust vehicle detection. Thus the appropriate parameters which are suitable for various environments can be set by the proposed method, and an effective traffic flow measurement is realized.
• A Note on 3D Active Grid Based Estimation of 3D Pass Region from Broadcasted Soccer Videos

  TAKAHASHI Sho, HASEYAMA Miki  IEICE technical report  109-  (415)  185  -190  2010/02/08  [Not refereed][Not invited]
   

  This paper proposes an estimation method of 3D pass regions from the actual broadcasted soccer videos, where the 3D pass region has a high probability in which the pass succeeds. We have proposed an estimation method of 2D pass regions by using player positions and a ball position, where the player positions and the ball position are obtained as 2D coordinate from the actual broadcasted soccer video. However, the pass regions exist on a 3D space as a region of a 3D shape. Therefore, the proposed method realizes the estimation of the 3D pass region from the 3D space by using the 3D Active Grid. In order to use the 3D Active Grid, new volume data are generated by using player and ball positions, where the volume data indicate a possibiliy in which players can reach positions in the 3D space. Consequently, this paper realizes the estimation of the pass regions by using the 3D Active Grid.
• 最新!感染看護Q&A

  塚本容子, 松澤由香里, 平岡康子, 原理加, 鎌田明, 秩父千春, 塩谷光彦, 唯野貢司, 齊藤正恵, 加藤美紀, 桐則行, 西朝江, 三宮乙恵, 高井裕美, 山田佳緒里, 長谷山雅美, 石山景子, 家入裕子, 平畠弘美, 小川伸, 三宅隆仁, 小山田玲子, 石上香, 掛水智子, 五十嵐孝, 北山由香里, 四宮聡, 飯島正平, 石角鈴華, 長倉竜也, 宮崎佳奈子, 斎藤道子, 網島優, 菊地義弘, 萩野貴志, 野田久美子, 満田年宏  看護技術  56-  (12)  2010
• 形態素解析およびTFIDF手法による新たな語り分析の試み-回復期にある脳血管障害患者への手浴ケア場面を通して-

  矢野理香, 長谷山美紀  日本看護技術学会学術集会講演抄録集  9th-  2010
• PLAYERS TRACKING APPROACH USING LEVEL‐SET METHOD BASED ON COLOR COMPONENTS OF PLAYERS IN SOCCER VIDEOS

  TAKAHASHI Sho, LIM Wonkuk, HASEYAMA Miki  画像符号化シンポジウム資料  25th-  67  -68  2010  [Not refereed][Not invited]
  
• MCMC-based scene segmentation method using structure of video

  Yan Song, Takahiro Ogawa, Miki Haseyama  ISCIT 2010 - 2010 10th International Symposium on Communications and Information Technologies  862  -866  2010  [Not refereed][Not invited]
   

  Video scene segmentation and classification are fundamental steps for multimedia retrieval, browsing and indexing. In this paper, we present a robust scene segmentation approach based on the Markov Chain Monte Carlo (MCMC) method using the structure of video sequences. In our method, there are two novel approaches to segment video sequences into scenes. The first approach is the use of the video structures to extract scene boundary candidates from shot boundaries. Then using the MCMC method to select the true scene boundaries from these candidates, highly-accurate scene segmentation becomes possible. It should be noted that when the prior probability concerning the number of scenes in a target video sequence is given correctly, the MCMC method can provide a more accurate scene segmentation result. Therefore, in the second approach of the proposed method, the parameter utilized in the prior probability is set to the optimal value by using Multiple Regression Analysis (MRA). Consequently, accurate scene segmentation becomes possible by using the above novel approaches. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method. ©2010 IEEE.
• Two-step error concealment method for wavelet video coding in wireless network

  Sunmi Kim, Hirokazu Tanaka, Takahiro Ogawa, Miki Haseyama  2010 International Conference on Intelligent and Advanced Systems, ICIAS 2010  2010  [Not refereed][Not invited]
   

  We propose a two-step error concealment method based on an error resilient three-dimensional discrete wavelet transform (3-D DWT) video encoding scheme. The proposed scheme consists of an error-resilient encoder duplicating the lowest sub-band bit-streams for dispersive grouped frames and an error concealment decoder. The error concealment method of this decoder is decomposed of two steps, the first step is replacement of erroneous coefficients in the low frequency subband by the duplicated coefficients, and the second step is interpolation of the missing wavelet coefficients by minimum mean square error (MMSE) estimation. The proposed scheme can achieve robust transmission over unreliable channels. Experimental results provide performance comparisons in terms of peak signal-to-noise ratio (PSNR) and visual quality under different packet loss conditions.
• Kernel CCA-based music recommendation according to human motion robust to temporal expansion

  Hiroyuki Ohkushi, Takahiro Ogawa, Miki Haseyama  ISCIT 2010 - 2010 10th International Symposium on Communications and Information Technologies  1030  -1034  2010  [Not refereed][Not invited]
   

  This paper proposes a method for kernel canonical correlation analysis (CCA) based music recommendation robust to temporal expansion. Kernel CCA is used to find the relationship between different data sets. Generally, since motions and music pieces in video sequences have various time lengths, it is necessary to allow the internal temporal expansion of the data. Our kernel CCA-based music recommendation method uses similarities of human motions and music pieces, which are robust to internal temporal expansions. Then this approach enables successful extraction of the relationship between these data to recommend a music piece suitable for human motions. Experimental results are shown to verify the performance of the proposed method. ©2010 IEEE.
• Effective pedestrian detection using SVDD-based criterion for region integration

  Marie Katsurai, Takahiro Ogawa, Miki Haseyama  ISCIT 2010 - 2010 10th International Symposium on Communications and Information Technologies  991  -996  2010  [Not refereed][Not invited]
   

  Pedestrian detection is one of the most important techniques for surveillance applications. This paper proposes an effective method for pedestrian detection in low-contrast images. The main characteristic of the proposed method is a twostage moving object extraction. In the first stage, the watershed algorithm is used to extract multiple regions of moving objects. In the second stage, a novel criterion is introduced to integrate the segmented moving object regions. Specifically, the criterion is calculated on the basis of the distance from a center of the support vector data description (SVDD), where its hypersphere is constructed by using pedestrian features. By monitoring this SVDD-based criterion for the region integration, the segmented regions are appropriately integrated based on pedestrian features. This two-stage approach can extract the moving objects in low-contrast images and improve the performance of the pedestrian detection. Experimental results have demonstrated the effectiveness of the proposed method. ©2010 IEEE.
• A genetic algorithm for path generation and its applications

  Jun Inagaki, Tomoaki Shirakawa, Tetsuo Shimono, Miki Haseyama  APSIPA ASC 2009 - Asia-Pacific Signal and Information Processing Association 2009 Annual Summit and Conference  278  -281  2009/12/01  [Not refereed][Not invited]
   

  Path generation is an optimization problem mainly performed on grid maps that combines generation of paths with minimization of their cost. Several methods that belong to the class of exhaustive searches are available; however, these methods are only able to obtain a single path as a solution for each iteration of the search. Conversely, while genetic algorithms involving a type of multipoint search methods have been proposed as suitable candidates for this problem with the goal of simultaneously searching for multiple candidate paths, these methods are limited to particular applications, and there are limitations on the types of paths that can be represented. This paper therefore proposes a path generation method that is applicable to more general-purpose applications compared to previous methods based on a new design of the genotype used in the genetic algorithm.
• Quality estimation of fractal coded images by using level of self-similarity

  Megumi Takezawa, Hirofumi Sanada, Kazuhisa Watanabe, Miki Haseyama  APSIPA ASC 2009 - Asia-Pacific Signal and Information Processing Association 2009 Annual Summit and Conference  837  -840  2009/12/01  [Not refereed][Not invited]
   

  Fractal image coding is a technique for coding digital images. It can provide the compressed images with higher quality than JPEG compression at ultra low bit-rates. However, we cannot use it as a practical coding technique. One of the reasons is that the quality of some of the compressed images is not sufficient for practical applications. Moreover, what is I more inconvenient is that we cannot know the compressed image quality of a given image unless we actually encode and decode it which takes a lot of time. Therefore, we decided to try to resolve this problem by implementing the following steps: (i) we find the important image features which relate to the quality of the compressed images, and we establish an estimation method of the compressed image quality by using the important image feature 1 values, (ii) Then, based on the estimated quality by this method, we apply the fractal image coding to each image adaptively. This paper presents the above step (i).
• Performance of Adaptive Trellis Coded Modulation Applied to MC-CDMA with Bi-orthogonal Keying

  Hirokazu Tanaka, Shoichiro Yamasaki, Miki Haseyama  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E92A-  (11)  2837  -2843  2009/11  [Not refereed][Not invited]
   

  A Generalized Symbol-rate-increased (GSRI) Pragmatic Adaptive Trellis Coded Modulation (ATCM) is applied to a Multi-carrier CDMA (MC-CDMA) system with bi-orthogonal keying is analyzed. The MC-CDMA considered in this paper is that the input sequence. of a bi-orthogonal modulator has code selection bit sequence and sign bit sequence. In [9], an efficient error correction code using Reed-Solomon (RS) code for the code selection bit sequence has been proposed. However, since BPSK is employed for the sign bit modulation, no error correction code is applied to it. In-order to realize a high speed wireless system, a multi-level modulation scheme (e.g. MPSK, MQAM, etc.) is desired. In this paper, we investigate the performance of the MC-CDMA with bi-orthogonal keying employing GSRI ATCM. GSRI TC-MPSK can arbitrarily set the bandwidth expansion ratio keeping higher coding gain than the conventional pragmatic TCM scheme. By changing the modulation scheme and the bandwidth expansion ratio (coding rate), this scheme can optimize the performance according to the channel conditions. The performance evaluations by simulations on an AWGN channel and multi-path fading channels are presented. It is shown that the proposed scheme has remarkable throughput performance than that of the conventional scheme.
• Performance of Adaptive Modulation and Coding (AMC) on multi-carrier CDMA for mobile multimedia transmission

  Hirokazu Tanaka, Shoichiro Yanasaki, Miki Haseyama  2009 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, BMSB 2009  761  -+  2009/10/29  [Not refereed][Not invited]
   

  In this paper, we investigate the performance of MC-CDMA with bi-orthogonal keying employing Generalized Symbol-rate-increased (GSRI) Pragmatic Adaptive Trellis Coded Modulation (ATCM). It is advantageous that the GSRI TC-MPSK can arbitrarily set the bandwidth expansion ratio (coding rate) and higher coding gain than the conventional pragmatic TCM scheme can be obtained. By changing the modulation scheme and the bandwidth expansion ratio, this scheme can optimize the performance according to the channel conditions. The performance evaluations by simulations on an AWGN channel and multi-path fading channels are presented. It is shown that the proposed scheme has remarkable throughput performance than that of the conventional scheme.
• A note distance between music pieces based on personal preferences and its application to music retrieval

  KOBAYASHI Kazuya, HASEYAMA Miki  ITE technical report  33-  (44)  3  -6  2009/10/26  [Not refereed][Not invited]
   

  This paper proposes a new calculation method of distance between music pieces based on personal preferences. In the proposed method, in order to utilize musical features for the distance calculation, feature vectors which represent melody, rhythm, and timbre, are firstly calculated from traning data sets. Then, by using the feature vectors, a metric having a general form of the Mahalanobis distance is defined and learned. Specifically, the metric is learned by minimizing a cost function designed so that the cost increases when the distance between dissimilar music pieces is shorter than the distance between similar music pieces. According to the distance based on the metric, the accurate retrieval of similar music can be expected.
• A note on accurate scene segmentation by video structure analysis An approach using new features obtained by PLSA

  YAMAMOTO Makoto, HASEYAMA Miki  ITE technical report  33-  (44)  7  -10  2009/10/26  [Not refereed][Not invited]
   

  A method for accurate scene segmentation based on statistical trend analysis of audiovisual features is proposed. In the proposed method, based on the assumption that audiovisual features in the same broadcast programs include similar trends, their trends are analyzed by using Probabilistic Latent Semantic Analysis (PLSA), the method for statistical trend analysis. Then the proposed method calculates new features representing these trends. By using these new features, degradation of scene segmentation accuracy, which results from the direct use of the features obtained from the audiovisual materials, can be avoided and thereby accurate scene segmentation becomes feasible. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.
• A Note on Visualization of Video Retrieval Based on Perceptual Grouping

  TOKUMOTO Ryota, HASEYAMA Miki  ITE technical report  33-  (44)  47  -50  2009/10/26  [Not refereed][Not invited]
   

  This paper proposes a visualization method of video retrieval based on perceptual grouping. The perceptual grouping is a perceptual phenomenon which a human perceives figures as unities in eyeshot. We have previously proposed a video retrieval system for providing the videos desired by the users if users do not have specific queries. However, in traditional method, the interface is not considered a past search history. Thus, users cannot distinguish precisely between a similar video and a video provided newly from past search results. And, it becomes difficult to confirm the change of the video search results in a series of flows along the search history. Therefore, the proposed method arranges the videos in the three-dimensional space. Furthermore, we visualize the video retrieval results based on the perceptual grouping. Consequently, the effective visualization of the video retrieval results is discussed.
• A Note on Improvement of Coding Efficiency of Distributed Video Coding Based on Frame Reconstruction : Reconstructon of Key and Wyner-Ziv Frames Based on Error Correction

  TAKAHASHI Shigeki, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (44)  51  -54  2009/10/26  [Not refereed][Not invited]
   

  This paper proposes a new method which improves coding efficiency of distributed video coding (DVC). Before the encoding process of key frames, the proposed method adds missing areas to the key frames and reduces the bit rate of the target video sequence. In the decoder, the proposed method interpolates the missing areas and reconstructs Wyner-Ziv (WZ) frames and successive key frames based on error correction results of the WZ frames. Then, since decoded WZ frames are accurately reconstructed, and interpolation errors of the missing ares are effectively reduced, the proposed method can achieve higher rate-distortion (RD) performance than the traditional DVC. Experimental results show the improvement of the proposed method over the traditional DVC.
• Webのリンク構造に注目したNear‐Duplicate映像検出に関する検討

  IKEDA KEISUKE, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  ROMBUNNO.157  2009/10/17  [Not refereed][Not invited]
  
• 映像の構造に基づいたシーン分割の高精度化に関する検討

  SO KEN, YAMAMOTO MAKOTO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  ROMBUNNO.155  2009/10/17  [Not refereed][Not invited]
  
• 気象データ解析のための動き推定法の高精度化に関する一考察

  RI NORITOMO, KAKUKO NORIHIRO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  ROMBUNNO.154  2009/10/17  [Not refereed][Not invited]
  
• 画像特徴量の適応的な重み付けを導入したロジスティック回帰モデルによる画像内容に基づく類似画像検索手法の検討

  KATSURAI MARIE, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  ROMBUNNO.156  2009/10/17  [Not refereed][Not invited]
  
• 手ぶれにより劣化した動画像の復元および高解像度化に関する一検討―原画像の事前確率に対する適応的な選択法の導入―

  IZUMI DAISUKE, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  ROMBUNNO.158  2009/10/17  [Not refereed][Not invited]
  
• Performance of Reed-Solomon Coded MC-DS-CDMA with Bi-orthogonal Modulation

  Hirokazu Tanaka, Kyung Woon Jang, Shoichiro Yamasaki, Miki Haseyama  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E92A-  (10)  2600  -2608  2009/10  [Not refereed][Not invited]
   

  In this paper, an error correction scheme suitable for MC-DS-CDMA system with bi-orthogonal modulation is proposed. The input sequence of a bi-orthogonal modulator consists of n - 1 bit code selection bit sequence and 1 bit sign bit sequence. In order to apply an efficient error correction code, the following points should be considered; (1) if the code selection bits can be protected sufficiently, the sign bit error can also be reduced sufficiently, (2) since a code selection bit sequence consists of n - 1 bits, employing a symbol error correcting code is more effective for encoding code selection bits, (3) the complexity of the error correction encoder and decoder implementations should be minimum. Based on these conditions, we propose to employ Reed-Solomon (RS) code for encoding the code selection bits and no error correction code for the sign bit, Additionally, detection algorithm at the bi-orthogonal demodulator is modified for compensating degradations of the sign bit error rate performance. The performance in an Additive White Gaussian Noise (AWGN) channel is evaluated by both theoretical analysis and computer simulations. The performance evaluations by simulations on multi-path fading channels are also shown. It is shown that the proposed scheme has remarkable improvement.
• Video Frame Interpolation by Image Morphing Including Fully Automatic Correspondence Setting

  Miki Haseyama, Makoto Takizawa, Takashi Yamamoto  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E92D-  (10)  2163  -2166  2009/10  [Not refereed][Not invited]
   

  In this paper, it new video frame interpolation method based on image morphing for frame rate up-conversion is proposed. In this method, image features are extracted by Scale-Invariant Feature Transform in each frame, and their correspondence in two contiguous frames is then computed separately in foreground and background regions. By using the above two functions, the proposed method accurately generates interpolation frames and thus achieves frame rate up-conversion.
• A Note on Accurate Scene Segmentation Utilizing Speech Recognition : Introduction of Video Structure Analysis Using Object Matching

  NARITA Yo, HASEYAMA Miki  ITE technical report  33-  (33)  17  -20  2009/08/03  [Not refereed][Not invited]
   

  This paper proposes an improvement method for scene segmentation which uses new features defined from speech recognition result. In the proposed method, the new features are defined based on the appearance distribution of terms in fixed-length short segments of video materials. Furthermore, similarities between two segments are calculated from the obtained features as similarities of topics. Then, the changes of topics are detected from the time transition of the similarities, and the problem of over segmentations caused in the traditional methods can be alleviated by combining both their methods and our method. Finally, the proposed method realizes accurate scene segmentations based on the topics of video materials.
• A Note on SVDD-Based Automatic Target Object Extraction for Image Retrieval

  OKADA Hiroshi, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (33)  13  -16  2009/08/03  [Not refereed][Not invited]
   

  This paper proposes an automatic extraction method of target objects from images based on the support vector data description (SVDD) for image retrieval. From training images including and not including target objects, the proposed method enables their automatic extraction. Specifically, probabilities which feature vectors of local areas belong to the object class is defined based on distances from center of the hypersphere, and positive areas including target objects and negative areas not including these objects can be obtained from the training images based on the probabilities. In this scheme, in order to obtain the hypersphere which can separate the positive areas and the negative ones more accurately, the feature vectors are selected adaptively. Furthermore, by iterating the calculation of the hypersphere and the extraction of the positive areas and the negative ones in various sizes, the detection of target objects in images can be realized without manually extracting areas of target objects. Therefore, image retrieval based on object extraction can be expected.
• A Novel Video Retrieval Method Based on Web Community Extraction Using Features of Video Materials

  Yasutaka Hatakeyama, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E92A-  (8)  1961  -1969  2009/08  [Not refereed][Not invited]
   

  A novel video retrieval method based on Web community extraction using audio and visual features and textual features of video materials is proposed in this paper. In this proposed method, canonical correlation analysis is applied to these three features calculated from video materials and their Web pages, and transformation of each feature into the same variate space is possible. The transformed variates are based on the relationships between visual, audio and textual features of video materials, and the similarity between video materials in the same feature space for each feature can be calculated. Next, the proposed method introduces the obtained similarities of video materials into the link relationship between their Web pages. Furthermore, by performing link analysis of the obtained weighted link relationship, this approach extracts Web communities including similar topics and provides the degree of attribution of video materials in each Web community for each feature. Therefore, by calculating similarities of the degrees of attribution between the Web communities extracted from the three kinds of features, the desired ones are automatically selected. Consequently, by monitoring the degrees of attribution of the obtained Web communities, the proposed method can perform effective video retrieval. Some experimental results obtained by applying the proposed method to video materials obtained from actual Web pages are shown to verify the effectiveness of the proposed method.
• Visualization interface for video scene retrieval

  HASEYAMA Miki  IEICE technical report  109-  (65)  153  -157  2009/05/21  [Not refereed][Not invited]
   

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• Visualization interface for video scene retrieval

  HASEYAMA Miki  IEICE technical report  109-  (63)  153  -157  2009/05/21  [Not refereed][Not invited]
   

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• Visualization interface for video scene retrieval

  HASEYAMA Miki  IEICE technical report  109-  (64)  153  -157  2009/05/21  [Not refereed][Not invited]
   

  This paper shows new schemes of video retrieval and their visualization interfaces which aim at successful information retrieval from tremendous number of multimedia contents. In traditional retrieval schemes, it becomes difficult to find desired multimedia contents if users do not have specific queries. On the other hand, the proposed schemes enable the video scene retrieval by utilizing the visualization interface that leads users to desired contents. Finally, the effectiveness of the proposed schemes is shown from the results retrieved by our interface. Furthermore, the necessity of the visualization interfaces is discussed in the point of view that multimedia contents have several ambiguities.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (6)  157  -162  2009/02/04  [Not refereed][Not invited]
   

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  ITE technical report  33-  (6)  93  -96  2009/02/04  [Not refereed][Not invited]
   

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD

  OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (6)  269  -274  2009/02/04  [Not refereed][Not invited]
   

  This paper presents a support vector data description (SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution (HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution (LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake

  TAKIZAWA Makoto, HASEYAMA Miki  ITE technical report  33-  (6)  265  -268  2009/02/04  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (6)  261  -264  2009/02/04  [Not refereed][Not invited]
   

  In this paper, an ER (Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution (LR) image as that of a high-resolution (HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A study on tracking player position and taking camera focusing area in soccer video

  LIM Wonkuk, HASEYAMA Miki  ITE technical report  33-  (6)  229  -232  2009/02/04  [Not refereed][Not invited]
   

  In this paper, we propose particle filters for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  33-  (6)  47  -51  2009/02/04  [Not refereed][Not invited]
   

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error (MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR (Peak Signal-to-Noise Ratio) and visual quality.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (425)  269  -274  2009/01/28  [Not refereed][Not invited]
   

  This paper presents a support vector data description(SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution(HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution(LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake

  TAKIZAWA Makoto, HASEYAMA Miki  IEICE technical report  108-  (425)  265  -268  2009/01/28  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (425)  261  -264  2009/01/28  [Not refereed][Not invited]
   

  In this paper, an ER(Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution(LR) image as that of a high-resolution(HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A study on tracking player position and taking camera focusing area in soccer video

  LIM Wonkuk, HASEYAMA Miki  IEICE technical report  108-  (425)  229  -232  2009/01/28  [Not refereed][Not invited]
   

  In this paper, we propose particle filters, for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (425)  157  -162  2009/01/28  [Not refereed][Not invited]
   

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report  108-  (425)  93  -96  2009/01/28  [Not refereed][Not invited]
   

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (425)  47  -51  2009/01/28  [Not refereed][Not invited]
   

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error(MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR(Peak Signal-to-Noise Ratio) and visual quality.
• A Note on Adaptive Resolution Enhancement of Still Images Based on SVDD

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (424)  269  -274  2009/01/28  [Not refereed][Not invited]
   

  This paper presents a support vector data description (SVDD) based adaptive resolution enhancement of still images. First, the proposed method estimates unknown high-frequency components of a target image in such a way that the distance between the estimation result and the center of hypersphere obtained from training high-resolution (HR) local images by the SVDD becomes minimum. In this scheme, since local images between two different levels of a pyramid structure are similar to each other, local images within the target low resolution (LR) image are utilized as the training data. Furthermore, in order to realize accurate resolution enhancement of images containing several kinds of textures, these training local images are previously clustered, and the above estimation scheme is performed by using the hypersphere of each cluster. In this procedure, the proposed method determines the final output based on the probability defined from the distance between the estimation result and the center of the hypersphere for each cluster. Therefore, this approach provides a solution to a problem of traditional methods of not being able to perform adaptive estimation of the high-frequency components, and accurate resolution enhancement can be expected by our method. Some experimental results are shown to verify the performance of the proposed method.
• A note on accurate restoration and resolution enhancement of a deteriorated sequence caused by camera shake

  TAKIZAWA Makoto, HASEYAMA Miki  IEICE technical report  108-  (424)  265  -268  2009/01/28  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
• A Note on Resolution Enhancement of Still Images Using a Phase Retrieval Method

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (424)  261  -264  2009/01/28  [Not refereed][Not invited]
   

  In this paper, an ER (Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution (LR) image as that of a high-resolution (HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
• A study on tracking player position and taking camera focusing area in soccer video

  LIM Wonkuk, HASEYAMA Miki  IEICE technical report  108-  (424)  229  -232  2009/01/28  [Not refereed][Not invited]
   

  In this paper, we propose particle filters, for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
• A Note on Extension of Particle Filter : Application for Flow Estimation of Image Sequence

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (424)  157  -162  2009/01/28  [Not refereed][Not invited]
   

  This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
• A note on an accurate video retrieval method based on features of video materials : Web community extraction using CCA

  HATAKEYAMA Yasutaka, OGAWA Takahiro, ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report  108-  (424)  93  -96  2009/01/28  [Not refereed][Not invited]
   

  An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
• A Study on Error Resilient Video Coding using 3D Wavelet Transform and its Error Concealment method

  KIM Sunmi, TANAKA Hirokazu, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  108-  (424)  47  -51  2009/01/28  [Not refereed][Not invited]
   

  In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error (MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR (Peak Signal-to-Noise Ratio) and visual quality.
• A gender identification using shoeprint images

  ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report  108-  (373)  209  -212  2009/01/05  

  This paper proposes a gender identification using shoeprint images. It is difficult for the proposed method to identify an individual if shoeprint images for identification leaked out. Because the proposed method identifies gender without the faces, the type of dress and the hair types images. Therefore we can use safely the proposed method in public place. In addition, a sensor mat which we developed is reasonable to use mechanical switches arranged in a matrix pattern without pressure switches. We had shoeprint images with the sensor mat. We measure feature parameters from shoeprint images. The feature parameters are length, width and area of shoeprint. Utilizing the feature parameters, we identified, gender. In order to verify the gender identification rate of the proposed method, we set up the sensor mat at an entrance of buildings and took shoeprint images of 100 men and 100 women. As a result, we achieved about 86 percent of the gender identification rate.
• アクティブグリッドを用いたサッカー映像におけるパス可能領域の推定に関する検討

  高橋翔, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  2009
• 顕微鏡画像からのアポトーシス細胞抽出における誤検出低減に関する考察

  稲垣潤, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2009-  2009
• A WEB COMMUNITY-BASED VIDEO RETRIEVAL METHOD USING CANONICAL CORRELATION ANALYSIS

  Yasutaka Hatakeyama, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama  2009 16TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-6  805  -+  2009  [Not refereed][Not invited]
   

  This paper presents a Web community-based video retrieval method using canonical correlation analysis (CCA). In the proposed method, two novel approaches are introduced into the retrieval scheme of video materials on the Web. First, the CCA is applied to three kinds of video features, visual and audio features of video materials and textual features obtained from Web pages containing those video materials. This approach provides a solution of problems of traditional methods of not being able to calculate similarities between different kinds of video features. Furthermore, from the obtained similarities and link relationships of Web pages, a new adjacency matrix is defined, and link analysis can be applied to this matrix. Then, the Web communities of the video materials whose topics are similar to each other can be automatically extracted based on their features. Therefore, by ranking the video materials in the obtained Web community, accurate video retrieval can be realized.
• AN EFFECTIVE FLOW ESTIMATION METHOD WITH PARTICLE FILTER BASED ON HELMHOLTZ DECOMPOSITION THEOREM

  Norihiro Kakukou, Takahiro Ogawa, Miki Haseyaam  2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS  949  -952  2009  [Not refereed][Not invited]
   

  This paper proposes a novel flow estimation method with a particle filter based on a Helmholtz decomposition theorem. The proposed method extends a model of the Helmholtz decomposition theorem and enables the decomposition of flows into rotational, divergent, and translational components. From the extended model, the proposed method defines a state transition model and an observation model of the particle filter. Furthermore, the proposed method derives an observation density of the particle filter from an energy function based on the Helmholtz decomposition theorem. By utilizing these novel approaches, the proposed method provides a solution to the problem in the traditional ones of not being able to realize an effective flow estimation with the particle filter based on rotation, divergence, and translation, which are important geometric features. Consequently, the proposed method can accurately estimate the flows.
• ADAPTIVE RECONSTRUCTION METHOD OF MISSING TEXTURES BASED ON KERNEL CANONICAL CORRELATION ANALYSIS

  Takahiro Ogawa, Miki Haseyama  2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS  1165  -1168  2009  [Not refereed][Not invited]
   

  This paper presents an adaptive reconstruction method of missing textures based on kernel canonical correlation analysis (CCA). The proposed method calculates the correlation between two areas, which respectively correspond to a missing area and its neighbor area, from known parts within the target image and realizes the estimation of the missing textures. In order to obtain this correlation, the kernel CCA is applied to each set containing the same kind of textures, and the optimal result is selected for the target missing area. Specifically, a new approach monitoring errors caused in the above estimation process enables the selection of the optimal result. This approach provides a solution to the problem in traditional methods of not being able to perform adaptive reconstruction of the target textures due to the missing intensities. Experimental results show subjective and quantitative improvement of the proposed reconstruction technique over previously reported reconstruction techniques.
• Semantic image retrieval based on POCS algorithm using Kernel PCA and its performance verification

  Takahiro Ogawa, Miki Haseyama  Digest of Technical Papers - IEEE International Conference on Consumer Electronics  582  -583  2009  [Not refereed][Not invited]
   

  This paper presents a projection onto convex sets (POCS)-based semantic image retrieval method and its performance verification. The main contributions of the proposed method are twofold: introduction of nonlinear eigenspace of visual and semantic features into the constraint of the POCSbased semantic image retrieval algorithm and adaptive selection of the annotated images utilized for this algorithm. Then, by combining these two approaches, the semantic features of the query image are successfully estimated, and accurate image retrieval can be expected. Finally, relationship between the performance of the proposed method and the kinds of the kernel functions utilized for the kernel PCA is shown in this paper. ©2009 IEEE.
• Accurate graph-based scene segmentation using object matching and audio feature

  Makoto Yamamoto, Miki Haseyama  Digest of Technical Papers - IEEE International Conference on Consumer Electronics  639  -640  2009  [Not refereed][Not invited]
   

  A method for accurate scene segmentation utilizing two kinds of directed graph obtained by object matching and by using audio features is proposed. Generally, in audiovisual materials, there are repeated appearances of shots that include frames of the same background, object or place. It should be assumed that such shots are included in a single scene. In the proposed method, by performing object matching between two frames that are each included in different shots, multiple shots for which frames include the same object can be successfully found and their repeated appearances are represented as a directed graph. The proposed method also generates another directed graph that represents the repeated appearances of shots with similar audio features. By the combined use of these two graphs, accurate scene segmentation can be realized. The effectiveness of the proposed method is verified by applying this method to news programs and another broadcast program. ©2009 IEEE.
• A new image retrieval interface and its practical use in "view search Hokkaido"

  Miki Haseyama, Toshifumi Murata, Hisashi Ukawa  Digest of Technical Papers - IEEE International Conference on Consumer Electronics  851  -852  2009  [Not refereed][Not invited]
   

  New image retrieval technology, which is used for a service demonstration project "View Search Hokkaido" in "Information Grand Voyage Project" conducted by Ministry of Economy, Trade and Industry, Japan, is presented in this paper. The new technology enables image retrieval based on low-level features without utilizing any tag-based scheme and realizes a 3D interface for image retrieval. This 3D interface can provide new user experience, which has not been provided by the conventional retrieval services. The effectiveness of the 3D interface has been verified in "View Search Hokkaido". ©2009 IEEE.
• A Study on Multi-carrier CDMA with Adaptive Trellis Coded Modulation

  Hirokazu Tanaka, Shoichiro Yamasaki, Miki Haseyama  2009 INTERNATIONAL SYMPOSIUM ON INTELLIGENT SIGNAL PROCESSING AND COMMUNICATION SYSTEMS (ISPACS 2009)  477  -+  2009  [Not refereed][Not invited]
   

  A Generalized Symbol-rate-increased (GSRI) Pragmatic Adaptive Trellis Coded Modulation (ATCM) applied to a Multi-carrier CDMA system with bi-orthogonal keying is analyzed. In [7], an efficient ATCM scheme using GSRI-TCM has been proposed. However, since the modulation and coding selection is based on the ideal SNR estimation method, performance evaluations employing practical channel estimation scheme is desired. In this paper, we investigate the performance of the Multi-carrier CDMA with GSRI ATCM employing practical SNR estimation. The performance evaluations on a multi-path fading channel are presented.
• WARP-AEC: A Stereo Acoustic Echo Canceller based on W-Adaptive Filters for Rapid Projection

  Shigenobu Minami, Hirokazu Tanaka, Miki Haseyama  2009 INTERNATIONAL SYMPOSIUM ON INTELLIGENT SIGNAL PROCESSING AND COMMUNICATION SYSTEMS (ISPACS 2009)  299  -+  2009  [Not refereed][Not invited]
   

  It is difficult for stereo acoustic echo cancellers to converge toward real stereo echo path characteristics in the case of strong cross-channel correlation in right and left channel signals. To overcome this problem, we propose the WARP (W-Adaptive filters for Rapid Projection) method, which uses a monaural adaptive filter in parallel with a stereo adaptive filter. In this method, stereo echo path estimation is realized by projecting from two monaural estimated echo path characteristics at a far-end talker change. Stereo echo path estimation error, which is gradually decreased in conventional methods, is quickly improved at a far-end talker change by adopting the WARP method. This method has been evaluated by using 5 talker position's stereo signals with /without 10% level drift in right and left channels. The results showed that there was a 15-20dB improvement in estimation error at a far-end talker change compared with estimation errors in conventional methods, such as MC-NLMS and Affine projection method with p=3 in the case of WGN, and that there were more than 10 dB echo path estimation error improvement and 20-30 dB residual echo level improvement in the case of voice source.
• Error correction scheme for multi-carrier CDMA with bi-orthogonal modulation

  Hirokazu Tanaka, Shoichiro Yamasaki, Miki Haseyama  2008 International Symposium on Information Theory and its Applications, ISITA2008  1061  -+  2008/12/01  [Not refereed][Not invited]
   

  In order to reduce out-of-band emission and PAPR for Multi-carrier (MC) CDMA system, bi-orthogonal modulation is one of the efficient schemes. However the performance of the scheme rapidly degrades if Hadamard code selection errors are occurred at the receiver. We propose and examine a frequency domain error correction scheme for the system employing Reed-Solomon (RS) code. This basic concept is that a Forward Error Correction (FEC) code corrects the code selection errors for all the subcarriers. An efficient decoding method that not only code selection error but the transmitted bit error can correct is also examined.
• 誤差伝搬の抑制を目的としたカルマンフィルタによる動画像のエラーコンシールメント手法に関する検討

  TAKAHASHI SHIGEKI, OGAWA TAKAHIRO, TANAKA HIROKAZU, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  ROMBUNNO.0137  2008/10/25  [Not refereed][Not invited]
  
• 映像特徴量に注目したWeb映像検索手法の提案

  HATAKEYAMA TAIKI, OGAWA TAKAHIRO, ASAMIZU SATOSHI, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  ROMBUNNO.0126  2008/10/25  [Not refereed][Not invited]
  
• SIFT特徴に基づく投票による類似パターン抽出に関する考察

  KAKUNO KORYO, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  ROMBUNNO.0125  2008/10/25  [Not refereed][Not invited]
  
• 撮像物体の自動抽出を可能とするSVMの実現に関する一考察

  OKADA KANEFUMI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  ROMBUNNO.0136  2008/10/25  [Not refereed][Not invited]
  
• 動画像における歩行者識別のための処理対象領域の設定に関する一考察

  CHIN U, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  ROMBUNNO.0135  2008/10/25  [Not refereed][Not invited]
  
• Low Complexity Speaker Identification in AAC Domain

  AI Haojun, HASEYAMA Miki  ITE technical report  32-  (46)  31  -34  2008/10/23  

  This paper presents an implementation of a low-complexity speaker identification algorithm working in the compressed audio domain. The goal is to perform speaker modeling and identification without decoding the AAC bitstream to extract speaker dependent features, thus saving important system resource. The silence detection and MFCC parameters are calculated from MDCT coefficient other than from the FFT spectrum. Each speaker is modeled by a GMM, which is trained using the EM algorithm to refine the weight and the parameters of each component. The recognition accuracies of our algorithm reach 97% for ARCTIC database with 16% CPU overload comparing to the algorithms based on the analysis of the decoded PCM signals.
• A Note on Flow Estimation based on Particle Filter : Introduction of Helmholtz Decomposition Theorem into Observation Model

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  32-  (46)  19  -22  2008/10/23  [Not refereed][Not invited]
   

  This paper proposes an effective flow estimation method based on a Helmholtz decomposition theorem with a particle filter. In a traditional method based on the Helmholtz decomposition theorem, since it does not consider translation and estimation errors included in previous flows, its performance tends to become degraded. Therefore, the proposed method extends the Helmholtz decomposition theorem with the translation. By using the extended theorem, the flow estimation considering the translation can be realized. Furthermore, the proposed method regards the flows as state variables in the particle filter. Then, the proposed method defines a process that the flows of the pixels in the current frame are provided from the ones of pixels in the previous frame as a state transition model. The proposed method defines the following two processes as an observation model. One is that intensities including additive noise in the current frame are generated from the flows of the pixels in the current frame. The other one is that the flows of the pixels in the current frame are decomposed into rotational and divergent components including the additive noise. By using the particle filter based on these models, the proposed method can estimate the flows of the pixels in the current frame in such a way that the estimation errors included in the previous flows do not affect the flow estimation scheme. Consequently, an accurate flow estimation can be achieved.
• A note on Similarity Evaluation of Images in Web Pages Based on Link Analysis

  MURAYAMA Shota, HASEYAMA Miki  ITE technical report  32-  (46)  23  -26  2008/10/23  [Not refereed][Not invited]
   

  This paper proposes a new similarity evaluation method for images in Web pages based on link analysis. The proposed method assumes that there are areas which contain common contents and have similar image features in the images, whose Web pages are connected by links. Based on the assumption, the proposed method defines similarity measure of the images. Consequently, with the similarity measure, the problem in the traditional methods of being affected by the image features of different contents can be solved, and an effective image retrieval is realized.
• A note on spatio-temporal resolution enhancement of video sequences based on super-resolution reconstruction : Introduction of clustering scheme based on feature point correspondence

  TAKIZAWA Makoto, HASEYAMA Miki  ITE technical report  32-  (46)  27  -30  2008/10/23  [Not refereed][Not invited]
   

  This paper proposes a new spatio-temporal resolution enhancement method of video sequences based on super-resolution reconstruction. The proposed method uses obtained homography matrix's parameters based on feature points correspondence for calculating the pixel correspondence between successive frames. In this procedure, the proposed method clusters the feature points correspondence and calculates the homography matrix's parameters from the feature points correspondence in the largest cluster. By using the obtained parameters, camera parameters are calculated accurately. Furthermore, from the camera parameters, we estimates the pixel correspondence between the successive frames and the interpolated one. With the obtained motion vector, we apply the super-resolution reconstruction. Consequently, we can obtain an accurate high-resolution and high frame rate video sequence. Simulation results are shown to confirm the high performance of the proposed method.
• POCS-based annotation method using kernel PCA for semantic image retrieval

  Takahiro Ogawa, Miki Haseyama  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E91A-  (8)  1915  -1923  2008/08  [Not refereed][Not invited]
   

  A projection onto convex sets (POCS)-based annotation method for semantic image retrieval is presented in this paper. Utilizing database images previously annotated by keywords, the proposed method estimates unknown semantic features of a query image from its known visual features based on a POCS algorithm, which includes two novel approaches. First, the proposed method semantically assigns database images to some clusters and introduces a nonlinear eigenspace of visual and semantic features in each cluster into the constraint of the POCS algorithm. This approach accurately provides semantic features for each cluster by using its visual features in the least squares sense. Furthermore, the proposed method monitors the error converged by the POCS algorithm in order to select the optimal cluster including the query image. By introducing the above two approaches into the POCS algorithm, the unknown semantic features of the query image are successfully estimated from its known visual features. Consequently, similar images can be easily retrieved from the database based on the obtained semantic features. Experimental results verify the effectiveness of the proposed method for semantic image retrieval.
• Estimating Number of People from a Video Sequence via Geometrical Model

  ARAI Hiroyuki, MIYAGAWA Isao, KOIKE Hideki, HASEYAMA Miki  ITE technical report  32-  (26)  33  -36  2008/06/23  [Not refereed][Not invited]
   

  We propose a novel technique for estimating the number of people in a video sequence ; it has the advantages of being stable even in crowded situations and needing no ground-truth data. By analyzing the geometrical relationships between image pixels and their intersection volumes in the real world quantitatively, a foreground image can be directly indicate the number of people. Because foreground detection can be done even in crowded situations, the proposed method can be applied to such situations. Also it can estimate the number of people in an a-priori manner, so it needs no ground-truth data which is necessary for existing feature-based estimating techniques. Experiments show the validity of the proposed method.
• A study on frequency domain error correction scheme for MC-CDMA

  TANAKA Hirokazu, JANG Kyung Woon, YAMASAKI Shochiro, HASEYAMA Miki  IEICE technical report  108-  (85)  65  -70  2008/06/12  [Not refereed][Not invited]
   

  Recently, multi-code MC CDMA has a lot of attention in wireless multimedia communications because of its high bit rate performance. On the other hand, considering hardware implementation, it requires high power amplifier (HPA) with highly linear characteristics. HPA nonlinearity causes high out-of-band spurious emission and high peak to average power ration (PAPR). In order to solve these problems, employing bi-orthogonal modulation instead of multi-code modulation has been studied. As for the error correction coding of the MC-CDMA with bi-orthogonal modulation, the probability of the erroneous Hadamard code selection at receiver is critical for the BER performance compared to the sign bit error. In this paper, we propose an efficient error correcting scheme for the code selection error employing Reed-Solomon (RS) code. Additionally, an efficient decoding structure that can also correct the uncoded sign bit error using feed back RS decoded codeword is proposed. The basic performance of the proposed scheme was evaluated by both theoretical analysis and computer simulations.
• A Study on Frequency Domain Error Correction Scheme for MC-CDMA

  TANAKA Hirokazu, JANG KyungWoon, YAMASAKI Shochiro, HASEYAMA Miki  電子情報通信学会技術研究報告. SIS, スマートインフォメディアシステム  108-  (85)  65  -70  2008/06/05  [Not refereed][Not invited]
  
• A note on Error Concealment Using Kalman Filter : Realization of State Transition Model Representing Video Decoding Process

  TAKAHASHI Shigeki, OGAWA Takahiro, TANAKA Hirokazu, HASEYAMA Miki  IEICE technical report  107-  (487)  13  -18  2008/02/19  [Not refereed][Not invited]
   

  This paper proposes an accurate error concealment technique using a Kalman filter. In order to realize the error concealment, the proposed technique regards the intensities in each frames as elements of the state variable of the Kalman filter and designs the following state transition and observation models. First, the state transition model is designed by using motion vectors and represents the video decoding process. Further, the observation model is designed as a new degradation model and enables the Kalman gain to be calculated for the error concealment. Then, the compensation of the estimation results by the state transition becomes feasible. Consequently, by utilizing these new state transition and observation models, the accurate error concealment based on the rule of least squares can be realized. Experimental results show that the proposed technique has better performance than that of the traditional techniques.
• A note on video frame interpolation based on image morphing : An accurate frame segmentation used in adaptive correspondence estimation

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  107-  (487)  19  -24  2008/02/19  [Not refereed][Not invited]
   

  This paper presents an image morphing based interpolation method of video frames using an adaptive correspondence estimation. The proposed method estimates the pairs of points which represent the correspondence of image features between two successive frames. Based on these pairs of points, the interpolation frames are calculated by using the image morphing. In the estimation scheme of pairs of points, the proposed method segments each frame into background and foreground regions. Furthermore, in order to estimate the foreground and background regions more accurately, the verification of the segmentation results is performed by using the previous interpolation frame and its segmentation results. Then, in each region, the pairs of points which represent the correspondence of the image features between the two successive frames is estimated adaptively. Furthermore, in each region, the interpolation frames are calculated by using the appropriate scheme, and the intensities of the interpolation frames are obtained accurately around boundaries between foreground and background regions. Consequently, the proposed method can generate interpolation frames successfully.
• A note on adaptive interpolation filter design method based on clustering using image features

  MATSUNO Takaya, KONDO Satoshi, HASEYAMA Miki  IEICE technical report  107-  (487)  25  -28  2008/02/19  [Not refereed][Not invited]
   

  This paper presents an adaptive interpolation filter design method based on classification using image features. The proposed method classifies macroblocks into some clusters representing the interpolation directions based on their motion vectors. For each cluster, the optimal interpolation filters are designed in the least-squares sense. Introducing the classification of the macroblocks into the filter design scheme, our method considers the difference between the horizontal and vertical correlation and can design the optimal filters. Thus, by using these filters for motion compensated prediction, the total bit rate with the same PSNR value is reduced compared to the conventional methods.
• A note on improvement of human detection based on multiple detectors

  KAGA Yosuke, HASEYAMA Miki  IEICE technical report  107-  (487)  51  -54  2008/02/19  [Not refereed][Not invited]
   

  This paper presents a human detection method based on multiple detectors. In the proposed method, a human detection result is obtained by voting the results of a pedestrian detector and a face detector in 2-dimensional plane. Then, by integrating the multiple detection results of the different detectors into a single human detection result, the accurate human detection can be realized compared to each single detector. Experimental results performed by applying the proposed method to actual images are shown to verify its effectiveness.
• A note on player tracking by using level-set method based on color component in soccer video

  LIM Wonkuk, HASEYAMA Miki  IEICE technical report  107-  (487)  67  -70  2008/02/19  [Not refereed][Not invited]
   

  This paper proposes an accurate player tracking method in soccer video using a level-set method. The proposed method introduces features of baundary surfaces and those of color components, which compose the player's uniforms, into the evaluation function of the level-set method. Then, our method realizes the accurate tracking, that can distinguish the players in different teams.
• A note on Error Concealment Using Kalman Filter : Realization of State Transition Model Representing Video Decoding Process

  TAKAHASHI Shigeki, OGAWA Takahiro, TANAKA Hirokazu, HASEYAMA Miki  IEICE technical report  107-  (489)  13  -18  2008/02/19  [Not refereed][Not invited]
   

  This paper proposes an accurate error concealment technique using a Kalman filter. In order to realize the error concealment, the proposed technique regards the intensities in each frames as elements of the state variable of the Kalman filter and designs the following state transition and observation models. First, the state transition model is designed by using motion vectors and represents the video decoding process. Further, the observation model is designed as a new degradation model and enables the Kalman gain to be calculated for the error concealment. Then, the compensation of the estimation results by the state transition becomes feasible. Consequently, by utilizing these new state transition and observation models, the accurate error concealment based on the rule of least squares can be realized. Experimental results show that the proposed technique has better performance than that of the traditional techniques.
• A note on video frame interpolation based on image morphing : An accurate frame segmentation used in adaptive correspondence estimation

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  107-  (489)  19  -24  2008/02/19  [Not refereed][Not invited]
   

  This paper presents an image morphing based interpolation method of video frames using an adaptive correspondence estimation. The proposed method estimates the pairs of points which represent the correspondence of image features between two successive frames. Based on these pairs of points, the interpolation frames are calculated by using the image morphing. In the estimation scheme of pairs of points, the proposed method segments each frame into background and foreground regions. Furthermore, in order to estimate the foreground and background regions more accurately, the verification of the segmentation results is performed by using the previous interpolation frame and its segmentation results. Then, in each region, the pairs of points which represent the correspondence of the image features between the two successive frames is estimated adaptively. Furthermore, in each region, the interpolation frames are calculated by using the appropriate scheme, and the intensities of the interpolation frames are obtained accurately around boundaries between foreground and background regions. Consequently, the proposed method can generate interpolation frames successfully.
• A note on adaptive interpolation filter design method based on clustering using image features

  MATSUNO Takaya, KONDO Satoshi, HASEYAMA Miki  IEICE technical report  107-  (489)  25  -28  2008/02/19  [Not refereed][Not invited]
   

  This paper presents an adaptive interpolation filter design method based on classification using image features. The proposed method classifies macroblocks into some clusters representing the interpolation directions based on their motion vectors. For each cluster, the optimal interpolation filters are designed in the least-squares sense. Introducing the classification of the macroblocks into the filter design scheme, our method considers the difference between the horizontal and vertical correlation and can design the optimal filters. Thus, by using these filters for motion compensated prediction, the total bit rate with the same PSNR value is reduced compared to the conventional methods.
• A note on improvement of human detection based on multiple detectors

  KAGA Yosuke, HASEYAMA Miki  IEICE technical report  107-  (489)  51  -54  2008/02/19  [Not refereed][Not invited]
   

  This paper presents a human detection method based on multiple detectors. In the proposed method, a human detection result is obtained by voting the results of a pedestrian detector and a face detector in 2-dimensional plane. Then, by integrating the multiple detection results of the different detectors into a single human detection result, the accurate human detection can be realized compared to each single detector. Experimental results performed by applying the proposed method to actual images are shown to verify its effectiveness.
• A note on player tracking by using level-set method based on color component in soccer video

  LIM Wonkuk, HASEYAMA Miki  IEICE technical report  107-  (489)  67  -70  2008/02/19  [Not refereed][Not invited]
   

  This paper proposes an accurate player tracking method in soccer video using a level-set method. The proposed method introduces features of baundary surfaces and those of color components, which compose the player's uniforms, into the evaluation function of the level-set method. Then, our method realizes the accurate tracking, that can distinguish the players in different teams.
• A note on music extraction based on similarity of melody line

  KOBAYASHI Kazuya, HASEYAMA Miki  IEICE technical report  107-  (486)  49  -53  2008/02/18  [Not refereed][Not invited]
   

  This paper verifies performance of a similar music extraction using the similarity measure computation method, which we have already proposed. This similarity measure is computed by comparing the melody lines of two different music pieces. In order to compare the melody lines between two music pieces, our method adopts Dynamic Time Warping (DTW). According to this similarity measure, the accurate extraction of similar music pieces can be expected. However, since this scheme had not been applied to a large music database yet, it was not sufficient to verify its performance for actual use. Therefore, this paper applies the proposed method to the larger database containing 284 music pieces and evaluates its extraction performance. Furthermore, some comparisons between our method and the traditional method are performed to evaluates its extraction performance.
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval

  MICHIYAMA Daigo, HASEYAMA Miki  IEICE technical report  107-  (486)  55  -59  2008/02/18  [Not refereed][Not invited]
   

  This paper proposes a method to calculate scene dissimilarities utilizing audio and visual signals for video retrieval. By utilizing Dynamic Time Warping, the proposed method calculates the scene dissimilarities by selecting the most similar features, which are selected from visual, audio, and their integrated features, and these selectings are perfomed between two shots in scenes. Furthermore, a new feature space which can match audio and visual features defined by utilizing canonical correlation analysis is introduced into these selectings. In conclusion, the proposed method can appropriately calculate the scene dissimilarities based on similar features between scenes. Experimental results performed by applying the proposed method to scene retrieval in actual video data are shown to verify its effectiveness.
• A note on scene segmentation based on video structure : An approach by combination of audio and visual processing

  YAMAMOTO Makoto, HASEYAMA Miki  IEICE technical report  107-  (486)  61  -66  2008/02/18  [Not refereed][Not invited]
   

  This paper proposes video scene segmentation method utilizing shot-based repetitive structures of video and audio signals. In the proposed method, two repetitive structures are respectively obtained by clustering shots based on their video and audio features. Specifically, these obtained structures are represented as directed graphs whose nodes are clusters containing similar shots and edges are formed based on time transition of shots. Thus, by calculating common cut edges in both of the two constructed graphs, video materials are effectively segmented into scenes from video and audio signals. Then, the proposed method can avoid over scene segmentation caused in the conventional method using only video signals.
• A note on semantic feature estimation method of still images using kernel PCA : An approach for similar image retrieval based on semantic concepts

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  107-  (486)  67  -72  2008/02/18  [Not refereed][Not invited]
   

  This paper presents a semantic feature estimation method using kernel PCA in order to realize semantic image retrieval. The proposed method performs keyword-based clustering of database images and derives a new map by using a nonlinear eigenspace of visual and semantic features in each cluster. This map provides accurate semantic features for the images belonging to each cluster from their visual features. Therefore, the proposed method adaptively selects the optimal cluster including a query image, whose keywords are unknown, by monitoring the errors of the visual features caused in its semantic feature estimation process. Then, since the semantic features of the query image can be successfully estimated, semantic image retrieval for the query image can be easily achieved.
• A note on music extraction based on similarity of melody line

  KOBAYASHI Kazuya, HASEYAMA Miki  IEICE technical report  107-  (488)  49  -53  2008/02/18  [Not refereed][Not invited]
   

  This paper verifies performance of a similar music extraction using the similarity measure computation method, which we have already proposed. This similarity measure is computed by comparing the melody lines of two different music pieces. In order to compare the melody lines between two music pieces, our method adopts Dynamic Time Warping (DTW). According to this similarity measure, the accurate extraction of similar music pieces can be expected. However, since this scheme had not been applied to a large music database yet, it was not sufficient to verify its performance for actual use. Therefore, this paper applies the proposed method to the larger database containing 284 music pieces and evaluates its extraction performance. Furthermore, some comparisons between our method and the traditional method are performed to evaluates its extraction performance.
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval

  MICHIYAMA Daigo, HASEYAMA Miki  IEICE technical report  107-  (488)  55  -59  2008/02/18  [Not refereed][Not invited]
   

  This paper proposes a method to calculate scene dissimilarities utilizing audio and visual signals for video retrieval. By utilizing Dynamic Time Warping, the proposed method calculates the scene dissimilarities by selecting the most similar features, which are selected from visual, audio, and their integrated features, and these selectings are perfomed between two shots in scenes. Furthermore, a new feature space which can match audio and visual features defined by utilizing canonical correlation analysis is introduced into these selectings. In conclusion, the proposed method can appropriately calculate the scene dissimilarities based on similar features between scenes. Experimental results performed by applying the proposed method to scene retrieval in actual video data are shown to verify its effectiveness.
• A note on scene segmentation based on video structure : An approach by combination of audio and visual processing

  YAMAMOTO Makoto, HASEYAMA Miki  IEICE technical report  107-  (488)  61  -66  2008/02/18  [Not refereed][Not invited]
   

  This paper proposes video scene segmentation method utilizing shot-based repetitive structures of video and audio signals. In the proposed method, two repetitive structures are respectively obtained by clustering shots based on their video and audio features. Specifically, these obtained structures are represented as directed graphs whose nodes are clusters containing similar shots and edges are formed based on time transition of shots. Thus, by calculating common cut edges in both of the two constructed graphs, video materials are effectively segmented into scenes from video and audio signals. Then, the proposed method can avoid over scene segmentation caused in the conventional method using only video signals.
• A note on semantic feature estimation method of still images using kernel PCA : An approach for similar image retrieval based on semantic concepts

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  107-  (488)  67  -72  2008/02/18  [Not refereed][Not invited]
   

  This paper presents a semantic feature estimation method using kernel PCA in order to realize semantic image retrieval. The proposed method performs keyword-based clustering of database images and derives a new map by using a nonlinear eigenspace of visual and semantic features in each cluster. This map provides accurate semantic features for the images belonging to each cluster from their visual features. Therefore, the proposed method adaptively selects the optimal cluster including a query image, whose keywords are unknown, by monitoring the errors of the visual features caused in its semantic feature estimation process. Then, since the semantic features of the query image can be successfully estimated, semantic image retrieval for the query image can be easily achieved.
• A note on music extraction based on similarity of melody line

  小林 和也, 長谷山 美紀  ITE technical report  32-  (9)  49  -53  2008/02  [Not refereed][Not invited]
  
• A note on definition of scene dissimilarities utilizing audio and visual signals for video retrieval

  道山 大悟, 長谷山 美紀  ITE technical report  32-  (9)  55  -59  2008/02  [Not refereed][Not invited]
  
• A note on scene segmentation based on video structure: an approach by combination of audio and visual processing

  山本 誠, 長谷山 美紀  ITE technical report  32-  (9)  61  -66  2008/02  [Not refereed][Not invited]
  
• A note on semantic feature estimation method of still images using kernel PCA: an approach for similar image retrieval based on semantic concepts

  小川 貴弘, 長谷山 美紀  ITE technical report  32-  (9)  67  -72  2008/02  [Not refereed][Not invited]
  
• A note on video frame interpolation based on image morphing: an accurate frame segmentation used in adaptive correspondence estimation

  山本 貴史, 長谷山 美紀  ITE technical report  32-  (10)  19  -24  2008/02  [Not refereed][Not invited]
  
• A note on adaptive interpolation filter design method based on clustering using image features

  松野 孝也, 近藤 敏志, 長谷山 美紀  ITE technical report  32-  (10)  25  -28  2008/02  [Not refereed][Not invited]
  
• A note on improvement of human detection based on multiple detectors

  加賀 陽介, 長谷山 美紀  ITE technical report  32-  (10)  51  -54  2008/02  [Not refereed][Not invited]
  
• A note on player tracking by using level-set method based on color component in soccer video

  林 原局, 長谷山 美紀  ITE technical report  32-  (10)  67  -70  2008/02  [Not refereed][Not invited]
  
• QUALITY ESTIMATION OF FRACTAL CODED IMAGES BASED ON IMAGE FEATURES

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  IEICE technical report  107-  (411)  289  -293  2008/01/07  

  Fractal image coding is one of the coding techniques used for digital images. It can provide the decoded images with a higher quality than the JPEG compression at ultra low bit-rates. However, we cannot adopt it practically. One of the reasons is that the decoded images are sometimes low quality. Moreover, what is more inconvenient is that we cannot know the decoded image quality of a given image unless we actually encode and decode it which takes a lot of time. Therefore, in this paper, we have investigated the relationship between the quality of the decoded image and various features of the image. By finding the important features which relate to the image quality, we can estimate the decoded image quality in a short amount of time without actually encoding the image.
• 次世代映像検索を考える-Cyber Space Navigator-

  長谷山美紀  映情学平成19年度メディア工学シンポジウム (Mar.)  2008
• 映像中に撮像された物体の幾何学的特徴に着眼したシーン分割手法の高精度化に関する一検討

  山本誠, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  2008
• モーフィングを用いたフレーム補間の改良-特徴点の高精度な対応付けによる検討-

  瀧澤誠, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  2008
• ファジィクラスタリングを用いた音響シーン分割の高精度化に関する考察-音響シーンに対する新たな特徴量の定義-

  成田容, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  2008
• 色ヒストグラムを用いた類似画像分類の高速化に関する検討

  徳本了太, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2008-  2008
• Media distribution and processing

  Kazuto Kamikura, Miki Haseyama, Kazuhito Murakami  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  62-  (8)  1251  -1254  2008  [Not refereed][Not invited]
  
• A Kalman filter based restoration method for in-vehicle camera images in foggy conditions

  Tomoki Hiramatsu, Takahiro Ogawa, Miki Haseyama  2008 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, VOLS 1-12  1245  -1248  2008  [Not refereed][Not invited]
   

  This paper proposes a Kalman filter based restoration method for images obtained by in-vehicle camera in foggy conditions. The proposed method introduces two novel approaches into the Kalman filter based restoration. The first one is an automatic determination of a fog deterioration model. A vanishing point in the foggy image is estimated by using cross ratio of lane marking, and automatic determination of all parameters of the fog deterioration model is realized. Furthermore, the obtained model is introduced into the Kalman filter. Specifically, our method regards each frame as a state variable and its observation model is defined by the fog deterioration model. Then, since the correlation between successive frame can be effectively utilized by the Kalman filter, the accurate restoration of foggy images is achieved. Experimental results show that the proposed method achieves higher performance than the traditional method based on the fog deterioration model.
• Kernel PCA-based resolution enhancement approach of still images using different levels of pyramid structure

  Takahiro Ogawa, Miki Haseyama  2008 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, VOLS 1-12  1293  -1296  2008  [Not refereed][Not invited]
   

  This paper presents a kernel PCA-based adaptive resolution enhancement method of still images. The proposed method introduces two novel approaches into the kernel PCA-based reconstruction of high-frequency components missed from a high-resolution (HR) image. First, since local images between two different resolution levels of a pyramid structure are similar to each other, nonlinear eigenspaces of local images in the target low-resolution (LR) image are utilized as those of local images in the HR image. Further, in the kernel PCA-based reconstruction process of the high-frequency components, our method monitors errors caused in the known low-frequency components and realizes the selection of the optimal eigenspace. Then, since the missing high-frequency components can be adaptively estimated, the accurate HR image can be obtained.
• Two-phased region integration approach for effective pedestrian detection in low contrast images

  Miki Haseyama, Yosuke Kaga  2008 DIGEST OF TECHNICAL PAPERS INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS  384  -385  2008  [Not refereed][Not invited]
   

  An accurate pedestrian detection method is proposed, which is effective especially in low-contrast images. The pedestrian detection method generally consists of two serial functions: One is moving object extraction, and the other is pedestrian decision. In particular, the moving objects need to be extracted precisely for the accurate pedestrian detection. However, since the moving objects are extracted as multiple small regions when the image contrast is low, the pedestrian detection often fails. To overcome this problem, the proposed method integrates the separated moving objects by two-phased region integration. In the first phase, the pieces of the moving objects are integrated by applying the watershed algorithm. Then, these regions are, in the second phase, further integrated by watching the outputs of support vector machines for the pedestrian detection. By using these integration schemes, the moving objects are appropriately extracted, and thereby the accurate pedestrian detection can be realized.
• A KALMAN FILTER-BASED APPROACH FOR ADAPTIVE RESTORATION OF IN-VEHICLE CAMERA FOGGY IMAGES

  Tomoki Hiramatsu, Takahiro Ogawa, Miki Haseyama  2008 15TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-5  3160  -3163  2008  [Not refereed][Not invited]
   

  In this paper, a Kalman filter-based approach for adaptive restoration of video images acquired by an in-vehicle camera in foggy conditions is proposed. In order to realize Kalman filter-based restoration, the proposed method regards the intensities in each frame as elements of the state variable of the Kalman filter and designs the following two models for restoration of foggy images. The first one is an observation model, which represents a fog deterioration model. The second one is a non-linear state transition model, which represents the target frame in the original video image from its previous frame based on motion vectors. By utilizing the observation and state transition models, the correlation between successive frames can be effectively utilized for restoration. Further, the proposed method introduces a new estimation scheme of the parameter, which determines the deterioration characteristic in foggy conditions, into the Kalman filter algorithm. Consequently, since automatic determination of the fog deterioration model, which specifies the observation model, from only the foggy images is realized, the accurate restoration can be achieved. Experimental results show that the proposed method has better performance than that of the traditional method based on the fog deterioration model.
• Performance Improvement of CMA Based FM Multipath Cancelling System in Automobile Application

  Masaichi Akiho, Tsuguhide Aoki, Kiyonori Takahashi, Miki Haseyama  ICSP: 2008 9TH INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING, VOLS 1-5, PROCEEDINGS  2670  -2673  2008  [Not refereed][Not invited]
   

  Constant Modulus Algorithm (CMA) is widely applied to improve distorted transmitting radio signal at the receiver. FM radio is still very popular equipment for in-car entertainment. Therefore, correcting FM multi-path distortion on conventional FM radio system is one of the greatly expected applications of CMA technique. In this paper, we discuss performance and improvement of the CMA based FM multi-path cancelling system in automobile application by demonstrating influence of impulse noise, which is very common in vehicle environment. Impulse noise, which is received at antenna, will not be corrected at all before it goes into CMA based multi-path canceling system as the input. the impulse noise on CMA input will make the system unstable so easily. Proposing limit error signal CAM, we can successfully achieve to avoid influence of the impulse noise.
• KERNEL PCA-BASED SEMANTIC FEATURE ESTIMATION APPROACH FOR SIMILAR IMAGE RETRIEVAL

  Takahiro Ogawa, Miki Haseyama  2008 15TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-5  965  -968  2008  [Not refereed][Not invited]
   

  A kernel PCA-based semantic feature estimation approach for similar image retrieval is presented in this paper. Utilizing database images previously annotated by keywords, the proposed method estimates unknown semantic features of a query image. First, our method performs semantic clustering of the database images and derives a new map from a nonlinear eigenspace of visual and semantic features in each cluster. This map accurately provides the semantic features for the images belonging to each cluster by using their visual features. Further, in order to select the optimal cluster including the query image, the proposed method monitors errors of the visual features caused by the semantic feature estimation process. Then, even if any semantics of the query image are unknown, its semantic features are successfully estimated by the optimal cluster. Experimental results verify the effectiveness of the proposed method for semantic image retrieval.
• <bold>Estimating the Number of People in a Video Sequence via Geometrical Model</bold>

  Hiroyuki Arai, Isao Miyagawa, Hideki Koike, Miki Haseyama  19TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOLS 1-6  3823  -3826  2008  [Not refereed][Not invited]
   

  We propose a novel technique for estimating the number of people in a video sequence; it has the advantages of being stable even in crowded situations and needing no ground-truth data. By analyzing the geometrical relationships between image pixels and their intersection volumes in the real world quantitatively a foreground image can be directly indicate the number of people. Because foreground detection can be done even in crowded situations, the proposed method can be applied to such situations. Also it can estimate the number of people in an a-priori manner so it needs no ground-truth data which is necessary for existing feature-based estimating techniques. Experiments show the validity of the proposed method.
• AN EFFECTIVE DETECTION METHOD OF ROTATIONAL AND DIVERGENT STRUCTURES IN STILL IMAGES BASED ON HELMHOLTZ DECOMPOSITION

  Norihiro Kakukou, Takahiro Ogawa, Miki Haseyama  2008 15TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-5  2336  -2339  2008  [Not refereed][Not invited]
   

  This paper proposes a novel detection method of rotational and divergent structures in still images based on Helmholtz decomposition. These structures are mathematical features in vector analysis. Traditionally, some detection methods of these structures in image sequences have been proposed. By using the Helmholtz decomposition, which can decompose flows into rotational and divergent components, the traditional methods can detect these structures in image sequences. However, the rotational and divergent structures in still images cannot be detected with the traditional methods. Therefore, the proposed method introduces a new criterion into the traditional schemes in order to realize the detection of the rotational and divergent structures in still images. This criterion is derived from two properties based on relation between still images and the flows, which are composed of the rotational and divergent components. Consequently, the detection of the rotational and divergent structures in still images can be achieved.
• Image recognition method using intensity gradient vectors

  Masao Hiramoto, Takahiro Ogawa, Miki Haseyama  Systems and Computers in Japan  38-  (13)  15  -27  2007/11/30  [Not refereed][Not invited]
   

  This paper proposes a method for general image recognition based on the progress of increasing the pixels of image sensors and improving image quality. This method can also adapt to images which have undergone geometric transformations such as the rotation and movement of images. The proposed method uses a voting system that utilizes vectors. This method uses vectors which represent intensity gradients and vectors which show position to express images and also defines voting vectors and similarity for the recognition. In addition, the proposed method has characteristics of concentrating voting locations at an origin point if there are identical images such that the voting results do not influence geometric transformations. When we performed experiments on natural images including original images which may have undergone image processing such as Gaussian or median filtering and JPEG compression, we understood that distinct differences appeared in the similarities and that recognition was possible even if an artificial process was added to the images. Even further, when we examined recognition of images using the greatest number of voting points as an application of this method, we were able to show that the recognition capability was high and that a partial image contained in one image could also be recognized. © 2007 Wiley Periodicals, Inc.
• Finding Representative of Cluster at Invariant Location by Discretized Clustering Method Dynamic Coalescence Model

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo  IEICE technical report  107-  (360)  1  -6  2007/11/22  [Not refereed][Not invited]
   

  We propose a discreted clustering method that founds representative of cluster found by the clustering method Dynamic Coalescence Model (DCM) at invariant location. In our proposed method, points are expressed by particles, and they are moved by their acceleration calculated by its gravity force from other particles. So, representative of cluster is founded near the centre of gravity of the cluster. Availability of our method to projection, region partition, characters recognition, etc. is pointed out.
• An Improved Third-order Discrimination System for Two-dimensional Point Distribution Using Its Topology and Similar Distribution Estimated Statistically

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo  IEICE technical report  107-  (360)  7  -12  2007/11/22  [Not refereed][Not invited]
   

  We propose an improved third-order system for the discrimination of two-dimensional point distributions. This has training pattern creator to transpose a training pattern into its representative distribution that expresses relative location invariance feature of the pattern. Its feature extractor uses second-order feature that expresses topology of combinations of two points without redundancy. This transposes a training pattern into its feature vectors that express its similar patterns estimated statistically and its statistical feature. Discrimination accuracy by its discriminator of single-layered perceptron is high even if there is only one training pattern.
• Finding Representative of Cluster at Invariant Location by Discretized Clustering Method Dynamic Coalescence Model

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo  IEICE technical report  107-  (359)  1  -6  2007/11/22  [Not refereed][Not invited]
   

  We propose a discreted clustering method that founds representative of cluster found by the clustering method Dynamic Coalescence Model (DCM) at invariant location. In our proposed method, points are expressed by particles, and they are moved by their acceleration calculated by its gravity force from other particles. So, representative of cluster is founded near the centre of gravity of the cluster. Availability of our method to projection, region partition, characters recognition, etc. is pointed out.
• An Improved Third-order Discrimination System for Two-dimensional Point Distribution Using Its Topology and Similar Distribution Estimated Statistically

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo  IEICE technical report  107-  (359)  7  -12  2007/11/22  [Not refereed][Not invited]
   

  We propose an improved third-order system for the discrimination of two-dimensional point distributions. This has training pattern creator to transpose a training pattern into its representative distribution that expresses relative location invariance feature of the pattern. Its feature extractor uses second-order feature that expresses topology of combinations of two points without redundancy. This transposes a training pattern into its feature vectors that express its similar patterns estimated statistically and its statistical feature. Discrimination accuracy by its discriminator of single-layered perceptron is high even if there is only one training pattern.
• 霧の劣化特性を考慮した車載カメラ画像の高精度な復元に関する一考察

  平松朋樹, 小川貴弘, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2007-  ROMBUNNO.206  2007/10/27  [Not refereed][Not invited]
  
• A note on detection of rotational and divergent structure in still image based on Helmoltz decomposition

  覚幸 典弘, 小川 貴弘, 長谷山 美紀  ITE technical report  31-  (47)  1  -4  2007/10  [Not refereed][Not invited]
  
• A note on adaptive resolution enhancement of still images using Kernel PCA

  小川 貴弘, 長谷山 美紀  ITE technical report  31-  (47)  5  -8  2007/10  [Not refereed][Not invited]
  
• Removal of impulse noise with iterated function system: an approach for improvement using regional division of a block

  西村 浩哉, 小川 貴弘, 長谷山 美紀  ITE technical report  31-  (47)  9  -12  2007/10  [Not refereed][Not invited]
  
• A note on coding of coefficients in adaptive interpolation filter based on vector quantization

  松野 孝也, 近藤 敏志, 長谷山 美紀  ITE technical report  31-  (47)  19  -22  2007/10  [Not refereed][Not invited]
  
• A note on spatial and temporal resolution enhancement of video sequences using super-resolution reconstruction

  瀧澤 誠, 長谷山 美紀  ITE technical report  31-  (47)  23  -26  2007/10  [Not refereed][Not invited]
  
• Development of shoeprint images acquisition system for a gender determination

  浅水 仁, 長谷山 美紀  ITE technical report  31-  (47)  51  -54  2007/10  [Not refereed][Not invited]
  
• POCS-Based texture reconstruction method using clustering scheme by kernel PCA

  Takahiro Ogawa, Milki Haseyama  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E90A-  (8)  1519  -1527  2007/08  [Not refereed][Not invited]
   

  A new framework for reconstruction of missing textures in digital images is introduced in this paper. The framework is based on a projection onto convex sets (POCS) algorithm including a novel constraint. In the proposed method, a nonlinear eigenspace of each cluster obtained by classification of known textures within the target image is applied to the constraint. The main advantage of this approach is that the eigenspace can approximate the textures classified into the same cluster in the least-squares sense. Furthermore, by monitoring the errors converged by the POCS algorithm, a selection of the optimal cluster to reconstruct the target texture including missing intensities can be achieved. This POCS-based approach provides a solution to the problem in traditional methods of not being able to perform the selection of the optimal cluster due to the missing intensities within the target texture. Consequently, all of the missing textures are successfully reconstructed by the selected cluster's eigenspaces which correctly approximate the same kinds of textures. Experimental results show subjective and quantitative improvement of the proposed reconstruction technique over previously reported reconstruction techniques.
• A note on accurate restoration of in-vehicle camera images in foggy conditions Application of fog deterioration model to Kalman filter

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  107-  (100)  1  -6  2007/06/21  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• A note on accurate restoration of in-vehicle camera images in foggy conditions : Application of fog deterioration model to Kalman filter

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  107-  (104)  1  -6  2007/06/14  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• A note on accurate restoration of in-vehicle camera images in foggy conditions : Application of fog deterioration model to Kalman filter

  HIRAMATSU Tomoki, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  107-  (102)  1  -6  2007/06/14  [Not refereed][Not invited]
   

  This paper proposes an accurate restoration method of in-vehicle camera images in foggy conditions based on Kalman filter. Utilizing a cross ratio, the proposed method estimates a vanishing point from the target images, and calculates parameters of a fog deterioration model. Further, this model is introduced into the observation model of the Kalman filter. In addition, a new model, which estimates intensities between successive frames by using motion vector and compensate their errors, is also introduced into the state machine model. Consequently, since the proposed method achieves the restoration using the correlation between the successive frames, the accurate performance can be realized.
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  106-  (535)  1  -4  2007/02/23  [Not refereed][Not invited]
   

  This paper proposes a new method for video frame interpolation. In the proposed method, the video frames are interpolated by image morphing. For accurate interpolation using the image morphing, it is effective to consider the characteristic of the objects in the frames. Thus, the proposed method introduces a new scheme that estimates the correspondence between consecutive frames based on the feature points extracted every frame. By using this scheme, the accurate video frame interpolation can be realized. Finally, simulation results are shown to demonstrate the validity of the proposed method.
• A Study on Precise Player Localization in a Basketball Video

  SAKUMA Daisuke, HASEYAMA Miki  IEICE technical report  106-  (535)  17  -22  2007/02/23  [Not refereed][Not invited]
   

  This paper proposes a new method for precise player localization in basketball video images. In the proposed method, the positions of the players on a basketball court are estimated by a Mean-Shift method that tracks moving objects using color histograms. Further, the proposed method introduces a new scheme, which identifies the players who are affected by occlusions, into the Mean-Shift method. Then, our method searches these players in the subsequent frames by using the color histograms that are obtained from their regions in the previous frames. Consequently, the robust player tracking can be realized, so that our method can estimate the positions of the players precisely. In this paper, experimental results are shown to verify the high performance of the proposed method.
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  106-  (537)  1  -4  2007/02/16  [Not refereed][Not invited]
   

  This paper proposes a new method for video frame interpolation. In the proposed method, the video frames are interpolated by image morphing. For accurate interpolation using the image morphing, it is effective to consider the characteristic of the objects in the frames. Thus, the proposed method introduces a new scheme that estimates the correspondence between consecutive frames based on the feature points extracted every frame. By using this scheme, the accurate video frame interpolation can be realized. Finally, simulation results are shown to demonstrate the validity of the proposed method.
• A Study on Precise Player Localization in a Basketball Video

  SAKUMA Daisuke, HASEYAMA Miki  IEICE technical report  106-  (537)  17  -22  2007/02/16  [Not refereed][Not invited]
   

  This paper proposes a new method for precise player localization in basketball video images. In the proposed method, the positions of the players on a basketball court are estimated by a Mean-Shift method that tracks moving objects using color histograms. Further, the proposed method introduces a new scheme, which identifies the players who are affected by occlusions, into the Mean-Shift method. Then, our method searches these players in the subsequent frames by using the color histograms that are obtained from their regions in the previous frames. Consequently, the robust player tracking can be realized, so that our method can estimate the positions of the players precisely. In this paper, experimental results are shown to verify the high performance of the proposed method.
• A Note on Improvement of Similar Image Clustering Method using Neocognitron : Introduction of Novel Structure for Extracting Color Features

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (534)  1  -6  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a new image clustering method. We have proposed an image clustering method using a neocognitron which is one of the neural networks. In this paper, we use the edge features of the images as the input stimulus of the neocognitron. However, since this method notes only the edge features of the images, misclassifications are inevitable by changes of the target object's direction. Therefore, this paper proposes a new image clustering method based on the edge and color features. The color histogram that is one of the color features is robust to these problems. Therefore, the proposed method classifies the images based on the color histogram by using the self-organizing map. Further, the structure of the neocognitron is adaptively set from the obtained clustering result to perform the classification based on edges. Therefore, since both the color and edge features of the images can be utilized, our method accurately classifies the images. Some experimental results show the proposed method achieves high classification performance.
• A Study on Accurate Image Clustering Using Color Edges

  WATANABE Takashi, HASEYAMA Miki  IEICE technical report  106-  (534)  7  -10  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a similar image clustering method. We have proposed a image clustering scheme using image structures and quadratic form distances between color histograms. However, since this scheme cannot effectively utilize the distribution of colors within the target images, misclassification of these images is inevitable. In order to utilize these features more effectively, the proposed method classifies the images by the K-means algorithm based on distances between color correlograms, which can be utilized for describing the relationship of colors to spatial distances. Furthermore, by reconstructing clusters based on colors at the edges and their neighborhoods, the accurate image clustering can be realized.
• A Study on Group Tactics Analysis in Soccer Videos

  KON Hirofumi, HASEYAMA Miki  IEICE technical report  106-  (534)  121  -124  2007/02/15  [Not refereed][Not invited]
   

  This paper presents a method for group tactics analysis in soccer videos. For this analysis, the proposed method classifies players into some groups. As a result of this classification, the players in one group are closely related in terms of soccer tactics. In the proposed method, the relations between the players in the groups are analyzed by using player positions. Based on these relations, the proposed method estimates group advantage, which represents the degree of team's superiority or inferiority in each group. The group advantage helps viewers to understand game situations.
• A Note on the similarity between music audio signals using the melody line and the rhythm

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki  IEICE technical report  106-  (534)  125  -128  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a novel similarity between music audio signals using rhythmic contents. The proposed method estimates the rhythmic content utilizing the autocorrelation functions of the signal. Then, by applying DTW to the functions, the similarity between music audio signals can be appropriately computed. Furthermore, our method visualizes the similarity using both melody line and rhythm to try to easily retrieve the similar music audio signals.
• An Accurate IFS-based Image Enlargement Method using Edge Continuity

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (534)  129  -134  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS). The proposed method introduces a line process into the IFS-based enlargement procedures to solve edge discontinuity of an enlarged image in the traditional IFS schemes. The line process can detect the edges which retain continuity. However, the line process cannot apply to the enlarged image whose intensities are not known. Therefore, the proposed method estimates the enlarged image's edges from the target image's edges obtained by using the line process. Then, since the estimated edges retain the continuity, the proposed method can calculate the enlarged image in such a way that its edges retain the continuity. Consequently, an accurate image enlargement can be achieved.
• A Note on Improvement of HITS using Clusters based on Link Analysis

  TAKEYOSHI Tomoya, HASEYAMA Miki  IEICE technical report  106-  (534)  159  -164  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes an effective Web page scoring method based on HITS, which is traditionally used for Web page scoring using a link analysis. Recently, some Web page scoring methods that modify the HITS have been proposed. However, they cannot solve the problem that many Web pages including almost the same links as one page and the pages linked by them have high scores. Therefore, the proposed method previously classifies the target Web pages based on similarity of link structure. Then each cluster tends to contain those controversial pages. Thus, our method introduces weights, that assign lower values to the links of the Web pages included in the same clusters than the others, into the HITS and calculates the scores. Consequently, the proposed method can solve the problem of the traditional schemes, and satisfies search requests of various users enough. The experimental results show the high performance of the proposed scoring method.
• A Note on Improvement of Similar Image Clustering Method using Neocognitron : Introduction of Novel Structure for Extracting Color Features

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (536)  1  -6  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a new image clustering method. We have proposed an image clustering method using a neocognitron which is one of the neural networks. In this paper, we use the edge features of the images as the input stimulus of the neocognitron. However, since this method notes only the edge features of the images, misclassifications are inevitable by changes of the target object's direction. Therefore, this paper proposes a new image clustering method based on the edge and color features. The color histogram that is one of the color features is robust to these problems. Therefore, the proposed method classifies the images based on the color histogram by using the self-organizing map. Further, the structure of the neocognitron is adaptively set from the obtained clustering result to perform the classification based on edges. Therefore, since both the color and edge features of the images can be utilized, our method accurately classifies the images. Some experimental results show the proposed method achieves high classification performance.
• A Study on Accurate Image Clustering Using Color Edges

  WATANABE Takashi, HASEYAMA Miki  IEICE technical report  106-  (536)  7  -10  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a similar image clustering method. We have proposed a image clustering scheme using image structures and quadratic form distances between color histograms. However, since this scheme cannot effectively utilize the distribution of colors within the target images, misclassification of these images is inevitable. In order to utilize these features more effectively, the proposed method classifies the images by the K-means algorithm based on distances between color correlograms, which can be utilized for describing the relationship of colors to spatial distances. Furthermore, by reconstructing clusters based on colors at the edges and their neighborhoods, the accurate image clustering can be realized.
• A Study on Group Tactics Analysis in Soccer Videos

  KON Hirofumi, HASEYAMA Miki  IEICE technical report  106-  (536)  121  -124  2007/02/15  [Not refereed][Not invited]
   

  This paper presents a method for group tactics analysis in soccer videos. For this analysis, the proposed method classifies players into some groups. As a result of this classification, the players in one group are closely related in terms of soccer tactics. In the proposed method, the relations between the players in the groups are analyzed by using player positions. Based on these relations, the proposed method estimates group advantage, which represents the degree of team's superiority or inferiority in each group. The group advantage helps viewers to understand game situations.
• A Note on the similarity between music audio signals using the melody line and the rhythm

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki  IEICE technical report  106-  (536)  125  -128  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes a novel similarity between music audio signals using rhythmic contents. The proposed method estimates the rhythmic content utilizing the autocorrelation functions of the signal. Then, by applying DTW to the functions, the similarity between music audio signals can be appropriately computed. Furthermore, our method visualizes the similarity using both melody line and rhythm to try to easily retrieve the similar music audio signals.
• An Accurate IFS-based Image Enlargement Method using Edge Continuity

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (536)  129  -134  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS). The proposed method introduces a line process into the IFS-based enlargement procedures to solve edge discontinuity of an enlarged image in the traditional IFS schemes. The line process can detect the edges which retain continuity. However, the line process cannot apply to the enlarged image whose intensities are not known. Therefore, the proposed method estimates the enlarged image's edges from the target image's edges obtained by using the line process. Then, since the estimated edges retain the continuity, the proposed method can calculate the enlarged image in such a way that its edges retain the continuity. Consequently, an accurate image enlargement can be achieved.
• A Note on Improvement of HITS using Clusters based on Link Analysis

  TAKEYOSHI Tomoya, HASEYAMA Miki  IEICE technical report  106-  (536)  159  -164  2007/02/15  [Not refereed][Not invited]
   

  This paper proposes an effective Web page scoring method based on HITS, which is traditionally used for Web page scoring using a link analysis. Recently, some Web page scoring methods that modify the HITS have been proposed. However, they cannot solve the problem that many Web pages including almost the same links as one page and the pages linked by them have high scores. Therefore, the proposed method previously classifies the target Web pages based on similarity of link structure. Then each cluster tends to contain those controversial pages. Thus, our method introduces weights, that assign lower values to the links of the Web pages included in the same clusters than the others, into the HITS and calculates the scores. Consequently, the proposed method can solve the problem of the traditional schemes, and satisfies search requests of various users enough. The experimental results show the high performance of the proposed scoring method.
• A note on improvement of similar image clustering method using neocognitron

  大原 貴都, 小川 貴弘, 長谷山 美紀  ITE technical report  31-  (10)  1  -6  2007/02  [Not refereed][Not invited]
  
• A study on accurate image clustering using color edges

  渡辺 隆志, 長谷山 美紀  ITE technical report  31-  (10)  7  -10  2007/02  [Not refereed][Not invited]
  
• A study on group tactics analysis in soccer videos

  今 宏史, 長谷山 美紀  ITE technical report  31-  (10)  121  -124  2007/02  [Not refereed][Not invited]
  
• A note on the similarity between music audio signals using the melody line and the rhythm

  今野 聡司, 二反田 直己, 長谷山 美紀  ITE technical report  31-  (10)  125  -128  2007/02  [Not refereed][Not invited]
  
• An accurate IFS-based image enlargement method using edge continuity

  覚幸 典弘, 小川 貴弘, 長谷山 美紀  ITE technical report  31-  (10)  129  -134  2007/02  [Not refereed][Not invited]
  
• A note on improvement of HITS using clusters based on link analysis

  武吉 朋也, 長谷山 美紀  ITE technical report  31-  (10)  159  -164  2007/02  [Not refereed][Not invited]
  
• An accurate video frame interpolation using image morphing

  山本 貴史, 長谷山 美紀  ITE technical report  31-  (11)  1  -4  2007/02  [Not refereed][Not invited]
  
• A study on precise player localization in a basketball video

  佐久間 大輔, 長谷山 美紀  ITE technical report  31-  (11)  17  -22  2007/02  [Not refereed][Not invited]
  
• An Accurate Video Frame Interpolation Using Image Morphing : A Note on Feature Point Extraction of Objects

  YAMAMOTO Takashi, HASEYAMA Miki  電子情報通信学会技術研究報告. SDM, シリコン材料・デバイス  106-  (521)  1  -4  2007/01/25  [Not refereed][Not invited]
  
• A Study on Precise Player Localization in a Basketball Video

  SAKUMA Daisuke, HASEYAMA Miki  電子情報通信学会技術研究報告. SDM, シリコン材料・デバイス  106-  (521)  17  -22  2007/01/25  [Not refereed][Not invited]
  
• サッカー映像における色成分を用いた選手追跡に関する検討

  林原局, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2007-  2007
• POCS-based iterative reconstruction algorithm of missing textures

  Takahiro Ogawa, Miki Haseyama  2007 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-7  3-  1229  -1232  2007  [Not refereed][Not invited]
   

  In this paper, a new framework for texture reconstruction of missing areas, which exist all over the target image, is presented. The framework is based on a projection onto convex sets (POCS) algorithm including a novel constraint. In the proposed method, a nonlinear eigenspace of each cluster obtained by texture classification is applied to the constraint. Furthermore, by monitoring the errors converged by the POCS algorithm, selection of the optimal cluster for the target texture including missing intensities is realized in order to reconstruct it adaptively. Then, iterating the POCS-based procedures, our method renews the nonlinear eigenspaces and the reconstruction image, and outputs the reliable result. This approach provides a solution to the problem in traditional methods of not being able to perform adaptive reconstruction of the target textures due to the missing intensities. Experimental results show subjective and quantitative improvement of the proposed reconstruction technique over previously reported reconstruction techniques.
• Phase retrieval based on a snake for image reconstruction

  Keiko Kondo, Miki Haseyama, Hideo Kitajima  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E90D-  (1)  283  -287  2007/01  [Not refereed][Not invited]
   

  A new phase retrieval method using an active contour model (snake) for image reconstruction is proposed. The proposed method reconstructs a target image by retrieving the phase from the magnitude of its Fourier transform and the measured area of the image. In general, the measured area is different from the true area where the target image exists. Thus a snake, which can extract the shape of the target image, is utilized to renew the measured area. By processing this renewal iteratively, the area obtained by the snake converges to the true area and as a result the proposed method can accurately reconstruct a target image even when the measured area is different from the true area. Experimental results show the effectiveness of the proposed method.
• Audio-based shot classification for audiovisual indexing using PCA, MGD and fuzzy algorithm

  Naoki Nitanda, Miki Haseyama  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E90-A-  (8)  1542  -1548  2007  [Not refereed][Not invited]
   

  An audio-based shot classification method for audiovisual indexing is proposed in this paper. The proposed method mainly consists of two parts, an audio analysis part and a shot classification part. In the audio analysis part, the proposed method utilizes both principal component analysis (PCA) and Mahalanobis generalized distance (MGD). The effective features for the analysis can be automatically obtained by using PCA, and these features are analyzed based on MGD, which can take into account the correlations of the data set. Thus, accurate analysis results can be obtained by the combined use of PCA and MGD. In the shot classification part, the proposed method utilizes a fuzzy algorithm. By using the fuzzy algorithm, the mixing rate of the multiple audio sources can be roughly measured, and thereby accurate shot classification can be attained. Results of experiments performed by applying the proposed method to actual audiovisual materials are shown to verify the effectiveness of the proposed method. Copyright © 2007 The Institute of Electronics, Information and Communication Engineers.
• Players clustering based on graph theory for tactics analysis purpose in soccer videos

  Hirofumi Kon, Miki Haseyama  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E90-A-  (8)  1528  -1533  2007  [Not refereed][Not invited]
   

  In this paper, a new method for clustering of players in order to analyze games in soccer videos is proposed. The proposed method classifies players who are closely related in terms of soccer tactics into one group. Considering soccer tactics, the players in one group are located near each other. For this reason, the Euclidean distance between the players is an effective measurement for the clustering of players. However, the distance is not sufficient to extract tactics-based groups. Therefore, we utilize a modified version of the community extraction method, which finds community structure by dividing a non-directed graph. The use of this method in addition to the distance enables accurate clustering of players. © 2007 The Institute of Electronics, Information and Communication Engineers.
• Adaptive reconstruction method of missing texture based on projection onto convex sets

  Takahiro Ogawa, Miki Haseyama  2007 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL I, PTS 1-3, PROCEEDINGS  1-  697  -700  2007  [Not refereed][Not invited]
   

  This paper presents a missing texture reconstruction method based on projection onto convex sets (POCS). The proposed method classifies textures within the target image into some clusters in a high-dimensional texture feature space. Further, for the target missing texture, our method performs a novel approach, that monitors the errors caused by the POCS algorithm in the feature space, and adaptively selects the optimal cluster including similar textures. Then, the missing texture is restored from these similar textures by a new POCS-based nonlinear subspace projection scheme. Consequently, since the proposed method realizes the nonconventional adaptive technique using the optimal nonlinear subspace, the accurate restoration result can be obtained. Experimental results show that our method achieves higher performance than the traditional method.
• A new audio signal analyzer using PCA and MGD

  Naoki Nitanda, Miki Haseyama  ICCE: 2007 DIGEST OF TECHNICAL PAPERS INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS  223  -+  2007  [Not refereed][Not invited]
   

  This paper presents a new audio signal analyzer Since our system can represent the possibility that the audio signal belongs to some audio classes, such as speech and music, it can be useful for several applications, such as audiovisual indexing and retrieval system.
• Image authentication system based on two-dimensional matrix codes without contamination of original images

  Takahiro Ogawa, Miki Haseyama  ICCE: 2007 DIGEST OF TECHNICAL PAPERS INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS  253  -+  2007  [Not refereed][Not invited]
   

  This paper proposes an authentication system without contamination of original images. The proposed system also realizes complete reconstruction of original signatures by introduction of 2-D matrix codes including error correction capability. Therefore, the system is suitable for the recent digital rights management.
• System for providing information about emergency rescue procedures to mobile phone users

  Takahiro Ogawa, Daisuke Sakuma, ShiN'Ichi Shiraishi, Miki Haseyama  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  61-  (12)  1818  -1827  2007  [Not refereed][Not invited]
   

  We propose the use of a system that provides information about emergency rescue procedures for mobile phone users. To enable the users to easily understand the procedures, avatars demonstrate how they are used. Since the avatars are efficiently constructed on a subset of scalable vector graphics (SVG), they can be quickly transmitted to the mobile phone and with low computational loads. In addition to the avatars, complementary voice and text data that explain the procedures are also transmitted via a synchronized multimedia integration language (SMIL, pronounced "smile") format. Moreover, to verify the performance and the effectiveness of the proposed system, we implemented a dedicated SMIL player suitable for use with mobile phones.
• Steady-state properties of a CORDIC-based adaptive ARMA lattice filter

  Shin'ichi Shiraishi, Miki Haseyama, Hideo Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E89A-  (12)  3724  -3729  2006/12  [Not refereed][Not invited]
   

  This paper analyzes the steady-state properties of a CORDIC-based adaptive ARMA lattice filter. In our previous study, the convergence properties of the filter in the non-steady state were clarified; however, its behavior in the steady state was not discussed. Therefore, we develop a distinct analysis technique based on a Markov chain in order to investigate the steady-state properties of the filter. By using the proposed technique, the relationship between step size and coefficient estimation error is revealed.
• A multi-objective service restoration method for power distribution systems

  Jun Inagaki, Jun Nakajima, Miki Haseyama  Proceedings - IEEE International Symposium on Circuits and Systems  1784  -1787  2006/12/01  [Not refereed][Not invited]
   

  Service restoration problem in distribution systems is formulated as a multi-objective optimization problem which is demanded not only for minimizing the amount of unrestored total loads but also for minimizing the number of the switching operations. The solution of the multi-objective optimization problem is usually obtained with a set of Pareto optimal solutions. The Pareto optimal solutions for the service restoration problem are useful for users to obtain their desired restoration by comparing a Pareto optimal solution with the others. However, the conventional methods cannot obtain several Pareto optimal solutions in one trial. Therefore, this paper proposes a method for obtaining the Pareto optimal set for the service restoration problem with a genetic algorithm. The genetic algorithm produces many possible solutions in its search process. By utilizing this feature, the proposed method can obtain the Pareto optimal set. © 2006 IEEE.
• 動き補償予測誤差に基づく適応内挿フィルタ選択手法に関する考察

  松野 孝也, 長谷山 美紀  映像情報メディア学会技術報告  30-  (55)  9  -12  2006/10/30  [Not refereed][Not invited]
  
• A Study on Image Clustering Using Color Histogram

  WATANABE Takashi, HASEYAMA Miki  ITE technical report  30-  (55)  37  -40  2006/10/30  [Not refereed][Not invited]
  
• A note on estimation of moving direction for omnidirectional images by using optical flows

  TAKAHASHI Hiroaki, HASEYAMA Miki  ITE technical report  30-  (55)  53  -56  2006/10/30  [Not refereed][Not invited]
  
• A Study on Estimation of Pass Courses in Soccer Videos with Active Net

  KON Hirofumi, HASEYAMA Miki  ITE technical report  30-  (55)  57  -60  2006/10/30  [Not refereed][Not invited]
  
• A similar image clustering method using a neocognitron

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki  ITE technical report  30-  (55)  61  -64  2006/10/30  [Not refereed][Not invited]
  
• A Note on POCS-based Restoration of Missing Areas in Still Images : Introduction of New Classification Scheme of Local Images

  OGAWA Takahiro, HASEYAMA Miki  ITE technical report  30-  (55)  83  -86  2006/10/30  [Not refereed][Not invited]
  
• PCA-based audio signal classification : An approach for accuracy improvement by using MGD

  NITANDA Naoki, HASEYAMA Miki  ITE technical report  30-  (55)  91  -94  2006/10/30  [Not refereed][Not invited]
  
• A Note on Melody Detection in Musical Audio Signal using Phrases

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki  ITE technical report  30-  (55)  95  -98  2006/10/30  [Not refereed][Not invited]
  
• IFSを用いたインパルス性雑音除去に関する一考察

  NISHIMURA HIROYA, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  ROMBUNNO.223  2006/10/28  [Not refereed][Not invited]
  
• ウィナーフィルタを用いた霧の存在する画像の復元に関する一考察

  HIRAMATSU TOMOKI, OGAWA TAKAHIRO, HASEYAMA MIKI  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  ROMBUNNO.222  2006/10/28  [Not refereed][Not invited]
  
• 拘束条件の適応的選択を導入した静止画像中に存在する消失テクスチャの再構成に関する検討

  小川 貴弘, 長谷山 美紀  映像情報メディア学会技術報告  30-  (42)  63  -66  2006/08/31  [Not refereed][Not invited]
  
• バスケットボール映像における選手位置推定に関する検討

  佐久間 大輔, 長谷山 美紀  映像情報メディア学会技術報告  30-  (42)  109  -112  2006/08/31  [Not refereed][Not invited]
  
• A note on video frame interpolation : Application of image morphing technique

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  106-  (114)  1  -5  2006/06/16  [Not refereed][Not invited]
   

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (114)  7  -12  2006/06/16  [Not refereed][Not invited]
   

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• A note on video frame interpolation : Application of image morphing technique

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  106-  (116)  1  -5  2006/06/16  [Not refereed][Not invited]
   

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (116)  7  -12  2006/06/16  [Not refereed][Not invited]
   

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• A note on video frame interpolation : Application of image morphing technique

  YAMAMOTO Takashi, HASEYAMA Miki  IEICE technical report  106-  (112)  1  -5  2006/06/16  [Not refereed][Not invited]
   

  In this paper, we propose a new frame interpolation method of image sequences using an image morphing technique. The proposed method establishes the correspondence between moving objects using the motion estimation between two consecutive frames into the morphing scheme. Further, using the morphing based on the correspondence between moving objects, new frames are interpolated between each two consective frames. Then the proposed method realizes the accurate frame interpolation since the morphing can generate new interpolated frames in such a way the moving objects change smoothly.
• A note on restoration of missing areas in still images using ER Algorithm : Introduction of new constraints using kernel principal component analysis

  OGAWA Takahiro, HASEYAMA Miki  IEICE technical report  106-  (112)  7  -12  2006/06/16  [Not refereed][Not invited]
   

  This paper proposes an accurate reconstruction method of missing textures in still images using an ER algorithm. It is assumed that, in the same texture image, different local areas have similar images and Fourier transform magnitudes mapped into a high-dimensional feature space each other. Further, the proposed method sets the constraints of the ER algorithm in such a way that the target local image and its Fourier transform magnitude mapped into the feature space are respectively in their eigenspaces generated from the known local areas in the target image. Then, the proposed method realizes the accurate calculation of the unknown Fourier transform magnitude and its phase of the target local area, so that the missing areas can be restored accurately.
• Restoration of Missing Intensities of Still Images Using GMRF Model

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  The IEICE transactions on information and systems  89-  (6)  1327  -1335  2006/06/01  [Not refereed][Not invited]
   

  本論文では,GMRFモデルを用いた静止画像における失われた輝度値の復元手法を提案する.画像中に存在する輝度値消失領域を高精度に復元するためには,エッジ部を精度良く再構成する必要がある.そこで,我々はGMRFモデルを用いて,特に失われたエッジ部を高精変に復元する手法の実現を試みる.このとき,提案手法において用いられるGMRFモデルは,画像が複数の領域により構成され,各領域は弱定常な確率過程であるという仮定のもとで導出される.したがって,提案手法のモデルは画像中の異なる領域間でエッジの表現が可能となり,エッジ部において高精度な復元結果を与える.更に,・このモデルは輝度値消失領域に複数の領域が合まれ,それらが複数のエッジを構成する場合についても,その適用が可能であり,このような場合に従来手法が抱えていた問題である再構成されたエッジ部の過剰な平滑化が抑制される.本論文の最後では,提案手法の有効性を示すために実画像を用いた比較実験を行い,エッジ部における復元性能を評価する.
• Audio signal segmentation and classification using fuzzy c-means clustering

  Naoki Nitanda, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  37-  (4)  23  -34  2006/04  [Not refereed][Not invited]
   

  This paper proposes a method of segmentation and classification of audio signals which is coded by MPEG Audio. The proposed method first detects the boundaries between two different audio signals, which are called audio-cuts, and then classifies segments, which are called audio-segments, and uses audio-cuts detected by fuzzy c-means clustering their boundaries. Since conventional methods detect audio-cuts by applying thresholding to audio features such as energy and zero-crossing rate, misdetection often occurs when they are applied to an audio signal which contains several audio effects, such as fade-in, fade-out, cross-fade, and the like. The proposed method represents the possibility that the audio-cut exists by a real number from 0 to 1, obtained by using fuzzy c-means clustering all of the possible candidates for the audio-cuts are detected. Since audio effects which are difficult to detect by using conventional methods are also detected as the candidates, misdetection can be reduced. Furthermore, the audio-segments, whose boundaries are the candidates, are subdivided into five audio classes, which are silence, speech, music, speech with music background, and speech with noise background. By using the classification results, unnecessary audio-cuts can be removed, and thereby both accurate audio-cut detection and accurate audio-segment classification can be attained. © 2006 Wiley Periodicals, Inc.
• Filter coefficient quantization method with genetic algorithm, including simulated annealing

  M Haseyama, D Matsuura  IEEE SIGNAL PROCESSING LETTERS  13-  (4)  189  -192  2006/04  [Not refereed][Not invited]
   

  A method based on a genetic algorithm (GA), including a simulated annealing (SA), is proposed for filter coefficient quantization. The proposed method uses the GA to search a population of the quantized filters of a digital filter for the optimal quantized filter. It retains the most accurate frequency characteristic of the original filter, which is either finite impulse response filter or an infinite impulse response filter. The initial population in the GA is generated by binomial distributions, which are not used for the other GAs. An SA is also embedded in the GA search, which can support the GA to converge to the optimum in the early generations. The experimental results verify that our method can provide a quantized filter with a better frequency characteristic than those obtained by the traditional quantization methods, such as rounding off, rounding up, and rounding down.
• Restoration of missing intensity of still images by using optical flow

  Takahiro Ogawa, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  37-  (3)  49  -57  2006/03  [Not refereed][Not invited]
   

  This paper proposes an accurate method for the restoration of missing intensities of still images by using the optical flow. It is important in restoration to reconstruct missing edges correctly. Therefore, this paper modifies the optical flow conventionally used for motion analysis in video images and applies it to the restoration of missing intensities. Further, the proposed method introduces a new index expressing the correlation of intensities between two pixels into the scheme for calculation of the optical flow in order to obtain a flow which gives more accurate estimated values. The optical flow calculated by this index provides the pixel from the neighborhood whose intensity is most similar to that of the target pixel, so that the estimated intensity is not affected by pixels whose intensities are quite different. Consequently, even when multiple edges pass through the missing area or the direction of the edge changes significantly inside the area, the proposed method can reconstruct the edges correctly. Some experimental results are presented in order to verify the high performance of the proposed method. © 2006 Wiley Periodicals, Inc.
• An extraction method for apoptotic cells from videomicroscopy images

  Miki Haseyama, Yukari Sasamura, Hideo Kitajima  Systems and Computers in Japan  37-  (3)  32  -40  2006/03  [Not refereed][Not invited]
   

  An automatic extraction method for apoptotic cells from videomicroscopy images is proposed. The edges of the apoptotic cells in the microscopy images are gentle and thick because of the fluorescence enzyme used for the observation their shapes are distorted circles the apoptotic cells also neighbor on the other cells. Since the apoptotic cells cannot be easily extracted, they are generally extracted by visual recognition. Therefore, we propose a method to automatically extract only apoptotic cells based on shape features. The proposed method, in the first stage, extracts areas with strong possibilities that apoptotic cells are included. In the second stage, the method applies shape-feature analysis to the extracted areas. By using these two stages, the proposed method can accurately extract the apoptotic cells without increasing computational costs. Further, when videomicroscopy images are acquired as an image sequence, the proposed method can extract the cells more accurately by comparing the extraction results in neighboring frames. Simulation results verify the effectiveness of the proposed method. © 2006 Wiley Periodicals, Inc.
• A Note on an Accurate IFS-based Image Enlargement Method using Line Process

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (14)  1  -6  2006/02/21  [Not refereed][Not invited]
  
• A note on accurate reconstruction method of missing textures using ER algorithm : A study on subspaces used in constraints of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (14)  7  -12  2006/02/21  [Not refereed][Not invited]
  
• Detection of Person who Moves in Prohibited Direction by using Video Camera

  ARAI Hiroyuki, YASUNO Takayuki, MIZUKAMI Midori, HSEYAMA Miki  ITE technical report  30-  (14)  29  -34  2006/02/21  [Not refereed][Not invited]
  
• High-Quality Image Compression with Region-Based Filtering

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  映像情報メディア学会技術報告  30-  (14)  85  -88  2006/02/21  [Not refereed][Not invited]
  
• Quality Improvement of Cell Animation Images Using a Region-based Algorithm

  YAMAGUCHI Takumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (14)  105  -108  2006/02/21  [Not refereed][Not invited]
  
• A Study on Quantification of Paper-Relations Using Co-occurrence and Its Visualization

  KAMAKURA Junichi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (14)  127  -130  2006/02/21  [Not refereed][Not invited]
  
• A New Application of Fractal Coding to Video Images

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  映像情報メディア学会技術報告  30-  (14)  131  -136  2006/02/21  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system (IFS), one of the coding techniques for digital images, compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the decoded image quality by the fractal image coding. However, the relationship between the self-similarity of the original image and the quality of its decoded image has not been clarified so far. Therefore, in the previous study, we have clarified the relationship for still images. Then, in this paper, we clarify the relationship for video images and consider a new approach to video coding based on the fractal image coding.
• A Note on an Accurate IFS-based Image Enlargement Method using Line Process

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (609)  1  -6  2006/02/21  [Not refereed][Not invited]
   

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS), which is traditionally used for image coding. The IFS can reconstruct an image of a different size from the coding target image's. Based on this property, some methods enlarging images have been proposed by using the IFS. However, their enlargement results suffer from block noise and edge discontinuity in the boundary between neighboring range blocks, which are units of the procedures in the IFS. The reasons for these problems are that they use non-overlapping range blocks and do not consider edge continuity in the boundary between the neighboring range blocks. Therefore, the proposed method allows selection of overlapping range blocks in order to avoid the block noise. Further, the proposed method introduces a line process, which is used for edge detection, into the enlargement procedures. The edges obtained by using the line process can retain the edge continuity. Therefore, the images enlarged by the proposed method retain the edge continuity. Consequently, an accurate image enlargement can be achieved.
• A note on accurate reconstruction method of missing textures using ER algorithm : A study on subspaces used in constraints of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (609)  7  -12  2006/02/21  [Not refereed][Not invited]
   

  This paper proposes a reconstruction method of missing textures in still images using an ER(Error Reduction) algorithm which is one of phase retrieval methods. The proposed method defines constraints used in this algorithm in such a way that the target local image and its Fourier transform magnitude are respectively in two subspaces generated from some clipped local images and their Fourier transform magnitudes. Then, the Fourier transform magnitude and phase of the target local image are estimated so that the missing area can be reconstructed. Further, the proposed method adaptively selects the local images and Fourier transform magnitudes utilized to generate these subspaces for the target local image. Consequently, the proposed method can accurately reconstruct the missing textures by utilizing the ER algorithm including the novel constraints. Some experimental results are shown to verify the high performance of the proposed method.
• Detection of Person who Moves in Prohibited Direction by using Video Camera

  ARAI Hiroyuki, YASUNO Takayuki, MIZUKAMI Midori, HSEYAMA Miki  IEICE technical report  105-  (609)  29  -34  2006/02/21  [Not refereed][Not invited]
   

  We propose a method for detection of person who moves in prohibited direction, by using a general video camera which is set along under-diagonal viewing angle. In crowded situation, tracking every person is difficult, but computing optical-flow is rather stable than tracking, and the flow vector with opposite direction is a key to detect opposite moving. However, flow vectors from one person appear fragmentary. Therefore, to integrate such fragmented flow-vectors as one motion is a problem. To solve this, we use "x-t map" which records x-position and its moving direction versus time, which is made from x-projection of background difference image and x-projection of x-component of optical flow. Finding opposite moving in "x-t map" is done by Hough transformation. Experimental result shows its effectiveness.
• High-Quality Image Compression with Region-Based Filtering

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  105-  (609)  85  -88  2006/02/21  [Not refereed][Not invited]
   

  Recently, MPEG-2 is widely used in many fields such as DVD, digital broadcasting, etc. The video coding technique achieves efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that blocking and ringing artifacts appear in the decoded images at low bit rates. Therefore, we have proposed a method utilizes a two-dimensional FIR filter to reduce these artifacts. The method designs the filter to minimize the error between the original image and the coded image for each frames, and its coefficients are added into the bitstream. This paper proposes a method of region segmentation using the decoded images and filter design for each regions. By applying the filters to the decoded image, the proposed method improves the decoded video quality.
• Quality Improvement of Cell Animation Images Using a Region-based Algorithm

  YAMAGUCHI Takumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (609)  105  -108  2006/02/21  [Not refereed][Not invited]
   

  MPEG-4 as video coding techniques is widely used in video distribution. This technique achieves efficient compression by using block-based DOT and motion compensation. Nevertheless, there is a problem that compression quality of cell animation images, which is one of artificial images, is deteriorated to lose high-frequency component in the decoded images at low bit rates. This is caused by the fact that the cell animation images have a feature that intensity values change on the edges more steeply than natural images. To improve the image quality, this paper proposes a method, adaptively processes each region obtained from the cell animation images based on their features the cell animation images. Experimental results are shown to verify the performance of the proposed method.
• A Study on Quantification of Paper-Relations Using Co-occurrence and Its Visualization

  KAMAKURA Junichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (609)  127  -130  2006/02/21  [Not refereed][Not invited]
   

  Some visualization methods of paper-relations are proposed for efficient retrieval of papers. However, these methods require a large number of papers to obtain the appropriate visualization results. Therefore, we have proposed an appropriate visualization method of paper-relations even though the number of papers is small. This method realizes the visualization by applying singular value decomposition to the matrix which consist of the feature vectors of the papers. The feature vector are defined using occurrence probability of the words. This paper defines the feature vectors using co-occurrence of the words in order to evaluate importance of the words. Some experimental results show that the propose method can obtain the appropriate visualization.
• A New Application of Fractal Coding to Video Images

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  ITE technical report  105-  (609)  131  -136  2006/02/21  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system (IFS), one of the coding techniques for digital images, compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the decoded image quality by the fractal image coding. However, the relationship between the self-similarity of the original image and the quality of its decoded image has not been clarified so far. Therefore, in the previous study, we have clarified the relationship for still images. Then, in this paper, we clarify the relationship for video images and consider a new approach to video coding based on the fractal image coding.
• A Note on an Accurate IFS-based Image Enlargement Method using Line Process

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (611)  1  -6  2006/02/21  [Not refereed][Not invited]
   

  This paper proposes an effective image enlargement method based on an Iterated Function System (IFS), which is traditionally used for image coding. The IFS can reconstruct an image of a different size from the coding target image's. Based on this property, some methods enlarging images have been proposed by using the IFS. However, their enlargement results suffer from block noise and edge discontinuity in the boundary between neighboring range blocks, which are units of the procedures in the IFS. The reasons for these problems are that they use non-overlapping range blocks and do not consider edge continuity in the boundary between the neighboring range blocks. Therefore, the proposed method allows selection of overlapping range blocks in order to avoid the block noise. Further, the proposed method introduces a line process, which is used for edge detection, into the enlargement procedures. The edges obtained by using the line process can retain the edge continuity. Therefore, the images enlarged by the proposed method retain the edge continuity. Consequently, an accurate image enlargement can be achieved.
• A note on accurate reconstruction method of missing textures using ER algorithm : A study on subspaces used in constraints of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (611)  7  -12  2006/02/21  [Not refereed][Not invited]
   

  This paper proposes a reconstruction method of missing textures in still images using an ER (Error Reduction) algorithm which is one of phase retrieval methods. The proposed method defines constraints used in this algorithm in such a way that the target local image and its Fourier transform magnitude are respectively in two subspaces generated from some clipped local images and their Fourier transform magnitudes. Then, the Fourier transform magnitude and phase of the target local image are estimated so that the missing area can be reconstructed. Further, the proposed method adaptively selects the local images and Fourier transform magnitudes utilized to generate these subspaces for the target local image. Consequently, the proposed method can accurately reconstruct the missing textures by utilizing the ER algorithm including the novel constraints. Some experimental results are shown to verify the high performance of the proposed method.
• Detection of Person who Moves in Prohibited Direction by using Video Camera

  ARAI Hiroyuki, YASUNO Takayuki, MIZUKAMI Midori, HSEYAMA Miki  IEICE technical report  105-  (611)  29  -34  2006/02/21  [Not refereed][Not invited]
   

  We propose a method for detection of person who moves in prohibited direction, by using a general video camera which is set along under-diagonal viewing angle. In crowded situation, tracking every person is difficult, but computing optical-flow is rather stable than tracking, and the flow vector with opposite direction is a key to detect opposite moving. However, flow vectors from one person appear fragmentary. Therefore, to integrate such fragmented flow-vectors as one motion is a problem. To solve this, we use "x-t map" which records x-position and its moving direction versus time, which is made from x-projection of background difference image and x-projection of x-component of optical flow. Finding opposite moving in "x-t map" is done by Hough transformation. Experimental result shows its effectiveness.
• High-Quality Image Compression with Region-Based Filtering

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (611)  85  -88  2006/02/21  [Not refereed][Not invited]
   

  Recently, MPEG-2 is widely used in many fields such as DVD, digital broadcasting, etc. The video coding technique achieves efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that blocking and ringing artifacts appear in the decoded images at low bit rates. Therefore, we have proposed a method utilizes a two-dimensional FIR filter to reduce these artifacts. The method designs the filter to minimize the error between the original image and the coded image for each frames, and its coefficients are added into the bitstream. This paper proposes a method of region segmentation using the decoded images and filter design for each regions. By applying the filters to the decoded image, the proposed method improves the decoded video quality.
• Quality Improvement of Cell Animation Images Using a Region-based Algorithm

  YAMAGUCHI Takumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (611)  105  -108  2006/02/21  [Not refereed][Not invited]
   

  MPEG-4 as video coding techniques is widely used in video distribution. This technique achieves efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that compression quality of cell animation images, which is one of artificial images, is deteriorated to lose high-frequency component in the decoded images at low bit rates. This is caused by the fact that the cell animation images have a feature that intensity values change on the edges more steeply than natural images. To improve the image quality, this paper proposes a method, adaptively processes each region obtained from the cell animation images based on their features the cell animation images. Experimental results are shown to verify the performance of the proposed method.
• A Study on Quantification of Paper-Relations Using Co-occurrence and Its Visualization

  KAMAKURA Junichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (611)  127  -130  2006/02/21  [Not refereed][Not invited]
   

  Some visualization methods of paper-relations are proposed for efficient retrieval of papers. However, these methods require a large number of papers to obtain the appropriate visualization results. Therefore, we have proposed an appropriate visualization method of paper-relations even though the number of papers is small. This method realizes the visualization by applying singular value decomposition to the matrix which consist of the feature vectors of the papers. The feature vector are defined using occurrence probability of the words. This paper defines the feature vectors using co-occurrence of the words in order to evaluate importance of the words. Some experimental results show that the propose method can obtain the appropriate visualization.
• A New Application of Fractal Coding to Video Images

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  IEICE technical report  105-  (611)  131  -136  2006/02/21  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system (IPS), one of the coding techniques for digital images, compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the decoded image quality by the fractal image coding. However, the relationship between the self-similarity of the original image arid the quality of its decoded image has not been clarified so far. Therefore, in the previous study, we have clarified the relationship for still images. Then, in this paper, we clarify the relationship for video images and consider a new approach to video coding based on the fractal image coding.
• A Note on Accurate Estimation of Chords in Musical Audio Signal : An Approach using Single Tone in Low-Frequency Band

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (13)  1  -4  2006/02/20  [Not refereed][Not invited]
  
• Improvement of Target Tracking using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (13)  23  -26  2006/02/20  [Not refereed][Not invited]
  
• A Study on Tracking Soccer Players using Active Contour Models

  WATANABE Tomoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (13)  27  -30  2006/02/20  [Not refereed][Not invited]
  
• A Method of Soccer Video Analysis by Using Active Net

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  映像情報メディア学会技術報告  30-  (13)  31  -34  2006/02/20  [Not refereed][Not invited]
  
• A study on accurate audiovisual indexing using audio signal

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  30-  (13)  77  -82  2006/02/20  [Not refereed][Not invited]
  
• A Note on Accurate Estimation of Chords in Musical Audio Signal : An Approach using Single Tone in Low-Frequency Band

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (608)  1  -4  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a method for the accurate estimation of major chords and minor chords in musical audio signals, using the rules of musical chords. Chords are important for analysing the compositon of musical audio signals, and need to be estimated accurately. Then, studies of the estimation of chords have been done, based on the estimation of single tones. However, it is difficult to estimate chords accurately when many single tones exist at a time. Then, first, the proposed method estimates the pitches of the bass sounds which form the basis of chords. Secondly the method calculates the energy of the frequency which corresponds to the major and the minor intervals, and compares them. Consequently, the proposed method can estimate chords without complete estimation of the single tones which are included in chords.
• Improvement of Target Tracking using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (608)  23  -26  2006/02/20  [Not refereed][Not invited]
   

  As an object moves through the field of view of a camera, it undergoes changes in pose relative to the viewing camera, in illumination relative to light sources, and may even become partially occluded. This causes erroneous object-tracking results. To deal with this problem, we have proposed a method using a shape-constraint-based splitting snake. In this method, the approximate 3-D shape of the target object is assumed to be given, and the snake extracts the target using a reference contour which is sorted from the several contours prepared from the given shape. However, this method could obtain erroneous results in the case where the prepared contours do not include any appropriate ones. To solve this problem, this paper proposes a more efficient method, which introduces a new term to the objective function defined for the snake. This term is calculated from the variation amount of the pose of the approximate shape. By optimising the objective function, the proposed method can obtain the appropriate reference contour and track moving objects in the image sequence more accurately.
• A Study on Tracking Soccer Players using Active Contour Models

  WATANABE Tomoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (608)  27  -30  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a soccer player tracking method using level sets in soccer video. The level set method is a kind of active contour model. The soccer video we see via TV has a lot of scenes showing the large area of the soccer field. In those scenes, the player's position in a frame is almost the same in the next one. In order to utilize the inter-frame relation, we regard the video as a three-dimensional array consisting of time-series images. By applying the level set method to the array, we can extract the soccer player from the soccer video as three-dimensional regions. Since the same player is extracted as one region all through the frames, we can track the players at the time of the player extraction.
• A Method of Soccer Video Analysis by Using Active Net

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  105-  (608)  31  -34  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a method of soccer video analysis by using Active Net, which is a lattice model based on minimizing energy. Generally, in order to soccer video analysis, image and audio processing algorithms are utilized. These methods can create summary and detect important events. However, these are not enough to satisfy in order to develop the system that makes the viewers enjoy. Therefore, the semantic analysis based on soccer tactics is needed. The proposed method focuses on the pass, which is an important action in soccer tactics, and visualizes the pass-routes.
• A study on accurate audiovisual indexing using audio signal

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (608)  77  -82  2006/02/20  [Not refereed][Not invited]
   

  An accurate audiovisual indexing method is proposed in this paper. The proposed method classifies the audio signal into the following five audio classes: silence, speech, music, speech with music background, and speech with noise background. For this audio classification, both principal component analysis (PCA) and fuzzy c-means clustering (FCM) are utilized. The effective features can be theoretically extracted by using PCA, and the reliability of the clustering results can be measured by using FCM. Moreover, combination use of the proposed method and video segmentation technique, which can accurately detect the segmentation point, make it possible to measure the similarity between two adjacent shots. According to the similarity, the scene transition can be accurately detected. Experimental results obtained by the combination approach to actual audiovisual materials are shown to verify its effectiveness.
• A Note on Accurate Estimation of Chords in Musical Audio Signal : An Approach using Single Tone in Low-Frequency Band

  IMANO Satoshi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (610)  1  -4  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a method for the accurate estimation of major chords and minor chords in musical audio signals, using the rules of musical chords. Chords are important for analysing the compositon of musical audio signals, and need to be estimated accurately. Then, studies of the estimation of chords have been done, based on the estimation of single tones. However, it is difficult to estimate chords accurately when many single tones exist at a time. Then, first, the proposed method estimates the pitches of the bass sounds which form the basis of chords. Secondly the method calculates the energy of the frequency which corresponds to the major and the minor intervals, and compares them. Consequently, the proposed method can estimate chords without complete estimation of the single tones which are included in chords.
• Improvement of Target Tracking using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (610)  23  -26  2006/02/20  [Not refereed][Not invited]
   

  As an object moves through the field of view of a camera, it undergoes changes in pose relative to the viewing camera, in illumination relative to light sources, and may even become partially occluded. This causes erroneous object-tracking results. To deal with this problem, we have proposed a method using a shape-constraint-based splitting snake. In this method, the approximate 3-D shape of the target object is assumed to be given, and the snake extracts the target using a reference contour which is sorted from the several contours prepared from the given shape. However, this method could obtain erroneous results in the case where the prepared contours do not include any appropriate ones. To solve this problem, this paper proposes a more efficient method, which introduces a new term to the objective function defined for the snake. This term is calculated from the variation amount of the pose of the approximate shape. By optimising the objective function, the proposed method can obtain the appropriate reference contour and track moving objects in the image sequence more accurately.
• A Study on Tracking Soccer Players using Active Contour Models

  WATANABE Tomoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (610)  27  -30  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a soccer player tracking method using level sets in soccer video. The level set method is a kind of active contour model. The soccer video we see via TV has a lot of scenes showing the large area of the soccer field. In those scenes, the player's position in a frame is almost the same in the next one. In order to utilize the inter-frame relation, we regard the video as a three-dimensional array consisting of time-series images. By applying the level set method to the array, we can extract the soccer player from the soccer video as three-dimensional regions. Since the same player is extracted as one region all through the frames, we can track the players at the time of the player extraction.
• A Method of Soccer Video Analysis by Using Active Net

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (610)  31  -34  2006/02/20  [Not refereed][Not invited]
   

  This paper proposes a method of soccer video analysis by using Active Net, which is a lattice model based on minimizing energy. Generally, in order to soccer video analysis, image and audio processing algorithms are utilized. These methods can create summary and detect important events. However, these are not enough to satisfy in order to develop the system that makes the viewers enjoy. Therefore, the semantic analysis based on soccer tactics is needed. The proposed method focuses on the pass, which is an important action in soccer tactics, and visualizes the pass-routes.
• A study on accurate audiovisual indexing using audio signal

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (610)  77  -82  2006/02/20  [Not refereed][Not invited]
   

  An accurate audiovisual indexing method is proposed in this paper. The proposed method classifies the audio signal into the following five audio classes: silence, speech, music, speech with music background, and speech with noise background. For this audio classification, both principal component analysis (PCA) and fuzzy c-means clustering (FCM) are utilized. The effective features can be theoretically extracted by using PCA, and the reliability of the clustering results can be measured by using FCM. Moreover, combination use of the proposed method and video segmentation technique, which can accurately detect the segmentation point, make it possible to measure the similarity between two adjacent shots. According to the similarity, the scene transition can be accurately detected. Experimental results obtained by the combination approach to actual audiovisual materials are shown to verify its effectiveness.
• A New Discrete Clustering Method Based on the Clustering Method Dynamic Coalescence Model

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  IEICE technical report  105-  (536)  159  -164  2006/01/13  [Not refereed][Not invited]
   

  We propose a new discrete clustering method based on concept of the field. It is supposed that a point in a feature space is a mass point, and location of a point is quantized to a lattice point. Points are moved to a lattice point of at most its adjacent iteratively, and it is speculated that a set of points gathered in a lattice point is a cluster as a class to which the points belong to. It is clarified that acceleration of a point is expressed by convolution of attractive force function that expresses the force of the point subject to other point under the supposition of mass of each point is 1. It is also clarified that the acceleration is determined by attractive force from points that expresses outline of point distribution observed from the location of the point and also in a sphere with certain radius of which center is the location of the point. On pattern distribution that is expressed by mixture normal and semicircular distributions with two classes in two dimension, rate of points of which classes are accurately speculated is 30.5% greater than that of continuous method Dynamic Coalescence Method, and duration is 0.0747%.
• A New Discrete Clustering Method Based on the Clustering Method Dynamic Coalescence Model

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  IEICE technical report  105-  (534)  159  -164  2006/01/13  [Not refereed][Not invited]
   

  We propose a new discrete clustering method based on concept of the field. It is supposed that a point in a feature space is a mass point, and location of a point is quantized to a lattice point. Points are moved to a lattice point of at most its adjacent iteratively, and it is speculated that a set of points gathered in a lattice point is a cluster as a class to which the points belong to. It is clarified that acceleration of a point is expressed by convolution of attractive force function that expresses the force of the point subject to other point under the supposition of mass of each point is 1. It is also clarified that the acceleration is determined by attractive force from points that expresses outline of point distribution observed from the location of the point and also in a sphere with certain radius of which center is the location of the point. On pattern distribution that is expressed by mixture normal and semicircular distributions with two classes in two dimension, rate of points of which classes are accurately speculated is 30.5% greater than that of continuous method Dynamic Coalescence Method, and duration is 0.0747%.
• Design of nonuniform transmission-line matching circuits using a genetic algorithm

  SANADA Hirofumi, ITOH Hiroaki, TAKEZAWA Megumi, WATANABE Kazuhisa, HASEYAMA Miki  IEICE technical report  105-  (503)  7  -12  2006/01/05  [Not refereed][Not invited]
   

  We present a new method for designing nonuniform transmission-line matching circuits. This method is based on a genetic algorithm and can effectively be applied to various nonuniform transmission-line design problems without increasing computational complexity. To evaluate this method, we design several different types of nonuniform transmission-line matching circuits and show that it is possible to design the circuits, taking restrictions on the shape and size of the circuits into account.
• A MODEL-BASED APPROACH FOR SOCCER TEAM ADVANTAGE MEASUREMENT

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (501)  317  -322  2006/01/03  

  This paper proposes a method of measuring team advantage to facilitate understanding of soccer game contents in TV images. In the proposed method, the team advantage is represented by possibility of getting a goal. The possibility is defined by using a novel player-action model, which is realized by a Markov-chain model. The Markov-chain model is suitable to realize the player-action model, because it denotes the transition probability between two states; that is, in the proposed method, player action states and their result ones, such as "Shoot", "Goal", etc. By using the player-action model, the proposed method can analyze the games more effectively than the other traditional methods, which utilize only the player locations. Consequently, through the proposed method, the users can not only understand the soccer game situations, such as the scenes that the players scrambled to get a ball, the favorite team of the users is getting a goal, etc, but also watch the games from the technical and tactical perspectives. Some experiments on real soccer videos have demonstrated the effectiveness of the proposed method.
• A SIMILAR IMAGE CLUSTERING METHOD INCLUDING AUTOMATIC SELECTION OF NUMBER OF CLUSTERS

  OHARA Takatoshi, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (501)  209  -212  2006/01/03  

  This paper proposes an image clustering method which includes an automatic cluster-number setting scheme. By using the proposed method, the user can effectively categorize similar images existing in an image database without presetting the total number of categories. Actually, the proposed method is based on a k-means algorithm which utilizes a color histogram of each image as a feature vector, and the number of the cluster is determined, according to the variation of the average of the clustering error in several number of the clusters. Consequently, the proposed method can automatically select the suitable number of the clusters for the image database, and then it provides the accurate image clustering result. Some experimental results show the proposed method achieves accurate clusterings.
• AN IMAGE ENLARGEMENT METHOD USING ITERATED FUNCTION SYSTEM

  KAKUKOU Norihiro, OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (500)  283  -286  2006/01/02  

  This paper proposes an effective image enlargement method based on an Iterated Function System (IFS), which is well-known as an image coding method. The IFS can reconstruct an image of a different size from the coding target image's. Based on this property, this paper expands the IFS for coding to a method for the image enlargement. However, if the IFS is simply expanded for the image enlargement use, it cannot achieve accurate enlargement. For example, the block-wised procedures of the IFS make a problem that the enlarged images suffer from block noise. Therefore, the proposed method is realized by not only the expansion but also some modifications. Some simulation results are shown to verify that the performance of the proposed method is higher than that of the conventional methods.
• EFFECTIVE PEDESTRIAN DETECTION IN LOW-CONTRAST IMAGES

  KAGA Yosuke, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report  105-  (500)  287  -290  2006/01/02  

  This paper proposes an accurate pedestrian detection method in low-contrast images by using Support Vector Machine (SVM). The proposed pedestrian detection is realized by the following two systems : (1) Moving object extraction system In the low contrast case, traditional methods cannot extract the moving objects successfully. Therefore, in order to accurately extract the objects, the proposed method adopts hierarchical procedures depending on the intensity levels of the low-contrast images. (2) Pedestrian decision system From the objects extracted by (1), two features, which are the ratio between the width and the height and the moving speed, are computed. Pedestrian decision is executed by inputting these features to the SVM. By using these two systems, the features can be accurately extracted, and the SVM using these features can successfully judge whether the moving object is a pedestrian or not. Consequently, the proposed method can accurately detect the pedestrian in low-contrast images. Some experimental results are shown to verify the effectiveness of the proposed method.
• A VIDEO CODING TECHNIQUE USING FRACTAL CODING

  TAKEZAWA Megumi, HASEYAMA Miki, SANADA Hirofumi, WATANABE Kazuhisa  IEICE technical report  105-  (500)  51  -54  2006/01/02  

  This paper proposes a new approach to video coding that utilizes the fractal image coding. The fractal image coding based on an iterated function system is one of the coding techniques for digital images. It utilizes the self-similarity of the images and thus realizes the high image-compression performance. However, we cannot adopt the fractal image coding as the standard coding technique because its decoded images are sometimes low quality. Therefore, in this paper, to promote the use of the fractal image coding, we clarify the difference between the suitable images, whose decoded images are high quality, and the unsuitable images, whose decoded images are low quality. Then, based on the results, we apply the fractal image coding adaptively to each image. In this paper, we extend this approach to encoding the video images and show some simulations.
• 動画像信号と音響信号を用いた映像の類似シーン検出に関する検討

  道山大悟, 二反田直己, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  2006
• 低コントラスト画像における動物体抽出の高精度化に関する一考察

  加賀陽介, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  2006
• SVMによる画像分類を用いた製造フィルムの欠陥種類の識別に関する検討

  瀧澤誠, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  2006
• 音楽信号におけるフレーズに着眼した重要旋律の推定に関する一考察

  今野聡司, 二反田直己, 長谷山美紀  電気・情報関係学会北海道支部連合大会講演論文集(CD-ROM)  2006-  2006
• 映像検索技術-その必要性の増大と技術革新について

  長谷山美紀  テクノカレント  (419)  2006
• Bandwidth-efficient information-providing system for mobile phone users

  Shin'ichi Shiraishi, Daisuke Sakuma, Miki Haseyama, Hideo Kitajima  Digest of Technical Papers - IEEE International Conference on Consumer Electronics  2006-  389  -390  2006  [Not refereed][Not invited]
   

  This paper proposes an information-providing system for mobile phone users. The proposed system only requires reduced bandwidth and then brings a low latency service. Therefore, the system is suitable for urgent usage such as an information service for emergency rescue. © 2006 IEEE.
• An improved GRAS algorithm using archimedes's spiral

  Yanjun Zhao, Miki Haseyama, Hideo Kitajima  ISSCAA 2006: 1ST INTERNATIONAL SYMPOSIUM ON SYSTEMS AND CONTROL IN AEROSPACE AND ASTRONAUTICS, VOLS 1AND 2  2006-  461  -+  2006  [Not refereed][Not invited]
   

  In this paper an improvement is proposed to solve a problem in the object recognition using GRAS: when an object is rotated, the change may not be acceptable by GRAS in most cases. The proposed algorithm searches the shape of a given object to define some nodes by Archimedes's spiral, and transforms the nodes into a directed graph. Then, it defines some interrelated graph characteristics to compare the given object with others. The performance of the proposed algorithm is presented via experimental results which compare the recognition accuracy of GRAS and proposed algorithm.
• An accurate ifs-based image enlargement method using line process

  Norihiro Kakukou, Takahiro Ogawa, Miki Haseyama, Hideo Kitajima  2006 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP 2006, PROCEEDINGS  2701  -+  2006  [Not refereed][Not invited]
   

  This paper proposes an effective image enlargement method based on an Iterated Function System(IFS), which is traditionally used for image coding. The IFS can reconstruct an image of a different size from the coding target image's. Based on this property, some methods enlarging images have been proposed by using the IFS. However, the images enlarged by the traditional methods suffer from block noise and edge discontinuity in the boundary between neighboring range blocks, which are units of the process in the IFS. The reasons for the problems of the traditional methods are that they use non-overlapping range blocks and do not consider edge continuity in the boundary between the neighboring range blocks. Therefore, the proposed method allows selection of overlapping range blocks in order to avoid the block noise. Further, the proposed method introduces a line process, which is used for edge detection, into the enlargement procedure. The edges obtained by using the line process can retain the edge continuity. Therefore, the images enlarged by the proposed method retain the edge continuity. Consequently, an accurate image enlargement can be achieved.
• A novel image recognition method by vector voting

  Masao Hiramoto, Miki Haseyama  WMSCI 2006: 10TH WORLD MULTI-CONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS, VOL V, PROCEEDINGS  5-  234  -+  2006  [Not refereed][Not invited]
   

  This paper introduces a new method for recognizing images and finding their sizes and angles. Using this method, an image is constructed from 2 types of vectors-vectors indicating positions and vectors denoting intensity gradients for those positions, and image recognition data is created from the vectors. When investigating the difference between a reference image and a sample one, we vote vectors and similarities are evaluated by calculating voting densities in the image space. Furthermore when two images are identified, we can find the size and angle of the sample image in comparison with the reference one by summating voted vectors at the origin. The proposed method is invariant to image rotation and size. Using this method, we carried out experimentation recognizing images and finding their sizes and angles, and the results show that the method is effective in retrieving images.
• A steady-state analysis of a CORDIC-Based adaptive ARMA lattice filter

  Shin'ichi Shiraishi, Miki Haseyama, Hideo Kitajima  TENCON 2005 - 2005 IEEE REGION 10 CONFERENCE, VOLS 1-5  2007-  2266  -+  2006  [Not refereed][Not invited]
   

  This paper analyzes steady-state properties of a CORDIC-based adaptive ARMA lattice filter. In our earlier work, convergence properties of the filter in the non-steady state have been clarified; however, its behavior in the steady-state is not discussed. Therefore, we develop a distinct analysis technique based on a Markov chain in order to investigate steady-state properties of the filter. By using the proposed technique, the relation between step size and coefficient estimation error is revealed. We can use the results of the analysis to obtain an efficient design of the filter.
• Highly Accurate Designing Motion Table for Table Lookup Motion Compensation

  ASAMIZU Satoshi, HASEYAMA Miki  The IEICE transactions on information and systems Pt. 2  88-  (9)  1982  -1985  2005/09/01  [Not refereed][Not invited]
   

  ブロックマッチング法により求めた各フレームの予測誤差を重みとしてGAの評価関数に導入し, 符号化対象画像に必要な動きを探索する手法を提案する. 提案手法を用いることにより, フレームごとの予測画像の画質は, ばらつきが小さくなることを確認した.
• 動的輪郭モデルを用いた位相回復手法に関する一考察 (サマーセミナー(若葉研究者の発表会))

  近藤 啓子, 長谷山 美紀, 北島 秀夫  映像情報メディア学会技術報告  29-  (46)  1  -4  2005/08  [Not refereed][Not invited]
  
• 静止画像中に存在する消去テクスチャの再構成に関する検討--ERアルゴリズムにおける拘束条件の考察 (サマーセミナー(若葉研究者の発表会))

  小川 貴弘, 長谷山 美紀, 北島 秀夫  映像情報メディア学会技術報告  29-  (46)  5  -8  2005/08  [Not refereed][Not invited]
  
• オーディオインデキシングにおける特徴量の選択に関する一考察 (サマーセミナー(若葉研究者の発表会))

  二反田 直己, 長谷山 美紀, 北島 秀夫  映像情報メディア学会技術報告  29-  (46)  9  -12  2005/08  [Not refereed][Not invited]
  
• 音楽信号の低周波数域における楽曲の構成に着眼した時間周波数解析に関数する検討 (サマーセミナー(若葉研究者の発表会))

  今野 聡司, 二反田 直己, 長谷山 美紀  映像情報メディア学会技術報告  29-  (46)  13  -16  2005/08  [Not refereed][Not invited]
  
• チュートリアル講演 ユーザが望む映像を提供するために--画像認識とクラスタリングそして意味理解への発展 (サマーセミナー(若葉研究者の発表会))

  長谷山 美紀  映像情報メディア学会技術報告  29-  (47)  49  -52  2005/08  [Not refereed][Not invited]
  
• Convergence properties of a CORDIC-based adaptive ARMA lattice filter

  S Shiraishi, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E88A-  (8)  2154  -2164  2005/08  [Not refereed][Not invited]
   

  This paper presents a theoretical convergence analysis of a CORDIC-based adaptive ARMA lattice filter. In previous literatures, several investigation methods for adaptive lattice filters have been proposed; however, they are available only for AR-type filters. Therefore, we have developed a distinct technique that can reveal the convergence properties of the CORDIC ARMA lattice filter. The derived technique provides a quantitative convergence analysis, which facilitates an efficient hardware design for the filter. Moreover, our analysis technique can be applied to popular multiplier-based filters by slight modifications. Hence, the presented convergence analysis is significant as a leading attempt to investigate ARMA lattice filters.
• A New Discrete Clustering Method Based on Filtering Theory

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  IEICE technical report. WBS, Wideband System  105-  (177)  39  -44  2005/07/08  [Not refereed][Not invited]
   

  A new discrete clustering method is proposed. Its procedure is as follows. First, transpose feature space into discrete one by lattices and correspond each pattern to a lattice point. Second, move every point to its neighbor lattice point through lattices by gravitational force from other points under the concept of the static field, then suppose every point a sphere and coalesce touched ones. Repeat these until the number of points decreses to the desired, then class of a pattern is given by one of the coalesced. Distribution is transposed into another with which its kind is identified owing to short movements of points, but distribution coalescing needs a lot of steps. For this reason, adjust values of parameters to move points long but not to pass others. To show effectiveness of the proposed method, it is shown that standard of distribution transposing is unique. On movement of a point, firstly it is shown displacement of a point is given by convolution of force function and point distribution function. From this new viewpoint, secondly the force function is shown as filter and its window is not depended upon the parameters.
• A New Discrete Clustering Method Based on Filtering Theory

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  IEICE technical report. Signal processing  105-  (175)  39  -44  2005/07/08  [Not refereed][Not invited]
   

  A new discrete clustering method is proposed. Its procedure is as follows. First, transpose feature space into discrete one by lattices and correspond each pattern to a lattice point. Second, move every point to its neighbor lattice point through lattices by gravitational force from other points under the concept of the static field, then suppose every point a sphere and coalesce touched ones. Repeat these until the number of points decreses to the desired, then class of a pattern is given by one of the coalesced. Distribution is transposed into another with which its kind is identified owing to short movements of points, but distribution coalescing needs a lot of steps. For this reason, adjust values of parameters to move points long but not to pass others. To show effectiveness of the proposed method, it is shown that standard of distribution transposing is unique. On movement of a point, firstly it is shown displacement of a point is given by convolution of force function and point distribution function. From this new viewpoint, secondly the force function is shown as filter and its window is not depended upon the parameters.
• Adaptive Interpolation-Filter Assignment which Needs No Additional Bits Using Block Motion and Block Image Features

  TOKUNAGA Yoshiaki, HASEYAMA Miki, KITAJIMA Hideo  The transactions of the Institute of Electronics, Information and Communication Engineers. D-II  88-  (7)  1313  -1317  2005/07/01  [Not refereed][Not invited]
   

  本論文では動画像符号化における適応内挿フィルタ手法において, 動きと画像内容に応じてフィルタを適応的に選択する手法を提案する.提案手法はフィルタ選択情報を必要とせずに動き補償性能を高めることができ, 符号化効率向上を実現する.
• Sample size dependence of estimation error of information carried by neuronal spike train

  Jun Murakami, Tomohide Kidachi, Miki Haseyama, Tateo Shimozawa  Systems and Computers in Japan  36-  (7)  84  -95  2005/06/30  [Not refereed][Not invited]
   

  The operating performance of a neural system should be represented by the amount of information carried by neurons. The neuron behaves as a pulse encoder, which converts a stimulus waveform into a discrete pulse (spike) train. In practice, the rate of information transmission is estimated on the basis of a finite number of spikes obtained in a finite observation period. Consequently, the result includes an error. This paper considers the case in which the rate of information transmission is derived from the stimulus-response coherence, and analyzes the relation between the estimation error and the sample size (total number of spikes). It is found that the estimation error of information carried by a single spike is inversely proportional to the total number of spikes used for the estimation. It is pointed out that the asymptotic value of inverse proportion should be used for the comparison of information obtained from different neurons. The validity of the relation between the number of samples and the information per spike is verified by using data obtained from the cricket cercal sensory cells. © 2005 Wiley Periodicals, Inc.
• Pattern Recognition with Higher-Order Neural Network

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  Technical report of IEICE. PRMU  104-  (741)  7  -12  2005/03/10  [Not refereed][Not invited]
   

  In this paper, systems to recognize patterns expressed by binary point distributions are proposed. The feature extractor transforms a pattern into a new feature vector that expresses the structure of distribution by relative location of two points and relative distance between them. The feature vector is shift, rotation and scale invariant. The extractor is second order and its amount of calculation is little. The extractor also corrects deformation and distortion of a training pattern at transforming into a feature vector. Thus, the classifier learns also unknown distorted and deformed patterns of a training pattern with only one feature vector. Training period is short and recognition accuracy is high. The classifier is a third-order and two-layer neural network. General purpose of these systems are shown experimentally by hand-written character recognition and distribution identification.
• Pattern Recognition with Higher-Order Neural Network

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  Technical report of IEICE. Thought and language  104-  (739)  7  -12  2005/03/10  [Not refereed][Not invited]
   

  In this paper, systems to recognize patterns expressed by binary point distributions are proposed. The feature extractor transforms a pattern into a new feature vector that expresses the structure of distribution by relative location of two points and relative distance between them. The feature vector is shift, rotation and scale invariant. The extractor is second order and its amount of calculation is little. The extractor also corrects deformation and distortion of a training pattern at transforming into a feature vector. Thus, the classifier learns also unknown distorted and deformed patterns of a training pattern with only one feature vector. Training period is short and recognition accuracy is high. The classifier is a third-order and two-layer neural network. General purpose of these systems are shown experimentally by hand-written character recognition and distribution identification.
• Analysis of the Clustering Method DCM

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  Technical report of IEICE. KBSE  104-  (724)  31  -36  2005/03/07  [Not refereed][Not invited]
   

  The clustering method Dynamic Coalescence Model (DCM) is analyzed by filtering theory in this paper. Any point, that expresses a pattern in feature space, is moved by mutual gravitational attraction from other points, and a cluster, i.e., a set of points that are gathered, is let into a class of the points upon DCM. In this paper, it is clarified that the function is a filter. It is also done that the displacement of a point is calculated with the output of the filter at the point when point distribution is input. From these new points of view, it is shown that the Gaussian attraction is the most useful for clustering among the proposed ones.
• A-1-3 Design of transmission-line filters using genetic algorithms

  Sanada Hirofumi, Takezawa Megumi, Watanabe Kazuhisa, Haseyama Miki  Proceedings of the IEICE General Conference  2005-  3  -3  2005/03/07  [Not refereed][Not invited]
  
• A genetic algorithm for routing with an upper bound constraint

  J Inagaki, M Haseyama  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E88D-  (3)  679  -681  2005/03  [Not refereed][Not invited]
   

  This paper presents a method of searching for the shortest route via the most designated points with the length not exceeding the preset upper bound. The proposed algorithm can obtain the quasi-optimum route efficiently and its effectiveness is verified by applying the algorithm to the actual map data.
• Improvement of Detection of Between-the-Eyes with Circle-Frequency Filter

  HASEYAMA Miki, KANEKO Chiaki, KITAJIMA Hideo  The IEICE transactions on information and systems Pt. 2  88-  (3)  661  -664  2005/03/01  [Not refereed][Not invited]
   

  リング周波数フィルタは, 高速かつ高精度に眉間の位置を検出できるが, 前髪を眉の近くまでおろしている人の顔に対しては, 適用ができない.本論文では, このような場合にも, 適用を可能とするためにフィルタの適用領域を限定する前処理を提案する.
• Adaptive Interpolation Filter Selection Using Block Motion and Block Image Features : A Note on Adaptive Filter Window Design

  TOKUNAGA Yoshiaki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (9)  51  -54  2005/02/04  [Not refereed][Not invited]
  
• A Method of High-Quality Image Compression Using a Two-Dimensional FIR Filter

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (9)  55  -58  2005/02/04  [Not refereed][Not invited]
  
• A Note on Quality Estimation of Reconstructed Images by Fractal Image Coding

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  ITE technical report  29-  (9)  59  -64  2005/02/04  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system is one of the coding techniques for digital images. It compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the image quality obtained by the fractal image coding. However, the relations have not been clarified so far. Therefore, in this paper, we clarify the relations of these two factors, the image quality and the self-similarity, based on some simulations.
• Adaptive Interpolation Filter Selection Using Block Motion and Block Image Features : A Note on Adaptive Filter Window Design

  TOKUNAGA Yoshiaki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (647)  51  -54  2005/02/04  [Not refereed][Not invited]
   

  In recent motion-compensated video coding, an adaptive interpolation filter method that can achieve high prediction performance in motion compensation has been proposed. In this method, a horizontal or vertical filter is used to interpolate a pixel at a half-pel position. However, there should still be room for further prediction improvement because the method leaves design freedom for the filter window. From the point, we propose a method that changes the shape of the filter window with the moving object adaptively according to the motion vector. Finally, we show that the proposed method can reduce the total bit rate with the same PSNR as compared to the conventional method.
• A Method of High-Quality Image Compression Using a Two-Dimensional FIR Filter

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (647)  55  -58  2005/02/04  [Not refereed][Not invited]
   

  The video coding techniques such as MPEG-1 and MPEG-2 are widely used in video distribution over the Internet, DVD, etc. These video coding techniques achieve efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that blocking and ringing artifacts appear in the decoded images at low bit rates. Therefore, the proposed method utilizes a two-dimensional FIR filter to reduce these artifacts. The filter is to minimize the error between the original image and the coded image, and its coefficients are added into the bitstream. By applying the filter to the decoded image, the proposed method improves the decoded video quality.
• A Note on Quality Estimation of Reconstructed Images by Fractal Image Coding

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  電子情報通信学会技術研究報告. ITS  104-  (647)  59  -64  2005/02/04  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system is one of the coding techniques for digital images. It compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the image quality obtained by the fractal image coding. However, the relations have not been clarified so far. Therefore, in this paper, we clarify the relations of these two factors, the image quality and the self-similarity, based on some simulations.
• Adaptive Interpolation Filter Selection Using Block Motion and Block Image Features : A Note on Adaptive Filter Window Design

  TOKUNAGA Yoshiaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (649)  51  -54  2005/02/04  [Not refereed][Not invited]
   

  In recent motion-compensated video coding, an adaptive interpolation filter method that can achieve high prediction performance in motion compensation has been proposed. In this method, a horizontal or vertical filter is used to interpolate a pixel at a half-pel position. However, there should still be room for further prediction improvement because the method leaves design freedom for the filter window. From the point, we propose a method that changes the shape of the filter window with the moving object adaptively according to the motion vector. Finally, we show that the proposed method can reduce the total bit rate with the same PSNR as compared to the conventional method.
• A Method of High-Quality Image Compression Using a Two-Dimensional FIR Filter

  YAMAURA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (649)  55  -58  2005/02/04  [Not refereed][Not invited]
   

  The video coding techniques such as MPEG-1 and MPEG-2 are widely used in video distribution over the Internet, DVD, etc. These video coding techniques achieve efficient compression by using block-based DCT and motion compensation. Nevertheless, there is a problem that blocking and ringing artifacts appear in the decoded images at low bit rates. Therefore, the proposed method utilizes a two-dimensional FIR filter to reduce these artifacts. The filter is to minimize the error between the original image and the coded image, and its coefficients are added into the bitstream. By applying the filter to the decoded image, the proposed method improves the decoded video quality.
• A Note on Quality Estimation of Reconstructed Images by Fractal Image Coding

  TAKEZAWA Megumi, SANADA Hirofumi, WATANABE Kazuhisa, HASEYAMA Miki  IEICE technical report. Image engineering  104-  (649)  59  -64  2005/02/04  [Not refereed][Not invited]
   

  Fractal image coding based on an iterated function system is one of the coding techniques for digital images. It compresses images by utilizing their self-similarity. Hence, it is conceivable that the self-similarity is closely related to the image quality obtained by the fractal image coding. However, the relations have not been clarified so far. Therefore, in this paper, we clarify the relations of these two factors, the image quality and the self-similarity, based on some simulations.
• Audio Signal Segmentation and Classification using Fuzzy C-Means Clustering : A Study on Definition of Distance between Speech and Music Class

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  51  -56  2005/02/03  [Not refereed][Not invited]
  
• Tracking Pedestrian Responding to Changing Posture

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  29  -34  2005/02/03  [Not refereed][Not invited]
  
• A measurement of team advantage for soccer video analysis

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  35  -40  2005/02/03  [Not refereed][Not invited]
  
• A Target Tracking Method using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  41  -44  2005/02/03  [Not refereed][Not invited]
  
• Audio Signal Segmentation and Classification using Fuzzy C-Means Clustering : A Study on Definition of Distance between Speech and Music Class

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  51  -56  2005/02/03  [Not refereed][Not invited]
  
• A study on reconstruction of missing texture in still images Application of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  75  -80  2005/02/03  [Not refereed][Not invited]
  
• A Note on Soccer Video Indexing Using Neural Networks

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  93  -98  2005/02/03  [Not refereed][Not invited]
  
• A Study on Audio Signal Indexing of TV Soccer Program Using Fuzzy C-Means Clustering

  WATANABE Takashi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  99  -104  2005/02/03  [Not refereed][Not invited]
  
• A Study on Extraction of Irises Using Color and Shape Features

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  105  -110  2005/02/03  [Not refereed][Not invited]
   

  A method which applies the separability filter to facial images to extract the facial parts, whose shape is a circular form, such as the irises has been proposed. However, this filter extracts not only the irises but also the other facial parts such as nostrils. Therefore, this paper presents a method which estimates the region including the irises and extracts the irises from the obtained region by the separability filter. In the proposed method, a region including the eyes is estimated by using the features of the color of the eyes. Further, the proposed method calculates the location of the middle of the eyebrows from the output of the circle-frequency filter. According to the middle of the eyebrows, the region including the irises is estimated ; and then applying the separability filter to the obtained region, the irises can be accurately extracted.
• A Mobile Phone-Based Information-Providing System for Emergency Rescue

  SAKUMA Daisuke, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  29-  (8)  137  -142  2005/02/03  [Not refereed][Not invited]
  
• Tracking Pedestrian Responding to Changing Posture

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  29  -34  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a robust pedestrian tracking method against their postural change. The proposed method prepares different postural templates extracted from pedestrian images under several environments. Further, the posture of a pedestrian in a tracking target frame is estimated by using both the templates and the posture of the non-occluded pedestrian in the previous frames. In addition, the method judges whether the extracted pedestrian is occluded or not by evaluating the difference between the extracted pedestrian and the template corresponding to the posture. Therefore, the robust tracking can be realized without bad influences by the postural change. Experimental results are shown to verify the high performance of the proposed method.
• A measurement of team advantage for soccer video analysis

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  35  -40  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a measurement of team advantage, which represents the possibility of scoring in a soccer game. In the proposed method, actions of a ball carrier are modeled by a Markov chain, and by using this model, the team advantage can be measured. By using the team advantage, viewers of a soccer game can easily understand the game situations. Experimental results are shown to verify its high performance.
• A Target Tracking Method using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  41  -44  2005/02/03  [Not refereed][Not invited]
   

  As an object moves through the field of view of a camera, it undergoes changes in pose relative to the viewing camera, in illumination relative to light sources, and may even become partially occluded. This causes erroneous object-tracking results. Although methods which try to suppress the effect using previous tracking results have been proposed, the improvement of the tracking accuracy is limited in the case where the change of the pose occurs. To solve this problem, this paper proposes a method using a shape-constraint-based snake with a splitting mechanism. This snake can extract the target object accurately by using a reference contour which approximates the contour of the object, even if a part of the contour can not be obtained clearly. In the proposed method, the approximate 3-D shape of the target object is assumed to be given, and the reference contour is obtained from the pose estimated by comparing the previous result and the approximate shape. Consequently, the proposed method can accurately track moving objects in the image sequence even if they undergo changes in the environment.
• Audio Signal Segmentation and Classification using Fuzzy C-Means Clustering : A Study on Definition of Distance between Speech and Music Class

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告  104-  (648)  51  -56  2005/02/03  [Not refereed][Not invited]
   

  Automatic segmentation and classification technique of audio signal is required for audiovisual indexing, and we have been proposed an audio signal segmentation and classification method. This method segments the audio signal into different audio signals at their boundaries, and classifies them into five audio classes, which are silence, speech, music, speech with music, and speech with noise. This paper defines a distance between speech and music class in order to judge that a speech with music class is similar to which speech or music class. The proposed method consists of three steps : (1) audio features, which represent the characteristic of speech, music, and speech with music signal, are extracted ; (2) principal component analysis is applied to the extracted audio features ; (3) fuzzy c-means clustering is applied to the principal components, and distance can be computed by using membership values, which are obtained from fuzzy clustering. Experimental results performed by applying the proposed method to real audio signal are shown to verify its high performance.
• A study on reconstruction of missing texture in still images Application of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  75  -80  2005/02/03  [Not refereed][Not invited]
   

  Several methods for reconstruction of missing textures in still images have been studied. However, they cannot restore the missing areas without suffering from degradations, so that their performance is not enough for accurate restoration. Therefore, this paper proposes an accurate reconstruction method using the ER (Error-Reduction) algorithm. Noting a feature of texture images, the proposed method can calculate the Fourier transform magnitude of the area including missing pixels from the area whose texture is similar in the target image. The similar texture is selected by a novel approach monitoring the errors of the ER algorithm. Further, the proposed method recovers the phase of the target area from the obtained Fourier transform magnitude by using the ER algorithm modified for the restoration and can accurately reconstruct the missing textures. Some experimental results are shown to verify its high performance of the proposed method.
• A Note on Soccer Video Indexing Using Neural Networks

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  93  -98  2005/02/03  [Not refereed][Not invited]
   

  For the soccer video indexing, long shots, which are mostly occupied by the soccer field, are frequently utilized. We have already proposed a method to detect the long shots by focusing on the occupancy rate of the soccer field, the size of the players included in the field, etc.. However, before using this method we have to appropriately set some thresholds depending on the data acquisition conditions, the soccer field conditions, etc., and they cannot be easily done. Therefore, in order to avoid this problem we propose a new detection method of the long shots by utilizing neural networks. The proposed method has the advantage that there is no need to set the thresholds manually and can detect the long shots more accurately than our conventional method does.
• A Study on Audio Signal Indexing of TV Soccer Program Using Fuzzy C-Means Clustering

  WATANABE Takashi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  99  -104  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes an indexing method of audio signals, which are obtained from TV soccer programs. Generally, cheers can be occurred at the event scene, such as the shoot scene, the foul scene, etc., and these cheers are useful for the event scene detection. However, the audio signal of the TV soccer program is a mixture of cheers, announcer's voice, and so on, and thereby it is difficult to obtain only cheers. Therefore, the proposed method defines a similarity between a soccer audio signal and a cheer signal. This similarity is computed by using memberships, which are obtained from fuzzy c-means clustering. By using this similarity, the event scene can be detected accurately.
• A Study on Extraction of Irises Using Color and Shape Features

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  105  -110  2005/02/03  [Not refereed][Not invited]
   

  A method which applies the separability filter to facial images to extract the facial parts, whose shape is a circular form, such as the irises has been proposed. However, this filter extracts not only the irises but also the other facial parts such as nostrils. Therefore, this paper presents a method which estimates the region including the irises and extracts the irises from the obtained region by the separability filter. In the proposed method, a region including the eyes is estimated by using the features of the color of the eyes. Further, the proposed method calculates the location of the middle of the eyebrows from the output of the circle-frequency filter. According to the middle of the eyebrows, the region including the irises is estimated ; and then applying the separability filter to the obtained region, the irises can be accurately extracted.
• A Mobile Phone-Based Information-Providing System for Emergency Rescue

  SAKUMA Daisuke, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  104-  (646)  137  -142  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a system to provide information about emergency rescue procedures for mobile phone users. In the proposed system, avatars demonstrate the procedures, so that the users understand them easily. The avatars are efficiently represented in a scalable vector graphics (SVG) format, and they are transmitted to the users in a shorter time. In addition to the avatars, the proposed system provides voice and text explaining the procedures. The users receive these multimedia data in an integrated form based on the synchronized multimedia integration language (SMIL, pronounced "smile"), and play them readily on their terminals. In this paper, a SMIL player for a mobile phone is prepared to verify the functionality of the proposed system.
• Tracking Pedestrian Responding to Changing Posture

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  29  -34  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a robust pedestrian tracking method against their postural change. The proposed method prepares different postural templates extracted from pedestrian images under several environments. Further, the posture of a pedestrian in a tracking target frame is estimated by using both the templates and the posture of the non-occluded pedestrian in the previous frames. In addition, the method judges whether the extracted pedestrian is occluded or not by evaluating the difference between the extracted pedestrian and the template corresponding to the posture. Therefore, the robust tracking can be realized without bad influences by the postural change. Experimental results are shown to verify the high performance of the proposed method.
• A measurement of team advantage for soccer video analysis

  KON Hirofumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  35  -40  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a measurement of team advantage, which represents the possibility of scoring in a soccer game. In the proposed method, actions of a ball carrier are modeled by a Markov chain, and by using this model, the team advantage can be measured. By using the team advantage, viewers of a soccer game can easily understand the game situations. Experimental results are shown to verify its high performance.
• A Target Tracking Method using a Shape-Constraint-Based Active Contour Model with a Splitting Mechanism

  YOKOYAMA Yukinori, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  41  -44  2005/02/03  [Not refereed][Not invited]
   

  As an object moves through the field of view of a camera, it undergoes changes in pose relative to the viewing camera, in illumination relative to light sources, and may even become partially occluded. This causes erroneous object-tracking results. Although methods which try to suppress the effect using previous tracking results have been proposed, the improvement of the tracking accuracy is limited in the case where the change of the pose occurs. To solve this problem, this paper proposes a method using a shape-constraint-based snake with a splitting mechanism. This snake can extract the target object accurately by using a reference contour which approximates the contour of the object, even if a part of the contour can not be obtained clearly. In the proposed method, the approximate 3-D shape of the target object is assumed to be given, and the reference contour is obtained from the pose estimated by comparing the previous result and the approximate shape. Consequently, the proposed method can accurately track moving objects in the image sequence even if they undergo changes in the environment.
• Audio Signal Segmentation and Classification using Fuzzy C-Means Clustering : A Study on Definition of Distance between Speech and Music Class

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  51  -56  2005/02/03  [Not refereed][Not invited]
   

  Automatic segmentation and classification technique of audio signal is required for audiovisual indexing, and we have been proposed an audio signal segmentation and classification method. This method segments the audio signal into different audio signals at their boundaries, and classifies them into five audio classes, which are silence, speech, music, speech with music, and speech with noise. This paper defines a distance between speech and music class in order to judge that a speech with music class is similar to which speech or music class. The proposed method consists of three steps : (1) audio features, which represent the characteristic of speech, music, and speech with music signal, are extracted ; (2) principal component analysis is applied to the extracted audio features ; (3) fuzzy c-means clustering is applied to the principal components, and distance can be computed by using membership values, which are obtained from fuzzy clustering. Experimental results performed by applying the proposed method to real audio signal are shown to verify its high performance.
• A study on reconstruction of missing texture in still images Application of ER algorithm

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  75  -80  2005/02/03  [Not refereed][Not invited]
   

  Several methods for reconstruction of missing textures in still images have been studied. However, they cannot restore the missing areas without suffering from degradations, so that their performance is not enough for accurate restoration. Therefore, this paper proposes an accurate reconstruction method using the ER (Error-Reduction) algorithm. Noting a feature of texture images, the proposed method can calculate the Fourier transform magnitude of the area including missing pixels from the area whose texture is similar in the target image. The similar texture is selected by a novel approach monitoring the errors of the ER algorithm. Further, the proposed method recovers the phase of the target area from the obtained Fourier transform magnitude by using the ER algorithm modified for the restoration and can accurately reconstruct the missing textures. Some experimental results are shown to verify its high performance of the proposed method.
• A Note on Soccer Video Indexing Using Neural Networks

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  93  -98  2005/02/03  [Not refereed][Not invited]
   

  For the soccer video indexing, long shots, which are mostly occupied by the soccer field, are frequently utilized. We have already proposed a method to detect the long shots by focusing on the occupancy rate of the soccer field, the size of the players included in the field, etc.. However, before using this method we have to appropriately set some thresholds depending on the data acquisition conditions, the soccer field conditions, etc., and they cannot be easily done. Therefore, in order to avoid this problem we propose a new detection method of the long shots by utilizing neural networks. The proposed method has the advantage that there is no need to set the thresholds manually and can detect the long shots more accurately than our conventional method does.
• A Study on Audio Signal Indexing of TV Soccer Program Using Fuzzy C-Means Clustering

  WATANABE Takashi, NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  99  -104  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes an indexing method of audio signals, which are obtained from TV soccer programs. Generally, cheers can be occurred at the event scene, such as the shoot scene, the foul scene, etc., and these cheers are useful for the event scene detection. However, the audio signal of the TV soccer program is a mixture of cheers, announcer's voice, and so on, and thereby it is difficult to obtain only cheers. Therefore, the proposed method defines a similarity between a soccer audio signal and a cheer signal. This similarity is computed by using memberships, which are obtained from fuzzy c-means clustering. By using this similarity, the event scene can be detected accurately.
• A Study on Extraction of Irises Using Color and Shape Features

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  105  -110  2005/02/03  [Not refereed][Not invited]
   

  A method which applies the separability filter to facial images to extract the facial parts, whose shape is a circular form, such as the irises has been proposed. However, this filter extracts not only the irises but also the other facial parts such as nostrils. Therefore, this paper presents a method which estimates the region including the irises and extracts the irises from the obtained region by the separability filter. In the proposed method, a region including the eyes is estimated by using the features of the color of the eyes. Further, the proposed method calculates the location of the middle of the eyebrows from the output of the circle-frequency filter. According to the middle of the eyebrows, the region including the irises is estimated ; and then applying the separability filter to the obtained region, the irises can be accurately extracted.
• A Mobile Phone-Based Information-Providing System for Emergency Rescue

  SAKUMA Daisuke, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (648)  137  -142  2005/02/03  [Not refereed][Not invited]
   

  This paper proposes a system to provide information about emergency rescue procedures for mobile phone users. In the proposed system, avatars demonstrate the procedures, so that the users understand them easily. The avatars are efficiently represented in a scalable vector graphics (SVG) format, and they are transmitted to the users in a shorter time. In addition to the avatars, the proposed system provides voice and text explaining the procedures. The users receive these multimedia data in an integrated form based on the synchronized multimedia integration language (SMIL, pronounced "smile"), and play them readily on their terminals. In this paper, a SMIL player for a mobile phone is prepared to verify the functionality of the proposed system.
• Audio Signal Segmentation and Classification Using Fuzzy Clustering

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  The IEICE transactions on information and systems Pt. 2  88-  (2)  302  -312  2005/02/01  [Not refereed][Not invited]
   

  本論文では, ファジーc-means法を用いてMPEG Audioで符号化されたオーディオ信号から, オーディオ信号の切り換わる時刻(オーディオカット)を検出し, それらを境界とするセグメント(オーディオセグメント)の分類を行う手法を提案する.従来の手法は, オーディオ信号から得られるエネルギーや零交差数等の特徴量に対し, しきい値処理を用いてオーディオカットを検出するため, フェードインやフェードアウト, クロスフェード等の音響効果が施された場合, 検出精度が低下する危険性があった.提案手法は, ファジーc-means法の導入により, その時刻にオーディオカットが存在する可能性を0から1までの実数で表現し, 存在する可能性がある時刻をオーディオカットの候補として検出する.このとき, 従来手法では検出が困難であった音響効果もオーディオカット候補として検出され, その結果, オーディオカットの未検出が軽減される.更に, それらの候補を境界とするオーディオセグメントを無音, 音声, 音楽, 音楽付き音声, 雑音付き音声の五つのクラスに分類し, これらの分類結果と併せて, 候補中の不要なオーディオカットを削除することにより, 高精度なオーディオカット検出を実現すると同時に, オーディオセグメントの分類を完了する.
• A SIMPLE WORD SPOTTING METHOD BASED ON TEMPLATE MATCHING FOR SPEECH RETRIEVAL

  KOMAKI Noriko, HASEYAMA Miki, YAMAMOTO Tsuyoshi  IEICE technical report. Image engineering  104-  (544)  61  -66  2005/01/03  

  This paper proposes a simple word spotting method for speech retrieval used for indexing and digesting for audio-visual contents. The proposed method retrieves desired words by using template-matching based on simple waveform features. Since the features can be computed with low computational costs, the proposed method executes the speech retrieval at high speed, which is actually 1/200 of the real-time speed. It is helpful in retrieving huge audio-visual database. Also experiments are performed by applying the proposed method to real audio signals, and the results are shown to verify its retrieval accuracy.
• A REGION MERGING METHOD FOR IMAGE SEGMENTATION

  ZHAO Yanjun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  104-  (544)  79  -83  2005/01/03  [Not refereed][Not invited]
   

  We propose in this paper a new merging method for image segmentation that reduces the number of generations over which intermediate, temporary results propagate. The proposed method consists of three major steps : preparation, batch merging and asynchronous update. In the step of preparation, the colors of two neighboring regions are compared. If their distance is smaller than a threshold, they are issued a merge label. The comparison is repeated for all neighboring pairs of regions. In the step of batch merging, a pair carrying a merge label is actually merged to create a new region. Some of the old merge labels will have to be cancelled according to the merged regions. In the step of asynchronous update, the merge labels between the new regions and their neighboring regions are updated. The steps of batch merging and asynchronous update will be repeated until there is no merge label.
• Accurate audio-segment classification using feature extraction matrix

  N Nitanda, M Haseyama, H Kitajima  2005 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1-5  III-  261  -264  2005  [Not refereed][Not invited]
   

  This paper proposes an accurate audio signal classification method using feature extraction matrix. The proposed method classifies the segments of the audio signal into the following five audio classes: silence, speech, music, speech with music background, and speech with noise background. In this classification, a diagonal matrix, which is called feature extraction matrix, is utilized in order to extract the effective audio features for the classification. By using this feature extraction matrix, the five audio classes are clearly separated each other in the feature space, and thereby highly precise classification can be attained. Experimental results performed by applying the proposed method to real audio signals are shown to verify its high performance.
• An algorithm for extracting lines directly from the input image to improve speed

  John W. Gates, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  59-  (2)  287  -295  2005  [Not refereed][Not invited]
   

  This paper presents a fast and accurate line extraction algorithm. Conventional line extraction algorithms first extract the edge pixels and then connect these edge pixels into lines. In contrast the proposed algorithm extracts the lines directly from the input image by tracing the perimeter of an intensity region and then uses simple linear regression to compute the equations of the lines. This proposed method allows the algorithm to use the intensity information in the image to correlate the edge pixels, thus simplifying the computation of the lines and increasing the robustness in the presence of noise. Experiments were performed on a variety of synthetic and natural images and demonstrated that the algorithm can extract lines reliably, while realizing a large improvement in speed over the conventional algorithms.
• The extraction of circles from arcs represented by extended digital lines

  Euijin Kim, Miki Haseyama, Hideo Kitajima  IEICE Transactions on Information and Systems  E88-D-  (2)  252  -267  2005  [Not refereed][Not invited]
   

  This paper presents a new fast and robust circle extraction method that is capable of extracting circles from images with complicated backgrounds. It is not based on the Hough transform (HT) that requires a time-consuming voting process. The proposed method uses a least-squares circle fitting algorithm for extracting circles. The arcs are fitted by extended digital lines that are extracted by a fast line extraction method. The proposed method calculates accurate circle parameters using the fitted arcs instead of evidence histograms in the parameter space. Tests performed on various real-world images show that the proposed method quickly and accurately extracts circles from complicated and heavily corrupted images. Copyright © 2005 The Institute of Electronics, Information and Communication Engineers.
• A new conic section extraction approach and its applications

  John Gates, Miki Haseyama, Hideo Kitajima  IEICE Transactions on Information and Systems  E88-D-  (2)  239  -251  2005  [Not refereed][Not invited]
   

  This paper presents a new conic section extraction approach that can extract all conic sections (lines, circles, ellipses, parabolas and hyperbolas) simultaneously. This approach is faster than the conventional approaches with a computational complexity that is O(n), where n is the number of edge pixels, and is robust in the presence of moderate levels of noise. It has been combined with a classification tree to produce an offline character recognition system that is invariant to scale, rotation, and translation. The system was tested with synthetic images and with images scanned from real world sources with good results. Copyright © 2005 The Institute of Electronics, Information and Communication Engineers.
• A robust human-eye tracking method in video sequences

  M Haseyama, C Kaneko  2005 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), VOLS 1-5  2-  1445  -1448  2005  [Not refereed][Not invited]
   

  An accurate tracking method of the human eyes in video sequence is proposed. The method consists of the following two systems. (1) The first system extracts a region including the both eyes by using a statistical processing. By the processing, the extracted region does not include the other features, especially such as the hair. (2) The second system, which is realized with the circle-frequency filter (CFF), locates each eye position from the region extracted by the first system. Since the output of the CFF is robust to luminescent noise near the eyes, it can successfully obtain the exact eye-location inside of the above region without any noise effect. Consequently, the combination of these two systems makes the proposed method correctly track the eyes without any learning schemes, templates, and geometric relations of the other facial features. Experimental results verify that the proposed method achieves accurate eye-tracking.
• Reconstruction method of missing texture using error reduction algorithm

  T Ogawa, M Haseyama, H Kitajima  2005 International Conference on Image Processing (ICIP), Vols 1-5  2-  1389  -1392  2005  [Not refereed][Not invited]
   

  This paper presents a novel reconstruction method of missing textures using an error reduction algorithm which is one of phase retrieval methods. The proposed method estimates the Fourier transform magnitude of the missing area from another area whose texture is similar in the obtained image. In order to realize this, a novel approach that monitors the errors caused by the error reduction algorithm is introduced into the selection scheme of the similar texture. Further, the proposed method estimates the phase of the target area by using the error reduction algorithm modified for the texture reconstruction and can restore the missing area accurately. Some experimental results show that the proposed method achieves more accurate restoration than that of the traditional methods.
• GA-based applications for routing with an upper bound constraint

  Jun Inagaki, Miki Haseyama  Proceedings - IEEE International Symposium on Circuits and Systems  2239  -2242  2005  [Not refereed][Not invited]
   

  This paper presents a method of searching for the shortest route via the most designated points among the routes whose lengths are less than the upper bound using a genetic algorithm (GA). If chromosomes whose route lengths exceed the upper bound are simply screened out in the GA process, the optimization probability gets worse. For the purpose of solving this problem, this paper proposes a new fitness function including an upper bound constraint which can be flexibly changed during the searching process. By using this function, the optimum is efficiently obtained and the optimization probability can be raised. Furthermore, the effectiveness of the proposed method is verified by the experiments applying it to the actual map data. © 2005 IEEE.
• Quality improvement technique for JPEG images with fractal image coding

  Megumi Takezawa, Hirofumi Sanada, Kazuhisa Watanabe, Miki Haseyama  Proceedings - IEEE International Symposium on Circuits and Systems  6320  -6323  2005  [Not refereed][Not invited]
   

  This paper proposes a quality improvement technique for JPEG images by using fractal image coding. JPEG coding is a commonly used standard method of compressing images. However, in its decoded images, quantization noise is sometimes visible in high frequency regions, such as the edges of objects. Hence, in order for the JPEG coding to become a more powerful image-coding method, the JPEG image quality must be improved. Therefore, our method solves this problem by adding the obtained codes by the fractal image coding to improve the image quality. Some simulation results verify that the proposed method achieved higher coding-performance than the traditional JPEG coding. © 2005 IEEE.
• Audio signal segmentation and classification for scene-cut detection

  Naoki Nitanda, Miki Haseyama, Hideo Kitajima  Proceedings - IEEE International Symposium on Circuits and Systems  4030  -4033  2005  [Not refereed][Not invited]
   

  A scene is regarded as a basic unit of audiovisual material, and thereby the boundaries between two adjacent scenes, which are called scene-cuts, must be detected in advance for audiovisual indexing. This paper proposes a scenecut detection method. Since scene-cuts are associated with a simultaneous change of visual and audio characteristics, both audio and visual analyses are required for the scene-cut detection. For the audio signal analysis, the proposed method utilizes an audio signal segmentation and classification method using fuzzy c-means clustering, which has been proposed by the authors. For the visual signal analysis, the proposed method utilizes some visual segmentation methods. By using these methods simultaneously, the proposed method can accurately detect the scene-cuts, and thereby it is highly valuable for the preprocessing of the audiovisual indexing. Experimental results performed by applying the proposed method to real audiovisual material are shown to verify its high performance. © 2005 IEEE.
• Restoration method of missing areas in still images using GMRF model

  Takahiro Ogawa, Miki Haseyama, Hideo Kitajima  Proceedings - IEEE International Symposium on Circuits and Systems  4931  -4934  2005  [Not refereed][Not invited]
   

  This paper proposes a GMRF-model based restoration method of missing areas in still images. The GMRF model used in the proposed method is realized by a new assumption that reasonably holds for an image source. This model can express important image features such as edges because of the use of the new assumption. Therefore, the proposed method restores the missing areas using the modified GMRF model and can correctly reconstruct the missing edges. Consequently, the proposed method achieves more accurate restoration than those of the traditional methods on both objective and subjective measures. Extensive experimental results demonstrate the improvement of the proposed method over the previous methods. © 2005 IEEE.
• Effective apoptotic cell extraction from video microscopy images

  M Haseyama, Y Sasamura  2005 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1-5  II-  461  -464  2005  [Not refereed][Not invited]
   

  This paper presents an automatic extraction method for apoptotic cells from video microscopy images. Apoptosis, which is also called programmed cell death, is well-known to be closely related with various diseases. Detection of apoptotic cells can help scientists to study apoptosis and to reveal its mechanism, so that new medical technologies can be exploited. Scientists often analyze the images acquired with video microscopes with some sort of image analysis tools. However, it is usual that the performance of the tools is not enough to thoroughly detect the apoptotic cells, thus the researchers have to manually detect the apoptotic cells from the images. This paper proposes a method to automatically detect the apoptotic cells based on their inherent shape features. The proposed method is applied to real world video microscopy images containing apoptotic cells and the results demonstrate its high performance to successfully extract the cells.
• Moving object extraction using a shape-constraint-based splitting active contour model

  M Haseyama, Y Yokoyama  2005 International Conference on Image Processing (ICIP), Vols 1-5  3-  2681  -2684  2005  [Not refereed][Not invited]
   

  This paper proposes an efficient moving object extraction method based on an active contour model, which is usually called a Snake. The previous Snakes cannot extract multiple objects inside one contour, since they require that one initial contour can only include one object. Such a restriction becomes a problem for moving object extraction application, because an initial contour may contain more than one object including not only the extraction target but also other objects. To deal with this kind of problem, the proposed method utilizes a new splitting mechanism. By incorporating this mechanism, the proposed Snake can successfully extract each of the multiple objects located inside one initial contour. The experiments of extracting moving vehicles from actual image sequences verify the effectiveness and high performance of the proposed method.
• An audio signal segmentation and classification using fuzzy c-means clustering

  Naoki Nitanda, Miki Haseyama, Hideo Kitajima  Proceedings of the Second International Conference on Information Technology and Applications (ICITA 2004)  332  -337  2004/12/27  [Not refereed][Not invited]
   

  This paper proposes an audio signal segmentation and classification method using fuzzy c-means clustering. Recently, high performance of the audio signal segmentation and classification is required for audio-visual indexing because of the popular use of the Internet, higher bandwidth access to the network, widespread of digital recording and storage; and several methods have been proposed. They segment the audio signal at boundaries between two different audio signals, which are called audio-cuts, and then classify the audio signal into basic audio classes such as speech, music, etc. However, since most of the methods utilize thresholding for the audio-cut detection, they cannot provide high accuracy because of several audio effects, such as fade-In, fade-out, cross-fade, etc. To overcome this problem, we utilize the fuzzy c-means clustering. The possibility that the audio-cut exists is represented by the fuzzy number, and thus we can detect audio-cuts accurately. After the segmentation, the audio signal is classified into audio classes. This classification results are utilized for verification processing of the audio-cuts, so that segmentation and classification errors are reduced. Experimental results performed by applying the proposed method to real audio signals are shown to verify its high performance.
• An Extraction Method for Apoptotic Cells from Videomicroscopy Images

  HASEYAMA Miki, SASAMURA Yukari, KITAJIMA Hideo  The IEICE transactions on information and systems Pt. 2  87-  (11)  2045  -2053  2004/11/01  [Not refereed][Not invited]
   

  ビデオ顕微鏡により得られた画像からアポトーシスに至る細胞のみを自動抽出する手法を提案する.ビデオ顕微鏡により取得された細胞像は,観察の際に使用される蛍光酵素によって,エッジが緩やかで太く,その形状は円形に類似しているものの各所にゆがみをもち,かつ,対象細胞以外の細胞が隣接して存在する.そのため,容易に自動検出することができず,その解析は目視に依存していた.そこで,本論文では,このような観察条件と固有の形状に適した形状解析を行うことによって,所望の細胞のみを検出する手法を提案する.また,提案手法は,画像全体に詳細な形状解析を施すのではなく,検出対象細胞を合む可能性がある領域を検出し,検出された領域の内部にのみ詳細な形状解析を行う.このような2段階の処理により,計算量の増加を避けながら,高い検出精度を維持する.また,時間軸上に連続した動画像として観察された場合に,各フレームの検出結果を比較することによって,抽出精度の更なる向上を図る.本論文の最後では実際に得られた顕微鏡像を用いて,アポトーシスに至る細胞の抽出を行い,提案手法の有効性について検討する.
• Restoration for Missing Intensity of Still Images Using the Optical Flow

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics, Information and Communication Engineers  87-  (9)  1786  -1795  2004/09/01  [Not refereed][Not invited]
   

  論文では,オプティカルフローを用いた静止画像の輝度値の高精度な復元手法を提案する.画像の失われた輝度値の復元において,エッジ部を正確に再構成することは重要な課題である.そこで本論文では,従来より動画像の動き解析に利用されているオプティカルフローを,静止画像に適用可能な形に変更し,これを用いることで輝度値の高精度な復元を行う.また,より高精度な推定輝度値を与えるオプティカルフローを算出するために,提案手法では画素間の輝度値の相関に着眼した新たな指標を導入する.この指標を用いて算出されるオプティカルフローは輝度値が失われた画素の近傍に存在する画素から,より真の輝度値に類似した画素を選択することを可能とし,輝度値が大きく異なる画素の影響を受けずに復元できる.これにより,特にエッジ部分における復元精度が向上するだけでなく,従来の復元手法では困難であった輝度値の失われた領域を複数のエッジが通過する場合や,通過するエッジの方向がその領域内で大きく変化する場合にも,正確に失われたエッジの再構成を行うことが可能となる.本論文の最後では実験を行い,提案手法の有効性を示す.
• I-023 Head Detection of a Pedestrian Using a Variable Observation Window

  Kawase Shinji, Kim Euijin, Fujii Toshifumi, Haseyama Miki, Kitajima Hideo  情報科学技術フォーラム一般講演論文集  3-  (3)  49  -51  2004/08/20  [Not refereed][Not invited]
  
• Quasi-realization of mechanism of Dynamic Coalescence Model upon Higher-Order Neural Network

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  Technical report of IEICE. HIP  103-  (744)  19  -24  2004/03/19  [Not refereed][Not invited]
   

  We propose a method to improve recognition accuracy of a system - that is constructed by a system to translate a pattern into a feature vector to express a pattern from the view of static characteristics of clustering method DCM (Dynamic Coalescence Model), a system to create similar and unknown training patterns from dynamic characteristics of DCM, and a modified discrimination system HONN (Higher-Order Neural Network) - with a small number of training pattern(s). Our proposed method with theoretical grounds easily creates vectors of each training pattern by focusing two points wide apart on it without exception. Furthermore, they express various characteristics of each given training pattern, thus our proposed method is superior to plural applying of the system to create similar and unknown training patterns with different values of parameters. Effectiveness of our proposed method is examined experimentally by distribution identification and handwritten character recognition.
• D-14-12 Study on Word Spotting Method for Continuous Speech

  Komaki Noriko, Haseyama Miki, Yamamoto Tsuyoshi  Proceedings of the IEICE General Conference  2004-  (1)  153  -153  2004/03/08  [Not refereed][Not invited]
  
• A cost-effective CORDIC-based architecture for adaptive lattice filters

  S Shiraishi, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E87A-  (3)  567  -576  2004/03  [Not refereed][Not invited]
   

  This paper presents a cost-effective CORDIC-based architecture for adaptive lattice filters. An implementation method for an ARMA lattice filter using the CORDIC algorithm has been proposed. The previously proposed method can provide a simple filter architecture; however, it has problems such as redundant structure and numerical inaccuracy. Therefore, by solving each problem we derive a new non-redundant filter architecture with improved numerical accuracy. The obtained filter architecture provides a low cost ARMA lattice filter in which high-precision data processing is feasible. In addition, the proposed architecture can be applied to AR-type lattice filters, so that it may have several applications in adaptive signal processing. The presented filter architecture is useful from a hardware point of view because it facilitates an effective VLSI design of various adaptive lattice filters.
• A significant property of mapping parameters for signal interpolation using fractal interpolation functions

  S Uemura, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E87A-  (3)  748  -752  2004/03  [Not refereed][Not invited]
   

  This letter presents a significant property of the mapping parameters that play a central role to represent a given signal in Fractal Interpolation Functions (FIF). Thanks to our theoretical analysis, it is derived that the mapping parameters required to represent a given signal are also applicable to represent the upsampled signal of a given one. Furthermore, the upsampled signal obtained by using the property represents the self-affine property more distinctly than the given signal. Experiments show the validity and usefulness of the significant property.
• An Application of a Circle-Frequency Filter for Detection of Human Eyes

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (8)  1  -6  2004/02/04  [Not refereed][Not invited]
  
• A Note on Texture Discrimination by Using Scale-Space Filtering

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (8)  25  -30  2004/02/04  [Not refereed][Not invited]
  
• A Note on Tracking Pedestrian with SVM

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (8)  45  -50  2004/02/04  [Not refereed][Not invited]
  
• Calculating the Error Produced by Rotational and Translational Changes of a Stereo Camera

  WAKI Hirohisa, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (8)  95  -100  2004/02/04  [Not refereed][Not invited]
  
• A Note on Solving Method for Nurse Scheduling Problem based on the Group Decision Making

  NAKAJIMA Jun, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (7)  43  -48  2004/02/03  [Not refereed][Not invited]
  
• A Note on Audiovisual Indexing for Scene-Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (7)  119  -124  2004/02/03  [Not refereed][Not invited]
  
• A Note on Long Shot Detection in Soccer Video

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (7)  125  -130  2004/02/03  [Not refereed][Not invited]
  
• A Note on Quality Improvement of Color JPEG Images

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (6)  7  -12  2004/02/02  [Not refereed][Not invited]
  
• A Study on Restoration of Missing Areas Application of GMRF Model

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  28-  (6)  19  -24  2004/02/02  [Not refereed][Not invited]
  
• A novel contour description with expansion ability using extended fractal interpolation functions

  S Uemura, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E87D-  (2)  453  -462  2004/02  [Not refereed][Not invited]
   

  In this paper, a novel description method of the contour of a shape using extended fractal interpolation functions (EFIFs) is presented. Although the scope of application of traditional FIFs has been limited to cases in which a given signal is represented by a single-valued function, the EFIFs derived by the introduction of a new parameter can describe a multiple-valued signal such as the contour of a shape with a high level of accuracy. Furthermore, the proposed description method possesses the useful property that once a given contour has been modeled by the proposed description method, the shape can be easily expanded at an arbitrary expansion rate. Experimental results show the effectiveness and usefulness of the proposed description method for representing contours.
• An Application of a Circle-Frequency Filter for Detection of Human Eyes

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (641)  1  -6  2004/01/28  [Not refereed][Not invited]
   

  A method has been proposed to detect the middle of the eyebrows by applying the circle-frequency filter to facial images. This paper proposes a method to detect the location of the eyes by using the filter. In the proposed method, a region including the eyes is estimated by using the color, and then the location of the middle of the eyebrows and the rotation angle of the face are calculated by using the output of the circle-frequency filter. According to the obtained location of the middle of the eyebrows and the rotation angle of the face, the location of the eyes can be accurately detected without using learning patterns and templates.
• A Note on Texture Discrimination by Using Scale-Space Filtering

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (641)  25  -30  2004/01/28  [Not refereed][Not invited]
   

  A method to extract texture features that are robust against blurring or additive noise by using scale-space filtering has been proposed. In the method, first, the images smoothed at various scales are obtained by applying scale-space filtering to a given texture image. Secondly the smoothed images representing a given texture image are selected. Lastly texture features are extracted from this selected images. However, there are no criteria to select images which the texture features are effectively extracted from. We, therefore, propose a novel method to select such an image by observing normal vectors of the curved surface corresponding to the intensity of the smoothed images obtained by applying the scale-space filtering.
• A Note on Tracking Pedestrian with SVM

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (641)  45  -50  2004/01/28  [Not refereed][Not invited]
   

  This paper proposes an accurate tracking method of pedestrians by using a Support Vector Machine (SVM). The SVM included in the proposed method judges whether a target pedestrian is occluded or not. The method extracts the targets by using the templates which are made from the pedestrians extracted from the previous frame where they were not occluded. Therefore, the accurate tracking is realized even if the pedestrians are occluded. Experimental results are shown to verify the performance of the proposed method.
• Calculating the Error Produced by Rotational and Translational Changes of a Stereo Camera

  WAKI Hirohisa, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (641)  95  -100  2004/01/28  [Not refereed][Not invited]
   

  This paper calculates the errors that are caused when measuring the exact position of an object by using an onboard stereo camera. In order to measure the exact position of an object by using such a camera, precise camera calibration is required. However, aging of the cameras and vibrations that occur when driving have a bad influence on the calibration. Therefore, it is important to calculate the errors that are caused by such aging and vibration.
• An Application of a Circle-Frequency Filter for Detection of Human Eyes

  KANEKO Chiaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (644)  1  -6  2004/01/28  [Not refereed][Not invited]
   

  A method has been proposed to detect the middle of the eyebrows by applying the circle-frequency filter to facial images. This paper proposes a method to detect the location of the eyes by using the filter. In the proposed method, a region including the eyes is estimated by using the color, and then the location of the middle of the eyebrows and the rotation angle of the face are calculated by using the output of the circle-frequency filter. According to the obtained location of the middle of the eyebrows and the rotation angle of the face, the location of the eyes can be accurately detected without using learning patterns and templates.
• A Note on Texture Discrimination by Using Scale-Space Filtering

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (644)  25  -30  2004/01/28  [Not refereed][Not invited]
   

  A method to extract texture features that are robust against blurring or additive noise by using scale-space filtering has been proposed. In the method, first, the images smoothed at various scales are obtained by applying scale-space filtering to a given texture image. Secondly the smoothed images representing a given texture image are selected. Lastly texture features are extracted from this selected images. However, there are no criteria to select images which the texture features are effectively extracted from. We, therefore, propose a novel method to select such an image by observing normal vectors of the curved surface corresponding to the intensity of the smoothed images obtained by applying the scale-space filtering.
• A Note on Tracking Pedestrian with SVM

  MATSUDA Kohei, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (644)  45  -50  2004/01/28  [Not refereed][Not invited]
   

  This paper proposes an accurate tracking method of pedestrians by using a Support Vector Machine (SVM). The SVM included in the proposed method judges whether a target pedestrian is occluded or not. The method extracts the targets by using the templates which are made from the pedestrians extracted from the previous frame where they were not occluded. Therefore, the accurate tracking is realized even if the pedestrians are occluded. Experimental results are shown to verify the performance of the proposed method.
• Calculating the Error Produced by Rotational and Translational Changes of a Stereo Camera

  WAKI Hirohisa, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (644)  95  -100  2004/01/28  [Not refereed][Not invited]
   

  This paper calculates the errors that are caused when measuring the exact position of an object by using an onboard stereo camera. In order to measure the exact position of an object by using such a camera, precise camera calibration is required. However, aging of the cameras and vibrations that occur when driving have a bad influence on the calibration. Therefore, it is important to calculate the errors that are caused by such aging and vibration.
• A Note on Solving Method for Nurse Scheduling Problem based on the Group Decision Making

  NAKAJIMA Jun, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (640)  43  -48  2004/01/27  [Not refereed][Not invited]
   

  Nurse Scheduling Problem (NSP) is aimed to provide a servise table for nurses under various constraints which are demanded avoiding the severe service and maintaining the required number of nurse. Some methods have been already proposed for this problem with genetic algorithm (GA). However, those conventional methods cannot provide the service table that has equality for all nurses and is suitable for the demand of each nurse. In order to solve this problem, this paper proposes a new method for NSP based on the Group Decision Making. By using proposed method, it is possible to provide a service table that has equality for all nurses. Moreover, the proposed method can provide a service table that is suitable for the demand of each nurse utilizing Analytic Network Process (AHP).
• A Note on Audiovisual Indexing for Scene-Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (640)  119  -124  2004/01/27  [Not refereed][Not invited]
   

  The shot can be defined as a single sequence of frames in a motion picture obtained by one camera without interruption, and is regarded as a basic unit of constructing/analyzing an audiovisual material. However, since the shot is short unit for the audiovisual retrieval, the semantically correlated shots, which are called the scene, can be regarded as a basic unit when retrieving. Therefore we propose an audiovisual indexing method in order to detect the boundaries between two different scenes, which are called scene-cut. The proposed method consists of three steps: (1) the boundaries between two different shots, which are called shot-cuts, are detected by using the video signal; (2) the boundaries between two different audio signals, which are called audio-cuts, are detected by using the audio signal; (3) the scene-cuts are detected by using the results of the shot-cut detection and the audio-cut detection. Experimental results performed by applying the proposed method to real audiovisual signals are shown to verify its high performance.
• A Note on Long Shot Detection in Soccer Video

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (640)  125  -130  2004/01/27  [Not refereed][Not invited]
   

  For the soccer video indexing, long shots, which are mostly occupied by the soccer field, are frequently utilized. Several indexing methods including the long shot detection functions have been proposed. However, their performances are not enough to thoroughly extract the long shots, because they are mainly proposed not for the long shot extraction, but for the soccer video indexing. Therefore, we propose a method to accurately detect the long shots by focusing on the occupancy rate of the soccer field in each frame, the size of the players included in the field, etc.
• A Note on Solving Method for Nurse Scheduling Problem based on the Group Decision Making

  NAKAJIMA Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (643)  43  -48  2004/01/27  [Not refereed][Not invited]
   

  Nurse Scheduling Problem (NSP) is aimed to provide a servise table for nurses under various constraints which are demanded avoiding the severe service and maintaining the required number of nurse. Some methods have been already proposed for this problem with genetic algorithm (GA). However, those conventional methods cannot provide the service table that has equality for all nurses and is suitable for the demand of each nurse. In order to solve this problem, this paper proposes a new method for NSP based on the Group Decision Making. By using proposed method, it is possible to provide a service table that has equality for all nurses. Moreover, the proposed method can provide a service table that is suitable for the demand of each nurse utilizing Analytic Network Process (AHP).
• A Note on Audiovisual Indexing for Scene-Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (643)  119  -124  2004/01/27  [Not refereed][Not invited]
   

  The shot can be defined as a single sequence of frames in a motion picture obtained by one camera without interruption, and is regarded as a basic unit of constructing/analyzing an audiovisual material. However, since the shot is short unit for the audiovisual retrieval, the semantically correlated shots, which are called the scene, can be regarded as a basic unit when retrieving. Therefore we propose an audiovisual indexing method in order to detect the boundaries between two different scenes, which are called scene-cut. The proposed method consists of three steps: (1) the boundaries between two different shots, which are called shot-cuts, are detected by using the video signal; (2) the boundaries between two different audio signals, which are called audio-cuts, are detected by using the audio signal; (3) the scene-cuts are detected by using the results of the shot-cut detection and the audio-cut detection. Experimental results performed by applying the proposed method to real audiovisual signals are shown to verify its high performance.
• A Note on Long Shot Detection in Soccer Video

  TSUTSUMI Kouki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (643)  125  -130  2004/01/27  [Not refereed][Not invited]
   

  For the soccer video indexing, long shots, which are mostly occupied by the soccer field, are frequently utilized. Several indexing methods including the long shot detection functions have been proposed. However, their performances are not enough to thoroughly extract the long shots, because they are mainly proposed not for the long shot extraction, but for the soccer video indexing. Therefore, we propose a method to accurately detect the long shots by focusing on the occupancy rate of the soccer field in each frame, the size of the players included in the field, etc.
• A Note on Quality Improvement of Color JPEG Images

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (639)  7  -12  2004/01/26  [Not refereed][Not invited]
   

  In JPEG coding, quantization errors in decoded images sometimes occur in high frequency regions. Especially, in the color images, the quantization errors of the luminance data are more visible than those of the chrominance data. Therefore, we propose a quality improvement technique for the high frequency regions by adding new codes to reduce the luminance quantization errors. Moreover, to make the number of bits for the new codes smaller, the proposed method calculates them by utilizing fractal image coding. Some simulation results verify that the proposed method can effectively improve the image quality in the high frequency regions and provide the high quality images.
• A Study on Restoration of Missing Areas Application of GMRF Model

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  103-  (639)  19  -24  2004/01/26  [Not refereed][Not invited]
   

  Several methods for restoration of missing areas in still images have been studied. However, their performances are not enough because they cannot accurately reconstruct missing edges. Therefore, this paper proposes an accurate restoration method by using a new GMRF model. Since the new GMRF model can be obtained by an assumption that one image consists of some weakly stationary processes, it can express images including edges more accurately than the traditional GMRF models; and thus, by using this model, the proposed method accurately estimates the missing intensities especially in the edges. Some experimental results are shown to verify its high performance of the proposed method.
• A Note on Quality Improvement of Color JPEG Images

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (642)  7  -12  2004/01/26  [Not refereed][Not invited]
   

  In JPEG coding, quantization errors in decoded images sometimes occur in high frequency regions. Especially, in the color images, the quantization errors of the luminance data are more visible than those of the chrominance data. Therefore, we propose a quality improvement technique for the high frequency regions by adding new codes to reduce the luminance quantization errors. Moreover, to make the number of bits for the new codes smaller, the proposed method calculates them by utilizing fractal image coding. Some simulation results verify that the proposed method can effectively improve the image quality in the high frequency regions and provide the high quality images.
• A Study on Restoration of Missing Areas Application of GMRF Model

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  103-  (642)  19  -24  2004/01/26  [Not refereed][Not invited]
   

  Several methods for restoration of missing areas in still images have been studied. However, their performances are not enough because they cannot accurately reconstruct missing edges. Therefore, this paper proposes an accurate restoration method by using a new GMRF model. Since the new GMRF model can be obtained by an assumption that one image consists of some weakly stationary processes, it can express images including edges more accurately than the traditional GMRF models; and thus, by using this model, the proposed method accurately estimates the missing intensities especially in the edges. Some experimental results are shown to verify its high performance of the proposed method.
• A Study of Designing Finite-Word-Length Filters with a GA and an SA

  MATSUURA Daiki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  103-  (568)  7  -12  2004/01/15  [Not refereed][Not invited]
   

  When digital filters are realized in finite-word-length, there are two problems : one is the output roundoff noise due to the quantization of multiplication results, and the other is the frequency error due to the quantization of filter coefficients. In this paper, we propose a method to solve them simultaneously. In the proposed method, cascade-form filters realized with the delta operators are used in order to solve the former problem, and a coefficient quantization method based on a GA and an SA is used in order to solve the latter problem. Experimental results show the validity of the proposed method.
• Convergence Properties of a CORDIC-Based ARMA Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  103-  (568)  13  -18  2004/01/15  [Not refereed][Not invited]
   

  This paper presents a convergence analysis of a CORDIC-based ARMA lattice filter. In previous works, several convergence analysis techniques have been proposed ; however, they are available only for AR-type lattice filters. On the other hand, we tackle the convergence analysis of an adaptive ARMA lattice filter, and then derive two types of models : a convergence model and a steady-state model. The results of the analysis facilitates an appropriate setting of the step size in a hardware implementation of the filter.
• GA-BASED MOTION TABLE REALIZATION FOR EFFICIENT VIDEO CODING

  ASAMIZU Satoshi, HASEYAMA Miki  IEICE technical report. Image engineering  103-  (539)  47  -51  2004/01/05  

  This paper proposes a motion-table realization method for efficiently and high quality motion compensation. The proposed motion table is realized by using a genetic algorithm (GA) based on a fitness function which is defined by both the prediction image quality and the bit rate to represent motion vectors. According to this fitness function, the GA searches all the possible motion vectors for the necessary vectors; that is to say, by using a motion table including only the necessary vectors, we can obtain a high quality prediction image without increasing the bit rate. Since the previous methods do not directly search for the motion vectors which provide the prediction image quality with the low bit rate; therefore, our approach is not only novel but also provide high quality prediction image. In this paper, we apply the proposed method to actual motion pictures to verify its effectiveness. Furthermore, some of the previous methods not including GA are applied to the same pictures, and the results are compared to our results. This comparison also shows that the proposed method is more effective than the others.
• 高能率な動き補償のためのGAを用いた高速探索法の検討

  浅水仁, 長谷山美紀  画像センシングシンポジウム講演論文集  10th-  2004
• A convergence model for a cordic-based ARMA lattice filter

  S Shiraishi, M Haseyama, H Kitajima  2004 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 3, PROCEEDINGS  3-  301  -304  2004  [Not refereed][Not invited]
   

  This paper presents a convergence model for a CORDIC-based ARMA lattice filter. In previous works, several convergence analyses have been discussed; however, they are available only for ARtype filters. On the other hand, our convergence model reveals convergence properties of an adaptive ARMA lattice filter. The results of the analysis facilitates an efficient ASIC design of the ARMA lattice filter. In addition, we can apply the presented analysis techniques to multiplier-based ARMA lattice filters, so that they can be utilized for many applications in the digital signal processing field.
• An audio-scene cut detection method using fuzzy c-means algorithm for audio-visual indexing

  N Nitanda, M Haseyama, H Kitajima  2004 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 2, PROCEEDINGS  2-  89  -92  2004  [Not refereed][Not invited]
   

  This paper proposes an accurate audio-scene cut detection method. The audio-scene denotes a segment which is constructed of semantically correlated audio-shots, where the audio-shot is a smaller segment than the audio-scene; and the boundary between two audio-scenes and that between two audio-shots are called the audio-scene cut and the audio-shot cut, respectively. Recently, high performance of the audio-scene cut detection methods is required for the audio-visual indexing; and several detection methods have been proposed. However, since most of the methods segment the audio signal in a fixed time interval before indexing, the users cannot obtain the exact time of the audio-scene cuts. Therefore, we propose an accurate audio-scene cut detection method. We utilize the fuzzy c-means algorithm so that the reliability of the audio-shot cut is represented by the fuzzy number. Afterwards, the semantially correlated audio-shots are merged into the same audio-scene, and thereby the audio-scene cuts are obtained.
• Audio-cut detection and audio-segment classification using fuzzy c-means clustering

  N Nitanda, M Haseyama, H Kitajima  2004 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL IV, PROCEEDINGS  4-  325  -328  2004  [Not refereed][Not invited]
   

  This paper proposes an audio-cut detection and audio-segment classification method using fuzzy c-means clustering. In the proposed method, the boundaries between two different audio signals, which are called audio-cuts, can be detected by the fuzzy c-means clustering. In the fuzzy c-means clustering, the fuzzy number represents the possibility that the audio-cut exists. Therefore, according to the possibility, qualified candidates for audio-cuts can be obtained even if audio effects such as fade-in, fade-out, etc. are included in the audio signal. The audio signal is segmented at the detected audio-cuts, and these segments are classified into the following five classes: silence, music, speech, speech with music background, and speech with noise background. This classification simultaneously deletes the wrongly detected audio-cuts. Consequently, we can obtain the accurate audio-cuts and the classification results.
• A GA-based realization method of optimal finite-precision system

  M Haseyama, D Matsuura  2004 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL II, PROCEEDINGS  2-  413  -416  2004  [Not refereed][Not invited]
   

  A GA-based realization method of the optimal finite-prccision system is proposed. The optimal realizations of the finite-precision systems are defined as those that minimize the error between the frequency characteristics of the original infinite-precision system and its finite-precision represented one and can be shown as the solutions of a nonlinear programming problem. Therefore, in this paper, GA-based optimization strategy is presented to provide an efficient method for solving this problem. The proposed realization method of the optimal finite-precision system is based on not only the GA but also an SA to prevent the GA from going into local minima. Some numerical examples and comparison simulations with the traditional quantization methods, such as rounding off, rounding up, and rounding down, and another SA-based method are given to verify the high performance of the proposed method.
• A novel image recognition method based on feature-extraction vector scheme

  M Hiramoto, T Ogawa, M Haseyama  ICIP: 2004 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1- 5  2-  3049  -3052  2004  [Not refereed][Not invited]
   

  This paper introduces a method for recognizing images using a new approach to expressing images as vectors. Using this expression method, an image is constructed from 2 types of vectors - vectors indicating positions and vectors denoting intensity gradients for those positions. When investigating the amount of difference between two images, similarities are evaluated by calculating voting densities in the image space, using the vectors making up the sample image in relation to the vectors expressing the reference image. The expression proposed is invariant to image rotation and by changing the resolution hierarchically, recognition using this expression is also adaptable to perspective and detail. Using this method, we carried out experimentation recognizing representative images from various fields and the results show that the method is effective in discriminating between them.
• A convergence analysis technique for a cordic-based ARMA lattice filter

  S Shiraishi, M Haseyama, H Kitajima  TENCON 2004 - 2004 IEEE REGION 10 CONFERENCE, VOLS A-D, PROCEEDINGS  A-  A487  -A490  2004  [Not refereed][Not invited]
   

  This paper presents an analysis technique for an adaptive ARMA lattice filter based on the CORDIC algorithm. The proposed technique brings a theoretical convergence analysis, which facilitates an efficient ASIC design of the ARMA lattice filter Although the proposed technique is dedicated to filters using a special arithmetic unit (CORDIC processor), we can apply it to popular multiplier-based fillers through slight modifications. Consequently, we can utilize the proposed technique for many applications in the signal processing.
• A soccer field tracking method with wire frame model from TV images

  T Watanabe, M Haseyama, H Kitajima  ICIP: 2004 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1- 5  3-  1633  -1636  2004  [Not refereed][Not invited]
   

  This paper proposes a tracking method of soccer field area in a soccer video captured from TV. The camera taking the soccer game video is controlled by three parameters: its mount position, the angle, and the magnification. In order to estimate these three parameters, the proposed method designs a wire frame model, which represents the official layout of the soccer field lines; and by the wire frame model matching with the field area in the video, the above three parameters can be estimated, even if the soccer video includes the camera panning, zooming, etc. By using the estimated parameters. We call accurately obtain where the field area corresponds in the actual soccer field. Some experiments in tracking the field area in actual soccer videos are performed and their results verify the high performance of the proposed method.
• An efficient phase retrieval method using snakes for image reconstruction

  K Kondo, M Haseyama, H Kitajima  ICIP: 2004 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1- 5  4-  2427  -2430  2004  [Not refereed][Not invited]
   

  This paper proposes a novel phase retrieval method using active contour models (snakes) for image reconstruction. The proposed method reconstructs a target image from the magnitude of fits Fourier transform and the measured area of the image. In general, the measured area is different from the true area where the target image exists. Thus the snake with an area term, which can extract the shape of the concave target image, is utilized to renew the measured area. By processing this renewal iteratively, the area obtained by the snake converges to the true area and as a result the proposed method can accurately reconstruct a target image even when the measured area is different from the true area. Experimental results show the effectiveness of the proposed method. This method has many applications Such as biomedical imaging using X-ray or optical detectors which have lost the phase information.
• A GA-based Multi-Objective Optimization for Service Restoration in Power Distribution Systems

  Jun Inagaki, Jun Nakajima, Miki Haseyama, Hideo Kitajima  IEEJ Transactions on Power and Energy  124-  (3)  404  -412  2004  [Not refereed][Not invited]
   

  Service restoration problem in distribution systems is formulated as a multi-objective optimization problem which is demanded not only for minimizing the amount of unrestored total loads but also for minimizing the number of the switching operations. The solution of the multi-objective optimization problem is usually obtained with a set of Pareto optimal solutions. The Pareto optimal solutions for the service restoration problem are useful for users to obtain their desired restoration by comparing a Pareto optimal solution with the others. However, the conventional methods cannot obtain plural Pareto optimal solutions in one trial. Therefore, this paper proposes a method for obtaining a Pareto optimal set for the service restoration problem with a genetic algorithm. The genetic algorithm produces many possible solutions in its search process. By utilizing this feature, the proposed method can obtain the Pareto optimal set. © 2004, The Institute of Electrical Engineers of Japan. All rights reserved.
• An efficient criterion for fractal image coding and its GA-based optimization

  Megumi Takezawa, Miki Haseyama  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  58-  (7)  937  -943  2004  [Not refereed][Not invited]
   

  In fractal image coding, optimal contractive mapping should be determined to minimize the difference between the original image and a reconstruction of it. Though this can be determined by an exhaustive search, this approach is impractical because of its heavy calculation cost. Therefore, to reduce calculation cost, conventional methods determine an suitable contractive mapping by minimizing the difference between the original image and the collage. However, the obtained mapping cannot necessarily reconstruct a high quality image and this does not minimize reconstruction errors. Hence, in order to always obtain high-quality reconstructed images, we propose a new fractal image coding to minimize the errors by using genetic algorithms.
• Quality improvement of JPEG images based on fractal image coding

  Megumi Takezawa, Miki Haseyama  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  58-  (9)  1317  -1323  2004  [Not refereed][Not invited]
   

  We proposed high accuracy JPEG coding by including fractal image coding. JPEG coding is a commonly used standard method of accurately compressing images. However, in its decoded images, quantization errors occur especially in the region of high frequencies, such as object edges. Our method solved this problem by adding the obtained codes by fractal image coding to improve image quality in high frequency regions. Thus, our method achieved higher coding-performance than the traditional JPEG coding.
• PARCOR係数を用いたオーディオカット検出法 (適応信号処理特集号)

  二反田 直己, 白石 真一, 長谷山 美紀  信号処理  7-  (6)  453  -460  2003/11  [Not refereed][Not invited]
  
• 複数モデルに基づく適応的動き推定のためのVLSIアーキテクチャ (適応信号処理特集号)

  白石 真一, 仲川 和志, 長谷山 美紀  信号処理  7-  (6)  461  -470  2003/11  [Not refereed][Not invited]
  
• Image authentication based on fractal image coding without contamination of original image

  Miki Haseyama, Isao Kondo  Systems and Computers in Japan  34-  (9)  1  -9  2003/08  [Not refereed][Not invited]
   

  Several digital watermarking methods proposed for the purpose of copyright protection embed signature data or images in the original images. Thus, they unavoidably produce degradation of the original images. Authors who create art works, however, desire to distribute their own originals unaltered. Therefore, this paper proposes a system for authentication of original images distributed without embedding any watermarks. The proposed method is based on fractal image coding, an image coding method. The IFS parameters obtained in fractal image coding are utilized as authentication parameters, and signature images for authentication are extracted using these parameters. © 2003 Wiley Periodicals, Inc.
• An Audio-Cut Detection Method Based on Fuzzy c-means Algorithm Using PARCOR Coefficients

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (38)  1  -4  2003/07/17  [Not refereed][Not invited]
  
• A Note on an Approach for Accurately Removing Random-valued Impulse Noise

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (38)  5  -8  2003/07/17  [Not refereed][Not invited]
  
• A Note on an Automatic Extract Method for Apoptotic cells from Videomicroscopy Images

  SASAMURA Yukari, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (38)  21  -24  2003/07/17  [Not refereed][Not invited]
  
• A Note on a Restoration Method for Missing Image Data Using a Correlation of Neighbor Pixels

  OGAWA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (38)  25  -28  2003/07/17  [Not refereed][Not invited]
  
• A line drawing simplification method which retains shape by using the fractal dimension

  Satoshi Uemura, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  34-  (5)  1  -13  2003/05  [Not refereed][Not invited]
   

  This paper presents a novel method for simplification of line drawings which can retain the shape of a given object. By noting that natural shapes innately possess the property designated self-similarity, the fractal dimension, one of the characteristics of fractal objects and a measure of complexity, is used as an index. With this index the degree of simplification can be quantified. In order to represent the approximation of a given object, the feature points are extracted by our analysis, whose basic idea was derived from Vector Tracer. Experimental results show the efficiency of our proposed method. © 2003 Wiley Periodicals, Inc.
• An Accurate Impulse Detector for Removing Impulse Noise

  KONDO Keiko, HASEYAMA Miki, KITAJYMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  86-  (5)  654  -667  2003/05/01  [Not refereed][Not invited]
   

  本論文では,インパルス性雑音によって劣化した画像の復元を目的とした高精度な雑音検出法を提案する。雑音検出と雑音除去の二つの処理によって画像復元を行う手法において,雑音除去は雑音検出の結果を用いて行われることから,復元結果の精度を向上させるためには,雑音の検出精度の向上か必須である。提案手法は,次の二つの処理により,高精度な雑音検出を実現する。(i)新たなフラグ画像(エッジノイズ画像)を導入し,それに基づいて処理点ごとに窓の大きさが異なるメジアンフィルタを選択し,雑音検出を行う。処理点ごとに適した大きさの窓のメジアンフィルタを用いることにより,精度良く雑音の位置を検出できる。(ii)(i)により得られた雑音検出結果に対して,インパルス性雑音か否かを検証する。この検証処理により,雑音検出結果の信頼性が高まり,(i)で得られた雑音検出の精度が更に向上する。提案手法は雑音の割合が高い場合にも適用が可能であり,本論文の最後では,実際の画像を用いた実験結果によりその有効性を示す。
• D-11-51 A Simple Performance Measure for Interframe Predictors for Video Coding

  Kitajima Hideo, Haseyama Miki  Proceedings of the IEICE General Conference  2003-  (2)  51  -51  2003/03/03
• Generalization of a Higher-Order Neural Network based upon Dynamic Coalescence Model as knowledge

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  IEICE technical report. Natural language understanding and models of communication  102-  (689)  7  -10  2003/02/27  [Not refereed][Not invited]
   

  We have theoretically investigated a clustering method Dynamic Coalescence Model from the viewpoint of filtering [1],[3], and have presented several new higher-order neural networks as learning and discriminating systems for distributions recognition and hand-written characters recognition with knowledge inherited from DCM. [2],[3] In this paper, we present new HONNs without any condition upon linear discrimination including a new technique to improve recognition rate even for applying a little amount of training patterns.
• Generalization of a Higher-Order Neural Network based upon Dynamic Coalescence Model as knowledge

  KAITA Takeshi, KITAJIMA Hideo, HASEYAMA Miki, TOMITA Shingo, YAMANAKA Junkichi  Technical report of IEICE. Thought and language  102-  (687)  7  -10  2003/02/27  [Not refereed][Not invited]
   

  We have theoretically investigated a clustering method Dynamic Coalescence Model from the view-point of filtering [1], [3], and have presented several new higher-order neural networks as learning and discriminating systems for distributions recognition and hand-written characters recognition with knowledge inherited from DCM. [2], [3] In this paper, we present new HONNs without any condition upon linear discrimination including a new technique to improve recognition rate even for applying a little amount of training patterns.
• A Note on Audio-Scene Cut Detection Using Fuzzy c-means Algorithm

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (8)  121  -126  2003/02/03  [Not refereed][Not invited]
   

  Since a kind of transition in the audio signals which is called audio-scene cut is generally utilized for segmentation of audio-visual materials, several audio-scene cut detection methods have been proposed. However, since most of the methods segment the audio signal in a fixed time interval before indexing, users cannot obtain the exact time of the audio-scene cut. In order to develop a more accurate audio-scene cut detection method, we utilize the fuzzy c-means algorithm. Our proposed method works as follows: (1) it detects all of the possible transitions in the audio signals without miss-detections; (2) it divides the audio signals into segments bounded by the transitions detected in (1); (3) it classifies the segments into the speech, music, and silence classes; (4) it merges the semantically similar adjacent segments by comparing their audio classes which are obtained in (3) and detects the audio-scene cuts. Further, in all of these steps, since the proposed method directly processes the sub-band data of the MPEG audio signal, the proposed method can be applied to the audio-visual indexing without any decoding procedures.
• Improvement of the Watershed Algorithm by Using Edge Features and Its Application to Hair Style Extraction

  YANJUN Zhao, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (8)  155  -160  2003/02/03  [Not refereed][Not invited]
   

  Watershed algorithms have been proposed as one type of segmentation method. This method can guarantee a closed and continuous boundary because the region is obtained by growing it from the center, called the mark, to the neighboring pixels. However the two problems of surplus segmentation and non-segmentation may exist at the same time. In order to solve these problems, a method is proposed that combines regions by making a comparison of their brightness after the watershed algorithm using edge features and marks has finished processing. To verify the efficiency of this method, the proposed algorithm is used to extract hair styles from digital images.
• A Note on Features Based on a Multiscale Analysis for Texture Classification

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (9)  77  -82  2003/02/03  [Not refereed][Not invited]
   

  This paper presents a novel approach to texture analysis using the features which are based on a multiscale analysis. Since edges or line elements in an image have been utilized as texture elements in the traditional approach, they cannot accurately analyze contaminated images by blur, additive noise, etc. In the proposed method, features of convex and concave surfaces obtained by applying Gaussian niters are defined as texture elements. Even when the texture images including the contamination are observed; the proposed method can provide the accurate analysis results, since the elements are robust against the contamination. Therefore, some experimental results are shown for the high performance of the proposed method to be verified.
• A Note on Expansion Technique for Self-Affine Fractal Objects Using Extended Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (9)  95  -100  2003/02/03  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for self-affine fractal objects using Extended Fractal Interpolation Functions (EFIF). When EFIF is applied to a given object, the map parameter, which play a significant role to represent the object, must be determined by solving the inverse problem. However, the inverse problem concerned with the expanded object which associates with the given object cannot be solved, since the expanded object includes unknown data which must be interpolated. Therefore, in this paper, we demonstrate a flexible property that the map parameter concerned with the expanded object can be simply described by using those concerned with the given object. Using such flexible property in our EFIF, a given self-affine fractal object can be easily expanded with an arbitrary rate.
• A Note on an Impulse Detection Method for Removing Fixed-valued and Random-valued Impulse Noises

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (9)  101  -106  2003/02/03  [Not refereed][Not invited]
   

  The conventional impulse detection methods tend to work well for fixed-valued impulse noise but poorly for random-valued impulse noise. This paper proposes an impulse detection method which can accurately detect both fixed-valued and random-valued impulse noise. The proposed method consists of two systems. The first system detects the impulse noise by considering the difference between the intensity of a target pixel and the output of the median filter. The second system verifies whether the noise detection results obtained by the first system are correct. By using these systems, the proposed method can accurately detect both the types of impulses and can be effectively used as a preprocessor for noise reduction filtering.
• Reducing the Calculation Cost of Binary Image Retrieval

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (9)  107  -112  2003/02/03  [Not refereed][Not invited]
   

  This paper presents a novel method for retrieving binary images by using a binary tree. In the proposed method, a binary tree, in which each leaf has a group of the images belonging to the database, is used. To retrieve the desired images, the tree is traced by comparing the similarities between the input image and the images which correspond to the leaves in the tree, and then the group, which corresponds to the reached leaf in the tree, of the images to be displayed for the user is determined. Moreover, the proposed method uses training and reclustering. The training assigns the Weights to the partial features based on the similarity between the input image and the desired image, and after training, the desired image is reclustered. By these methods, the time for the computation of the similarities to retrieve the desired images is decreased.
• A Determination Method for IFS Parameters in Fractal Image Coding : Reduction in Computational Costs

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  27-  (9)  149  -154  2003/02/03  [Not refereed][Not invited]
   

  In fractal image coding, it is actually impossible to find the optimal IFS parameters, which provide the highest quality reconstructed image, by minimizing the error in the reconstructed image because of heavy computation. Hence, most of the fractal image coding determines the suitable IFS parameters by minimizing the error in the collage image instead of the reconstructed image in short time. However, these obtained IFS parameters do not necessarily provide a high-quality reconstructed image whose error is small. Therefore, in order to obtain the high-quality reconstructed image, we have before proposed a fractal image coding method which can find the optimal IFS parameters in low computational costs by using a genetic algorithm and simulated annealing. In this paper, to achieve further reduction in its computational costs, our previous method is improved by adopting a new procedure.
• A Kalman Filter Using Adaptive Image Modeling for Noise Reduction

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  86-  (2)  212  -222  2003/02/01  [Not refereed][Not invited]
   

  本論文では,カルマンフィルタによる効果的な雑音除去を実現するための画像のモデル化法を提案する.カルマンフィルタにより画像を処理する際,原画像は注目画素とその隣接画素との相関を考慮して,自己回帰(autoregressive : AR)モデルで記述される.一般に,画像の種類やその画像内の平たん部やエッジ部によって,注目画素とその隣接画素間の相関は大きく異なることから,カルマンフィルタによる効果的な雑音除去のためには,処理対象画像,更にはその画像内の各部分に適応したモデル化が必要とされる.提案手法では,劣化状態にある処理対象画像から平たん部及びエッジ部を記述するモデルを高精度に推定することにより,画像の性質及び劣化状態を考慮したモデル化を実現する.画像内の各部分に適応したモデル化が実現されることにより,そのモデルから設計されるカルマンフィルタは,エッジ保存及び効果的なノイズの平滑化が可能となる.本論文の最後では,提案手法の有効性を示すために実験を行い,エッジ保存及びノイズの低減効果を確認している.
• A Note on Features Based on a Multiscale Analysis for Texture Classification

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (630)  77  -82  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel approach to texture analysis using the features which are based on a multiscale analysis. Since edges or line elements in an image have been utilized as texture elements in the traditional approach, they cannot curately nalyze contaminated images by blur, additive noise, etc. In the proposed method, features of convex and concave surfaces obtained by applying Gaussian filters are defined as texture elements. Even when the texture images including the contamination are observed ; the proposed method can provide the accurate analysis results, since the elements are robust against the contamination. Therefore, some experimental results are shown for the high performance of the proposed method to be verified.
• A Note on Expansion Technique for Self-Affine Fractal Objects Using Extended Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (630)  95  -100  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for self-affine fractal objects using Extended Fractal Interpolation Function(EFIF). When EFIF is applied to a given object, the map parameter, which play a significant role to represent the object, must be determined by solving the inverse problem. However, the inverse problem concerned with the expanded object which associates with the given object cannot be solved, since the expanded object includes unknown dada which must be interpolated. Therefore, in this paper, we demonstrate a flexible property that the map parameter concerned with thc expanded object can be simply described by using those concerned with the given object. Using such flexible property in our EFIF,a given self-affine fractal object can be easily expanded with an arbitrary rate.
• A Note on an Impulse Detection Method for Removing Fixed-valued and Random-valued Impulse Noises

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (630)  101  -106  2003/01/28  [Not refereed][Not invited]
   

  The conventional impulse detection methods tend to work well for fixed-valued impulse noise but poorly for random-valued impulse noise. This paper proposes an impulse detection method which can accurately detect both fixed-valued and random-valued impulse noise. The proposed method consists of two systems. The first system detects the impulse noise by considering the difference between the intensity of a target pixel and the output of the median filter. The second system verifies whether the noise detection results obtained by the first system are correct. By using these systems, the proposed method can accurately detect both the types of impulses and can be effectively used as a preprocessor for noise reduction filtering.
• Reducing the Calculation Cost of Binary Image Retrieval

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (630)  107  -112  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel method for retrieving binary images by using a binary tree. In the proposed method, a binary tree, in which each leaf has a group of the images belonging to the database, is used. To retrieve the desired images, the tree is traced by comparing the similarities between the input image and the images which correspond to the leaves in the tree. and then the group,which corresponds to the reached leaf in the tree, of the images to be displayed for the user is deterrnined. Moreover, the proposed method uses training and reclustering. The training assigns the Weights to the partial features based on the similarity between the input image and the desired image, and after training, the desired image is reclustered. By these methods, the time for the computation of the similarities to retrieve the desired images is decreased.
• A Determination Method for IFS Parameters in Fractal Image Coding : Reduction in Computational Costs

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (630)  149  -154  2003/01/28  [Not refereed][Not invited]
   

  In fractal image coding, it is actually impossible to find the optimal IFS parameters, which provide the highest quality reconstructed image, by minimizing the error in the reconstructed image because of heavy computation. Hence, most of the fractal image coding determines the suitable IFS parameters by minimizing the error in the collage image instead of the reconstructed image in short time. However, these obtained IFS parameters do not necessarily provide a high-quality reconstructed image whose error is small. Therefore, in order to obtain the high-quality reconstructed image, we have before proposed a fractal image coding method which can find the optimal IFS parameters in low comoutational costs by using a genetic algorithm and simulated annealing. In this paper, to achieve further reduction in its computational costs, our previous method is improved by adopting a new procedure.
• A Note on Features Based on a Multiscale Analysis for Texture Classification

  ISHIDA Tomoya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (632)  77  -82  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel approach to texture analysis using the features which are based on a niultiscalc analysis. Since edges or line elements in an image have been utilized as texture elements in the traditional approach, they cannot accurately analyze contaminated images by blur, additive noise, etc. In the proposed method, Features of convex and concave surfaces obtained by applying Gaussian filters are defined as texture elements. Even when the texture images including the contamination arc observed; the proposed method can provide the accurate analysis results, since the elements are robust against the contamination. Therefore, some experimental results are shown for the high performance of the proposed method to be verified.
• A Note on Expansion Technique for Self-Affine Fractal Objects Using Extended Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (632)  95  -100  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for self-affine fraetal objects using Extended Fractal Interpolation Functions (EFIF). When EFIF is applied to a given object, the map parameter, which play a significant role to represent the object, must be determined by solving the inverse problem. However, the inverse problem concerned with the expanded object which associates with the given object cannot be solved, since the expanded object includes unknown data which must be interpolated. Therefore, in this paper, we demonstrate a flexible property that the map parameter concerned with the expanded object can be simply described by using those concerned with the given object. Using such flexible property in our EFTF, a given .self-affine fractal object can be easily expanded with an arbitrary rate.
• A Note on an Impulse Detection Method for Removing Fixed-valued and Random-valued Impulse Noises

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (632)  101  -106  2003/01/28  [Not refereed][Not invited]
   

  The conventional impulse detection methods tend to work well for fixed-valued impulse noise but poorly for random-valued impulse noise. This paper proposes an impulse detection method which can accurately detect both fixed-valued and random-valued impulse noise. The proposed method consists of two systems. The first system detects the impulse noise by considering the difference between the intensity of a target pixel and the output of the median filter. The second system verifies whether the noise detection results obtained by the first system arc correct. By using these systems, the proposed method can accurately detect both the types of impulses and can be effectively used as a preprocessor for noise reduction filtering.
• Reducing the Calculation Cost of Binary Image Retrieval

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (632)  107  -112  2003/01/28  [Not refereed][Not invited]
   

  This paper presents a novel method for retrieving binary images by using a binary tree. In the proposed method, a binary tree, in which each leaf has a group of the images belonging to the database, is used. To retrieve the desired images, the tree is traced by comparing the similarities between the input image and the images which correspond to the leaves in the tree, and then the group, which corresponds to the reached leaf in the tree, of the images to be displayed for the user is determined. Moreover, tin' proposed method uses training and reclustering. The training assigns the Weights to the partial features based on the similarity between the input image and the desired image, and after training, the desired image is reclustered. By these methods, the time for the computation of the similarities to retrieve the desired images is decreased.
• A Determination Method for IFS Parameters in Fractal Image Coding : Reduction in Computational Costs

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (632)  149  -154  2003/01/28  [Not refereed][Not invited]
   

  In fractal image coding, it is actually impossible to find the optimal IFS parameters, which provide the highest quality reconstructed image, by minimizing the error in the reconstructed image because of heavy computation. Hence, most of the fractal image coding determines the suitable IFS parameters by minimizing the error in the collage image instead of the reconstructed image in short time. However, these obtained IFS parameters do not necessarily provide a high-quality reconstructed image whose error is small. Therefore, in order to obtain the high-quality reconstructed image, we have before proposed a fractal image coding method which can find the optimal IFS parameters in low computational costs by using a genetic algorithm and simulated annealing. In this paper, to achieve further reduction in its computational costs, our previous method is improved by adopting a new procedure.
• A Note on Audio-Scene Cut Detection Using Fuzzy c-means Algorithm

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (629)  121  -126  2003/01/27  [Not refereed][Not invited]
   

  Since a kind of transition in the audio signals which is called audio-scene cut is generally utilized for segmentation of audio-visual materials, several audio-scene cut detection methods have been proposed. However, since most of the methods segment the audio signal in a fixed time interval before indexing, users cannot obtain the exact time of the audio-scene cut. In order to develop a more accurate audio-scene cut detection method, we utilize the fuzzy c-means algorithm. Our proposed method works as follows: (1) it detects all of the possible transitions in the audio signals without miss-detections; (2) it divides the audio signals into segments bounded by the transitions detected in (1); (3) it classifies the segments into the speech, music, and silence classes; (4) it merges the semantically similar adjacent segments by comparing their audio classes which are obtained in (3) and detects the audio-scene cuts. Further, in all of these steps, since the proposed method directly processes the sub-band data of the MPEG audio signal, the proposed method can be applied to the audio-visual indexing without any decoding procedures.
• Improvement of the Watershed Algorithm by Using Edge Features and Its Application to Hair Style Extraction

  ZHAO YANJUN, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  102-  (629)  155  -160  2003/01/27  [Not refereed][Not invited]
   

  Watershed algorithms have been proposed as one type of segmentation method. This method can guarantee a closed and continuous boundary because the region is obtained by growing it from the center, called the mark, to the neighboring pixels. However the two problems of surplus segmentation and non-segmentation may exist at the same time. In order to solve these problems, a method is proposed that combines regions by making a comparison of their brightness after the watershed algorithm using edge features and marks has finished processing. To verify the efficiency of this method, the proposed algorithm is used to extract hair styles from digital images.
• A Note on Audio-Scene Cut Detection Using Fuzzy c-means Algorithm

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (631)  121  -126  2003/01/27  [Not refereed][Not invited]
   

  Since a kind of transition in the audio signals which is called audio-scene cut is generally utilized for segmentation of audio-visual materials, several audio-scene cut detection methods have been proposed. However, since most of the methods segment the audio signal in a fixed time interval before indexing, users cannot obtain the exact time of the audio-scene cut. In order to develop a more accurate audio-scene cut detection method, we utilize the fuzzy c-means algorithm. Our proposed method works as follows:(1) it detects all of the possible transitions in the audio signals without miss-detections;(2) it divides the audio signals into segments bounded by the transitions detected in (1);(3) it classifies the segments into the speech, music, and silence classes;(4) it merges the semantically similar adjacent segments by comparing their audio classes which are obtained in (3) and detects the audio-scene cuts. Further, in all of these steps, since the proposed method directly processes the sub-band data of the MPEG audio signal, the proposed method can be applied to the audio-visual indexing without any decoding procedures.
• Improvement of the Watershed Algorithm by Using Edge Features and Its Application to Hair Style Extraction

  ZHAO YANJUN, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (631)  155  -160  2003/01/27  [Not refereed][Not invited]
   

  Watershed algorithms have been proposed as one type of segmentation method. This method can gnarantec a closed and continuous boundary because the region is obtained by growing it from the center, called the mark, to the neighboring pixels. However the two problems of surplus segmentation and non-segmentation may exist at the same time. In order to solve these problems, a method is proposed that combines regions by making a comparison of their brightness after the watershed algorithm using edge features and marks has finished processing. To verify the efficiency of this method, the proposed algorithm is used to extract hair styles from digital images.
• A SHAPE-CONSTRAINT-BASED SPLITTING ACTIVE CONTOUR MODEL AND ITS APPLICATIONS

  YONEYAMA Shogo, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (571)  83  -86  2003/01/15  

  This paper presents a new splitting active contour model which can extract the shapes of multiple objects included in one initial contour individually. Most of the approaches based on active contour models (Snakes) cannot extract the shapes of multiple objects, since they require that one initial contour includes one object. To overcome this problem, the proposed method utilizes a splitting mechanism which works according to two types of Snake energy functions. Based on the difference in the extraction results obtained by minimizing their energy functions, an initial contour is split into multiple contours. After iterative splitting, each contour can include one object. Therefore, our splitting Snake can extract the desired object even if other objects are included within the same initial contour together. For evaluating its extraction performance, we apply our splitting Snake to extracting facial features from an image sequence taken from TV news program. The extraction results show the effectiveness of the proposed method.
• AN ACCURATE NOISE DETECTOR FOR THE REMOVAL OF IMPULSIVE NOISE

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (571)  21  -24  2003/01/15  [Not refereed][Not invited]
   

  This paper proposes a new noise-detection method for the restoration of images corrupted by impulsive noise. The proposed method consists of two systems. In the first system, we generate a new flag image to classify the pixels of an observed image into two types: the pixels in flat areas and edge areas, and then different noise detectors process each type of pixels classified according to the flag image. The second system verifies the pixels once detected as impulsive noise by the first system. The combination of these two systems can accurately detect the location of impulsive noise and be effectively used as a preprocessor for a noise reduction filter. Experiments show that the proposed method can accurately detect the location of impulsive noise even if the observed images are highly corrupted.
• A GA AND SA-BASED FAST SEARCH ALGORITHM FOR IFS PARAMETERS IN FRACTAL IMAGE CODING

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (571)  25  -28  2003/01/15  [Not refereed][Not invited]
   

  This paper proposes a fast search technique using a genetic algorithm (GA) and simulated annealing (SA) for the optimal parameters of the iterated function system (IPS) utilized in fractal image coding. Heavy computation cost for finding the optimal IPS parameters is conventionally a serious problem, and thus several techniques to reduce its computation cost have been reported. One of them utilizes the GA to find them in shorter time than the other techniques However, its reduction is not enough for the practical use. Therefore, the proposed method achieves further reduction by including an SA search in a GA search. By utilizing the advantages of both the GA and the SA, we can search for the optimal IPS parameters more effectively. Moreover, some simulation results show that the proposed method achieves more reduction in the computation cost than the only GA-based method does.
• A CONTENT-BASED IMAGE RETRIEVAL METHOD BASED ON FRACTAL IMAGE CODING

  KONDO Isao, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (570)  89  -92  2003/01/14  

  In this paper, we present a new content-based image retrieval method based on the fractal image coding. Several content-based image retrieval methods have been reported. However, most of them require high computation cost because they perform multiplication when computing the similarity between a query image and database images. There-fore, in order to achieve image retrieval with low computation cost, the proposed method introduces a new similarity measurement. Moreover, to effectively distinguish similar images to the query image from the others, the proposed method searches the database images for the similar images based on a novel approach which utilizes the advantage of the fractal image coding. The experimental results show that the proposed method can search for the similar images not only quickly but also accurately.
• サッカー映像における撮影フィールド領域の検出に関する一考察

  渡辺友樹, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• ブロックマッチングの高精度化 ブロックサイズが小さい場合の考察

  徳永佳晃, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• 有限語長ディジタルフィルタの設計に関する考察-周波数特性の劣化と丸め誤差の軽減-

  松浦大樹, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• GA及びSAを用いた対応点決定法の高精度化に関する一考察

  蛸島康平, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• 多重スケール解析を用いたテクスチャ画像の識別に関する考察

  石田智也, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• FIFを用いた音声信号のサンプリングレート変換に関する考察

  上村郷志, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• 映像インデキシングを目的としたMPEG Audioデータのクラスタリングに関する一考察

  二反田直己, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• リング周波数フィルタを用いた眉間検出の高精度化に関する一考察

  金子千晶, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• ビデオ顕微鏡におけるアポトーシス細胞の自動抽出に関する一考察

  笹村ゆかり, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• 上限値制約を伴う経路探索手法のGAによる実現

  稲垣潤, 長谷山美紀, 北島秀夫  情報科学技術フォーラム  FIT 2003-  2003
• Restoration method of missing image data using modified optical flow

  T Ogawa, M Haseyama, H Kitajima  IEEE EMBS APBME 2003  130  -131  2003  [Not refereed][Not invited]
   

  This paper proposes a restoration method of still images including missing pixels by a modified optical flow, which is different from the traditional flow. For successful restoration, the proposed method introduces a new parameter into the selection scheme of the optimum optical flow, which can represent how much reliable the obtained flow is; and based on this parameter, highly reliable optical flows are mainly used in the image restoration. The proposed method consequently achieves more accurate restoration than the previous method does especially in the edge area.
• Ultra low bit-rate image coding algorithm based on fractal image coding

  M Takezawa, M Haseyama, H Kitajima  ISPA 2003: PROCEEDINGS OF THE 3RD INTERNATIONAL SYMPOSIUM ON IMAGE AND SIGNAL PROCESSING AND ANALYSIS, PTS 1 AND 2  2-  1013  -1017  2003  [Not refereed][Not invited]
   

  This paper proposes a low bit-rate fractal image coding algorithm for wireless communication. The fractal image coding is well known for its high compression performance. This performance is useful for the future wireless communication with narrow-band. However, in order to become a powerful image coding technique for the wireless communication, its reconstructed image quality must be improved. Therefore, to achieve further quality improvement, a novel fractal image coding algorithm is proposed. Some simulation results verify that the proposed method can reconstruct the images with 1.4dB higher quality than JPEG at low bit-rates below 0.12bits pixel and also higher quality than JPEG2000 at very low bit-rates below 0.035-0.05bits pixel.
• Fast line extraction from digital images using line segments

  Euijin Kim, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  34-  (10)  76  -89  2003  [Not refereed][Not invited]
   

  This paper presents a fast line extraction method using the line segments found in digital images. A digital line can be decomposed into line segments, which consist of continuous edge pixels, in four directions. The directions of the line segments are varied and limited by the relationship between the line segments and the slopes of analog lines. The proposed method attains high speed and accuracy by tracking each line segment in the same direction which comes from the relationship. Experimental results are included to show that the proposed method can achieve high accuracy with a large reduction in the computation time and has robustness in the presence of noise.
• A trainable retrieval system for cartoon character images

  M Haseyama, A Matsumura  2003 INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, VOL II, PROCEEDINGS  2-  393  -396  2003  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve cartoon character images in a database or network. In this method, partial features of an image, defined as Regions and Aspects, are used as keys to identify cartoon character images. The similarities between a query cartoon character image and the images in the database are computed by using these features. Based on the similarities, the cartoon images same or similar to the query image are identified and retrieved from the database. Moreover, our method adopts a training scheme to reflect the user's subjectivity. The training emphasizes the significant Regions or Aspects by assigning more weight based on the user's preferences and actions, such as selecting a desired image or an area of an image. These processes make the retrieval more effective and accurate. Experiment results verify the effectiveness and retrieval accuracy of the method.
• 2-D functional AR Model for image identification

  M Haseyama, Kondo, I  2003 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL III, PROCEEDINGS  2-  605  -608  2003  [Not refereed][Not invited]
   

  This paper proposes a 2-D Functional AR Model for image identification. The definition of the proposed model includes functions that can exploit the self-similarity nature in images to throughly extract image features. By introducing the functional scheme into the model, only a few number of parameters, which are called 2-D Functional AR parameters, can describe the image features simply and accurately. These characteristics make the model suitable for image identification applications. Some experiments of image identification are performed, and the results verify that the proposed model accurately represents the image feature, and the image can be correctly identified. The calculation time is fast enough for practical use in image retrieval.
• An automatic extraction method for apoptotic cell from video microscopy images

  Y Sasamura, M Barboric, M Haseyama  IEEE EMBS APBME 2003  210  -211  2003  [Not refereed][Not invited]
   

  This paper presents an automatic extraction method for apoptotic cells from video microscopy images. Since the apoptosis, which is called cell death, is well-known to he related with various disease. Therefore, several researchers try to reveal its mechanism and often analyze the images acquired under the video microscopes. However, the performance of software used in the image analysis is not enough for the auto-detection of the apoptotic cells, and thus the researchers have to manually detect the apoptotic cells from the images. Therefore, this paper proposes a method to automatically detect the apoptotic cells based on their inherent shape features.
• A GA-based routing method with an upper bound constraint

  J Inagaki, M Haseyama, H Kitajima  PROCEEDINGS OF THE 2003 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL II  2-  488  -491  2003  [Not refereed][Not invited]
   

  This paper presents a method of searching for the shortest route via the most designated points among the routes whose lengths are less than the upper bound and its effectiveness is verified by applying the algorithm to map data. The proposed algorithm uses a fitness function which includes weight parameters based on the building block hypothesis. By using this function, the optimum is efficiently obtained. This kind of routing is difficult in the case of using existing methods such as Dijkstra's method. Also, the proposed method is more useful.
• Convergence analysis of a cordic-based gradient adaptive lattice filter

  S Shiraishi, M Haseyama, H Kitajima  PROCEEDINGS OF THE 2003 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL IV  4-  456  -459  2003  [Not refereed][Not invited]
   

  This paper presents a theoretical analysis of a CORDIC-based gradient adaptive lattice filter. First, we provide a convergence model and reveal convergence properties of filter coefficients. Second, we derive a steady-state model based on a Markov chain. By using the steady-state model, the relation between the step size parameter and the variance of the estimation error is clarified. The results of the analysis facilitate an efficient hardware design of the filter.
• 2-D functional AR Model for image identification

  M Haseyama, Kondo, I  2003 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL III, PROCEEDINGS  3-  605  -608  2003  [Not refereed][Not invited]
   

  This paper proposes a 2-D Functional AR Model for image identification. The definition of the proposed model includes functions that can exploit the self-similarity nature in images to throughly extract image features. By introducing the functional scheme into the model, only a few number of parameters, which are called 2-D Functional AR parameters, can describe the image features simply and accurately. These characteristics make the model suitable for image identification applications. Some experiments of image identification are performed, and the results verify that the proposed model accurately represents the image feature, and the image can be correctly identified. The calculation time is fast enough for practical use in image retrieval.
• A trainable retrieval system for cartoon character images

  M Haseyama, A Matsumura  2003 INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, VOL II, PROCEEDINGS  3-  393  -396  2003  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve cartoon character images in a database or network. In this method, partial features of an image, defined as Regions and Aspects, are used as keys to identify cartoon character images. The similarities between a query cartoon character image and the images in the database are computed by using these features. Based on the similarities, the cartoon images same or similar to the query image are identified and retrieved from the database. Moreover, our method adopts a training scheme to reflect the user's subjectivity. The training emphasizes the significant Regions or Aspects by assigning more weight based on the user's preferences and actions, such as selecting a desired image or an area of an image. These processes make the retrieval more effective and accurate. Experiment results verify the effectiveness and retrieval accuracy of the method.
• A Kalman filter using texture for noise reduction in SAR images

  M Tsuchida, M Haseyama, H Kitajima  ELECTRONICS AND COMMUNICATIONS IN JAPAN PART I-COMMUNICATIONS  86-  (10)  21  -32  2003  [Not refereed][Not invited]
   

  In this paper, we propose a noise reduction method for synthetic aperture radar (SAR) images that uses a Kalman filter. The noise in SAR images is primarily generated by speckle and appears as salt-and-pepper noise. Speckle noise is known to have the multiplicative statistical property, but this property may not hold depending on the region. Consequently, adequate noise removal is difficult. In radar, the expected values of the backscattered waves change due to the type of reflecting object. The proposed method segments an image into several regions based on the texture and varies the parameters in Kalman filtering in response to the region. By varying the parameters, the processing can account for the noise generated primarily by speckle in actual SAR images and the changes in the statistical properties of the backscattered. waves. In this paper, we apply the proposed method to synthetic test images and to actual SAR images to quantitatively evaluate the noise reduction and present the results. (C) 2003 Wiley Periodicals, Inc.
• A cartoon character retrieval system including trainable scheme

  M Haseyama, A Matsumura  2003 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOL 3, PROCEEDINGS  3-  37  -40  2003  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve cartoon character images in a database or network. In this method, partial features of an cartoon image, defined as Regions and Aspects, are used as keys to identify cartoon characters. The similarities between a query cartoon character image and the images in the database are computed by using these features. Based on the similarities, the cartoon images same or similar to the query image are identified and retrieved from the database. Moreover, our method adopts a training scheme to reflect the user's subjectivity. The training emphasizes the significant Regions or Aspects based on the user's preferences and actions, such as selecting a desired image or an area of an image. These processes make the cartoon-character retrieval more effective and accurate. Experimental results verify the effectiveness and retrieval accuracy of the method.
• A shape-constraint-based active contour model with a splitting mechanism for face-feature extraction

  Miki Haseyama, Shogo Yoneyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  57-  (6)  729  -735  2003  [Not refereed][Not invited]
   

  The shape-constraint-based active contour model (SC-ACM), one of the Snakes, is robust against noise. However, an SC-ACM's extraction ability depends on initial contour locations so that it often fails to extract face features, because they are located too close to each other. Therefore, in order to resolve this problem, we use a splitting mechanism with the SC-ACM, whereby the proposed Snakes are split according to the difference between the results obtained by using the previous SC-ACM and other Snakes with an area term. The proposed Snakes with the splitting mechanism can extract every feature, even though the initial contour includes multiple objects. We simulated extracting face-features from TV news images of the upper part of an announcer maintaining a full face position. Our results show that the proposed method can extract each of the eyes with the brows and the mouth including its neighboring areas without getting trapped by other face features.
• A GA and SA-based fractal image coding

  Megumi Takezawa, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  57-  (7)  864  -867  2003  [Not refereed][Not invited]
   

  This paper proposes a fast search technique using a genetic algorithm (GA) including a simulated annealing (SA) algorithm for the optimal parameters of the iterated function system (IPS) utilized in fractal image coding. The heavy computational costs of the previous methods to find the optimal IFS parameters are a serious problem, and in order to overcome it, we have already proposed a GA-based technique to find them in short time. However, its reduction is not enough for the practical use. Therefore, the proposed method achieves further reduction by including an SA search in the GA search. Some simulation results show that the proposed method achieves more reduction in the computational costs than the only GA-based method does.
• A Note on an Extract Method for Apoptotic cell from Videomicroscopy Images

  SASAMURA Yukari, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2002-  (120)  63  -68  2002/12/16  [Not refereed][Not invited]
   

  This paper presents an effective method which can automatically extract the apoptotic cells from image sequence obtained by using videomicroscopy. Since apoptosis, which called cell death is well-known to be related with various disease. Therefore, several researchers try to reveal its mechanism. Further, for the purpose they often analyze images acquired in the videomicroscopy, by using simple software or visual recognition. However, even using the software, they have to manually process or check the images, because its performance is not enough for the auto-detection of the apototic cells. Thus this paper proposes a method to automatically detect the apototic cells, which are based on the inherent features of the images of the apototic cells.
• A Note on a Sloving Method for Nurse Scheduling Problem Using GA and Game Theory

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2002-  (120)  81  -86  2002/12/16  [Not refereed][Not invited]
   

  This paper presents a method using a genetic algorithm and game theory for Nurse Scheduling Problem (NSP), which is one of the multiobjective optimization problems. The object for the traditional approach of the NSP has to be definitely formulated by a performance function. The proposed method provides an optimization scheme for solving problems including objects like nurses' preferences, which cannot be formulated by any performance functions by utilizing a game where each player has its own strategy.
• A Novel Expansion Technique for One-Dimensional Discrete Signals Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2002-  (120)  87  -92  2002/12/16  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for one-dimensional discrete signals derived by using Fractal Interpolation Functions (FIF) known as one of the fractal models. When FIF is applied to a given discrete signal, the map parameters, which are required to represent the signal in FIF, must be determined by solving the inverse problem set as depending on the signal. However, the inverse problem concerned with the expanded signal cannot be easily solved, since the expanded signal includes unknown data points which must be interpolated. Therefore we demonstrate that the inverse problem concerned with the expanded signal can be easily solved by using the map parameters obtained by solving the problem of the given signal and thereby a given signal can be efficiently expanded with an arbitrary rate.
• A Note on an Extract Method for Apoptotic cell from Videomicroscopy Images

  SASAMURA Yukari, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (82)  63  -68  2002/12/12  [Not refereed][Not invited]
   

  This paper presents an effective method which can automatically extract the apoptotic cells from image sequence obtained by using videomicroscopy. Since apoptosis, which called cell death is well-known to be related with various disease. Therefore, several researchers try to reveal its mechanism. Further, for the purpose they often analyze images acquired in the videomicroscopy, by using simple software or visual recognition. However, even using the software, they have to manually process or check the images, because its performance is not enough for the auto-detection of the apototic cells. Thus this paper proposes a method to automatically detect the apototic cells, which are based on the inherent features of the images of the apototic cells.
• A Note on a Solving Method for Nurse Scheduling Problem Using GA and Game Theory

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (82)  81  -86  2002/12/12  [Not refereed][Not invited]
   

  This paper presents a method using a genetic algorithm and game theory for Nurse Scheduling Problem (NSP), which is one of the multiobjective optimization problems. The object for the traditional approach of the NSP has to be definitely formulated by a performance function. The proposed method provides an optimization scheme for solving problems including objects like nurses' preferences, which cannot be formulated by any performance functions by utilizing a game where each player has its own strategy.
• A Novel Expansion Technique for One-Dimensional Discrete Signals Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (82)  87  -92  2002/12/12  [Not refereed][Not invited]
   

  This paper presents a novel exoansion technique for one-dimensional discrete signals derived by using Fractal Interpolation Functions (FIF) known as one of the fractal models. When FIF is applied to a given discrete signal, the map parameters, which are required to represent the signal in FIF must be determined by solving the inverse problem set as depending on the signal. However, the inverse problem concerned with the expanded signal cannot be easily solved, since the expanded signal includes unknown data points which must be interpolated. Therefore we demonstrate that the inverse problem concerned with the expanded signal can be easily solved by using the map parameters obtained by solving the problem of the given signal and thereby a given signal can be efficiently expanded with an arbitrary rate.
• A Note on an Extract Method for Apoptotic cell from Videomicroscopy Images

  SASAMURA Yukari, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (518)  63  -68  2002/12/09  [Not refereed][Not invited]
   

  This paper presents an effective method which can automatically extract the apoptotic cells from image sequence obtained by using videomicroscopy. Since apoptosis, which called cell death is well-known to be related with various disease. Therefore, several researchers try to reveal its mechanism. Further, for the purpose they often analyze images acquired in the videomicroscopy, by using simple software or visual recognition. However, even using the software, they have to manually process or check the images, because its performance is not enough for the auto-detection of the apototic cells. Thus this paper proposes a method to automatically detect the apototic cells, which are based on the inherent features of the images of the apototic cells.
• A Note on a Solving Method for Nurse Scheduling Problem Using GA and Game Theory

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (518)  81  -86  2002/12/09  [Not refereed][Not invited]
   

  This paper presents a method using a genetic algorithm and game theory for Nurse Scheduling Problem (NSP), which is one of the multiobjective optimization problems. The object for the traditional approach of the NSP has to be definitely formulated by a performance function. The proposed method provides an optimization scheme for solving problems including objects like nurses' preferences, which cannot be formulated by any performance functions by utilizing a game where each player has its own strategy.
• A Novel Expansion Technique for One-Dimensional Discrete Signals Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (518)  87  -92  2002/12/09  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for one-dimensional discrete signals derived by using Fractal Interpolation Functions (FIF) known as one of the fractal models. When FIF is applied to a given discrete signal, the map parameters, which are required to represent the signal in FIF, must be determined by solving the inverse problem set as depending on the signal. However, the inverse problem concerned with the expanded signal cannot be easily solved, since the expanded signal includes unknown data points which must be interpolated. Therefore we demonstrate that the inverse problem concerned with the expanded signal can be easily solved by using the map parameters obtained by solving the problem of the given signal and thereby a given signal can be efficiently expanded with an arbitrary rate.
• A Note on an Extract Method for Apoptotic cell from Videomicroscopy Images

  SASAMURA Yukari, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  102-  (516)  63  -68  2002/12/09  [Not refereed][Not invited]
   

  This paper presents an effective method which can automatically extract the apoptotic cells from image sequence obtained by using videomicroscopy. Since apoptosis, which called cell death is well-known to be related with various disease. Therefore, several researchers try to reveal its mechanism. Further, for the purpose they often analyze images acquired in the videomicroscopy, by using simple software or visual recognition. However, even using the software, they have to manually process or check the images, because its performance is not enough for the auto-detection of the apototic cells. Thus this paper proposes a method to automatically detect the apototic cells, which are based on the inherent features of the images of the apototic cells.
• A Note on a Solving Method for Nurse Scheduling Problem Using GA and Game Theory

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  102-  (516)  81  -86  2002/12/09  [Not refereed][Not invited]
   

  This paper presents a method using a genetic algorithm and game theory for Nurse Scheduling Problem(NSP), which is one of the multiobjective optimization problems. The object for the traditional approach of the NSP has to be definitely formulated by a performance function. The proposed method provides an optimization scheme for solving problems including objects like nurses' preferences, which cannot be formulated by any performance functions by utilizing a game where each player has its own strategy.
• A Novel Expansion Technique for One-Dimensional Discrete Signals Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  102-  (516)  87  -92  2002/12/09  [Not refereed][Not invited]
   

  This paper presents a novel expansion technique for one-dimensional discrete signals derived by using Fractal Interpolation Functions (FIF) known as one of the fractal models. When FIF is applied to a given discrete signal, the map parameters, which are required to represent the signal in FIF, must be determined by solving the inverse problem set as depending on the signal. However, the inverse problem concerned with the expanded signal cannot be easily solved, since the expanded signal includes unknown data points which must be interpolated. Therefore we demonstrate that the inverse problem concerned with the expanded signal can be easily solved by using the map parameters obtained by solving the problem of the given signal and thereby a given signal can be efficiently expanded with an arbitrary rate.
• Fast and Robust Ellipse Extraction from Complicated Images

  Euijin Kim, Miki Haseyama, Hideo Kitajima  Proceedings of the First International Conference on Information Technology and Applications (ICITA 2002)  357  -362  2002/12/01  [Not refereed][Not invited]
   

  In this paper, we present a fast and robust ellipse extraction method. The proposed method can extract ellipses with high accuracy and speed from images with complicated backgrounds. It consists of two parts. First, we extract arc segments from an ellipse approximated by short straight lines that are extracted by a fast line extraction algorithm. Second, the arc segments are used to calculate accurate ellipse parameters with a least squares ellipse fitting algorithm. Tests performed on synthetic and real-world images show that the method quickly and accurately extracts circles from complicated images.
• Extraction of Circles from Arcs Segmented into Short Straight Lines

  Euijin Kim, Miki Haseyama, Hideo Kitajima  Proceedings of the First International Conference on Information Technology and Applications (ICITA 2002)  363  -368  2002/12/01  [Not refereed][Not invited]
   

  This paper presents a new method that is capable of extracting circles from complicated and heavily corrupted images, which is not based on the Hough transform (HT). The proposed method consists of three parts. First, we approximately detect short straight lines from the image by using a fast line extraction algorithm. Second, it uses a least squares fitting algorithm for arc segments. The arc segments are used to calculate accurate circle parameters instead of using a parameter space. Finally, by using the segmented arcs we extract circles. Tests performed on synthetic and real-world images show that the algorithm quickly and accurately extracts circles from complicated and heavily corrupted images.
• Image Authentication Based on Fractal Image Coding without Contamination of Original Image

  HASEYAMA Miki, KONDO Isao  The Transactions of the Institute of Electronics,Information and Communication Engineers.  85-  (10)  1513  -1521  2002/10/01  [Not refereed][Not invited]
   

  ディジタル画像の著作権保護を目的として提案されている,各種電子透かし手法は,著作画像に,署名データや画像を埋め込む.そのため,著作画像の品質の劣化は避けられない.しかしながら,芸術性が追求される画像データは,もとの画質を保ったままで配信することが希望される場合がある.そこで,本論文では,透かし情報の埋込みを行わずに画像データを配布した場合でも,著者であることを主張することが可能な著者認証システムを提案する.提案手法は,画像符号化手法の一つとして知られているフラクタル画像符号化を応用して実現されており,符号化の際のIFSパラメータを認証パラメータとして,著作画像から,著者であることを主張する署名画像を抽出することを可能としている.
• An Actualization of a Genetic Algorithm for Routing with an Upper Bound Constraint

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  102-  (296)  25  -30  2002/08/28  [Not refereed][Not invited]
   

  The authors have proposed a method of searching the shortest route via several designated points with a genetic algorithm. By applying this method, this paper proposes a method ot searching tor the snortest route via the most designated points among the routes whose lengths are less than the upper bound. However, this kind of routing has a drawback of making optimization accuracy deteriorate when using a method of simply screening individuals whose route lengths are more than the upper bound. In order to solving this proplem, two new methods are proposed and their validity is verified by several experiments.
• A Study of Quantizing Filter Coefficients with a GA and an SA

  MATSUURA Daiki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  102-  (296)  31  -36  2002/08/28  [Not refereed][Not invited]
   

  This paper presents efficient quantization method for coefficients using a GA and an SA. When digital filters are realized, their coefficients need to be quantized, so that computational circuits and registers are constructed with limited word length. Therefore the characteristic of the digital filter which is actually realized and used is different from its characteristic theoretically obtained. In order to improve such a decayed characteristic, the method with a GA and an SA has been proposed. In this paper, we improve the design of GA and SA and thereby the filter coefficients which give smaller decay can be obtained. Several experiments show the efficiency of our proposed method. Furthermore, the other experiment indicates that the proposed method is independent of types of filters and quantization intervals of coefficients.
• A Note on Solving Method for Service Restoration Problem in Distribution System with a Genetic Algorithm

  NAKAJIMA Jun, INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  102-  (296)  37  -42  2002/08/28  [Not refereed][Not invited]
   

  Service restoration problem in distribution system is fomulated as a multi-objective optimization problem which is demanded not only maximizing the amount of total load to be restored but also minimizing the number of the switching operations. This paper proposes a method to obtain a pareto-optimal set for the service restoration ploblem utilizing a Genetic Algorithm. By using our method, it is possible to compare an optimal solution with the others, so that a user can obtain his desired optimal solution. The conventional methods cannot realize this kind of application, thereforethe proposed method is useful.
• A Note on Convergence Properties of a CORDIC-Based Adaptive Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  102-  (296)  43  -48  2002/08/28  [Not refereed][Not invited]
   

  This paper presents an analysis of convergence properties of a CORDIC-based adaptive lattice filter. First, we derive a convergence model of filter coefficients (rotation angles). Next, we analyze the behavior of the angles in steady states. According to these analysises, the convergence of the CORDIC lattice filter is proved.
• An Actualization of a Genetic Algorithm for Routing with an Upper Bound Constraint

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Nonlinear problems  102-  (298)  25  -30  2002/08/28  [Not refereed][Not invited]
   

  The authors have proposed a method of searching the shortest route via several designated points with a genetic algorithm. By applying this method, this paper proposes a method of searching for the shortest route via the most designated points among the routes whose lengths are less than the upper bound. However, this kind of routing has a drawback of making optimization accuracy deteriorate when using a method of simply screening individuals whose route lengths are more than the upper bound. In order to solving this proplem, two new methods are proposed and their validity is verified by several experiments.
• A Study of Quantizing Filter Coefficients with a GA and an SA

  MATSUURA Daiki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Nonlinear problems  102-  (298)  31  -36  2002/08/28  [Not refereed][Not invited]
   

  This paper presents efficient quantization method for coefficients using a GA and an SA. When digital filters are realized, their coefficients need to be quantized, so that computational circuits and registers are constructed with limited word length. Therefore the characteristic of the digital filter which is actually realized and used is different from its characteristic theoretically obtained. In order to improve such a decayed characteristic, the method with a GA and an SA has been proposed. In this paper, we improve the design of GA and SA and thereby the filter coefficients which give smaller decay can be obtained. Several experiments show the efficiency of our proposed method. Furthermore, the other experiment indicates that the proposed method is independent of types of filters and quantization intervals of coefficients.
• A Note on Solving Method for Service Restoration Problem in Distribution System with a Genetic Algorithm

  NAKAJIMA Jun, INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Nonlinear problems  102-  (298)  37  -42  2002/08/28  [Not refereed][Not invited]
   

  Service restoration problem in distribution system is fomulated as a multi-objective optimization problem which is demanded not only maximizing the amount of total load to be restored but also minimizing the number of the switching operations. This paper proposes a method to obtain a pareto-optimal set for the service restoration ploblem utilizing a Genetic Algorithm. By using our method, it is possible to compare an optimal solution with the others, so that a user can obtain his desired optimal solution. The conventional methods cannot realize this kind of application, therefore the proposed method is useful.
• A Note on Convergence Properties of a CORDIC-Based Adaptive Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Nonlinear problems  102-  (298)  43  -48  2002/08/28  [Not refereed][Not invited]
   

  This paper presents an analysis of convergence properties of a CORDIC-based adaptive lattice filter. First, we derive a convergence model of filter coefficients (rotation angles). Next, we analyze the behavior of the angles in steady states. According to these analysises, the convergence of the CORDIC lattice filter is proved.
• A Trainable Retrieval System for Binary Images

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  102-  (220)  1  -6  2002/07/12  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve binary images. In this method, partial features, defined as Regions and Aspects, are used to retrieve the desired images. By using these features, the similarities between an input image and the images in a database are computed. Based on the computed results, the desired image is obtained. Moreover, in the processes of obtaining the desired image, our method adopts training to reflect the user's subjectivity. The training assigns the weights for the significant Regions and Aspects based on the user's actions, such as selecting the desired image. These processes make the retrieval more effective. This paper also shows the effectiveness of the proposed method by several experiments.
• A Trainable Retrieval System for Binary Images

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  102-  (218)  1  -6  2002/07/12  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve binary images. In this method, partial features, defined as Regions and Aspects, are used to retrieve the desired images. By using these features, the similarities between an input image and the images in a database are computed. Based on the computed results, the desired image is obtained. Moreover, in the processes of obtaining the desired image, our method adopts training to reflect the user's subjectivity. The training assigns the weights for the significant Regions and Aspects based on the user's actions, such as selecting the desired image. These processes make the retrieval more effective. This paper also shows the effectiveness of the proposed method by several experiments.
• A Trainable Retrieval System for Binary Images

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  IEICE Technical Report  102-  (216)  1  -6  2002/07/12  [Not refereed][Not invited]
   

  This paper proposes a novel method to retrieve binary images. In this method, partial features, defined as Regions and Aspects, are used to retrieve the desired images. By using these features, the similarities between an input image and the images in a database are computed. Based on the computed results, the desired image is obtained. Moreover, in the processes of obtaining the desired image, our method adopts training to reflect the user's subjectivity. The training assigns the weights for the significant Regions and Aspects based on the user's actions, such as selecting the desired image. These processes make the retrieval more effective. This paper also shows the effectiveness of the proposed method by several experiments.
• A Note on Classification of MPEG Audio Data for Scene Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  102-  (219)  7  -12  2002/07/11  [Not refereed][Not invited]
   

  Transitions in audio signals are generally utilized for scene cut detection. However, since silence audio segments are usually inserted into the audio signals, the transitions in the audio signals include not only the scene cuts but also boundaries between silence and non-silence segments. Therefore, the scene cuts can not be detected by using only the transitions. In this paper, we propose an effective audio classification method for scene cut detection. Our proposed method works as follows: (1) it detects the transitions in the audio singals automatically; (2) it divides the audio signals into segments bounded by the transitions detected in (1); (3) it classifies the segments into the silence and non-silence audio classes. By using our method, the scene cuts can be detcted accurately because the the silence class contains the scene cuts. In addition, since our method can directly process MPEG Audio data without decoding procedures, detection of the scene cuts in MPEG compressed data can be applied easily.
• A Note on Classification of MPEG Audio Data for Scene Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  102-  (217)  7  -12  2002/07/11  [Not refereed][Not invited]
   

  Transitions in audio signals are generally utilized for scene cut detection. However, since silence audio segments are usually inserted into the audio signals, the transitions in the audio signals include not only the scene cuts but also boundaries between silence and non-silence segments. Therefore, the scene cuts can not be detected by using only the transitions. In this paper, we propose an effective audio classification method for scene cut detection. Our proposed method works as follows: (1) it detects the transitions in the audio singals automatically; (2) it divides the audio signals into segments bounded by the transitions detected in (1); (3) it classifies the segments into the silence and non-silence audio classes. By using our method, the scene cuts can be detcted accurately because the the silence class contains the scene cuts. In addition, since our method can directly process MPEG Audio data without decoding procedures, detection of the scene cuts in MPEG compressed data can be applied easily.
• A Note on Classification of MPEG Audio Data for Scene Cut Detection

  NITANDA Naoki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  102-  (215)  7  -12  2002/07/11  [Not refereed][Not invited]
   

  Transitions in audio signals are generally utilized for scene cut detection. However, since silence audio segments are usually inserted into the audio signals, the transitions in the audio signals include not only the scene cuts but also boundaries between silence and non-silence segments. Therefore, the scene cuts can not be detected by using only the transitions. In this paper, we propose an effective audio classification method for scene cut detection. Our proposed method works as follows: (1) it detects the transitions in the audio singals automatically; (2) it divides the audio signals into segments bounded by the transitions detected in (1); (3) it classifies the segments into the silence and non-silence audio classes. By using our method, the scene cuts can be detcted accurately because the the silence class contains the scene cuts. In addition, since our method can directly process MPEG Audio data without decoding procedures, detection of the scene cuts in MPEG compressed data can be applied easily.
• Sample Size Dependence of Estimation Error of Information Carried by Neuronal Spike Train

  MURAKAMI Jun, KIDACHI Tomohide, HASEGAWA Miki, SHIMOZAWA Tateo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  85-  (7)  1233  -1243  2002/07/01  [Not refereed][Not invited]
   

  神経系の動作性能は,神経細胞が運ぶ情報量で記述すべきである.神経細胞は刺激波形を離散的なパルス(スパイク)列に変換するパルス符号化器として振舞うため,実際の情報伝送速度は,有限長の観測時間内に得られる有限個のスパイクから推定することになり,誤差を含むことになる.本論文では,刺激-応答間のコヒーレンスから情報伝送速度を求める場合について,この推定誤差と標本サイズ(総スパイク数)との関係を解析した.その結果,スパイク当りの情報量の推定誤差は,推定に用いる総スパイク数に反比例することが得られた.また,異なる神経細胞より計測した情報量を比べるには,反比例の収束値である真の情報量を用いるべきであることを述べる.また,上記の標本数とスパイク当りの情報量の関係が成立することを,実物のコオロギの気流感覚細胞から得たデータにより確認する.
• Experimental mobile video communication using PHS

  Satoshi Miyaji, Osamu Sugimoto, Shuichi Matsumoto, Fumihide Kojima, Katsuyoshi Sato, Masayuki Fujise, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  56-  (3)  455  -462  2002/03/01  [Not refereed][Not invited]
   

  Real-time video communication was tested in a mobile environment using PHS and a very low bit-rate video codec. A vehicle equipped with PHS and the codec performed video communication with an indoor system while running along a given course. The communication protocols were PIAFS 32k (actually 29.2 kb/s) and UDP/IP. How throughput and picture quality were affected by the speed of the vehicle and how the electromagnetic-field intensity fluctuated with vehicle movement were measured. Results from this experiment revealed parameters of the video coding and packetizing ideal for use in mobile communication. The experiments showed that PHS could provide real-time mobile video communication at low to middle speeds within a service area.
• A Simplification Method for Line Drawings which Retains the Shape by Using the Fractal Dimension

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE transactions on information and systems  85-  (3)  601  -601  2002/03/01  [Not refereed][Not invited]
   

  We present a novel simplification method for line drawings which retains the shape by using the fractal dimension as an index. By using such an index, it is possible to evaluate the degree of simplificaton. Hence we can realize the simplification that retains the approximate shape of original object by observing the degree of simplification with the fractal dimension. In order to simplify retaining the approximate shape of the original object, we extract the feature points based on the state of unevenness analyzed using an approach taken from a vector tracer method.
• Experimental Moblie Video Communication using PHS

  MIYAJI Satoshi, SUGIMOTO Osamu, MATSUMOTO Shuichi, KOJIMA Fumihide, SATO Katsuyoshi, FUJISE Masayuki, HASEYAMA Miki, KITAJIMA Hideo  The Journal of the Institute of Television Engineers of Japan  56-  (3)  455  -462  2002/03/01  [Not refereed][Not invited]
   

  Real-time video communication was tested in a mobile environment using PHS and a very low bit-rate video codec. A vehicle equipped with PHS and the codec performed video communication with an indoor system while running along a given course. The communication protocols were PIAFS 32k (actually 29.2 kb/s) and UDP/IP. How throughput and picture quality were affected by the speed of the vehicle and how the electromagnetic-field intensity fluctuated with vehicle movement were measured. Results from this experiment revealed parameters of the video coding and packetizing ideal for use in mobile communication. The experiments showed that PHS could provide real-time mobile video communication at low to middle speeds within a service area.
• A Simplification Method for Line Drawings which Retains the Shape by Using the Fractal Dimension

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  The transactions of the Institute of Electronics, Information and Communication Engineers. D-II  85-  (3)  435  -447  2002/03/01  [Not refereed][Not invited]
   

  本論文では,自然形状がもつ統計的自己相似性に注目し,従来,複雑さの尺度として用いられてきたフラクタル次元を指標とした線画像の単純化手法を提案する.指標を導入することで,単純化図形の単純化の度合を定量的に評価することが可能となる.提案手法では,この指標を用いて単純化の度合を観察しながら,原図形の概形を保持した単純化を実現する.原図形の概形を保持するために提案手法では,ベクトルトレーサ法に着想を得た手法を用いて,原図形の凹凸状態を解析することによって単純化後に概形を残すために必要なデータ点を特徴点として抽出する.本論文の最後では,提案手法の有効性を示すためにいくつかの実験結果を示す.更に,客観的評価方法を導入し,既存の単純化手法により得られた図形と比較を行い,その結果も併せて示す.
• A Kalman Filter using Texure for Noise Reduction in SAR Images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers B  85-  (2)  266  -277  2002/02/01  [Not refereed][Not invited]
   

  本論文では,カルマンフィルタを用いた合成開口レーダ(SAR)画像中のノイズ低減のための手法を提案する.SAR画像中に含まれるノイズは主にスペックルより生じ,ゴマ塩状の濃淡となって現れる.スペックルはその統計的な性質が乗法性であることが知られているが,領域によってはこの性質が成り立たないことが指摘されており,十分な除去が難しい.またレーダにおいて,後方散乱波は,反射する対象物体の種類によりその期待値が変化する.提案手法は,画像をそのテクスチャによりいくつかの領域に分類し,カルマンフィルタ処理の際にパラメータを領域に応じ変化させる.このようなパラメータの変化により,実際のSAR画像中での主にスペックルにより生じるノイズ及び後方散乱波の統計的性質の変化を考慮した処理が可能となる.本論文では,ノイズの低減効果を定量的に評価するために作成したテスト画像及び実際のSAR画像に提案手法を適用し,その結果を示す.
• A Note on Impulse Detection by Using Nonlinear Filters for Removing Impulsive Noise

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (9)  43  -48  2002/01/30  [Not refereed][Not invited]
   

  This papaer proposes a new noise-detection method for restoration of images corrupted by impulsive noise. Restoration methods, which consist of an impulse detection and a noise filtering, remove the impulsive noise from corrupted according to the results of the impulse detection. Therefore, a more improved impulse detector is required for more accurate image restoration. The proposed method verifies whether each pixel once detected as an impulse is a noise or not. By using the proposed method, high accurate detection results can obtained.
• A Note on a Representation Method for Contour Shape Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (9)  49  -54  2002/01/30  [Not refereed][Not invited]
   

  This paper presents a novel representation method for contour shape using Fractal Interpolation Functions (FIF). In the traditional idea of the FIF, the scope of its application has been limited to the case where the signal is represented by single-valued function. Therefore the traditional FIF cannot be applicable to multiple-valued signals. This proposed method can model a multiple-valued signal with an extended FIF derived by introducing new parameters to the traditional one. Furthermore the fractal dimension known as a measure of complexity is used to determine the parameters needed by the FIF and thereby the signal can be modeled based on its complexity. Experimental results show the validity of the proposed method.
• A Note on Solving Method for Nurse Scheduling Problem with a Genetic Algorithm

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (9)  125  -130  2002/01/30  [Not refereed][Not invited]
   

  An optimization method of a genetic algorithm for the problem of scheduling nurses has been proposed. This method optimizes the performance index that includes the weighted sum of the objective functions. However, the optimal solutions to the nurse scheduling problem are very difficult to achieve when each objective function has a different dynamic range. Therefore, this paper proposes a novel method to set the weighting factor with the standard deviations of the objective functions. The proposed method makes it possible to adopt any objective function and to actualize a practical scheduling system.
• Improvement of a Search Algorithm Using a Genetic Algorithm and Simulated Annealing for a Correspondence Problem

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (9)  131  -136  2002/01/30  [Not refereed][Not invited]
   

  This paper proposes a method which improves a search algorithm using a genetic algorithm and simulated annealing for a correspondence problem. The previous method may get caught in a local minimum. Due to an improvement in the search algorithm the proposed method detects good genes of each locus in search process of the previous method.
• A High-Accuracy Fractal Image Coding : A Note on a Decision Method for IFS Parameters

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  26-  (9)  137  -142  2002/01/30  [Not refereed][Not invited]
   

  In the fractal image coding, it is significant to determine a suitable contractive mapping, because the mapping has a serious influence on reconstructed images. The traditional fractal image coding methods determine it by minimizing the error (collage error) between an original image and an image obtained by once applying the mapping to the original. Therefore, minimizing the collage error dose not correspond to minimizing the error between the original image and the reconstructed image, that is the image obtained by decoding in the fractal image coding. Hence, in order to obtain the high-accuracy reconstructed image, we propose a new criterion-based method by using GA and SA to determine suitable IFS parameters that represent the best contractive mapping. The proposed method always provide better reconstructed images than the other conventional methods.
• A Note on Impulse Detection by Using Nonlinear Filters for Removing Impulsive Noise

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  101-  (625)  43  -48  2002/01/23  [Not refereed][Not invited]
   

  This paper proposes a new noise-detection method for restoration of images corrupted by impulsive noise. Rcstoration methods, which consist of an impulse detection and a noise filtering, remove the impulsive noise from corrupted images according to the results of the impulse detection. Therefore, a more improved impulse detector is required for more accurate image restoration. The proposed method verifies whether each pixel once detected as an impulse is a noise or not. By using the proposed method, high accurate detection results can be obtained.
• A Note on a Representation Method for Contour Shape Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  101-  (625)  49  -54  2002/01/23  [Not refereed][Not invited]
   

  This paper presents a novel representation method for contour shape using Fractal Interpolation Functions (FIF). In the traditional idea of the FIF, the scope of its application has been limited to the case where the signal is represented by single-valued function. Therefore the traditional FIF cannot be applicable to multiple-valued signals. The proposed method can model a multiple-valued signal with an extended FIF derived by introducing new parameters to the traditional one. Furthermore the fractal dimension known as a measure of complexity is used to determine the parameters needed by the FIF and thereby the signal can be modeled based on its complexity. Experimental results show the validity of the proposed method.
• A Note on Solving Method for Nurse Scheduling Problem with a Genetic Algorithm

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  101-  (625)  125  -130  2002/01/23  [Not refereed][Not invited]
   

  An optimization method of a genetic algorithm for the problem of scheduling nurses has been proposed. This method optimizes the performance index that includes the weighted sum of the objective functions. However, the optimal solutions to the nurse scheduling problem are very difficult to achieve when each objective function has a different dynamic range. Therefore, this paper proposes a novel method to set the weighting factor with the standard deviations of the objective functions. The proposed method makes it possible to adopt any objective function and to actualize a practical scheduling system.
• Improvement of a Search Algorithm Using a Genetic Algorithm and Simulated Annealing for a Correspondence Problem

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  101-  (625)  131  -136  2002/01/23  [Not refereed][Not invited]
   

  This paper proposes a method which improves a search algorithm using a genetic algorithm and simulated annealing for a correspondence problem. The previous method may get caught in a local minimum. Due to an improvement in the search algorithm the proposed method detects good genes of each locus in search process of the previous method.
• A High-Accuracy Fractal Image Coding : A Note on a Decision Method for IFS Parameters

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. ITS  101-  (625)  137  -142  2002/01/23  [Not refereed][Not invited]
   

  In the fractal image coding, it is significant to determine a suitable contractive mapping, because the mapping has a serious influence on reconstructed images. The traditional fractal image coding methods determine it by minimizing the error (collage error) between an original image and an image obtained by once applying the mapping to the original. Therefore, minimizing the collage error dose not correspond to minimizing the error between the original image and the reconstructed image, that is the image obtained by decoding in the fractal image coding. Hence, in order to obtain the high-accuracy reconstructed image, we propose a new criterion-based method by using GA and SA to determine suitable IFS parameters that represent the best contractive mapping. The proposed method always provide better reconstructed images than the other conventional methods.
• A Note on Impulse Detection by Using Nonlinear Filters for Removing Impulsive Noise

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (627)  43  -48  2002/01/23  [Not refereed][Not invited]
   

  This paper proposes a new noise-detection method for restoration of images corrupted by impulsive noise. Restoration methods, which consist of an impulse detection and a noise filtering, remove the impulsive noise from corrupted images according to the results of the impulse detection. Therefore, a more improved impulse detector is required for more accurate image restoration. The proposed method verifies whether each pixel once detected as an impulse is a noise or not. By using the proposed method, high accurate detection results can be obtained.
• A Note on a Representation Method for Contour Shape Using Fractal Interpolation Functions

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (627)  49  -54  2002/01/23  [Not refereed][Not invited]
   

  This paper presents a novel representation method for contour shape using Fractal Interpolation Functions (FIF). In the traditional idea of the FIF, the scope of its application has been limited to the case where the signal is represented by single-valued function. Therefore the traditional FIF cannot be applicable to multiple-valued signals. The proposed method can model a multiple-valued signal with an extended FIF derived by introducing new parameters to the traditional one. Furthermore the fractal dimension known as a measure of complexity is used to determine the parameters needed by the FIF and thereby the signal can be modeled based on its complexity. Experimental results show the validity of the proposed method.
• A Note on Solving Method for Nurse Scheduling Problem with a Genetic Algorithm

  OOYA Shin, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (627)  125  -130  2002/01/23  [Not refereed][Not invited]
   

  An optimization method of a genetic algorithm for the problem of scheduling nurses has been proposed. This method optimizes the performance index that includes the weighted sum of the objective functions. However, the optimal solutions to the nurse scheduling problem are very difficult to achieve when each objective function has a different dynamic range. Therefore, this paper proposes a novel method to set the weighting factor with the standard deviations of the objective functions. The proposed method makes it possible to adopt any objective function and to actualize a practical scheduling system.
• Improvement of a Search Algorithm Using a Genetic Algorithm and Simulated Annealing for a Correspondence Problem

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (627)  131  -136  2002/01/23  [Not refereed][Not invited]
   

  This paper proposes a method which improves a search algorithm using a genetic algorithm and simulated annealing for a correspondence problem. The previous method may get caught in a local minimum. Due to an improvement in the search algorithm the proposed method detects good genes of each locus in search process of the previous method.
• A High-Accuracy Fractal Image Coding : A Note on a Decision Method for IFS Parameters

  TAKEZAWA Megumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (627)  137  -142  2002/01/23  [Not refereed][Not invited]
   

  In the fractal image coding, it is significant to determine a suitable contractive mapping, because the mapping has a serious influence on reconstructed images. The traditional fractal image coding methods determine it by minimizing the error (collage error) between an original image and an image obtained by once applying the mapping to the original. Therefore, minimizing the collage error dose not correspond to minimizing the error between the original image and the reconstructed image, that is the image obtained by decoding in the fractal image coding. Hence, in order to obtain the high-accuracy reconstructed image, we propose a new criterion-based method by using GA and SA to determine suitable IFS parameters that represent the best contractive mapping. The proposed method always provide better reconstructed images than the other conventional methods.
• Estimation of the vanishing point for automatic driving system using a cross ratio

  Masato Kazui, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  33-  (9)  31  -40  2002  [Not refereed][Not invited]
   

  This paper proposes a new method to estimate the vanishing point used as the vehicle heading, which is essential in automatic driving systems. The proposed method uses a cross ratio comprised of a ratio of lengths from four collinear points for extracting the edges that shape the vanishing point. Then, lines that intersect at one point are fitted to the edges in a Hough space. Consequently, the vanishing point is estimated robustly even when the lane markings are occluded by other vehicles. In the presence of lane markings, the road boundaries are also estimated at the same time. Experimental results from images of a real road scene show the effectiveness of the proposed method.
• Efficient contour shape description by using fractal interpolation functions

  S Uemura, M Haseyama, H Kitajima  2002 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOL I, PROCEEDINGS  1-  485  -488  2002  [Not refereed][Not invited]
   

  This paper presents a novel representation method for contour shape using Fractal Interpolation Functions (FIF). In the traditional idea of the FIF, the scope of its application has been limited to the case where the signal is represented by a single-valued function. Therefore, the traditional FIF cannot be applicable to multiple-valued signals. The proposed method can model a multiple-valued signal with an extended FIF derived by introducing new parameters to the traditional one. Furthermore, the proposed method utilizes the fractal dimension known as a measure of complexity to determine the parameters in the FIF and thereby can model the signal based on its complexity. Experimental results show the validity of the proposed method.
• An accurate noise detector for image restoration

  K Kondo, M Haseyama, H Kitajima  2002 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOL I, PROCEEDINGS  1-  321  -324  2002  [Not refereed][Not invited]
   

  This paper proposes a new noise-detection method for restoration of images corrupted by impulsive noise. The proposed method consists of two stages. In the first stage, the pixels classified according to a new flag image are processed by different noise detectors. They are realized by using two median filters with different sizes of windows. In the second stage, each pixel once detected as an impulse in the first stage is verified by using a new system. According to the above stages, the proposed method can accurately detect the location of the impulsive noise and be effectively used as a preprocessor for noise reduction filtering. Experiments show that the proposed method can effectively detect impulsive noises in noisy images even when they are very highly corrupted.
• A cost-effective and high-precision architecture for cordic-based adaptive lattice filters

  S Shiraishi, M Haseyama, H Kitajima  2002 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL V, PROCEEDINGS  5-  297  -300  2002  [Not refereed][Not invited]
   

  This paper presents a CORDIC-based architecture for adaptive lattice filters. The proposed filter architecture consists of simple components: a CORDIC processor and an adder, so that it can be implemented with a reduced amount of hardware. Moreover, the proposed architecture is useful for ASIC design because it has a regular, modular, and locally-connected structure. In addition, since our architecture is effective even in case of ARMA lattice filters, it can be utilized for many applications in the digital signal processing field.
• A fast, multi-class algorithm for classifying aerial images

  John W. Gates, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  56-  (5)  826  -835  2002  [Not refereed][Not invited]
   

  An approach is described for classifying images that does not require first dividing an image into small blocks and then classifying it based on the features of the individual blocks, as do traditional approaches. Instead the line features of the entire image are extracted and used, along with the pixel intensity, to classify each pixel in the image. To increase the classification speed, a fast line-extraction algorithm has been developed that extracts the line features directly from the original image without pre-processing. A classification tree with single variable splits is used to classify the image. Testing of a five-class aerial-image classification algorithm showed that it had an average error rate of 17.6%. Running on a 600-MHz Pentium III processor, it had an average classification time of 2.18 seconds for 512 × 512 grayscale images. This approach can be used for many different applications by training the classification tree with the desired classes.
• An IFS-parameter quantization method for fractal image coding with genetic algorithms

  Megumi Takezawa, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  56-  (10)  1633  -1642  2002  [Not refereed][Not invited]
   

  This paper proposes a high-accuracy quantization method for IFS parameters in fractal image coding by using genetic algorithms (GA). The development of IFS-parameter quantization methods is significant for image coding because quantization errors have a negative influence on reconstructed image quality. The conventional method quantizes the IFS parameters to the nearest possible values and thus minimizes the quantization errors. However, it does not necessarily minimize the errors in images reconstructed from an original. Therefore, a new quantization method which minimizes that error is proposed. The proposed method consists of two GAs because if a simple GA searches for the optimal quantization-parameter set, it converges to local optima because of the complexity of the search space. Experimental results verify that the proposed method can effectively find the optimal quantization-parameter set and provide high-quality reconstructed images.
• A Hardware Architecture for an Affine-Based Motion Estimation

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (84)  1  -6  2001/12/14  [Not refereed][Not invited]
   

  This paper proposes an architecture for an affine-based motion estimation algorithm. Although an affine-based motion estimation algorithm can estimate accurate motion from image sequences, it requires high computational power. In order to avoid such computational complexity, we propose a hardware architecture for the motion estimator. Furthermore, we propose a method to achieve high throughput and low hardware complexity in the motion estimator. In this paper, we describe the motion estimator in VHDL and then show the result of design using an EPGA.
• A Method to Define Motion Vectors for Improving the Rate-Distortion Performance of Video Coding

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (84)  7  -12  2001/12/14  [Not refereed][Not invited]
   

  In ordinary Block Matching Algorithms(BMA), motion vectors are defined at the point to minimize the Mean Squared Error(MSE)or the Mean Absolute Error(MAE)of the intensity. However, in certain background images, the motion vectors whose MSE is minimum may not yield the best rate-distortion performance. Therefore, we propose a method to define the motion vectors which yield the best rate-distortion performance which is composed of three elements, the bit rate required for motion compensated predictive coding, the bit-rate required for prediction error transforming coding, and the reconstruction distortion. If the codes is H.263, the proposed method can reduce the total bit rate approximately 10%-50% with the same PSNR as compared to the conventional methods.
• A Hardware Architecture for an Affine-Based Motion Estimation

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2001-  (121)  61  -66  2001/12/13  [Not refereed][Not invited]
   

  This paper proposes an architecture for an affine-based motion estimation algorithm. Although an affine-based motion estimation algorithm can estimate accurate motion from image sequences, it requires high computational power. In order to avoid such computational complexity, we propose a hardware architecture for the motion estimator. Furthermore, we propose a method to achieve high throughput and low hardware complexity in the motion estimator. In this paper, we describe the motion estimator in VHDL and then show the result of design using an FPGA.
• A Method Define Motion Vectors for Improving the Rate-Distortion Performance of Video Coding

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2001-  (121)  67  -72  2001/12/13  [Not refereed][Not invited]
   

  In ordinary Block Matching Algorithms((BMA), motion vectors are defined at the point to minimize the Mean Squared Error(MSE) or the Mean Absolute Error(MAE) of the intensity. However, in certain background images, the motion vectors whose MSE is minimum may not yield the best rate-distortion performance. Therefore, we propose a method to define the motion vectors which yield the best rate-distortion performance which is composed of three elements, the bit rate required for motion compensated predictive coding, the bit rate required for prediction error transforming coding, and the reconstruction distortion. If the codec is H.263, the proposed method can reduce the total bit rate approximately 10%-50% with the same PSNR as compared to the conventional methods.
• A Hardware Architecture for an Affine-Based Motion Estimation

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (497)  1  -6  2001/12/07  [Not refereed][Not invited]
   

  This paper proposes an architecture for an affine-based motion estimation algorithm. Although an affine-based motion estimation algorithm can estimate accurate motion from image sequences, it requires high computational power. In order to avoid such computational complexity, we propose a hardware architecture for the motion estimator. Furthermore, we propose a method to achieve high throughput and low hardware complexity in the motion estimator. In this paper, we describe the motion estimator in VHDL and then show the result of design using an FPGA.
• A Method to Define Motion Vectors for Improving the Rate-Distortion Performance of Video Coding

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  101-  (497)  7  -12  2001/12/07  [Not refereed][Not invited]
   

  In ordinary Block Matching Algorithms(BMA), motion vectors are defined at the point to minimize the Mean Squared Error(MSE)or the Mean Absolute Error(MAE)of the intensity. However, in certain background images, the motion vectors whose MSE is minimum may not yield the best rate-distortion performance. Therefore, we propose a method to define the motion vectors which yield the best rate-distortion performance which is composed of three elements, the bit rate required for motion compensated predictive coding, the bit rate required for prediction error transforming coding, and the reconstruction distortion. If the codec is H.263, the proposed method can reduce the total bit rate approximately 10%〜50% with the same PSNR as compared to the conventional methods.
• A Hardware Architecture for an Affine-Based Motion Estimation

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  101-  (495)  1  -6  2001/12/07  [Not refereed][Not invited]
   

  This paper proposes an architecture for an affine-based motion estimation algorithm. Although an Affine-based motion estimation algorithm can estimate accurate motion from image sequences, it requires high computational power. In order to avoid such computational complexity, we propose a hardware architecture for the motion estimator. Furthermore, we propose a method to achieve high throughput and low hardware complexity in the motion estimator. Inthis paper, we describe the motion estimator in VHDL and then show the result of design using an FPGA.
• A Method to Define Motion Vectors for Improving the Rate-Distortion Performance of Video Coding

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  101-  (495)  7  -12  2001/12/07  [Not refereed][Not invited]
   

  In ordinary Block Matching Algorithms(BMA), motion vectors are defined at the point to minimize the Mean Squared Error(MSE)or the Mean Absolute Error(MAE)of the intensity. However, in certain background images, the motion vectors whose MSE is minimum may not yield the best rate-distortion performance. Therefore, we propose a method to define the motion vectors which yield the best rate-distortion performance which is composed of three elements, the bit rate required for motion compensated predictive coding, the bit rate required for prediction error transforming coding, and the reconstruction distortion. If the codec is H.263, the proposed method can reduce the total bit rate approximately 10%-50% with the same PSNR as compared to the conventional methods.
• A new fitness function of a genetic algorithm for routing applications

  Jun Inagaki, Miki Haseyama, Hideo Kitajima  IEICE Transactions on Information and Systems  E84-D-  277  -280  2001/12/01  [Not refereed][Not invited]
   

  SUMMARY This paper presents a method of determining a fitness function in a genetic algorithm for routing the shortest route via several designated points. We can search for the optimum route efficiently by using the proposed fitness function and its validity is verified by applying it to the actual map data.
• Two-dimensional Analysis of the Magnetic Microstructure of Media Using MFM Images

  TAKEKUMA I, HASEYAMA M, SUEOKA K, MUKASA K  Journal of Magnetics Society of Japan  25-  (10)  1535  -1540  2001/10/01  [Not refereed][Not invited]
   

  This paper analyzes two-dimensional features of the magnetic microstructure of longitudinal thin-film media. Several MFM analyses of the magnetic cluster size have been presented, but the two-dimensional features of the clusters are not well understood. In order to determine these features, we focused on the distribution of magnetic poles at the boundaries of magnetic clusters, and quantitatively analyzed it by using two kinds of image processing methods: binarization and thinning methods. The results were as follows: (1) we found that some media showed dependence on the specific direction in the distribution of magnetic poles in the dc-erased state, whose features are closely related to the magnetization fluctuations in transition regions, and (2) media with larger fluctuations in the transition region had larger N_m/S_o values and partial erasure (PE) probabilities.
• Fast Line Extraction form Digital Images Using Line Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  84-  (8)  1566  -1579  2001/08/01  [Not refereed][Not invited]
   

  本論文は,ディジタル直線を構成する線分の方向性を用いた高速直線抽出法を提案する.画像中の物体の特徴を抽出する際,その輪郭を線分の組合せとして抽出することは重要である.短い線分と長い線分とが混在している場合,従来手法は,抽出精度が不十分であり,処理時間も長い.本論文においては,(1)1本のアナログ直線は,対応する離散画像においては,離散化された複数の線分が連結されたものとして表現されること,(2)与えられるアナログ直線の傾きの範囲によって,対応するディジタル画像中の線分の方向が異なることに着目する.各々の線分の4方向に対応するアナログ直線の傾きの範囲内で,線分を追跡することにより直線抽出の高速化を図る.評価実験の結果により,処理速度,ロバスト性,結果の精度において提案手法が従来手法に対して優れていることを示す.
• Adaptive Use of Long-Term Memory for Motion Compensation

  Rivero Jorge F., Kitajima Hideo, Haseyama Miki  研究会講演予稿  187-  11  -16  2001/07/25
• A Recgnition and Retrieval System for Binary Images

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  101-  (204)  99  -104  2001/07/11  [Not refereed][Not invited]
   

  This paper proposes a new recognition and retrieval system for binary images. In this system, an image is segmented into several areas that have several different elements(e.g. circular measure, degree of occlusion, etc.). The similarity between an input image and images in the database is calculated by the elements. One of the significant points in this system is that it is a trainable system, that dignifies several points that are principal in the images. After training, the data in the image database system are renewed to get recognition rate better. This paper shows effectiveness of the proposed method and the training by some experiments.
• Improvement of a Search Algorithm using a Genetic Algorithm and Simulated Annealing for the Correspondence Problem

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  101-  (204)  119  -124  2001/07/11  [Not refereed][Not invited]
   

  The authors have proposed a method which determines the correspondence of the same object in two different images by using a genetic algorithm and simulated annealing. In the case where the algorithm has no knowledge about the scale of the object, the previous method needed to search many times. Due to an improvement in the search algorithm the proposed method can find the optimum solution with a single search. This paper also verifies the validity of the proposed method by several experiments.
• Image Restoration Using an Adaptive Kalman Filter with Edge Preservation

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  101-  (204)  125  -130  2001/07/11  [Not refereed][Not invited]
   

  In this paper, we propose an edge-adaptive image modeling approach to preserve edges with greater noise reduction. The general Kalman filter for image restoration removes most of the high-frequency components from the image thus reducing the noise. However, this also has the undersirable effect of smoothing edge resulting in blurring or ringing. The proposed method classifies the pixels of an observed image in the flat region and the four edge regions, and estimates the image model for each region. It is possible to restore an image more accurately without excessive smoothing. The five Kalman filters which are designed for each part perform the restoration adaptively in each region. As a result, we obtain an accurate image while preserving the edges.
• A Note on Impulse Noise Detector for Image Restoration

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  101-  (204)  131  -136  2001/07/11  [Not refereed][Not invited]
   

  This paper proposes a new impulse noise detection-method. PSM(progressive switching median) filter, which is composed of an impulse noise detector and a noise filter, has been reported to restore images corrupted by impulse noise. However, the impulse noise detector sometimes fails in noise deteciton to pixels in edges. Therefore, we propose a new impulse noise detection-method by using two median filters with different sizes of woindows. By using the proposed detection-method, we can accurately restore the corrupted images in the edges.
• Active Contour Models for Extracting Facial Images

  YONEYAMA Shogo, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. Multimedia and virtual environment  101-  (204)  137  -142  2001/07/11  [Not refereed][Not invited]
   

  Many different methods for extracting contours using active contour models (Snakes) have been proposed. One method is to use a snake that tries to conform to an a priori shape. This method is used to reduce the sensitivity of the snake to noise, however, one of problems in this method is that if the initial snake includes some objects, the snake can not extract the expected one. To overcome the problem of including objects, we propose to split the snake by using the area term method which has been proposed for extracting concave objects.
• A Recgnition and Retrieval System for Binary Images

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  101-  (202)  99  -104  2001/07/11  [Not refereed][Not invited]
   

  This paper proposes a new recognition and retrieval system for binary images. In this system, an image is segmented into several areas that have several different elements(e.g.circular measure, degree of occlusion, etc.). The similarity between an input image and images in the database is calculated by the elements. One of the significant points in this system is that it is a trainable system, that dignifies several points that are principal in the images. After training, the data in the image database system are renewed to get recognition rate better. This paper shows effectiveness of the proposed method and the training by some experiments.
• Improvement of a Search Algorithm using a Genetic Algorithm and Simulated Annealing for the Correspondence Problem

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  101-  (202)  119  -124  2001/07/11  [Not refereed][Not invited]
   

  The authors have proposed a method which determines the correspondence of the same object in two different images by using a genetic algorithm and simulated annealing. In the case where the algorithm has no knowledge about the scale of the object, the previous method needed to search many times. Due to an improvement in the search algorithm the proposed method can find the optimum solution with a single search. This paper also verifies the validity of the proposed method by several experiments.
• Image Restoration Using an Adaptive Kalman Filter with Edge Preservation

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  101-  (202)  125  -130  2001/07/11  [Not refereed][Not invited]
   

  In this paper, we propose an edge-adaptive image modeling approach to preserve edges with greater noise reduction. The general Kalman filter for image restoration removes most of the high-frequency components from the image thus reducing the noise. However, this also has the undersirable effect of smoothing edge resulting in blurring or ringing. The proposed method classifies the pixels of an observed image in the flat region and the four edge regions, and estimates the image model for each region. It is possible to restore an image more accurately without excessive smoothing. The five Kalman filters which are designed for each part perform the restoration adaptively in each region. As a result, we obtain an accurate image while preserving the edges.
• A Note on Impulse Noise Detector for Image Restoration

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  101-  (202)  131  -136  2001/07/11  [Not refereed][Not invited]
   

  This paper proposes a new impulse noise detection-method. PSM(progressive switching median) filter, which is composed of an impulse noise detector and a noise filter, has been reported to restore images corrupted by impulse noise. However, the impulse noise detector sometimes fails in noise deteciton to pixels in edges. Therefore, we propose a new impulse noise detection-method by using two median filters with different sizes of woindows. By using the proposed detection-method, we can accurately restore the corrupted images in the edges.
• Active Contour Models for Extracting Facial Images

  YONEYAMA Shogo, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. PRMU  101-  (202)  137  -142  2001/07/11  [Not refereed][Not invited]
   

  Many different methods for extracting contours using active contour models (Snakes) have been proposed. One method is to use a snake that tries to conform to an a priori shape. This method is used to reduce the sensitivity of the snake to noise, however, one of problems in this method is that if the initial snake includes some objects, the snake can not extract the expected one. To overcome the problem of including objects, we propose to split the snake by using the area term method which has been proposed for extracting concave objects.
• The Estimation of the Vanishing Point for Automatic Driving Systems Using a Cross Ratio

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  84-  (7)  1319  -1327  2001/07/01  [Not refereed][Not invited]
   

  本論文では, 自動走行システムが直線路において車両の進行方向を決定するための消失点を推定する新たな手法を提案する. 本手法では一直線上の4点から計算される複比を用いて, 画像から消失点を形成するエッジを抽出する. その後, 抽出されたエッジに対して, 1点で交差する直線群をHough空間においてフィッティングさせることにより消失点を推定する. 従来の代表的な手法では, 道路構造やレーン形状の大域的な拘束条件を用いることにより, 安定した道路境界推定を可能としているが, 他車両によってレーンマーキングが隠ぺいされた状況における消失点の推定に問題が残っている. これに対し, 提案手法はレーンマーキングの位置や形状に関する拘束条件を用いないので, 抽出すべき道路上の特徴をレーンマーキングに限定する必要がなく, 路肩や背景のエッジを用いて消失点を推定することができる. したがって, 提案手法では, レーンマーキングの隠ぺい等が生じても限られたエッジから消失点を安定して推定できる. 一般道路を撮影した画像を用いた実験結果により提案手法の有効性を示す.
• ICUにおけるカテーテル敗血症の発生状況についての検討 : 輸液ラインにクローズドシステムを使用して

  長谷山 雅美, 大槻 直子, 長谷山 美紀  日本看護研究学会雑誌  24-  (3)  307  -307  2001/06/27  [Not refereed][Not invited]
  
• A Gradient Adaptive Lattice Filter Algorithm for Efficient CORDIC Implementation

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  101-  (142)  25  -30  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes a new gradient adaptive lattice filter algorithm for efficient CORDIC implementation. Since the proposed filter algorithm directly updates rotation angles rather than reflection coefficients, it can be efficiently implemented with CORDIC processors. Moreover, the proposed filter architecture consists of simple components: adders and CORDIC processors, so that it can be realized with a reduced amount of hardware.
• A Note on an FPGA Design of a Motion Estimator

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  101-  (142)  45  -50  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes an architecture for an accurate motion estimation algorithm. The accurate motion estimation algorithm has a high computational cost. In order to avoid such computational complexity, we propose a method to realize the motion estimator in hardware. In this paper, we describe the motion estimator in VHDL and implement it on an FPGA device. To verify our design, we compute motion parameters from real image sequences by using the FPGA-implemented motion estimator.
• A Gradient Adaptive Lattice Filter Algorithm for Efficient CORDIC Implementation

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  101-  (146)  25  -30  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes a new gradient adaptive lattice filter algorithm for efficient CORDIC implementation. Since the proposed filter algorithm directly updates rotation angles rather than reflection coefficients, it can be efficiently implemented with CORDIC processors. Moreover, the proposed filter architecture consists of simple components: adders and CORDIC processors, so that it can be realized with a reduced amount of hardware.
• A Note on an FPGA Design of a Motion Estimator

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  101-  (146)  45  -50  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes an architecture for an accurate motion estimation algorithm. The accurate motion estimation algorithm has a high computational cost. In order to avoid such computational complexity, we propose a method to realize the motion estimator in hardware. In this paper, we describe the motion estimator in VHDL and implement it on an FPGA device. To verify our design, we compute motion parameters from real image sequences by using the FPGA-implemented motion estimator.
• A Gradient Adaptive Lattice Filter Algorithm for Efficient CORDIC Implementation

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  101-  (144)  25  -30  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes a new gradient adaptive lattice filter algorithm for efficient CORIC implementation. Since the proposed filter algorithm directly updates rotation angles rather than reflection coefficients, it can be efficiently implemented with CORDIC proccssors. Morcover, the proposed filter architecture consists of simple componcnts : adders and CORDIC processors, so that it can be realized with a reduced amount of hardware.
• A Note on an FPGA Design of a Motion Estimator

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  101-  (144)  45  -50  2001/06/22  [Not refereed][Not invited]
   

  This paper proposes an architecture for an accurate motion estimation algorithm. The accurate motion estimation algorithm has a high computational cost. In order to avoid such computational complexity, we propose a method to realize the motion estimator in hardware. In this paper, we describe the motion estimator in VHDL and implement it on an FPGA device. To verify our design, we compute motion parameters from real image sequences by using the FPGA-implemented motion estimator.
• A Study on Image Authentication Using the Affine Transfomation

  KONDO Isao, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  101-  (141)  151  -156  2001/06/21  [Not refereed][Not invited]
   

  With the advance in digital techniques and the prevalence of multimedia data in our electronic world, copyright protection of the digital contents has become an important issue. Therefore, in this paper we propose a new approach to image authentication using the affine transformation, to restrain illicit copying of images. The proposed method does not embed watermark data into original image, and thus it maintains the quality of original image.
• A Study on Image Authentication Using the Affine Transfomation

  KONDO Isao, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  101-  (145)  151  -156  2001/06/21  [Not refereed][Not invited]
   

  With the advance in digital techniques and the prevalence of multimedia data in our electronic world, copyright protection of the digital contents has become an important issue. Therefore, in this paper we propose a new approach to image authentication using the affine transformation, to restrain illicit copying of images. The proposed method does not embed watermark data into original image, and thus it maintains the quality of original image.
• A Study on Image Authentication Using the Affine Transfomation

  KONDO Isao, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  101-  (143)  151  -156  2001/06/21  [Not refereed][Not invited]
   

  With the advance in digital techniques and the prevalene of multimedia data in our electronic world, copyright protection of the digital contents has become an important issue. Therefore, in this paper we propose a new approach to image authentication using the affine transformation, to restrain illicit copying of images. The proposed method does not embed watermark data into original image, and thus it maintains the quality of original image.
• The Estimation of Curved Lane Markings Based on the Estimation of the Vanishing Point

  KAZU Masato, HASEYAMA Miki, KITAJIMA Hideo  The Journal of the Institute of Television Engineers of Japan  55-  (6)  894  -897  2001/06/20  [Not refereed][Not invited]
   

  本研究は, 自動走行システムが曲線路においてHough変換を用いてレーンマーキングを推定するための新たな手法について提案する.提案手法では, レーン形状を2次曲線で近似できる場合において, 消失点推定に基づきレーンパラメータを段階的に推定する.これにより, 曲線パラメータの推定にHough変換を用いる際の計算量を抑えることができる.計算機シミュレーションで関数フィッティングによる手法と比較を行い, 提案手法の有効性を示す.
• The estimation of curved lane markings based on the estimation of the vanishing point

  M. Kazui, M. Haseyama, H. Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  55-  894  2001/06/01  [Not refereed][Not invited]
  
• An ARMA order selection method with fuzzy reasoning

  M Haseyama, H Kitajima  SIGNAL PROCESSING  81-  (6)  1331  -1335  2001/06  [Not refereed][Not invited]
   

  A fuzzy reasoning based approach for ARMA order selection is discussed in this communication. The proposed method attempts to select the optimal ARMA order of a time-varying ARMA model. This method improves model validity-criterion based order selection, such as the Akaike's information criterion and the minimum description length, by applying a fuzzy recursive reasoning method and a fuzzy c-means clustering method. (C) 2001 Elsevier Science B.V. All rights reserved.
• A new fitness function of a genetic algorithm for routing applications

  J Inagaki, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E84D-  (2)  277  -280  2001/02  [Not refereed][Not invited]
   

  This paper presents a method of determining a fitness function in a genetic algorithm for routing the shortest route via several designated points. We can search for the optimum route efficiently by using the proposed fitness function and its validity is verified by applying it to the actual map data.
• A Simplification Method for Line Drawings with Fractal Dimension as Index

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (8)  1  -6  2001/01/30  [Not refereed][Not invited]
   

  A simplification method for line drawings is proposed. The proposed method is possible to keep significant features of the object. It is well known that natural shapes, such as shoreline and clouds, have statistical self-similarity. The fractal dimension is one of the characteristics of the object which have self-similarity, so that we have already proposed simplification method using it as index. This method have achieved to simplify the object with remaining features of them by extracting significant points from the object. However, this method cannot distinguish the projected parts from curved parts, which is important in image analysis. Therefore, this paper proposes feature analysis method which can distinguish projected parts from curved parts. It is possible to simplify the object with satisfying wider demand by applying this proposed method.
• A Nonlinear Image Restoration Method and Its Prameter Setting

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (8)  7  -12  2001/01/30  [Not refereed][Not invited]
   

  This paper proposes a new median filter-based method which can effectively remove impulse noise from highly corrupted images. The proposed method consists of two median filters with different sizes of windows. Moreover, the proposed method iteratively processes the corrupted images and consequently restore them effectively. The experimental results demonstrate that the proposed filter is better than conventional methods. This paper also discusses how to determine a parameter used in the proposed method.
• Matching Feature - Points Using the Scale of the Object with a GA and SA

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (8)  13  -18  2001/01/30  [Not refereed][Not invited]
   

  This paper proposes a method which determines the correspondence of the same objects in two different image. The previous methods which use an elastic contour model cannot determine it without knowledge about the scale of the object. The proposed method enable searching for the object without knowledge about its scale by an improvement of the genotype in a genetic algorithm. This paper also verifies the validity of proposed method by several experiments.
• A Note on Fractal Image Coding Using a GA and SA

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  25-  (8)  47  -52  2001/01/30  [Not refereed][Not invited]
   

  In fractal image coding, quantization errors of IFS parameters corrupt reconstructed images seriously. Even if the quantization errors are minimized, high-quality reconstructed images are not necessarily obtained. Hence, in order to obtain them, high-accuracy quantization methods are required. Therefore, we propose a novel method to quantize the IFS parameters accurately using a GA and SA. The proposed method finds the best set of the quantized IFS parameters, and thus it reconstructs images with high-quality.
• A Simplification Method for Line Drawings with Fractal Dimension as Index

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  100-  (604)  1  -6  2001/01/23  [Not refereed][Not invited]
   

  A simplification method for line drawings is proposed. The proposed method is possible to keep significant features of the object. It is well known that natural shapes, such as shoreline and clouds, have statistical self-similarity. The fractal dimension is one of the characteristics of the object which have self-similarity, so that we have already proposed simplification method using it as index. This method have achieved to simplify the object with remaining features of them by extracting significant points from the object. However, this method cannot distinguish the projected parts from curved parts, which is important in image analysis. Therefore, this paper proposes feature analysis method which can distinguish projected parts from curved parts. It is possible to simplify the object with satisfying wider demand by applying this proposed method.
• A Nonlinear Image Restoration Method and Its Prameter Setting

  KONDO Keiko, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  100-  (604)  7  -12  2001/01/23  [Not refereed][Not invited]
   

  This paper proposes a new median filter-based method which can effectively remove impulse noise from highly corrupted images. The proposed method consists of two median filters with different sizes of windows. Moreover, the proposed method iteratively processes the corrupted images and consequently restore them effectively. The experimental results demonstrate that the proposed filter is better than conventional methods. This paper also discusses how to determine a parameter used in the proposed method.
• Matching Feature-Points Using the Scale of the Object with a GA and SA

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  100-  (604)  13  -18  2001/01/23  [Not refereed][Not invited]
   

  This paper proposes a method which determines the correspondence of the same objects in two different images. The previous methods which use an elastic contour model cannot determine it without knowledge about the scale of the object. The proposed method enable searching for the object without knowledge about its scale by an improvement of the genotype in a genetic algorithm. This paper also verifies the validity of proposed method by several experiments.
• A Note on Fractal Image Coding Using a GA and SA

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  100-  (604)  47  -52  2001/01/23  [Not refereed][Not invited]
   

  In fractal image coding, quantization errors of IFS parameters corrupt reconstructed images seriously. Even if the quantization errors are minimized, high-quality reconstructed images are not necessarily obtained. Hence, in order to obtain them, high-accuracy quantization methods are required. Therefore, we propose a novel method to quantize the IFS parameters accurately using a GA and SA. The proposed method finds the best set of the quantized IFS parameters, and thus it reconstructs images with high-quality.
• Hopfield neural networks for edge detection

  Y. Uchiyama, M. Haseyama, H. Kitajima  Proceedings - IEEE International Symposium on Circuits and Systems  3-  608  -611  2001/01/01  [Not refereed][Not invited]
   

  In this paper, new parameters of Hopfield Neural Network (HNN) for edge detection of image are proposed. In conventional methods, two kinds of edge parameters, which detect horizontal and vertical direction edges, are introduced. In the proposed method, other four kinds of edge parameters are introduced for more accurate edge detection.
• A new genetic algorithm for routing the shortest route via several designated points

  Jun Inagaki, Miki Haseyama, Hideo Kitajima  ISCAS 2001 - 2001 IEEE International Symposium on Circuits and Systems, Conference Proceedings  2-  217  -220  2001  [Not refereed][Not invited]
   

  This paper presents a method of routing the shortest route via several designated points using a genetic algorithm and its effectiveness verified by applying the algorithm to map data. The proposed algorithm uses a fitness function which includes weight parameters based on the building block hypothesis. By using this function, the optimum is efficiently obtained. This kind of routing is difficult in the case of using existing methods such as Dijkstra's method. Also, the proposed method is more useful. © 2001 IEEE.
• An implementation of a normalized ARMA lattice filter with a CORDIC algorithm

  S Shiraishi, M Haseyama, H Kitajima  ELECTRONICS AND COMMUNICATIONS IN JAPAN PART III-FUNDAMENTAL ELECTRONIC SCIENCE  84-  (4)  76  -86  2001  [Not refereed][Not invited]
   

  In this paper we present a method of implementation of a normalized ARMA lattice filter using a CORDIC algorithm. Expansion of a previously proposed normalized AR lattice filter using the CORDIC algorithm for the purposes of its application to the normalized ARMA lattice filter is difficult due to problems related to the convergence magnitude of the rotation angle and the convergence speed. In this paper we demonstrate the possibility of overcoming the above-mentioned problems and of implementing a CORDIC ARMA lattice filter. Since the filter proposed in this work is implemented based on the CORDIC algorithm specifically adapted for the realization of the normalized ARMA lattice filter, it can be expected that it can be successively implemented in the form of a VLSI array. In addition, since such a filter performs the filtering by predicting the input signal, it can be used for the processing of unknown signals. This article also contains estimations of the effectiveness of the proposed method by experiments. (C) 2000 Scripta Technica.
• An implementation of a normalized ARMA lattice filter with a CORDIC algorithm

  S Shiraishi, M Haseyama, H Kitajima  ELECTRONICS AND COMMUNICATIONS IN JAPAN PART III-FUNDAMENTAL ELECTRONIC SCIENCE  84-  (4)  76  -86  2001  [Not refereed][Not invited]
   

  In this paper we present a method of implementation of a normalized ARMA lattice filter using a CORDIC algorithm. Expansion of a previously proposed normalized AR lattice filter using the CORDIC algorithm for the purposes of its application to the normalized ARMA lattice filter is difficult due to problems related to the convergence magnitude of the rotation angle and the convergence speed. In this paper we demonstrate the possibility of overcoming the above-mentioned problems and of implementing a CORDIC ARMA lattice filter. Since the filter proposed in this work is implemented based on the CORDIC algorithm specifically adapted for the realization of the normalized ARMA lattice filter, it can be expected that it can be successively implemented in the form of a VLSI array. In addition, since such a filter performs the filtering by predicting the input signal, it can be used for the processing of unknown signals. This article also contains estimations of the effectiveness of the proposed method by experiments. (C) 2000 Scripta Technica.
• A Study of the Estimation of Curved Lane Markings Based on a Hough Transform

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2000-  (116)  85  -90  2000/12/14  [Not refereed][Not invited]
   

  This paper proposes a new method to estimate curved lane markings for vheicle steering. The proposed method is decomposed into two steps. First, it estimates the linear parameters of curved lane markings, then it fits lines that intersect at one point to the edges in a Hough space using cross ratio. Second, an angle histogram based on the Hough transform is used to estimate the curvature of the lane markings. We study the stability of the proposed method from experimental results using images of a real road scene.
• A study of noise reduction considering the nonstationarity of speckle in SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2000-  (116)  91  -96  2000/12/14  [Not refereed][Not invited]
   

  Synthetic aperture radar(SAR)images are affected by speckle noise, and numerous methods have been proposed for reducting the noise. Speckle is assumed to have the characteristics of a random multiplicative noise whose intensity obeys a negative exponential distribution, but these characteristics are not true on some scenes, such as urban areas, in the image. In this paper, we propose a Kalman filter adapted to the regions which are classified according to texture in the SAR images.
• Vector Quantization for Images Using both a Specific Codebook and a Universal Codebook

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2000-  (116)  97  -102  2000/12/14  [Not refereed][Not invited]
   

  A method of generating the representative vectors for encoding still images efficiently using a vector quantizer, is proposed in this paper. A conventional codebook is generated by clustering many training images. But, many representative vectors are geristered in the codebook for obtaining the high quality reconstructed image, thus the index of the representative vector becomes long. To solve this problem, the conventional methods encode the index efficiently. The proposed method is another approach for reducing the number of representative vectors by using two codebooks adapted to edges and flat regions of the image respectively. It uses a specific codebook which is generated by filtering the representative vectors from the source image, and a universal codebook which is generated by using random numbers. It is shown that the coding efficiency of the proposed method is comparable to the conventional methods without encoding the index efficiently.
• A Consideration of Parameter Estimation for Image Restoration Using a Kalman Filter

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2000-  (116)  103  -108  2000/12/14  [Not refereed][Not invited]
   

  In this paper, we propose an estimation method of the parameters for accurate image restoration using a Kalman filter. If it is possible to estimate the parameters(i.e. an image model and a noise variance)with high accuracy, then an image can be restored accurately using a Kalman filter. In traditional parameter estimation methods, parameters are estimated from the whole image, and a change of the local statistics in an image is not considered. Therefore, detail and whole regions of the image are not restored accurately with the previous methods. The proposed method estimates parameters adapted to every part of image according to a change in the local statistics, and can thus restore an image more accurately. We have verified the validity of proposed method by several experiments.
• An Improvement of the Fitness Function to Increase the Accuracy of the Correspondence Problem Using a Genetic Algorithm and Simulated Anneling

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  2000-  (116)  109  -114  2000/12/14  [Not refereed][Not invited]
   

  The authors have proposed a method which determines the correspondence of feature points in two different images by using a genetic algorithm and simulated annealing. Our previous method may get caught in a local minimum. The proposed method enables serching for suitable correspondence by an improvement of the fitness function.
• A Study of the Estimation of Curved Lane Markings Based on a Hough Transform

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (78)  85  -90  2000/12/14  [Not refereed][Not invited]
   

  This paper proposes a new method to estimate curved lane markings for vheicle steering. The proposed method is decomposed into two steps. First, it estimates the linear parameters of curved lane markings, then it fits lines that intersect at one point to the edges in a Hough space using cross ratio. Second, an angle histogram based on the Hough transform is used to estimate the curvature of the lane markings. We study the stability of the proposed method from experimental results using images of a real road scene.
• A study of noise reduction considering the nonstationarity of speckle in SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (78)  91  -96  2000/12/14  [Not refereed][Not invited]
   

  Synthetic aperture radar(SAR)images are affected by speckle noise, and numerous methods have been proposed for reducting the noise. Speckle is assumed to have the characteristics of a random multiplicative noise whose intensity obeys a negative exponential distribution, but these characteristics are not true on some scenes, such as urban areas, in the image. In this paper, we propose a Kalman filter adapted to the regions which are classified according to texture in the SAR images.
• Vector Quantization for Images Using both a Specific Codebook and a Universal Codebook

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (78)  97  -102  2000/12/14  [Not refereed][Not invited]
   

  A method of generating the representative vectors for encoding still images efficiently using a vector quantizer, is proposed in this paper. A conventional codebook is generated by clustering many training images. But, many representative vectors are registered in the codebook for obtaining the high quality reconstructed image, thus the index of the representative vector becomes long. To solve this problem, the conventional methods encode the index efficiently. The proposed method is another approach for reducing the number of representative vectors by using two codebooks adapted to edges and flat regions of the image respectively. It uses a specific codebook which is generated by filtering the representative vectors from the source image, and a universal codebook which is generated by using random numbers. It is shown that the coding efficiency of the proposed method is comparable to the conventional methods without encoding the index efficiently.
• A Consideration of Parameter Estimation for Image Restoration Using a Kalman Filter

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (78)  103  -108  2000/12/14  [Not refereed][Not invited]
   

  In this paper, we propose an estimation method of the parameters for accurate image restoration using a Kalman filter. If it is possible to estimate the parameters(i.e. an image model and a noise variance)with high accuracy, then an image can be restored accurately using a Kalman filter. In traditional parameter estimation methods, parameters are estimated from the whole image, and a change of the local statistics in an image is not considered. Therefore, detail and whole regions of the image are not restored accurately with the previous methods. The proposed method estimates parameters adapted to every part of image according to a change in the local statistics, and can thus restore an image more accurately. We have verified the validity of proposed method by several experiments.
• An Improvement of the Fitness Function to Increase the Accuracy of the Correspondence Problem Using a Genetic Algorithm and Simulated Anneling

  SUZUKI Kennji, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (78)  109  -114  2000/12/14  [Not refereed][Not invited]
   

  The authors have proposed a method which determines the correspondence of feature points in two different images by using a genetic algorithm and simulated annealing. Our previous method may get caught in a local minimum. The proposed method enables serching for suitable correspondence by an improvement of the fitness function.
• A Study of the Estimation of Curved Lane Markings Based on a Hough Transform

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  100-  (501)  85  -90  2000/12/07  [Not refereed][Not invited]
   

  This paper proposes a new methodto estimate curved lane markings for vheicle steering. The proposed method is decomposed into two steps. First, it estimates the linear parameters of curved lane markings, then it fits lines that intersect at one point to the edges in a Hough space using cross ratio.Second, an angle histogram based on the Hough transform is used to estimate the curvature of the laen markings. We study the stability of the proposed method from experimental results using images of a real road scene.
• A study of noise reduction considering the nonstationarity of speckle in SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  100-  (501)  91  -96  2000/12/07  [Not refereed][Not invited]
   

  Synthetic aperture radar (SAR) images are affected by spekle noise, and numerous methods have been proposed for reducting the noise. Speckle is assumed to have the characteristics of a random multiplicative noise whose intensity obeys a negative exponential distribution, but these characteristics are not true on some scenes, such as urban areas, in the image. In this paper, we propose a Kalman filter adapted to the regions which are classified according to texture in the SAR images.
• Vector Quantization for Images Using both a Specific Codebook and a Universal Codebook

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  100-  (501)  97  -102  2000/12/07  [Not refereed][Not invited]
   

  A method of generating the representative vectors for encoding still images efficiently using a vector quantizer, is proposed in this paper. A conventional codibook is generated by clustering many training images. But, many representative vectors are registered in the codebook for obtaining the high quality reconstructed image, thus the index of the representative vector becomes long. To solve this problem, the conventional methods encode the index efficiently. The proposed method is another approach for reducing the number of representative vectors by using two codebooks adapted to edges and flat regions of the image respectively. It uses a specific codebook which is generated by filtering the representative vectors from the source image, and a universal codebook which is generated by using random numbers. It is shown that the coding efficiency of the proposed method is comparable to the conventional methods without encoding the index efficiently.
• A Consideration of Parameter Estimation for Image Restoration Using a Kalman Filter

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  100-  (501)  103  -108  2000/12/07  [Not refereed][Not invited]
   

  In this paper, we propose an estimation method of the parameters for accurate image restoration using a Kalman filter. If it is possible to estimate the parameters (i.e. an image model and a noise variance) with high accuracy, then an image can be restored accurately using a Kalman filter. In traditional parameter estimation methods, parameters are estimated from the whole image, and a change of the local statistics in an image is not considered. Therefore, detail and whole regions of the image are not restored accurately with the previous methods. The proposed method estimates parameters adapted to every part of image according to a change in the local statistics, and can thus restore an image more accurately. We have verified the validity of proposed method by several experiments.
• An Improvement of the Fitness Function to Increase the Accuracy of the Correspondence Problem Using a Genetic Algorithm and Simulated Anneling

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  100-  (501)  109  -114  2000/12/07  [Not refereed][Not invited]
   

  The authors have proposed a methodwhich determines the correspondence of feature points in two different images by using a genetic algorithm and simulated annealing. Our previous method may get caught in a local minimum. The proposed method enables serching for suitable correspondence by an improvement of the fitness function.
• A Study of the Estimation of Curved Lane Markings Based on a Hough Transform

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  100-  (499)  85  -90  2000/12/07  [Not refereed][Not invited]
   

  This paper proposes a new method to estimate curved lane markings for vheicle steering. The proposed method is decomposed into two steps. First, it estimates the linear parameters of curved lane markings, then it fits lines that intersect at one point to the edges in a Hough space using cross ratio. Second, an angle histogram based on the Hough transform is used to estimate the curvature of the lane markings. We study the stability of the proposed method from experimental results using images of a real road scene.
• A study of noise reduction considering the nonstationarity of speckle in SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  100-  (499)  91  -96  2000/12/07  [Not refereed][Not invited]
   

  Synthetic aperture radar(SAR)images are affected by speckle noise, and numerous methods have been proposed for reducting the noise. Speckle is assumed to have the characteristics of a rondom multiplicative noise whose intensity obeys a negative exponential distribution, but these characteristics are not true on some scenes, such as urban areas, in the image. In this paper, we propose a Kalman filter adapted to the regions which are classified according to texture in the SAR images.
• Vector Quantization for Images Using both a Specific Codebook and a Universal Codebook

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  100-  (499)  97  -102  2000/12/07  [Not refereed][Not invited]
   

  A method of generating the representative vectors for encoding still images efficiently using a vector quantizer, is proposed in this paper. A conventional codebook is generated by clustering many training images. But, many representative vectors are registered in the codebook for obtaining the high quality reconstructed image, thus the index of the representative vector becomes long. To solve this problem, the conventional methods encode the index efficiently. The proposed method is another approach for reducing the number of representative vectors by using two codebooks adapted to edges and flat regions of the image respectively. It uses a specific codebook which is generated by filtering the representative vectors from the source image, and a universal codebook which is generated by using random numbers. It is shown that coding efficiency of the proposed method is comparable to the conventional methods without encoding the index efficiently.
• A Consideration of Parameter Estimation for Image Restoration Using a Kalman Filter

  MATSUMURA Atsushi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  100-  (499)  103  -108  2000/12/07  [Not refereed][Not invited]
   

  In this paper, we propose an estimation method of the parameters for accurate image restoration using a Kalman filter. If it is possible to estimate the parameters(i.e.an image model and a noise variance)with high accuracy, then an image can be restored accurately using a Kalman Filter. In traditional parameter estimation methods, parameters are estimated from the whole image, and a change of the local statistics in an image is not considered. Therefore, detail and whole regions of the image are not restored accurately with the previous methods. The proposed method estimates parameters adapted to every part of image according to a change in the local statistics, and can thus restore an image more accurately. We have verified the validity of proposed method by several experiments.
• An Improvement of the Fitness Function to Increase the Accuracy of the Correspondence Problem Using a Genetic Algorithm and Simulated Anneling

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  100-  (499)  109  -114  2000/12/07  [Not refereed][Not invited]
   

  The authors have proposed a method which determines the correspondence of feature points in two different images by using a genetic algotirhm and simulated annealing. Our previous method may get caught in a local minimum. The proposed method enables serching for suitable correspondence by an improvement of the fitness function.
• IFS optimization using discrete parameter pools

  Hiroyuki Honda, Miki Haseyama, Hideo Kitajima  IEICE Transactions on Information and Systems  E83-D-  233  -241  2000/12/01  [Not refereed][Not invited]
   

  This paper proposes an Iterated Function System (IFS) which can reduce effects of quantization errors of the IFS parameters. The proposed method skips conventional analog-parameter search and directly selects optimum IFS parameters from pools of discrete IFS parameters. In conventional IFS-based image coding the IFS parameters are quantized after their analog optimum values are determined. The image reconstructed from the quantized parameters is degraded with errors that are traced back to quantization errors amplified in the iterated mappings. The effectiveness of this new realistic approach is demonstrated by simulation results over the conventional method.
• Trial of MRSA-infection control in ICU: Use of Hi-Clo-Soft-Water for mouth care

  HASEYAMA Masami, KAWAMOTO Mayuko, HATAYAMA Nobuko, TAKEUCHI Akiko, HASEYAMA Miki, SAKURAI Shigeko  日本集中治療医学会雑誌 = Journal of the Japanese Society of Intensive Care Medicine  7-  (4)  387  -388  2000/10/01  [Not refereed][Not invited]
  
• A sound radiation modeling of loudspeaker based on Inverse Numerical Acoustics

  TAMURA Masakatsu, AKIHO Masaichi, HASEYAMA Miki, KITAJIMA Hideo  日本音響学会研究発表会講演論文集  2000-  (2)  395  -396  2000/09/01  [Not refereed][Not invited]
  
• Two-dimensional analysis of magnetic microstructure of media

  TAKEKUMA I, HASEYAMA M, SUEOKA K, MUKASA K  日本応用磁気学会学術講演概要集 = Digest of ... annual conference on magnetics in Japan  24-  11  -11  2000/09/01  [Not refereed][Not invited]
  
• Application of Genetic Algorithm and Simulated Annealing for the Correspondence Problem Using an Elastic Contour Model

  SUZUKI Kenji, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (210)  1  -6  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes a method which determines the correspondence of feature points in two different images by using an elastic contour model. The previous method is not practial from the view point of the calculation cost when the number of the feature points extracted from the images increases and the object has moved a large distance from the initial position. The proposed method enables searching for suitable correspondence by application of a genetic algorithms and simulated annealing.
• A fast line detecting method using line-segments in Digital images

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (210)  7  -12  2000/07/25  [Not refereed][Not invited]
   

  We describe a fast line deteting method using the line-segments of a digital line in a digital image. An analogue line is represented as a discretized digital line in a digital image. Thus, We can analyze segments, which consist of pixels, with the properties of digital lines. When an analog line's slope and range are represented as a digital line, the line segment's diretion is changed. This algorithm attains high speed by tracking each line segment's direction which comes from the analog line's slope and range. Experimental results are included to show that the proposed method is fast and can reduce line fragments.
• A Note on Simplification for 2-D Shapes Using Vector Tracer Method

  UEMURA Satoshi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (210)  13  -18  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes a new simplification method using the fractal dimension, which is one of the characteristics of an object which has a statistical self-similarity, as an indicator of the level of simplification. It is possible to simplify the object keeping relatively strong features, because feature points are extracted by the result of the analysed state of unevenness of the original images that have strong relations to the fractal dimension.
• A Study Concerned with Choosing Motion Vectors for Low Bit-rate Coding of Moving Pictures

  NATSUI Tomoyoshi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (210)  19  -24  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes an original method concerned with choosing motion vectors for block matching to realize low bit-rate coding of moving pictures. We show the possibility of improving the conventional method for choosing motion vectors by analyzing the translational model of block matching. The proposed method chooses the motion vectors which contribute to lowering the bit rate. To evaluate its performance we calculate the mean of the bit rate and the quality of the pictures by hybrid coding of the plural frames of the moving pictures. The simulation results show it can reduce the total bit-rate to about two thirds of the bit rate of the conventional method in some cases. It is effective for moving pictures having little motion and a small value for the variance.
• A Note on a Design of a Motion Estimator using VHDL

  NAKAGAWA Kazushi, SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (210)  25  -30  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes an architecture for an accurate motion estimation algorithm. The accurate motion estimation algorithm requires high computational power. In order to avoid such computational complexity, we propose a method to realize the motion estimator in hardware. In this paper, we describe the motion estimator in VHDL which is one of the hardware description languages and verify ourk design.
• Application of Genetic Algorithm and Simulated Annealing for the Correspondence Problem Using an Elastic Contour Model

  Suzuki Kenji, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (212)  1  -6  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes a method which determines the correspondence of feature points in two different images by using an elastic contour model. The previous method is not practical from the view point of the calculation cost when the number of the feature points extracted from the images increases and the object has moved a large distance from the initial position. The proposed method enables searching for suitable correspondence by application of a genetic algorithms and simulated annealing.
• A fast line detecting method using line-segments in Digital images

  Kim Euijin, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (212)  7  -12  2000/07/25  [Not refereed][Not invited]
   

  We describe a fast line deteting method using the line-segments of a digital line in a digital image. An analogue line is represented as a discretized digital line in a digital image. Thus, We can analyze segments, which consist of pixels, with the properties of digital lines. When an analog line's slope and range are represented as a digital line, the line segment's direction is changed. This algorithm attains high speed by tracking each line segment's direction which comes from the analog line's slope and range. Experimental results are included to show that the proposed method is fast and can reduce line fragments.
• A Note on Simplification for 2-D Shapes Using Vector Tracer Method

  Uemura Satoshi, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (212)  13  -18  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes a new simplification method using the fractal dimension, which is one of the characteristics of an object which has a statistical self-similarity, as an indicator of the level of simplification. It is possible to simplify the object keeping relatively strong features, because feature points are extracted by the result of the analysed state of unevenness of the original images that have strong relations to the fractal dimension.
• A Study Concerned with Choosing Motion Vectors for Low Bit-rate Coding of Moving Pictures

  Natsui Tomoyoshi, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (212)  19  -24  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes an original method concerned with choosing motion vectors for block matching to realize low bit-rate coding of moving pictures. We show the possibility of improving the conventional method for choosing motion vectors by analyzing the translational model of block matching. The proposed method chooses the motion vectors which contribute to lowering the bit rate. To evaluate its performance we calculate the mean of the bit rate and the quality of the pictures by hybrid coding of the plural frames of the moving pictures. The simulation results show it can reduce the total bit-rate to about two thirds of the bit rate of the conventional method in some cases. It is effective for moving pictures having little motion and a small value for the variance.
• A Note on a Design of a Motion Estimator using VHDL

  Nakagawa Kazushi, Shiraishi'Shin ichi, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (212)  25  -30  2000/07/25  [Not refereed][Not invited]
   

  This paper proposes an architecture for an accurate motion estimation algorithm. The accurate motion estimation algorithm requires high computational power. In order to avoid such computational complexity, we propose a method to realize the motion estimator in hardware. In this paper, we describe the motion estimator in VHDL which is one of the hardware description languages and verify ourk design.
• A Note on an Implementation of a Normalized ARMA Lattice Filter Using the Double Rotation CORDIC Algorithm

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (209)  19  -24  2000/07/24  [Not refereed][Not invited]
   

  This paper proposes a new CORDIC-based architecture for a normalized ARMA lattice filter. A CORDIC implementation method for the normalized ARMA lattice filter has been reported. However, the method has a problem that its CORDIC lattice elementary section cannot execute the rotation operation by small angles. Therefore, we propose a new lattice elementary section using the double rotation CORDIC algorithm without the problem. By using the proposed CORDIC lattice elementary section, we can realize an accurate CORDIC ARMA lattice filter.
• A study of noise reduction using scene homogeneity in SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (209)  79  -84  2000/07/24  [Not refereed][Not invited]
   

  Synthetic aperture rader(SAR) images are affected by speckle noise, and numerous methods have been proposed for the noise reduction. The statistical nature of speckle noise is well known in optics, but these characteristics are not true on some scenes, such as urban areas in the image. In this paper, we propose the a Kalman filter adapted to the regions which are classified according to the local coefficient of variation which measures the image homogeneity in the SAR images.
• A Note on Convergence of a Genetic Algorithm for Routing Application

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  100-  (209)  85  -90  2000/07/24  [Not refereed][Not invited]
   

  The authors have proposed a method of searching the shortest route through several designated points with a genetic algorithm. It uses a fitness function including weight parameters which are controlled based on a building block hypothesis. This function makes it possible to reduce the calculation cost by letting good schemata even in the lethal chromosomes survive. This paper proposes a method which can converge to the optimum more efficiently by using a new approach: each weight parameter is varied independently in a GA process according to the number of designated points of individuals; and some chromosomes with many designated points are added to an initial population generated at random. This paper also verifies its validity by several experiments.
• A Note on an Implementation of a Normalized ARMA Lattice Filter Using the Double Rotation CORDIC Algorithm

  Shiraishi Shin'ichi, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (211)  19  -24  2000/07/24  [Not refereed][Not invited]
   

  This paper proposes a new CORDIC-based architecture for a normalized ARMA lattice filter. A CORDIC implementation method for the normalized ARMA lattice filter has been reported. However, the method has a problem that its CORDIC lattice elementary section cannot execute the rotation operation by small angles. Therefore, we propose a new lattice elementary section using the double rotation CORDIC algorithm without the problem. By using the proposed CORDIC lattice elementary section, we can realize an accurate CORDIC ARMA lattice filter.
• A study of noise reduction using scene homogeneity in SAR images

  Tsuchida Masayoshi, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (211)  79  -84  2000/07/24  [Not refereed][Not invited]
   

  Synthetic aperture radar(SAR)images are affected by speckle noise, and numerous methods have been proposed for the noise reduction. The statistical nature of speckle noise is well known in optics, but these characteristics are not true on some scenes, such as urban areas in the image. In this paper, we propose the a Kalman filter adapted to the regions which are classified according to the local coefficient of variation which measures the image homogeneity in the SAR images.
• A Note on Convergence of a Genetic Algorithm for Routing Application

  Inagaki Jun, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. SST  100-  (211)  85  -90  2000/07/24  [Not refereed][Not invited]
   

  The authors have proposed a method of searching the shortest route through several designated points with a genetic algorithm. It uses a fitness function including weight parameters which are controlled based on a building block hypothesis. This function makes it possible to reduce the calculation cost by letting good schemata even in the lethal chromosomes survive. This paper proposes a method which can converge to the optimum more efficiently by using a new approach : each weight parameter is varied independently in a GA process according to the number of designated points of individuals ; and some chromosomes with many designated points are added to an initial population generated at random. This paper also verifies its validity by several experiments.
• A Method of Routing the Shortest Route through Several Designated Points Using a Genetic Algorithm

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  83-  (5)  504  -507  2000/05/25  [Not refereed][Not invited]
   

  本論文では, あらかじめ指定された複数の点の全部並びに一部を経由する最短経路を遺伝的アルゴリズムを用いて検索する手法を提案する.また実際の地図データに適用した実験を行うことにより, 提案手法の有効性を確認する.
• Genetic-Algorithm-Based Analysis of the Magnetic Microstructure of Media

  TAKEKUMA I, HASEYAMA M, SUEOKA K, MUKASA K  Journal of the Magnetics Society of Japan  24-  (4)  303  -306  2000/04/15  [Not refereed][Not invited]
   

  Magnetization fluctuation in the transition region is one of the important factors causing transition noise. The fluctuation is closely related to magnetic microstructures such as magnetic clusters, and it is therefore necessary to quantitatively analyze these clusters. This paper proposes an analysis method that can extract two-dimensional features of clusters. The proposed method is based on a genetic algorithm (GA), and investigates the distribution of the magnetic poles that exist at the boundaries of magnetic clusters. By considering the obtained distribution, we estimate the cluster size and confirm the validity of the proposed method.
• A sound radiation modeling of loudspeaker based on vibration of cones by finite element method

  TAMURA Masakatsu, AKIHO Masaichi, HASEYAMA Miki, KITAJIMA Hideo  日本音響学会研究発表会講演論文集  2000-  (1)  503  -504  2000/03/01  [Not refereed][Not invited]
  
• Residual noise characteristics of the MEFX LMS based active noise cancellation system

  AKIHO Masaichi, HASEYAMA Miki, KITAJIMA Hideo  日本音響学会研究発表会講演論文集  2000-  (1)  (87)  -(88)  2000/03/01  [Not refereed][Not invited]
  
• A Study of Kalman filtering for noise reduction on SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (611)  1  -6  2000/02/04  [Not refereed][Not invited]
   

  This paper presents the Kalman filter of Synthetic Aperture Radar (SAR) images for noise reductin. Numerical methods have been proposed to suppress speckle in SAR images. SAR images have not only speckles which have the characteristic of a randam multiplicative noise but also additive noise such as receiver thermal noise. We used the Kalman filter which takes into acount both multiplicative noise and additive noise. The experiment results on the test images proved effectivity of the proposed method for both multiplivative and additive noise reduction. This method were also applied on SAR image.
• Estimation of Lane Markers and The Vanishing Point based on A Dynamics Model

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (611)  7  -12  2000/02/04  [Not refereed][Not invited]
   

  Recently, many algorithms about automatic driving systems or driver assitance systems have been proposed. Estimation of lane markers and the vanishing point is an important reserch field for these systems. In this paper, we propose a new method to estimate lane markers and the vanishing point based on a dynamics model. In this method, lines that intersect at one point are put at an arbitrary position of the image, and are attracted to lane markers by the force of edge potential. The vanishing point is also estimated at the same time. Experimental results from a real image sequence show the effectiveness of the proposed method.
• A Study of Kalman filtering for noise reduction on SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (609)  1  -6  2000/02/04  [Not refereed][Not invited]
   

  This paper presents the Kalman filter of Synthetic Aperture Radar(SAR)images for noise reduction. Numerical methods have been proposed to suppress speckle in SAR images. SAR images have not only speckles which have the characteristic of a randam multiplicative noise but also additive noise such as receiver thermal noise. We used the Kalman filter which takes into acount both multiplicative noise and additive noise. The experiment results on the test images proved effectivity of the proposed method for both multiplivative and additive noise reduction. This method were also applied on SAR image.
• Estimation of Lane Markers and The Vanishing Point based on A Dynamics Model

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (609)  7  -12  2000/02/04  [Not refereed][Not invited]
   

  Recently, many algorithms about automatic driving systmes or driver assitance systems have been proposed. Estimation of lane markers and the vanishing point is an important reserch field for these systems. In this paper, we propose a new method to estimate lane markers and the vanishing point based on a dynamics model. In this method, lines that intersect at one point are put at an arbitrary position of the image, and are attracted to lane markers by the force of edge potential. The vanishing point is also estimated at the same time. Experimental results from a real image sequence show the effectiveness of the proposed method.
• A Study of Kalman filtering for noise reduction on SAR images

  TSUCHIDA Masayoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (11)  1  -6  2000/02/04  [Not refereed][Not invited]
   

  This paper presents the Kalman filter of Synthetic Aperture Radar (SAR) images for noise reduction. Numerical methods have been proposed to suppress speckle in SAR images. SAR images have not only speckles which have the characteristic of a randam multiplicative noise but also additive noise such as receiver thermal noise. We used the Kalman filter which takes into acount both multiplicative noise and additive noise. The experiment results on the test images proved effectivity of the proposed method for both multiplivative and additive noise reduction. This method were also applied on SAR image.
• Estimation of Lane Markers and The Vanishing Point based on A Dynamics Model

  KAZUI Masato, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (11)  7  -12  2000/02/04  [Not refereed][Not invited]
   

  Recently, many algorithms about automatic driving systems or driver assitance systems have been proposed. Estimation of lane markers and the vanishing point is an important reserch field for these systems. In this paper, we propose a new method to estimate lane markers and the vanishing point based on a dynamics model. In this method, lines that intersect at one point are put at an arbitrary position of the image, and are attracted to lane markers by the force of edge potential. The vanishing point is also estimated at the same time. Experimental results from a real image sequence show the effectiveness of the proposed method.
• A Method of Similar Region Search on IFS image Coding with Cost function

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (610)  1  -6  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to reduce error of reconstruted image by changeing domain block search in IFS image coding with the cost function. The IFS image coding with the cost function reduces error of the image by considering errors and bit-rates when a range block is divided. On the other hand we have proposed a method to improve the image quality by the extent of similar region search for domain blocks before the range block division. The proposed method improve further the image quality by combining the above two methods. The experimental results show the effectiveness of the proposed method.
• A Note on GA-Based Fractal Block Coding

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (610)  7  -12  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a new method to quantize IFS parameters in fractal block coding by using a genetic algorithm (GA). In the fractal block coding how to quantize the IFS parameters is significant, because the nearest quantization values do not necessarily provide high-quality reconstructed images. Therefore, in order to obtain them, we report a novel GA-based quantization method. The proposed method searches for the best set of the quantized IFS parameters in short time, moreover provides the high-quality reconstructed images.
• Image Improvement for Full-Frame DCT Image Coding

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (610)  19  -24  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes image improvment for full-frame DCT image coding. Conventionally an image is divided into small blocks of pixels ; DCT is applied to each block. However, block distortion arises under a low bit rate. Block distortion is completely removed by using full-frame DCT, but quantization errors are spread over the entire image. Especially, we easily remark high-frequency noise in flat rgions. This paper poposes image improvement using adaptive post-filtering.
• A New Circle Extraction Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (610)  49  -54  2000/02/03  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time circle extraction algorithm. The algorithm is capable of extracting both complete circles and circular arcs from complex images. The algorithm was tested with a variety of natural and synthetic 256 × 256 grayscale images and an average frame rate of more than 69 frames per second was obtained using a 450MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• A Method of Similar Region Search on IFS image Coding with Cost function

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (608)  1  -6  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to reduce error of reconstruted image by changeing domain block search in IFS image coding with the cost function. The IFS image coding with the cost function reduces error of the image by considering errors and bit-rates when a range block is divided. On the other hand we have proposed a method to improve the image quality by the extent of similar region search for domain blocks before the range block division. The proposed method improve further the image quality by combining the above two methods. The experimental results show the effectiveness of the proposed method.
• A Note on GA-Based Fractal Block Coding

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (608)  7  -12  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a new method to quantize IFS parameters in fractal block coding by using a genetic algorithm(GA). In the fractal block coding how to quantize the IFS parameters is significant, because the nearest quantization values do not necessarily provide high-quality reconstructed images. Therefore, in order to obtain them, we report a novel GA-based quantization method. The proposed method searches for the best set of the quantized IFS parameters in short time, moreover provides the high-quality reconstructed images.
• Image Improvment for Full-Frame DCT Image Coding

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (608)  19  -24  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes image improvment for full-frame DCT image coding. Conventionally an image is divided into small blocks of pixels ; DCT is applied to each block. However, block distortion arises under a low bit rate. Block distortion is completely removed by using full-frame DCT, but quantization errors are spread over the entire image. Especially, we easily remark high-frequency noise in flat rgions. This paper poposes image improvement using adaptive post-filtering.
• A New Circle Extraction Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  99-  (608)  49  -54  2000/02/03  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time circle extraction algorithm. The algorithm is capable of extracting both complete circles and circular arcs from complex images. The algorithm was tested with a variety of natural and synthetic 256×256 grayscale images and an average frame rate of more than 69 frames per second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• A Method of Similar Region Search on IFS image Coding with Cost function

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (10)  1  -6  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to reduce error of reconstruted image by changeing domain block search in IFS image coding with the cost function. The IFS image coding with the cost function reduces error of the image by considering errors and bit-rates when a range block is divided. On the other hand we have proposed a method to improve the image quality by the extent of similar region search for domain blocks before the range block division. The proposed method improve further the image quality by combining the above two methods. The experimental results show the effectiveness of the proposed method.
• A Note on GA - Based Fractal Block Coding

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (10)  7  -12  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes a new method to quantize IFS parameters in fractal block coding by using a genetic algorithm(GA). In the fractal block coding how to quantize the IFS parameters is significant, because the nearest quantization values do not necessarily provide high-quality reconstructed images. Therefore, in order to obtain them, we report a novel GA-based quantization method. The proposed method searches for the best set of the quantized IFS parameters in shorttime, moreover provides the high-quality reconstructed images.
• Image Improvment for Full - Frame DCT Image Coding

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (10)  19  -24  2000/02/03  [Not refereed][Not invited]
   

  This paper proposes image improvement for full-frame DCT image coding. Conventionally an image is divided into small blocks of pixels ; DCT is applied to each block. However, block distortion arises under a low bit rate. Block distortion is completely removed by using full-frame DCT, but quantization errors are spread over the entire image. Especially, we easily remark high-frequency noise in flat rgions. This paper poposes image improvement using adaptive post-filtering.
• A New Circle Extraction Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  24-  (10)  49  -54  2000/02/03  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time circle extraction algorithm. The algorithm is capable of extracting both complete circles and circular arcs from complex images. The algorithm was tested with a variety of natural and synthetic 256 × 256 grayscale images and an average frame rate of more than 69 frames per second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• Performance improvements on MEFX-LMS based noise cancellation system in vehicle cabin

  M Akiho, M Tamura, M Haseyama, H Kitajima  ISCAS 2000: IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS - PROCEEDINGS, VOL IV  4-  353  -356  2000  [Not refereed][Not invited]
   

  Multiple Error Filtered X (MEFX) LMS (Least Mean Squares) based Booming noise or Engine noise cancellation system has been implemented to the automobile and the reciprocate engine airplane cabin to provide practical solutions to improve the annoying noisy environment. The noises are well known that they behave as resonance noise of the fuselage. A significant difference in noise level or power spectrum between control points is often observed. For the cabin noise cancellation system, all passenger areas are expected to be quiet. Because of the MEFX LMS algorithm minimizes the total squared power of the error signals, degradation of increase in the noise level is observed in certain situations. In this paper, we demonstrate such a behavior of the algorithm and propose a method to improve a practical performance of the system.
• Real-time polygon extraction from complex images

  JW Gates, M Haseyama, H Kitajima  ISCAS 2000: IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS - PROCEEDINGS, VOL V  5-  309  -312  2000  [Not refereed][Not invited]
   

  This paper presents an algorithm that is capable of extracting polygons from complex digital images for real-time video-applications. The algorithm uses a high-accuracy, real-time (HART) line-feature extraction algorithm to extract the line segments from the digital image. Then these line segments are grouped into triangles and quadrilaterals using an efficient scanning method that can be easily extended to n-sided polygons. The algorithm can process both convex and concave polygons. Tears were performed on a variety of 256 x 256 grayscale synthetic and natural images and an average frame rate of 58 frames per second was achieved using a 450 MHz. Pentium II processor. Experiments show that the algorithm accurately extracts the polygons from complex images.
• An IFS-based image restoration method for impulsive noise contamination

  M Haseyama, M Takezawa, H Honda, H Kitajima  ISCAS 2000: IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS - PROCEEDINGS, VOL IV  4-  593  -596  2000  [Not refereed][Not invited]
   

  This paper proposes an Iterated Function System (IFS) based image-restoration method. The proposed method can restore images corrupted by impulsive noise according to self-similarity represented by the IFS. Since the IFS is usually used for image coding, it has never been applied to image restoration; and it cannot be utilized for the image restoration as it is. In order to adapt the IFS for image restoration, this paper reforms the conventional criterion for the computation of the IFS parameters to suit for image restoration, and as preprocessing we apply a median filter to the contaminated images prior to computing the IFS parameters. Some simulation results are presented to demonstrate the effectiveness of this method.
• Performance improvements on MEFX-LMS based noise cancellation system in vehicle cabin

  M Akiho, M Tamura, M Haseyama, H Kitajima  ISCAS 2000: IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS - PROCEEDINGS, VOL IV  4-  353  -356  2000  [Not refereed][Not invited]
   

  Multiple Error Filtered X (MEFX) LMS (Least Mean Squares) based Booming noise or Engine noise cancellation system has been implemented to the automobile and the reciprocate engine airplane cabin to provide practical solutions to improve the annoying noisy environment. The noises are well known that they behave as resonance noise of the fuselage. A significant difference in noise level or power spectrum between control points is often observed. For the cabin noise cancellation system, all passenger areas are expected to be quiet. Because of the MEFX LMS algorithm minimizes the total squared power of the error signals, degradation of increase in the noise level is observed in certain situations. In this paper, we demonstrate such a behavior of the algorithm and propose a method to improve a practical performance of the system.
• A highly accurate pipelined architecture for a CORDIC ARMA lattice filter

  S Shiraishi, M Haseyama, H Kitajima  ISCAS 2000: IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS - PROCEEDINGS, VOL V  5-  369  -372  2000  [Not refereed][Not invited]
   

  This paper presents a method to improve implementation accuracy of a recently proposed CORDIC ARMA lattice filter. Since the CORDIC ARMA lattice filter algorithm has a problem in its shift sequence, it cannot implement a lattice filter accurately. Therefore, in this paper we apply the shift sequence proposed by Walther without the problem to the CORDIC ARMA lattice filter, and then we realize an accurate lattice filter. In addition, we propose a new pipelined architecture for the CORDIC ARMA lattice filter. The proposed lattice filter architecture consists of CORDIC processors implemented in the pipelined fashion, it can achieve a high data processing throughput rate.
• An image restoration method using IFS

  M Haseyama, M Takezawa, K Kondo, H Kitajima  2000 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOL III, PROCEEDINGS  3-  774  -777  2000  [Not refereed][Not invited]
   

  This paper proposes a new image-restoration method based on Iterated Function System (IFS). The proposed method can restore images contaminated by impulsive noise according to self-similarity represented by the IFS parameters. Since the IFS is usually used for image coding, it has never been applied to image restoration; and it cannot be utilized for image restoration as it is. In order to adapt the IFS for image restoration, this paper reforms the conventional criterion for the computation of the IFS parameters to suit for image restoration, and as preprocessing we apply an E-filter in which a median filter is embedded to contaminated images prior to computing the IFS parameters. Some simulation results are presented to demonstrate the effectiveness of this method.
• Determination of Parameters of Active Contour Models

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  23-  (79)  1  -6  1999/12/16  [Not refereed][Not invited]
   

  Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
• The Improvement of Hough Transform using the Value of Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  23-  (79)  43  -48  1999/12/16  [Not refereed][Not invited]
   

  We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
• Determination of Parameters of Active Contour Models

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. CS, 通信方式  99-  (510)  1  -6  1999/12/16  [Not refereed][Not invited]
  
• The Improvement of Hough Transform using the Value of Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. CS, 通信方式  99-  (510)  43  -48  1999/12/16  [Not refereed][Not invited]
   

  We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
• Introduction of new edge parameters for with a Hopfield NN image restoration

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. CS, 通信方式  99-  (510)  61  -66  1999/12/16  [Not refereed][Not invited]
   

  This paper proposes new parameters of a Hopfield Neural Network(HNN)for edge defection of images. The HNN has been used as a preprocessor of an image restoration HNN. Therefore, edge detection accuracy is closely related with the image restoration performance. The proposed new parameters can make the previous method detect edges more accurately.
• A Real-time Polygon Extraction and Classification Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  電子情報通信学会技術研究報告. CS, 通信方式  99-  (510)  73  -78  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time polygon extraction and classification algorithm. The algorithm is capable of extracting both convex and concave polygons from complex images. The algorithm can classify triangles into five distinct classes and can classify quadrilaterals into six distinct classes. The algorithm was tested with a variety of natural and synthetic 256×256 grayscale images and an average frame rate of more than 69 frames / second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• Determination of Parameters of Active Contour Models

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  27-  1  -6  1999/12/16  [Not refereed][Not invited]
   

  Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
• The Improvement of Hough Transform using the Value of Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (107)  43  -48  1999/12/16  [Not refereed][Not invited]
   

  We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven. that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
• Introduction of new edge parameters for with a Hopfield NN image restoration

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (107)  61  -66  1999/12/16  [Not refereed][Not invited]
   

  This paper proposes new parameters of a Hopfield Neural Network (HNN) for edge detection of images. The HNN has been used as a preprocessor of an image restoration HNN. Therefore, edge detection accuracy is closely related with the image restoration performance. The proposed new parameters can make the previous method detect edges more accurately.
• Multi-Agent Based Method to Extract Closed Smooth Curve in Image Sequences

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (107)  67  -72  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a multi-agent based method to extract the same objects in image sequences. In the proposed method, the objects are divided into line segments which are considered to be agents. Further, the method is based on a multi-agent algorithm of which the utility function is defined with a motion continuity and a shape classification of line segments. By optimizing the utility, the proposed method searches for a suitable line segment which corresponds to the target segment in a different frame. According to the group rationality, the proposed method can correctly provide correspondence among line segments in different frames. Further, the computational cost can be reduced by utilizing the group rationality. The proposed method can also extract occluded objects by tracking utilities in multi frames.
• A Real-time Polygon Extraction and Classification Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (107)  73  -78  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time polygon extraction and classification algorithm. The algorithm is capable of extracting both convex and concave polygons from complex images. The algorithm can classify triangles into five distinct classes and can classify quadrilaterals into six distinct classes. The algorithm was trsted with a variety of natural and synthetic 256×256 grayscale images and an average frame rate of more than 69 frames / second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• Determination of Parameters of Active Contour Models

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (512)  1  -6  1999/12/16  [Not refereed][Not invited]
   

  Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
• The Improvement of Hough Transform using the Value of Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (512)  43  -48  1999/12/16  [Not refereed][Not invited]
   

  We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
• Introduction of new edge parameters for with a Hopfield NN image restoration

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (512)  61  -66  1999/12/16  [Not refereed][Not invited]
   

  This paper proposes new parameters of a Hopfield Neural Network (HNN) for edge detection of images. The HNN has been used as a preprocessor of an image restoration HNN. Therefore, edge detection accuracy is closely related with the image restoration performance. The proposed new parameters can make the previous method detect edges more accurately.
• Multi-Agent Based Method to Extract Closed Smooth Curve in Image Sequences

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (512)  67  -72  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a multi-agent based method to extract the same objects in image sequences. In the proposed method, the objects are divided into line segments which are considered to be agents. Further, the method is based on a multi-agent algorithm of which the utility function is defined with a motion continuity and a shape classification of line segments. By optimizing the utility, the proposed method searches for a suitable line segment which corresponds to the target segment in a different frame. According to the group rationality, the proposed method can correctly provide correspondence among line segments in different frames. Further, the computational cost can be reduced by utilizing the group rationality. The proposed method can also extract occluded objects by tracking utilities in multi frames.
• A Real-time Polygon Extraction and Classification Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (512)  73  -78  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time polygon extraction and classification algorithm. The algorithm is capable of extracting both convex and concave polygons from complex images. The algorithm can classify triangles into five distinct classes and can classify quadrilaterals into six distinct classes. The algorithm was tested with a variety of natural and synthetic 256 × 256 grayscale images and an average frame rate of more than 69 frames / second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• Multi-Agent Based Method to Extract Closed Smooth Curve in Image Sequences

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  99-  (510)  67  -72  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a multi-agent based method to extract the same objects in image sequences. In the proposed method, the objects are divided into line segments which are considered to be agents. Further, the method is based on a multi-agent algorithm of which the utility function is defined with a motion continuity and a shape classification of line segments. By optimizing the utility, the proposed method searches for a suitable line segment which corresponds to the target segment in a different frame. According to the group rationality, the proposed method can correctly provide correspondence among line segments in different frames. Further, the computational cost can be reduced by utilizing the group rationality. The proposed method can also extract occluded objects by tracking utilities in multi frames.
• Determination of Parameters of Active Contour Models

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  23-  (79)  1  -6  1999/12/16  [Not refereed][Not invited]
   

  Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
• The Improvement of Hough Transform using the Value of Segments

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  23-  (79)  43  -48  1999/12/16  [Not refereed][Not invited]
   

  We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
• Introduction of new edge parameters for with a Hopfield NN image restoration

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  23-  (79)  61  -66  1999/12/16  [Not refereed][Not invited]
   

  This paper proposes new parameters of a Hopfield Neural Network (HNN) for edge detection of images. The HNN has been used as a preprocessor of an image restoration HNN. Therefore, edge detection accuracy is closely related with the image restoration performance. The proposed new parameters can make the previous method detect edges more accurately.
• Multi-Agent Based Method to Extract Closed Smooth Curve in Image Sequences

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  23-  (79)  67  -72  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a multi-agent based method to extract the same objects in image sequences. In the proposed method, the objects are divided into line segments which are considered to be agents. Further, the method is based on a multi-agent algorithm of which the utility function is defied with a motion continuity and a shape classification of line segments. By optimizing the utility, the proposed method searches for a suitable line segment which corresponds to the target segment in a different frame. According to the group rationality, the proposed method can correctly provide correspondence among line segments in different frames. Further, the computational cost can be reduced by utilizing the group rationality. The proposed method can also extract occluded objects by tracking utilities in multi frames.
• A Real-time Polygon Extraction and Classification Algorithm

  GATES John, HASEYAMA Miki, KITAJIMA Hideo  ITE technical report  23-  (79)  73  -78  1999/12/16  [Not refereed][Not invited]
   

  This paper presents a high-accuracy real-time polygon extraction and classification algorithm. The algorithm is capable of extracting both convex and concave polygons from complex images. The algorithm can classify triangles into five distinct classes and can classify quadrilaterals into six distinct classes. The algorithm was tested with a variety of natural and synthetic 256 × 256 grayscale images and an average frame rate of more than 69 frames / second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
• Determination of Parameters of Active Contour Models

  鐘文, 長谷山 美紀, 北島 秀夫  情報処理学会研究報告オーディオビジュアル複合情報処理（AVM）  1999-  (107)  1  -6  1999/12/16  [Not refereed][Not invited]
   

  動的輪郭モデルは輪郭抽出の有効な手法である。この手法においては、試行誤差によって適当なパラメータを設定する必要があり、これが具体的応用への障害である。本論文では、強化学習を用いて輪郭モデルパラメータを設定する方法を提案する。本提案方式においては既存の強化学習手法の流れを修正して、ロバストではないという欠点を改善する。また、ヒューリスティックルールを加えて、探索時間を従来法より大幅に減少させる。動的輪郭モデルパラメータを設定する実験結果から、提案手法は従来法より有効であることを示す。Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
• ディジタル直線を構成するセグメントの値に注目したHough変換の改善法

  金 義鎮, 長谷山 美紀, 北島 秀夫  電子情報通信学会技術研究報告  99-  (510)  43  -48  1999/12/16  [Not refereed][Not invited]
  
• 画像復元Hopfield NNにおける新しいエッジ抽出パラメータの導入

  内山 賢史, 長谷山 美紀, 北島 秀夫  電子情報通信学会技術研究報告  99-  (510)  61  -66  1999/12/16  [Not refereed][Not invited]
  
• マルチエージェントを用いた複数フレーム線分画像からの同一物体の抽出

  牛丸 主税, 長谷山 美紀, 北島 秀夫  電子情報通信学会技術研究報告  99-  (510)  67  -72  1999/12/16  [Not refereed][Not invited]
  
• A Real-time Polygon Extraction and Classification Algorithm

  Gates John, 長谷山 美紀, 北島 秀夫  電子情報通信学会技術研究報告  99-  (510)  73  -78  1999/12/16  [Not refereed][Not invited]
  
• Recurrent network model for range processing of the mustached bat

  Gang Wang, Miki Haseyama, Nobuo Suga  Proceedings of the International Joint Conference on Neural Networks  1-  301  -304  1999/12/01  [Not refereed][Not invited]
   

  A computational model, which simulates the neural processing of range information in the mustached bat is proposed. The internal states and the learning characteristics of the neural model as well as the mechanism for the processing of range information are explored. The neural model is capable of interpolation, so it can simulate the range information processing in bat's auditory system. The following important properties are found in the trained network: (1) five types of hidden units are created, and they have their counterparts in bat's brain based on the similarity of the response patterns; (2) the hidden units show characteristics of plasticity; and (3) a dis-inhibition mechanism is found, which has not yet been found in neurophysiological study.
• Study of magnetization fluctuations in transition region using MFM image analysis

  I. Takekuma, M. Yasui, M. Haseyama, K. Sueoka, K. Mukasa  Digests of the Intermag Conference  1999/12/01  [Not refereed][Not invited]
   

  Proposed is a quantitative analysis method to reveal the influence of the magnetization fluctuation on recorded states by processing MFM images. The proposed method processes the MFM images based on the concept of stochastic signals. The use of this concept can achieve to quantitatively analyze the magnetization fluctuation.
• Fractal interpolation for natural images

  Hiroyuki Honda, Miki Haseyama, Hideo Kitajima  IEEE International Conference on Image Processing  3-  657  -661  1999/12/01  [Not refereed][Not invited]
   

  This paper proposes a fractal interpolation for natural images. Generally, linear interpolation and spline interpolation are used for image interpolation. However, an image interpolated by the above conventional methods lose some high-frequency components of an original image. The loss of components lower fidelity of the interpolated images. Since the proposed method reduces the loss, an interpolated image generated by the proposed method has higher fidelity than the one generated by the conventional method. The reduction of the loss is realized by using the fractional Brownian motion (FBM) in a process of the interpolation. The proposed method uses a characteristic that the fractal dimension is strongly correlated with a sense of roughness.
• Genetic-algorithm based quantization method for fractal image coding

  Megumi Takezawa, Hiroyuki Honda, Jun Miura, Miki Haseyama, Hideo Kitajima  IEEE International Conference on Image Processing  1-  458  -461  1999/12/01  [Not refereed][Not invited]
   

  This paper proposes a high-accuracy quantization method for IFS parameters in fractal image coding by using a genetic algorithm (GA). The development of IFS-parameter quantization techniques is significant for the image coding because its errors make more serious problems in the iteration procedures than the other quantization topics. Even if the errors are small, high-quality reconstructed images are not necessarily obtained. Therefore, the high-accuracy quantization methods are required for the parameters. The proposed method provides higher quality reconstructed images than a conventional method which merely minimizes the errors.
• A Genetic Algorithm Based Picture Segmentation Method Using MSE as a Criterion

  HASEYAMA Miki, KUMAGAI Masateru, MIYAMOTO Toshihiro, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  82-  (11)  1903  -1911  1999/11/25  [Not refereed][Not invited]
   

  遺伝的アルゴリズムを用いた領域分割手法を提案する.提案手法では,領域分割の精度を示す評価関数を領域内画素値を近似値で置き換えて得られる画像と元の画像との平均2乗誤差で定義する.その結果,領域分割は,要求される領域数のもとで評価関数を最小化する問題となる.この評価関数は,領域内近似式のパラメータだけでなく,分割領域の情報(領域形状とその位置)を用いて表現されており,従来の最適化手法では最小化が容易ではない.したがって,提案手法はGAを用いることにより,各領域に適した領域内画素の近似式を決定しながら,平均2乗誤差を最小にする領域分割結果の探索を可能とする.更に,提案手法では,マルコフ確率場モデルの導入によって,探索空間に存在する小領域のみが異なる複数の分割結果により局所解に陥るという問題を回避している.本論文の最後では提案手法の有効性を示すためにいくつかの実験結果を示す.
• An analysis of magnetic cluster using genetic algorithms

  TAKEKUMA I, HASEYAMA M, SUEOKA K, MUKASA K  日本応用磁気学会学術講演概要集 = Digest of ... annual conference on magnetics in Japan  23-  307  -307  1999/10/01  [Not refereed][Not invited]
  
• A sound radiation modeling of loudspeaker based on vibration of cones by finite element method

  TAMURA Masakatsu, AKIHO Masaichi, HASEYAMA Miki, KITAJIMA Hideo  日本音響学会研究発表会講演論文集  1999-  (2)  525  -526  1999/09/01  [Not refereed][Not invited]
  
• A Routing Application with a Genetic Algorithm using a Fitness Function with Weight Parameters

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo, TAMURA Masakatsu  Proceedings of the Society Conference of IEICE  1999-  (0)  183  -183  1999/08/16  [Not refereed][Not invited]
  
• A Method of Determining Various Solutions for Routing Application with a Genetic Algorithm

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  82-  (8)  1102  -1111  1999/08  [Not refereed][Not invited]
   

  本論文は遺伝的アルゴリズム(GA : Genetic Algorithm)を用いた経路探索手法を提案する. 提案手法は, GAの特徴を生かし, 最短経路を選択するだけでなく, それに準ずる複数の解候補の決定が可能である. したがって本手法を用いれば, 最短経路と異なる短い経路が要求された場合に, 新たに探索し直すことなく他の経路候補を得ることができる. このような応用は従来法では容易に実現することができない. 更に提案手法は, 経路に依存した重みを遺伝子型に組み込むことにより, 経路長が短く, かつ経由するノードが類似しない経路を決定することが可能である.
• Considerations on the Hough Transform for Digital Images

  KIM Euijin, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  99-  (180)  89  -96  1999/07/16  [Not refereed][Not invited]
   

  The Hough transform is used to detect lines in complex images. However, it requires a long computation time because each pixel is mapped to a parameter space. We consider a Hough transform that reduces the computation time by exploiting the properties of digital images. We analyze segments, which consist of pixels, with the properties of digital lines. First, we use the inherent and shared direction to determine the initial range and then the segment width to determine the final range. We propose a new segment parameter space where voting only occurs in the final range. Experimental results are included to show that the proposed method can reduce the number of computations.
• Table-Look-Up Motion Compensation for Video Coding

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  82-  (6)  1001  -1008  1999/06/25  [Not refereed][Not invited]
   

  本論文では, 動画像符号化のためテーブル参照による動き補償を提案する. 様々な動きを格納した'動きテーブル'を複数フレームごとに作成し, 対象ブロックごとに動きテーブル内を探索し, ブロック内の動き補償予測誤差を最小にする動きをテーブル内から選択する. 本手法のポイントは動きに関する符号量削減のために動きベクトル数を削減することにある. 動きベクトル数の減少による画質の悪化を抑えるために, 次の二つの動きテーブル作成法を提案する. 一つ目は, 動き補償を行った際に使用される動きベクトルの出現頻度により, 動きベクトル数を削減する方法. 二つ目は, 遺伝的アルゴリズムを用いて, 動き補償予測誤差が最小になるように動きベクトル数を削減し, 動きテーブルを作成する方法である. こうして作成された動きテーブルを用いることにより, 動き補償予測画像の画質を劣化を低く抑えつつ符号量を削減することを示す. 本論文では, 動画像には動きの種類に偏りがあることを利用して, テーブルサイズを限定したテーブル参照による動き補償を用いることによって, 画質の劣化を抑えつつ従来法のブロックマッチング法より符号量の削減を実現する.
• A Note on Determining a Fitness Function for Routing with a Genetic Algorithm : An Approach Based on Building Block Hypothesis

  INAGAKI Jun, HASEGAWA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  99-  (104)  89  -96  1999/06/10  [Not refereed][Not invited]
   

  A Method of searching the shortest route through several designated points with a genetic algorithm has been proposed. This method generates many lethal genes in the process of generating an initial population and crossover. Hence it has a high calculation cost for processing the lethal genes. This paper proposes a new fitness function for the purpose of reducing the calculation cost. The proposed fitness function includes weight parameters which are controlled based on the building block hypothesis. This function makes it possible to reduce the calculation cost by using cost by using good schemata which the lethal genes in the current method have. This paper also examines a method determining the weight parameters for effective search for the optimum route.
• A Note on a Pipelined Architecture for a CORDIC ARMA Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  99-  (104)  123  -128  1999/06/10  [Not refereed][Not invited]
   

  This paper proposes a new pipelined architecture for a CORDIC ARMA lattice filter. Since the proposed filter consists of CORDIC processors implemented in the pipelined fashion, it can achieve a high data processing throughput rate. In addition, processing modules are made up of simple components: an addersubstracter and a shifter, so that an effective hardware implementation of the filter is expecte.
• A Note on Determining a Fitness Function for Routing with a Genetic Algorithm : An Approach Based on Building Block Hypothesis

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  99-  (110)  89  -96  1999/06/10  [Not refereed][Not invited]
   

  A Method of searching the shortest route through several designated points with a genetic algorithm has been proposed. This method generates many lethal genes in the process of generating an initial population and crossover. Hence it has a high calculation cost for processing the lethal genes. This paper proposes a new fitness function for the purpose of reducing the calculation cost. The proposed fitness function includes weight parameters which are controlled based on the building block hypothesis. This function makes it possible to reduce the calculation cost by using cost by using good schemata which the lethal genes in the current method have. This paper also examines a method determining the weight parameters for effective search for the optimum route.
• A Note on a Pipelined Architecture for a CORDIC ARMA Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  99-  (110)  123  -128  1999/06/10  [Not refereed][Not invited]
   

  This paper proposes a new pipelined architecture for a CORDIC ARMA lattice filter. Since the proposed filter consists of CORDIC processors implemented in the pipelined fashion, it can achieve a high data processing throughput rate. In addition, processing modules are made up of simple components: an addersubstracter and a shifter, so that an effective hardware implementation of the filter is expecte.
• A Note on Determining a Fitness Function for Routing with a Genetic Algorithm : An Approach Based on Building Block Hypothesis

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  99-  (107)  89  -96  1999/06/10  [Not refereed][Not invited]
   

  A Method of searching the shortest route through several designated points with a genetic algorithm has been proposed. This method generates many lethal genes in the process of generating an initial population and crossover. Hence it has a high calculation cost for processing the lethal genes. This paper proposes a new fitness function for the purpose of reducing the calculation cost. The proposed fitness function includes weight parameters which are controlled based on the building block hypothesis. This function makes it possible to reduce the calculation cost by using cost by using good schemata which the lethal genes in the current method have. This paper also examines a method determining the weight parameters for effective search for the optimum route.
• A Note on a Pipelined Architecture for a CORDIC ARMA Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  99-  (107)  123  -128  1999/06/10  [Not refereed][Not invited]
   

  This paper proposes a new pipelined architecture for a CORDIC ARMA lattice filter. Since the proposed filter consists of CORDIC processors implemented in the pipelined fashion, it can achieve a high data processing throughput rate. In addition, processing modules are made up of simple components: an addersubstracter and a shifter, so that an effective hardware implementation of the filter is expecte.
• Image Coding Using Full-Frame DCT

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (52)  43  -48  1999/06/04  [Not refereed][Not invited]
   

  This paper proposes image coding using full-frame DCT. Conventionally an image divided into small blocks of pixels; DCT is applied to each block. However, block distortion arises under a low bit-rate. Block distortion is completely removed by using full-frame DCT, but it is diffcult to adapt quantization to local image structure. This paper proposes a quantization strategy that adapts to transform-domain image features.
• Motion Table Optimization for Video Coding

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IPSJ SIG Notes  1999-  (52)  49  -54  1999/06/04  [Not refereed][Not invited]
   

  This paper proposes the estimation of a motion vector for motion-parameter identification in motion table. The motion compensation based on table lookup selects the motion to minimize prediction error from the motion table. Therefore, the quality of the prediction image changes making the motion table. The motion table identified by the conventional method the frequency of motion vectors. The proposed method used both frequency of motion vector and mean square error of block. A pair of motion tables sets up by distinct methods are compared.
• Analysis of MFM Images as Stochastic Signals and Interpretation of their Variances

  TAKEKUMA I, YASUI M, HASEYAMA M, SUEOKA K, MUKASA K  Journal of Magnetics Society of Japan  23-  (4)  1073  -1076  1999/04/15  [Not refereed][Not invited]
   

  Magnetic force microscopy (MFM) has been widely used to investigate magnetized states of recording media, and a method of analyzing MFM images as stochastic signals has been proposed. Through the analysis of the variance computed by this method, a parameter L/W was found. However, the relation between the variance variation and the recorded state of media is not well understood. In this study, in order to determine the relation, MFM output signals are simulated by using the micro-track model, and the variation in the variance around the transition region is considered. The relation between the L/W and some media parameters is investigated by meaus of the above simulation.
• A Method to Quantize IFS Parameters in Fractal Block Coding Using A Genetic Algorithm

  竹沢 恵, 本多 博之, 長谷山 美紀, 北島 秀夫  全国大会講演論文集  58-  (0)  285  -286  1999/03/09  [Not refereed][Not invited]
  
• Table Design for Video Coding Using Motion Compensation

  浅水 仁, 長谷山 美紀, 北島 秀夫  全国大会講演論文集  58-  (0)  287  -288  1999/03/09  [Not refereed][Not invited]
  
• D-11-32 A Negative View on IFS-Based Image Coding

  Kitajima Hideo, Haseyama Miki  Proceedings of the IEICE General Conference  1999-  (2)  32  -32  1999/03/08  [Not refereed][Not invited]
  
• A Note on High Accuracy Quantization of IFS Parameters Using Genetic Algorithm

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (576)  75  -80  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to quantize IFS parameters accurately. The conventional method quantizes them to the nearest values. However the method does not necessarily provide high-quality reconstructed images. Therefore, a new quantization method searches for the optimal IFS-parameter set by using a genetic algorithm to obtain the high-quality ones.
• A Study of the Correspondence Problem Using an Elastic Model

  MUTOU Takeshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (576)  81  -86  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method which determines correspondence of feature points in two different images by using an elastic model. The previous method cannot do it accurately, when the users do not know which object each feature point exists in. The proposed method can acquire suitable correspondence, by comparison of elastic energies of expression by the proposed elastic model.
• A Study of Closed Smooth Curve Extraction in Image Sequences : An Approach Using Process-Grammar

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (576)  87  -92  1999/02/03  [Not refereed][Not invited]
   

  This paper presents a method to extract the same closed smooth curves in different frames. Since the proposed method is based on not only the Process-Grammar but also localities of fragments of curves, it can detect the same curves even modified by rotation and magnification. The previous method cannot achieve such detection, because it is based on only the similarity of their motion vectors: the motion vectors of a rotated or magnified curve are different from the vectors of the original one.
• A Note on High Accuracy Quantization of IFS Parameters Using Genetic Algorithm

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  98-  (574)  75  -80  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to quantize IFS parameters accurately. The conventional method quantizes them to the nearest values. However the method does not necessarily provide high-quality reconstructed images. Therefore, a new quantization method searches for the optimal IFS-parameter set by using a genetic algorithm to obtain the high-quality ones.
• A Study of the Correspondence Problem Using an Elastic Model

  MUTOU Takeshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  98-  (574)  81  -86  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method which determines correspondence of feature points in two different images by using an elastic model. The previous method cannot do it accurately, when the users do not know which object each feature point exists in. The proposed method can acquire suitable correspondence, by comparison of elastic energies of expression by the proposed elastic model.
• A Study of Closed Smooth Curve Extraction in Image Sequences : An Approach Using Process-Grammar

  USHIMARAU Chikara, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  98-  (574)  87  -92  1999/02/03  [Not refereed][Not invited]
   

  This paper presents a method to extract the same closed smooth curves in different frames. Since the proposed method is based on not only the Process-Grammar but also localities of fragments of curves, it can detect the same curves even modified by rotation and magnification. The previous method cannot achieve such detection, because it is based on only the similarity of their motion vectors; the motion vectors of a rotated or magnified curve are different from the vectors of the original one.
• A Note on High Accuracy Quantization of IFS Parameters Using Genetic Algorithm

  TAKEZAWA Megumi, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  23-  (9)  75  -80  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method to quantize IFS parameters accurately. The conventional method quantizes them to the nearest values. However the method does not necessarily provide high-quality reconstructed images. Therefore, a new quantization method searches for the optimal IFS-parameter set by using a genetic algorithm to obtain the high-quality ones.
• A Study of the Correspondence Problem Using an Elastic Model

  MUTOU Takeshi, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  23-  (9)  81  -86  1999/02/03  [Not refereed][Not invited]
   

  This paper proposes a method which determines correspondence of feature points in two different images by using an elastic model. The previous method cannot do it accurately, when the users do not know which object each feature point exists in. The proposed method can acquire suitable correspondence, by comparison of elastic energies of expression by the proposed elastic model.
• A Study of Closed Smooth Curve Extraction in Image Sequences - An Approach Using Process-Grammar -

  USHIMARU Chikara, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  23-  (9)  87  -92  1999/02/03  [Not refereed][Not invited]
   

  This paper presents a method to extract the same closed smooth curves in different frames. Since the proposed method is based on not only the Process-Grammar but also localities of fragments of curves, it can detect the same curves even modified by rotation and magnification. The previous method cannot achieve such detection, because it is based on only the similarity of their motion vectors ; the motion vectors of a rotated or magnified curve are different from the vectors of the original one.
• A transformation method of a CORDIC ARMA lattice filter for signal synthesis

  S Shiraishi, M Haseyama, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E82A-  (2)  230  -237  1999/02  [Not refereed][Not invited]
   

  This paper proposes a method to transform a CORDIC ARMA lattice filter, which is originally realized for signal analysis, into a signal synthesis lattice filter (CORDIC ARMA lattice synthesis filter). In order to perform such a transformation and then obtain the CORDIC ARMA lattice synthesis filter, we must implement the followings with CORDIC: (1) the structure of the altered lattice filter; and (2) an angle calculation module. However, we cannot achieve such an implementation as an extension of the CORDIC ARMA lattice filter algorithm. Therefore, this paper proposes CORDIC implementation schemes for both the structure and module, and then we realize the CORDIC ARMA lattice synthesis filter. By using CORDIC processors, the elementary sections of the CORDIC ARMA lattice synthesis filter are efficiently implemented without any multipliers. Since the obtained signal synthesis lattice filter consists of dedicated CORDIC processors, it keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Two Cases with Mumps-associated Neurological Impairment.

  佐藤俊哉, 長谷山圭司, 井上美紀, 川浪貢, 中川俊一, 佐藤謙二, 伊藤希美  臨床小児医学  47-  (5)  1999
• Hopfield NNを用いた画像復元に関する考察

  内山賢史, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• フラクタル次元の変更による線画の単純化

  上村郷志, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• パイプラインCORDIC ARMAラチスフィルタの実現に関する考察

  白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• FMBを用いた画像補間法に関する一考察

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• 特徴psfを用いた直線検出に関する考察

  数井誠人, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• 遺伝的アルゴリズムを用いたIFSパラメータ決定の高速化に関する一考察

  竹沢恵, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• カルマンフィルタを用いたSAR画像のノイズ低減法に関する研究

  土田正芳, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• 低ビットレート動画像符号化のための新しいブロックマッチング法

  夏井知義, 浅水仁, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• エージェントを用いた線分画像のフレーム間対応に関する考察

  牛丸主税, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• ディジタル直線の方向性に基づいた逐次的直線抽出

  金義鎮, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• 弾性モデルを用いた対応点探索への遺伝的アルゴリズムの応用

  鈴木研二, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• コスト関数を用いたIFS符号化法に関する考察

  鈴木康夫, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• VHDLを用いた動き検出器の実現に関する考察

  仲川和志, 白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• GA経路探索における評価関数の重みに関する検討

  稲垣潤, 長谷山美紀, 北島秀夫, 吉野光則  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• CORDICアルゴリズムを用いた動き検出器の実現に関する検討

  山下雅人, 白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• フルフレームDCTによる静止画像符号化の改良

  季平浩治, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• ファジィ推論による物体抽出のためのメンバーシップ関数作成に関する考察

  岩井範幸, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1999-  1999
• Removal of additive white noise using an adaptive wiener filter with edge retention

  Masato Tsukahara, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  30-  (1)  81  -89  1999  [Not refereed][Not invited]
   

  In this paper, an adaptive Wiener filter for removal of additive white noise is proposed. Images are partitioned into a set of blocks of pixels, divided into five subsets of blocks according to their edge contents and directions, namely, shade, horizontal, vertical, and two diagonal classes. Each subset of blocks is used to define a covariance matrix, from which a Wiener filter is derived. For classification of blocks in the presence of noise, five eigenvectors calculated from the five covariance matrices are used. For a block classified into the appropriate class, five inner products are calculated between five eigenvectors and the block in the presence of noise. After classification, by switching the Wiener filter according to the input block, edge-preserving image filtering is useful. Experimental results are included to verify the usefulness of the proposed method. © 1999 Scripta Technica.
• A filter-coefficient quantization method with genetic algorithm

  M Haseyama, H Kitajima  ISCAS '99: PROCEEDINGS OF THE 1999 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 3  3-  399  -402  1999  [Not refereed][Not invited]
   

  A method with a genetic algorithm (GA) to quantize filter coefficients is proposed. The proposed method searches a population of the quantized filters for the optimal quantized filter which retains the frequency characteristic of a reference filter, by using the GA. In order to establish this method, we decide how to execute the following GA procedures: representation of strings; fitness (objective function); reproduction; crossover and mutation. The experimental results are included to verify that the proposed method can obtain the filter with better frequency characteristic than ones which the traditional quantization methods, such as to round off, round up, and round down, provides.
• A criterion-based image segmentation method with a genetic algorithm

  M Haseyama, N Iwai, H Kitajima  ISCAS '99: PROCEEDINGS OF THE 1999 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 4  4-  94  -97  1999  [Not refereed][Not invited]
   

  This paper proposes a new genetic algorithm (GA) based image segmentation method for image analysis. This method can segment an observed image into some regions based on a criterion, The criterion is defined as MSE caused by interpolating each region of the image with a parametric model. Since the criterion is expressed with not only the parameters of the model but also shape and location of the regions, the criterion can not be easily minimized by the usual optimization methods, the proposed method minimizes the criterion by a GA. The proposed method also includes a processor to eliminate small fragments with the Markov random field (MRF) model. Though the thresholds of the existent region-segmentation methods negatively affect image segmentation results; since no thresholds are required in the proposed method, it segments images more accurately than the existent methods.
• Genetic algorithm based image segmentation for image analysis

  Miki Haseyama, Masateru Kumagai, Hideo Kitajima  ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings  6-  3445  -3448  1999/01/01  [Not refereed][Not invited]
   

  In this paper a new genetic algorithm (CA) based image segmentation method is proposed for image analysis. This method using a mean square error (MSE) based criterion can segment an image into some regions, while estimating a suitable region representation. The criterion is defined as MSE caused by interpolating each region of an observed image with a parametric model. Since the criterion is expressed with not only the parameters of the model but also shape and location of the regions, the criterion can not be easily minimized by the usual optimization methods, the proposed method minimizes the criterion by a GA. The proposed method also includes a processor to eliminate fragile regions with the Markov random field (MRF) model. Though the thresholds of the existent methods negatively affect image segmentation results; since no thresholds are required in the proposed method, it segments images more accurately than the existent methods.
• A genetic algorithm for determining multiple routes and its applications

  J Inagaki, M Haseyama, H Kitajima  ISCAS '99: PROCEEDINGS OF THE 1999 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 6  6-  137  -140  1999  [Not refereed][Not invited]
   

  This paper presents a genetic algorithm approach for routing applications, The genetic algorithm produces many possible solutions in its search process. BS utilizing this feature, our method determines both the shortest route and multiple semi-shortest routes in one search. The existent methods: such as Dijkstra's algorithm, search only for the shortest route, but they cannot determine any other short, routes, Therefore proposed algorithm is useful for this kind of routing, Furthermore, the method can determine the routes which are short and not, similar depending on geographical conditions since the genotype structure of our method has a weighting factor which controls route length.
• A real-time line extraction algorithm

  J Gates, M Haseyama, H Kitajima  ISCAS '99: PROCEEDINGS OF THE 1999 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOL 4  4-  68  -71  1999  [Not refereed][Not invited]
   

  This paper presents an algorithm that is capable of extracting lines from digital images in real-time. The algorithm uses a radial search method that quickly converges to the line segment and then the detected line is verified using the least-mean-squares (LMS) algorithm. Tests were performed on 256 x 256-grayscale images using a 166 MHz Pentium processor and an average frame-rate of more than twenty frames/second was achieved. The line extraction performance of the algorithm is comparable with that produced by the well known but much slower, Hough Transform. The algorithm is approximately two orders of magnitude faster than the Hough Transform [1][2].
• A shape-constraint-based active contour model

  Wen Zhong, Miki Haseyama, Hideo Kitajima  Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers  53-  (10)  1421  -1429  1999  [Not refereed][Not invited]
   

  An active contour model Snake and its modifications are widely used for image analysis and computer vision problems, especially for object contour extraction. To overcome some drawbacks associated with existing active contour models used for object contour extraction, this paper presents a shape-constraint-based active contour model (SC-ACM) which uses information of the contour shape of the target object to constrain the snake and controls the snake shape when the snake moves at every step, where the object contour shape is assumed to be roughly known a priori. Successful application of the SC-ACM to extraction of contours on real and synthetic images demonstrates its effectiveness.
• A study of magnetization fluctuations in transition region using MFM image analysis

  I. Takekuma, M. Haseyama, K. Sueoka, K. Mukasa, M. Yasui  IEEE Transactions on Magnetics  35-  (5)  2685  -2687  1999  [Not refereed][Not invited]
   

  This paper proposes a method for analysis of MFM (magnetic force microscopy) images. By using the proposed method, the influence of the magnetization fluctuations in transition regions can be analyzed. In the proposed method, the MFM images are processed as stochastic signals, and their stochastic parameters are obtained. The variance profile (the variance of MFM signals across the track direction), one of the parameters, is closely related to the magnetization fluctuations in the transition region. Therefore, by observation of the variance profile, the fluctuations can be quantitatively analyzed. An example of the analysis, which is based on the variance profile, is presented. From the analysis, it is found that the parameter L/W, which is introduced to characterize the variance profile, has the same trend as the S/Nm (Nm:medium noise) does. © 1999 IEEE.
• Removal of additive white noise using an adaptive wiener filter with edge retention

  Masato Tsukahara, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  30-  (1)  81  -89  1999  [Not refereed][Not invited]
   

  In this paper, an adaptive Wiener filter for removal of additive white noise is proposed. Images are partitioned into a set of blocks of pixels, divided into five subsets of blocks according to their edge contents and directions, namely, shade, horizontal, vertical, and two diagonal classes. Each subset of blocks is used to define a covariance matrix, from which a Wiener filter is derived. For classification of blocks in the presence of noise, five eigenvectors calculated from the five covariance matrices are used. For a block classified into the appropriate class, five inner products are calculated between five eigenvectors and the block in the presence of noise. After classification, by switching the Wiener filter according to the input block, edge-preserving image filtering is useful. Experimental results are included to verify the usefulness of the proposed method. © 1999 Scripta Technica.
• Performance analysis and improvement of a Hopfield NN image restoration method

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (484)  79  -84  1998/12/17  [Not refereed][Not invited]
   

  One of image restoration methods using Neural Networks(NNs) utilized two kinds of NNs: edge detection and restoration NNs. This paper, modifies the edge detection Hopfield NN(HNN) to improve the restoration performance. For the improvement, we introduce new edge parameters. In the previous method, two kinds of edge parameters are prepared; they detect vertical direction and horizontal direction edges. In this paper, other two kinds of edge parameters are prepared which detect light slanting direction and left slanting direction edges. Therefore the proposed modification results more accurate edge detection. New criteria to accurately edge parameters are also proposed. They can reduce the calculation costs.
• Improving Active Contour Model Using Shape Information

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (484)  85  -90  1998/12/17  [Not refereed][Not invited]
   

  In this paper we present a shape-constraint-based active contour model (SC-ACM) to address the drawbacks associated with existing active contour models that some parameters must be set by elaborate human intervention. The SC-ACM use information of the contour approximate shape of the target object to controls the snake moving under the assumption that the object contour shape is approximately known. The SC-ACM is applied to detect the lip contours of announcers for TV is more effective than the previous methods.
• Lossless Image Coding Using an Approximate DCT Transform

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (484)  91  -96  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using an approximate DCT transform that is an interger matrix which approximates a DCT matrix. A lossless image coding method using an integer transform has already been proposed. The method has a problem in that the inverse transform is not its transpose matrix. This paper proposes an approximate DCT that is able to use the associated transpose matrix in the inverse transform and more closely approximates the DCT transform. Experimental results from the proposed method are also presented.
• A Method of Similar Region Search on Fractal Block Coding

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (484)  97  -102  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes an improved fractal block coding by adopting the overlap search for domain blocks. Conventionally, the quad-tree partitioning method is usually used in the case of a variable block size. In this method, the bitrate becomes high for obtaining high image quality, because of a large number of small blocks. The proposed method improves the image quality while suppressing the bitrate increase by searching the overlapped domain blocks before the quad-tree division. In addition we consider a threshold classifying range blocks.
• Performance analysis and improvement of a Hopfield NN image restoration method

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  98-  (482)  79  -84  1998/12/17  [Not refereed][Not invited]
   

  One of image restoration methods using Neural Networks(NNs) utilizes two kinds of NNs:edge detection and restoration NNs. This paper, modifies the edge detection Hopfield NN(HNN)to improve the restoration performance. For the improvement, we introduce new edge parameters. In the previous method, two kinds of edge parameters are prepared; they detect vertical direction and horizontal direction edge. In this paper, other two kinds of edge parameters are prepared which detect light slanting direction and left slanting direction edges. Therefore the proposed modification results more accurate edge detection. New criteria to accurately edge parameters are also proposed. they can reduce the calculation costs.
• Improving Active Contour Model Using Shape Information

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  98-  (482)  85  -90  1998/12/17  [Not refereed][Not invited]
   

  In this paper we present a shape-constraint-based active contour model (SC-ACM) to address the drawbacks associated with existing active contour models that some parameters must be set by elaborate human intervention. The SC-ACM use information of the contour approximate shape of the target object to controls the snake moving under the assumption that the object contour shape is approximately known. THE SC-ACM is applied to detect the lip contours of announcers for TV face-shoulder images. Successful application of the SC-ACM domenstrates that the presented method is more effective than the previous methods.
• Lossless Image Coding Using an Approximate DCT Transform

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  98-  (482)  91  -96  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using an approximate DCT transform that is an interger matrix which approximates a DCT matrix. A lossless image coding method using an integer transform has already been proposed. The method has a problem in that the inverse transform is not its transpose matrix. This paper proposes an approximate DCT that is able to use the associated transpose matrix in the inverse transform and more closely approximates the DCT transform. Experimental results from the proposed method are also presented.
• A Method of Similar Region Search on Fractal Block Coding

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  98-  (482)  97  -102  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes an improved fractal block coding by adopting the overlap search for domain blocks. Conventionally, the quad-tree partitioning method is usually used in the case of variable block size, In this method, the bitrate becomes high for obtaining high image quality, because of a large number of small blocks. The proposed method improves the image quality while suppressing the bitrate increase by searching the overlapped domain blocks before the quad-tree division. In addition we consider a threshold classifying range blocks.
• Performance analysis and improvement of a Hopfield NN image restoration method

  UCHIYAMA Yoshifumi, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (71)  79  -84  1998/12/17  [Not refereed][Not invited]
   

  One of image restoration methods using Neural Networks (NNs) utilizes two kinds of NNs : edge detection and restoration NNs. This paper, modifies the edge detection Hopfield NN(HNN) to improve the restoration performance. For the improvement, we introduce new edge parameters. In the previous method, two kinds of edge parameters are prepared ; they detect vertical direction and horizontal direction edges. In this paper, other two kinds of edge parameters are prepared which detect light slanting direction and left slanting direction edges. Therefore the proposed modification results more accurate edge detection. New criteria to accurately edge parameters are also proposed. They can reduce the calculation costs.
• Improving Active Contour Model Using Shape Information

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (71)  85  -90  1998/12/17  [Not refereed][Not invited]
   

  In this paper we present a shape-constraint-based active contour model (SC-ACM) to address the drawbacks associated with existing active contour models that some parameters must be set by elaborate human intervention. The SC-ACM use information of the contour approximate shape of the target object to controls the snake moving under the assumption that the object contour shape is approximately known. The SC-ACM is applied to detect the lip contours of announcers for TV face-shoulder images. Successful application of the SC-ACM domenstrates that the presented method is more effective than the previous methods.
• Lossless Image Coding Using an Approximate DVT Transform

  SUEHIRA Kouji, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (71)  91  -96  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using an approximate DCT transform that is an interger matrix which approximates a DCT matrix. A lossless image coding method using an integer transform has already been proposed. The method has a problem in that the inverse transform is not its transpose matrix. This paper proposes an approximate DCT that is able to use the associated transpose matrix in the inverse transform and more closely approximates the DCT transform. Experimental results from the proposed method are also presented.
• A Method of Similar Region Search on Fractal Block Coding

  SUZUKI Yasuo, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (71)  97  -102  1998/12/17  [Not refereed][Not invited]
   

  This paper proposes an improved fractal block coding by adopting the overlap search for domain blocks. Conventionally, the quad-tree partitioning method is usually used in the case of a variable block size. In this method. the bitrate becomes high for obtaining high image quality, because of a large number of small blocks. The proposed method improves the image quality while suppressing the bitrate increase by searching the overlapped domain blocks before the quad-tree division. In addition we consider a threshold classifying range blocks.
• Transformation of normalized ARMA lattice filters for the purpose of signal synthesis

  Miki Haseyama  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E81-A-  1529  -1532  1998/12/01  [Not refereed][Not invited]
   

  This letter proposes a method to transform normalized ARMA lattice filters, which are originally realized for signal analysis, into signal synthesis lattice filters. Although the transformation method has been proposed for normalized ARMA lattice filters with the MA order which is greater than or equal to the AR order, it has not been done when the AR order is greater than the MA order. With the proposed method, once an ARMA lattice filter with the AR order greater than the MA order is realized, then it can be transformed to the signal synthesis filter.
• An analysis of MFM images of recording media as stochastic signals

  武隈 育子, 安井 雅彦, 末岡 和久, 長谷山 美紀, 武笠 幸一  日本応用磁気学会学術講演概要集 = Digest of ... annual conference on magnetics in Japan  22-  474  -474  1998/09/01  [Not refereed][Not invited]
  
• A Method for Analyzing MFM Images as Stochastic Signals

  TAKEKUMA I, YASUI M, OKUMURA Y, AKITA M, SUEOKA K, HASEYAMA M, MUKASA K  Journal of Magnetics Society of Japan  22-  (9)  1251  -1256  1998/09/01  [Not refereed][Not invited]
   

  Magnetic force microscopy(MFM) is widely used to directly observe magnetized states with sub-micron resolution. The average profile, which is commonly used in the conventional quantitative analysis of MFM images, does not contain enough information to evaluate recording media. We introduce the concept of a stochastic process into MFM image analysis. By acquiring the statistical parameters-the average, the variance, and the histogram of the image-we reveal that the distribution of the fluctuation in the magnetic transition region is different from the Gaussian distribution. We also suggest a new evaluation value, L/W, which is related to the S/N of the read-back signal. With our model, it may be possible to obtain higher-order statistical parameters, which are important for achieving more precise characterization in the magnetic transition region.
• A Note on a High Accuracy Implementation of a CORDIC ARMA Lattice Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  98-  (232)  1  -6  1998/07/31  [Not refereed][Not invited]
   

  This paper proposes a scheme for a high accuracy implementation of a CORDIC ARMA lattice filter. A CORDIC implementation method for a normalized ARMA lattice filter has been reported. However, the method has a problem that its CORDIC lattice elementary section cannot execute the rotation operation by small angles. Therefore, we propose a new scheme to implement another CORDIC lattice elementary section using the shift sequence proposed by Walther without the problem. By using the new CORDIC lattice elementary section, we can realize the CORDIC ARMA lattice filter accurately.
• Transformation of normalized ARMA lattice filters for the purpose of signal synthesis

  M Haseyama, S Shiraishi, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E81A-  (7)  1529  -1532  1998/07  [Not refereed][Not invited]
   

  This letter proposes a method to transform normalized ARMA lattice filters, which are originally realized for signal analysis, into signal synthesis lattice filters. Although the transformation method has been proposed for normalized ARMA lattice filters with the MA order which is greater than or equal to the AR order, it has not been done when the AK order is greater than the MA order. With the proposed method, once an ARMA lattice filter with the AR order greater than the MA order is realized, then it can be transformed to the signal synthesis filter.
• 1)適応KL基底によるエッジ保存を考慮した画像符号化(3次元画像処理および一般)(〔情報ディスプレイ研究会 ヒューマンインフォメーション研究会 マルチメディア情報処理研究会 ネットワーク映像メディア研究会 画像情報システム研究会〕合同)

  塚原 正人, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  1998/06/20  [Not refereed][Not invited]
  
• 2)テーブル参照による動き補償のための動きモデルの検討(3次元画像処理および一般)(〔情報ディスプレイ研究会 ヒューマンインフォメーション研究会 マルチメディア情報処理研究会 ネットワーク映像メディア研究会 画像情報システム研究会〕合同)

  浅水 仁, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  830  -830  1998/06/20  [Not refereed][Not invited]
  
• 3)3次元モデルを用いた動き補償における符号量の検討(3次元画像処理および一般)(〔情報ディスプレイ研究会 ヒューマンインフォメーション研究会 マルチメディア情報処理研究会 ネットワーク映像メディア研究会 画像情報システム研究会〕合同)

  小池 理一郎, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  1998/06/20  [Not refereed][Not invited]
  
• 4)反復関数系による画像処理法の検討

  本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  832  -832  1998/06/20  [Not refereed][Not invited]
  
• 5)音声信号を利用したカット点検出手法に関する考察 : その改良と実験結果

  薄田 隆博, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  1998/06/20  [Not refereed][Not invited]
  
• 6)CORDIC ARMA ラチス合成フィルタの任意次数実現法

  白石 真一, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (6)  832  -832  1998/06/20  [Not refereed][Not invited]
  
• Motion Table Designed for Motion Compensation

  ASAMIZU Satoshi, HASEGAWA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  98-  (136)  63  -70  1998/06/19  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for motion-parameter identification in motion table. The motion compensation based on table lookup selects the motion to minimize perdiction error from the motion table. Therefore, the quality of the prediction image changes making the motion table. The parameters identified by the conventional method the LBG algorithm have quantization errors which make the decoded images are contaminated. The proposed method decrease the effect of this quantization error. A pair of motion tables sets up by distinct methods are compared.
• 8)画像符号化におけるエッジ保存を考慮した適応KL基底(放送方式研究会)

  塚原 正人, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  673  -674  1998/05/20  [Not refereed][Not invited]
  
• 9)動きテーブル参照による動き補償方式に関する考察(放送方式研究会)

  浅水 仁, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 10)フラクタル次元を用いた画素補間法に関する考察(放送方式研究会)

  本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 11)ファジィ推論を用いた物体抽出のための効果的な初期領域の決定法(放送方式研究会)

  新井 哲也, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 12)複数経由点指定を伴う経路探索に関する考察(放送方式研究会)

  稲垣 潤, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 1)遺伝的アルゴリズムを用いたフラクタルブロック符号化法(放送方式研究会)

  小林 亜令, 本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 2)アダマール変換の剰余パターンの類別を用いた静止画像のロスレス符号化(放送方式研究会)

  星野 亮, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 3)動画像のショット切換え種別の判定に関する考察(放送方式研究会)

  槙 宏晃, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 4)ファジィ理論を用いた音声データによるカット点検出の精度向上に関する考察(放送方式研究会)

  薄田 隆博, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -674  1998/05/20  [Not refereed][Not invited]
  
• 5)物体の3次元的回転を考慮した動き補償のための動きモデル(放送方式研究会)

  小池 理一郎, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  674  -675  1998/05/20  [Not refereed][Not invited]
  
• 6)分散が異なる雑音による劣化画像の復元に関する考察(放送方式研究会)

  熊谷 政輝, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  675  -675  1998/05/20  [Not refereed][Not invited]
  
• 7)信号合成のためのCORDIC ARMAラチスフィルタの実現法に関する一考察(放送方式研究会)

  白石 真一, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  675  -675  1998/05/20  [Not refereed][Not invited]
  
• 8)IFSを用いたパルス性雑音による劣化画像の復元法(放送方式研究会)

  岡美 佳子, 本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  52-  (5)  675  -675  1998/05/20  [Not refereed][Not invited]
  
• Removal of Additive White Noise Using an Adaptive Wiener Filter with Edge Retention

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  81-  (4)  623  -631  1998/04/25  [Not refereed][Not invited]
   

  本論文では, 画像のエッジを保存しつつ加法性白色雑音除去を行うための適応ウィーナーフィルタを提案する.学習用の複数の画像を用意した上で, 各画像を小画像(以下ではブロックと呼ぶ)に分割して, ブロックをその内部のエッジの性質により平たん, 垂直, 水平, 斜め方向(45度, 135度)の五つのクラスに分類する.そして, クラスごとに共分散行列を求め, ウィーナーフィルタを設計する.雑音除去におけるブロックのクラス分類は, 各クラスの共分散行列の固有ベクトルを用いて行われる.各クラスの固有ベクトルとノイズの付加された観測画像のブロックとの内積をとり, 得られた内積値に基づいて入力ブロックを適切なクラスへと分類する.クラス分類後, 入力ブロックのクラスに対応したウィーナーフィルタへ切換えを行うことにより, エッジを保存した高精度な画像復元が可能となる.本論文の最後では, さまざまな実験を行い提案手法の有用性を確認する.
• A Note on the Residues of Hadamard Transforms of Integer Signals

  Kitajima Hideo, Haseyama Miki  Proceedings of the IEICE General Conference  1998-  (2)  82  -82  1998/03/06  [Not refereed][Not invited]
  
• An Implementation of a Normalized ARMA Lattice Filter with a CORDIC Algorithm

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers. A  81-  (2)  198  -207  1998/02/25  [Not refereed][Not invited]
   

  本論文ではCORDICアルゴリズムを用いて規格化ARMAラチスフィルタを実現する方法を提案する.既に提案されているCORDICアルゴリズムを用いた規格化ARラチスフィルタの実現法は, 回転角の収束値, 収束速度の2点において問題があり, それを拡張し, 規格化ARMAラチスフィルタに応用することは困難である.そこで, 本論文ではこの問題を改善し, CORDIC ARMAラチスフィルタの実現を可能とする.本論文で提案するフィルタは, 規格化ARMAラチスフィルタの実現に特化したCORDICアルゴリズムを用いて実現されるため, VLSIで効果的に実現することが期待できる.更に, このフィルタは, 入力信号の推定を行いながらフィルタリングを行うことにより, 入力信号が未知の場合にも用いることができる.本論文の最後では, 実験により, 提案手法の有効性を確認する.
• Image Analysis Through Iterated Function Systems

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (529)  23  -28  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes an image analysis through iterated function systems (IFSs).On the IFSs, an image can be represented by few parameters.Therefore, the parameters contain some characteristics of the image.First, we examine some statistical characteristics of the IFS parameters which represent the location of Domain Block.Next, an edge detection using IFSs is shown. And a statistical characteristic of the image obtained by the above IFSs is examined.These results show the availability of this method.
• A Note of Cut Detection Using Audio Signals : Its Improvement and Results

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (529)  29  -34  1998/02/04  [Not refereed][Not invited]
   

  This paper presents a study of improvement for cut detection of video signals using audio ones.The existent method assists the cut detection effectively by fuzzy c-means clustering of PARCOR coefficients computed from audio signals corresponding to the video ones.However, this method has a problem which is how to select the number of clusters.To solve this problem, this paper proposes the improved clustering method for, the PARCOR coefficients computed from audio signals.
• An Arbitrary Order Implementation of a CORDIC ARMA Lattice Synthesis Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (529)  35  -42  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes a method which transforms CORDIC ARMA lattice filters for signal analysis into signal synthesis filters in case where their AR orders are larger than the MA orders. By using CORDIC processors, the proposed signal synthesis filter can be effectively implemented without any multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, it keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Image Coding for Edge Retention using Adaptive KL Bases

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (526)  1  -6  1998/02/04  [Not refereed][Not invited]
   

  Image coding for edge retention using adaptive KL transform is investigated. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Five covariance matrices are calculated and five matrices of KL transform(adaptive KL transform)are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. The DCT is used for image coding of the low frequency components. Adaptive KLT is used for edge images coding(high frequency components). The method for edge images coding is compared with a method using the DCT and the effectiveness of using adaptive KL transform is examined.
• A Motion Model for Motion Compensation Based on Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (526)  7  -12  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes a motion model using a DCT for motion compensation based on table lookup. In this paper the coefficients of DCT finds from a movement of every pixel. It is decided that the number of the coefficients in IDCT. The proposed method represented motion that the coefficients of IDCT. The proposed method make a table using the coefficients of DCT number can be set up freely. Various motion patterns are put into a table. The motion compensation based on table lookup selects the motion to minimize prediction error from the table.
• Bitrate in Motion-Compensated Prediction Using a 3-D Model

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (526)  13  -18  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes multi-flame motion-compensated prediction using 3-D model. Motion-compensated prediction using 3-D model has much quantity of coding. The proposed method improves coding efficiency by using same parameter in multi-flame. Results of some experiments show to verify the effect of proposed method.
• Image Analysis Through Iterated Function Systems

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (527)  23  -28  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes an image analysis through iterated function systems(IFSs). On the IFSs, an image can be represented by few parameters. Therefore, the parameters contain some characteristics of the image. First, we examine some statistical characteristics of the IFS parameters which represent the location of Domain Block. Next, an edge detection using IFSs is shown. And a statistical characteristic of the image obtained by the above IFSs is examined. These results show the availability of this method.
• A Note of Cut Detection Using Audio Signals : Its Improvement and Results

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (527)  29  -34  1998/02/04  [Not refereed][Not invited]
   

  This paper presents a study of improvement for cut detection of video signals using audio ones. The existent method assists the detection effectively by fuzzy c-means clustering of PARCOR coefficients computed from audio signals corresponding to the video ones. However, this method has a problem which is how to select the number of clusters. To solve this problem, this paper proposes the improved clustering method for the PARCOR coefficients computed from audio signals.
• An Arbitrary Order Implementation of a CORDIC ARMA Lattice Synthesis Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Electronic information displays  97-  (527)  35  -42  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes a method which transforms CORDIC ARMA lattice filters for signal analysis into signal synthesis filters in case where their AR orders are larger than the MA orders. BY using CORDIC processors, the proposed signal synthesis filter can be effectively implemented without any multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, it keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Image Analysis Through Iterated Function Systems

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA hideo  ITE Technical Report  22-  (9)  23  -28  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes an image analysis through iterated function systems (IFSs). On the IFSs, an image can be represented by few parameters. Therefore, the parameters contain some characteristics of the image. First, we examine some statistical characteristics of the IFS parameters which represent the location of Domain Block. Next, an edge detection using IFSs is shown. And a statistical characteristic of the image obtained by the above IFSs is examined. These results show the availability of this method.
• A Note of Cut Detection Using Audio Signals -Its Improvement and Results-

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA hideo  ITE Technical Report  22-  (9)  29  -34  1998/02/04  [Not refereed][Not invited]
   

  This paper presents a study of improvement for cut detection of video signals using audio ones. The existent method assists the cut detection effectively by fuzzy c-means clustering of PARCOR coefficients computed from audio signals corresponding to the video ones. However, this method has a problem which is how to select the number of clusters. To solve this problem, this paper proposes the improved clustering method for the PARCOR coefficients computed from audio signals.
• An Arbitrary Order Implementation of a CORDIC ARMA Lattics Synthesis Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA hideo  ITE Technical Report  22-  (9)  35  -42  1998/02/04  [Not refereed][Not invited]
   

  This paper proposes a method which transforms CORDIC ARMA lattice filters for signal analysis into signal synthesis filters in case where their AR orders are larger than the MA orders. By using CORDIC processors, the proposed signal synthesis filter can be effectively implemented without any multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, it keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Image Coding for Edge Retention using Adaptive KL Bases

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (528)  1  -6  1998/02/03  [Not refereed][Not invited]
   

  Image coding for edge retention using adaptive KL transform is investigated. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Five covariance matrices are calculated and five matrices of KL transform(adaptive KL transform) are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. The DCT is used for image coding of the low frequency components. Adaptive KLT is used for edge images coding(high frequency components). The method for edge images coding is compared with a method using the DCT and the effectiveness of using adaptive KL transform is examined.
• A Motion Model for Motion Compensation Based on Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (528)  7  -12  1998/02/03  [Not refereed][Not invited]
   

  This paper proposes a motion model using a DCT for motion compensation based on table lookup. In this paper the coefficients of DCT finds from a movement of every pixel. It is decided that the number of the coefficients in IDCT. The proposed method represented motion that the coefficients of IDCT. The proposed method make a table using the coefficients of DCT number can be set up freely. Various motion patterns are put into a table. The motion compensation based on table lookup selects the motion to minimize prediction error from the table.
• Bitrate in Motion-Compensated Prediction Using a 3-D Model

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (528)  13  -18  1998/02/03  [Not refereed][Not invited]
   

  This paper proposes multi-flame motion-compensated prediction using 3-D model. Motion-compensated prediction using 3-D model has much quantity of coding. The proposed method improves coding efficiency by using same parameter in multi-flame. Results of some experiments show to verify the effect of proposed method.
• Image Coding for Edge Retention using Adaptive KL Bases.

  TSUKAHARA.Masato, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (8)  1  -6  1998/02/03  [Not refereed][Not invited]
   

  Image coding for edge retention using adaptive KL transform is investigated. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Five covariance matrices are calculated and five matrices of KL transform (adaptive KL transform) are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. The DCT is used for image coding of the low frequency components. Adaptive KLT is used for edge images coding (high frequency components). The method for edge images coding is compared with a method using the DCT and the effectiveness of using adaptive KL transform is examined.
• A Motion Model for Motion Compensation Based on Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (8)  7  -12  1998/02/03  [Not refereed][Not invited]
   

  This paper proposes a motion model using a DCT for motion compensation based on table lookup. In this paper the coefficients of DCT finds from a movement of every pixel. It is decided that the number of the coefficients in IDCT. The proposed method represented motion that the coefficients of IDCT. The proposed method make a table using the coefficients of DCT number can be set up freely. Various motion patterns are put into a table. The motion compensation based on table lookup selects the motion to minimize prediction error from the table.
• Bitrate in Motion-Compensated Prediction Using a 3-D Model

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  22-  (8)  13  -18  1998/02/03  [Not refereed][Not invited]
   

  This paper proposes multi-flame motion-compensated prediction using 3-D model. Motion-compensated prediction using 3-D model has much quantity of coding. The proposed method improves coding efficiency by using same parameter in multi-flame. Results of some experiments show to verify the effect of proposed method.
• Consideration on realization of motion detection device using CORDIC algorithm.

  山下雅人, 白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• A Problem of IFS Parameter Determination based on the Collage Theorem.

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on edge extraction by Hopfield NN for image restoration.

  内山賢史, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Application of genetic algorithm to round path search with multiple point pass appointment.

  稲垣潤, 長谷山美紀, 北島秀夫, 高堰茂  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Eextraction of lip contour using dynamic contour model based on the shape.

  鐘文, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on reversible encoding of static image using DCT.

  季平浩治, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on corresponding point search using elastic model.

  武藤剛史, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Initial domain determination method for object extraction by fuzzy reasoning considering color components.

  岩井範幸, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on closed curve tracking using Process Grammar in moving image.

  牛丸主税, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Code quantity reduction using motion compensation by moving table reference.

  浅水仁, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on high precision realization of CORDIC ARMA lattice filter.

  白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on variable block size IFS considering overlapping Domain Block.

  鈴木康夫, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Improvement of cut point detection technique using frequency characteristics of audio signal.

  薄田隆博, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Multiple frame motion compensation using three-dimensional model.

  小池理一郎, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Parameter evaluation of radar image reproduction processing technique.

  土田正芳, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on domain division technique using fundamental processing of GA.

  熊谷政輝, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• Consideration on linear quantization method for IFS parameter using genetic algorithm.

  竹沢恵, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1998-  1998
• A Note on Realization of ARMA Lattice Filters for Signal Synthesis.

  長谷山美紀, 白石真一, 北島秀夫  電子情報通信学会大会講演論文集  1998-  1998
• A fuzzy reasoning based arma order selection method

  Miki Haseyama, Hideo Kitajima  European Signal Processing Conference  1998-January-  1  -4  1998/01/01  [Not refereed][Not invited]
   

  A fuzzy reasoning based approach for ARMA order selection is discussed in this paper. The proposed method attempts to select the optimal ARMA order of a time-varying ARMA model. This method improves model validity-criterion based order selection, such as the AIC (Akaike's Information Criterion) and the MDL (Minimum Description Length), etc with applying both of a fuzzy reasoning method and a fuzzy c-means clustering method. These fuzzy methods are incorporated in the proposed method as the follows: (1) Suppose the ARMA order of the reference time-varying model changes, the suitable ARMA order is selected by utilizing a recursive fuzzy reasoning method. (2) By using the fuzzy c-means clustering method, we detect the time at which the ARMA order of the reference model changes, and the clustering values are used for adaptively setting the forgetting factor in the recursive fuzzy reasoning method. The experimental results show that the proposed method effectively selects the ARMA orders of a time-varying ARMA model.
• An implementation of a normalized ARMA lattice filter with a CORDIC algorithm

  S Shiraishi, M Haseyama, H Kitajima  ISCAS '98 - PROCEEDINGS OF THE 1998 INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-6  5-  D253  -D256  1998  [Not refereed][Not invited]
   

  This paper presents a CORDIC (COordinate Rotation DIgital Computer) implementation of a normalized ARMA lattice filter. Since the proposed method uses a simplified CORDIC algorithm it is possible to implement the CORDIC ARMA lattice filter with a reduced amount of hardware. In addition, we propose a method which transforms the CORDIC ARMA lattice filter realized for signal analysis into a signal synthesis filter without an increase in the implementation complexity. Moreover, the proposed filter can identify unknown-input models so that it can be widely used in the digital signal processing.
• A fuzzy reasoning based approach for ARMA order selection

  M Haseyama, M Emura, H Kitajima  ISCAS '98 - PROCEEDINGS OF THE 1998 INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-6  3-  B259  -B262  1998  [Not refereed][Not invited]
   

  A fuzzy reasoning based approach for ARMA order selection is discussed in this paper. The proposed method attempts to select the optimal ARMA order of a time-varying ARMA model. This method improves model validity-criterion based order selection, such as the AIC (Akaike's Information Criterion) and the MDL (Minimum Description Length), etc with applying both of a fuzzy reasoning method and a fuzzy c-means clustering method. These fuzzy methods are incorporated in the proposed method as follows: (1) Suppose the ARMA order of the reference time-varying model changes. The suitable ARMA order is selected by utilizing a recursive fuzzy reasoning method. (2) By using a fuzzy c-means clustering method, we detect the time at which the ARMA order of the reference model changes, and the clustering values are used for adaptive setting the forgetting factor in the recursive fuzzy reasoning method. The experimental results show that the proposed method effectively selects the ARMA orders of a time-varying ARMA model.
• Image modeling and parameter identification for image restoration using a Kalman filter

  Takashi Jo, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  29-  (3)  1  -9  1998  [Not refereed][Not invited]
   

  This paper proposes a Kalman filter based method for high accuracy image restoration. When a Kalman filter is applied to image restoration, the model of the original image affects the accuracy of the restoration. An effective model for restoration depends on the characteristics of the image or the condition of the observed image. On the other hand, the correlation of the original image and the variance of the noise are necessary for image restoration with a Kalman filter. If these parameters are unknown, they must be identified from an observed image which has been contaminated with additive noise. To address the above problems, a method is proposed that identifies the number of pixels used for estimation and their positions. A method to estimate the unknown parameters in the image restoration process is also proposed. In this paper, the performance of the proposed algorithm is verified by simulations. ©1998 Scripta Technica.
• Virtual reference signals for active noise cancellation system

  Masaichi Akiho, Miki Haseyama, Hideo Kitajima  Journal of the Acoustical Society of Japan (E) (English translation of Nippon Onkyo Gakkaishi)  19-  (2)  95  -103  1998  [Not refereed][Not invited]
   

  This paper proposes a method that provides a new set of linear fixed coefficients filters to define and synthesize a new set of reference signals that are fewer in number than the original set of reference signals. It also can be used as the reference signals for an active noise cancellation system. To optimize the performance of the active noise cancellation system, it is very important to find proper reference signals that are coherent with the target noise. In a case of active road noise cancellation in a vehicle cabin, for satisfying this particular requirement, many accelerometers are installed to detect vibration on the vehicle's suspension mechanism. In most cases, these reference signals contain information about fewer number of noise sources compared to the number of accelerometers. Proposing the linear filtering method that is based on Singular Value Decomposition (SVD) of the power spectrum matrix of the reference signals enables the synthesis of the virtual reference signals. After defining the frequency response functions of the filters, we design the ARMA direct form digital filters. After designing the fixed coefficients filters for the virtual reference signals, we evaluate the performance of the ANC system through experimental tests and its results by comparing the original reference signals and the virtual reference signals to confirm the proposed method.
• Image modeling and parameter identification for image restoration using a Kalman filter

  Takashi Jo, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  29-  (3)  1  -9  1998  [Not refereed][Not invited]
   

  This paper proposes a Kalman filter based method for high accuracy image restoration. When a Kalman filter is applied to image restoration, the model of the original image affects the accuracy of the restoration. An effective model for restoration depends on the characteristics of the image or the condition of the observed image. On the other hand, the correlation of the original image and the variance of the noise are necessary for image restoration with a Kalman filter. If these parameters are unknown, they must be identified from an observed image which has been contaminated with additive noise. To address the above problems, a method is proposed that identifies the number of pixels used for estimation and their positions. A method to estimate the unknown parameters in the image restoration process is also proposed. In this paper, the performance of the proposed algorithm is verified by simulations. ©1998 Scripta Technica.
• Fractal Block Image Coding using a Genetic Algorithm

  KOBAYASHI Arei, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  1  -6  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for fractal block image coding. The conventional method effects of quantization error increase when the image is reconstructed. A genetic algorithm is constructed to cope with the problem. The proposed method quantizes fractal parameters their statistical properties. The proposed method uses a various block size. Results of experiments are included to show the effectiveness of the proposed method.
• Lossless Image Coding Using Hadamard Transform Residue Patterns

  HOSHINO Makoto, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  7  -12  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using Hadamard transform residue patterns. Hadamard transform residue patterns for N×N block of pixels are computed by dividing its coefficients by N×N. The proposed method improves coding efficiency by using the fact that the number of residue patterns is limited. The paper also proposes quasilossless coding by further limiting residue patterns.
• Identification of Shot Change Types for Image Sequences

  MAKI Hiroaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  13  -18  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes identification of scene change types for image sequences. The proposed method evaluates frame similarities through motion estimation. It can detect gradual shot changes such as dissolve and wipe; where other methods often fail. The proposed method identifies shot change types based on similarities. The effectiveness of the proposed method is verified by experiments.
• Improvement of Cut Detection by Audio Signals Analysis with a Fuzzy Theory

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  19  -24  1997/12/12  [Not refereed][Not invited]
   

  This paper describes a study of improvement for cut detection of video signals. The existent methods often fail to detect cut positions at fading in, out, or wiping scenes. The proposed algorithm improves them by using a fuzzy c-means clustering method. The clustered targets are PARCOR coefficients computed from audio signals corresponding to the video ones. The obtained cluster values are varied regardless of the failures if the audio signals are alternated. Therefore, it is expected that the alteration of the cluster values can assist the cut detection effectively.
• A Motion Model for Motion-Compensated Prediction Using a Sphere Moving in a 3-D Space

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  25  -30  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes motion-compensated prediction using a new motion model. The conventional block matching method does not consider motion in a 3-D space. The proposed method expresses motions using a sphere moving in a 3-D space. Results of some experiments show to verify the effect of proposed method.
• A Note of Restoration of corrupted image with variant variance of noise

  KUMAGAI Masateru, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  31  -36  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes a restoration method of images of which areas are contaminated by different variances of noises. When an image is restored by the Kalman filter algorithm, an image model and parameters (i.e. a noise variance, and a variance of the original image) are used. Therefore, when some areas of the image are contaminated by noises with different variances, the Kalman filter algorithm cannot restore accurately. The proposed method can estimate the noise variances in the areas of the image by using region segmentation results, and restore it more accurately than the previous Kalman filter algorithm does.
• A Note on an Implementation of a CORDIC ARMA Lattice Synthesis Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  37  -42  1997/12/12  [Not refereed][Not invited]
   

  This paper presents a method for a transformation of a CORDIC ARMA lattice filter into a signal synthesis filter. By using CORDIC processors, the proposed signal synthesis filter can be efficiently implemented without multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, this filter keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Image Restoration for Images Contaminated with Pulse Noise Using IFS

  ICHIOKA Mikako, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (430)  43  -47  1997/12/12  [Not refereed][Not invited]
   

  One restoration filter for images contaminated with pulse noise is a median filter. This filter often smoothes complex structure in the image. When the density of pulse noise is low, IFS image restoration methods which we previously proposed can keep complexity of structure in the image. However when the density of pulse noise is high, the restorated image is affected the noise. We propose in this paper an improved IFS image restoration method to change.
• Fractal Block Image Coding using a Genetic Algorithm

  KOBAYASHI Arei, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  1  -6  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for fractal block image coding. The conventional method effects of quantization error increase when the image is reconstructed. A genetic algorithm is constructed to cope with the problem. The proposed method quantizes fractal parameters their statistical properties. The proposed method uses a various block size. Results of experiments are included to show the effectiveness of the proposed method.
• Lossless Image Coding Using Hadamard Transform Residue Patterns

  HOSHINO Makoto, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  7  -12  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using Hadamard transform residue patterns. Hadamard transform residue patterns for N×N block of pixels are computed by dividing its coefficients by N×N. The proposed method improves coding efficiency by using the fact that the number of residue patterns is limited. The paper also proposes quasilossless coding by further limiting residue patterns.
• Identification of Shot Change Types for Image Sequences

  MAKI Hiroaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  13  -18  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes identification of scene change types for image sequences. The proposed method evaluates frame similarities through motion estimation. It can detect gradual shot changes such as dissolve and wipe; where other methods often fail. The proposed method identifies shot change types based on similarities. The effectiveness of the proposed method is verified by experiments.
• Improvement of Cut Detection by Audio Signals Analysis with a Fuzzy Theory

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  19  -24  1997/12/12  [Not refereed][Not invited]
   

  This paper describes a study of improvement for cut detection of video signals. The existent methods often fail to detect cut positions at fading in, out, or wiping scenes. The proposed algorithm improves them by using a fuzzy c-means clustering method. The clustered targets are PARCOR coefficients computed from audio signals corresponding to the video ones. The obtained cluster values are varied regardless of the failures if the audio signals are alternated. Therefore, it is expected that the alteration of the cluster values can assist the cut detection effectively.
• A Motion Model for Motion-Compensated Prediction Using a Sphere Moving in a 3-D Space

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  25  -30  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes motion-compensated prediction using a new motion model. The conventional block matching method does not consider motion in a 3-D space. The proposed method expresses motions using a sphere moving in a 3-D space. Results of some experiments show to verify the effect of proposed method.
• A Note of Restoration of corrupted image with variant variance of noise

  KUMAGAI Masateru, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  31  -36  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes a restoration method of images of which areas are contaminated by different variances of noises. When an image is restored by the Kalman filter algorithm, an image model and parameters (i.e. a noise variance, and a variance of the original image) are used. Therefore, when some areas of the image are contaminated by noises with different variances, the Kalman filter algorithm cannot restore accurately. The proposed method can estimate the noise variances in the areas of the image by using region segmentation results, and restore it more accurately than the previous Kalman filter algorithm does.
• A Note on an Implementation of a CORDIC ARMA Lattice Synthesis Filter

  SHIRAISHI Shin'ichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  37  -42  1997/12/12  [Not refereed][Not invited]
   

  This paper presents a method for a transformation of a CORDIC ARMA lattice filter into a signal synthesis filter. By using CORDIC processors, the proposed signal synthesis filter can be efficiently implemented without multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, this filter keeps the advantage of the CORDIC ARMA lattice filter, that is a simple structure.
• Image Restoration for Images Contaminated with Pulse Noise Using IFS

  ICHIOKA Mikako, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (428)  43  -47  1997/12/12  [Not refereed][Not invited]
   

  One restoration filter for images contaminated with pulse noise is a median filter. This filter often smoothes complex structure in the image. When the density of pulse noise is low, IFS image restoration methods which we previously proposed can keep complexity of structure in the image. However when the density of pulse noise is high, the restorated image is affected the noise. We propose in this paper an improved IFS image restoration method to change.
• Fractal Block Image Coding using a Genetic Algorithm

  KOBAYASHI Arei, HASEYAMA Miki, HONDA Hiroyuki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  1  -6  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for factal block image coding. The conventional method effects of quantization error increase when the image is reconstructed. A genetic algorithm is constructed to cope with the problem. The proposed method quantizes fractal parameters their statistical properties. The proposed method uses a various block size. Results of exeriments are included to show the effectiveness of the proposed method.
• Lossless Image Coding Using Hadamard Transform Residue Patterns

  HOSHINO Makoto, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  7  -12  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using Hadamard transform residue patterns. Hadamard transform reidue patterns for N×N block of pixels are computed by dividing its coefficients by N×N. The proposed method improves coding efficiency by using the fact that the number of residue patterns is limited. The paper also proposes quasilossless coding by further limiting residue patterns.
• Identification of Shot Change Types for Image Sequences

  MAKI Hiroaki, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  13  -18  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes identification of scene change types for image sequences. The proposed method evaluates frame similarities through motion estimation. It can detect gradual shot changes such as dissolve and wipe, where other methods often fail. The proposed method identifies shot change types based on similarities. The effectiveness of the proposed method is verified by experiments.
• Improvement of Cut Detection by Audio Signals Analysis with a Fuzzy Theory

  USUDA Takahiro, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  19  -24  1997/12/12  [Not refereed][Not invited]
   

  This paper describes a study of improvement for cut detection of video signals. The existent methods often fail to detect cut positions at fading in, out, or wiping scenes. The proposed algorithm improves them by using a fuzzy c-means clustering method. The clustered targets are PARCOR coefficients computed from audio signals corresponding to the video ones. The obtained cluster values are varied regardless of the failures if the audio signals are alternated. Therefore, it is expected that the alteration of the cluster values can assist the cut detection effectively.
• A Motion Model for Motion-Compensated Prediction Using a Sphere Moving in a 3-D Space

  KOIKE Riichiroh, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  25  -30  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes motion-compensated prediction using a new motion model. The conventional block matching method does not consider motion in a 3-D space. The proposed method expresses motions using a sphere moving in a 3-D space. Results of some experiments show to verify the effect of proposed method.
• A Note of Restoration of corrupted image with variant variance of noise

  KUMAGAI Masateru, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  31  -36  1997/12/12  [Not refereed][Not invited]
   

  This paper proposes a restoration method of images of which areas are contaminated by different variances of noises. When an image is restored by the Kalman filter algorithm, an image model and parameters (i.e. a noise variance, and a variance of the original image) are used. Therefore, when some areas of the image are contaminated by noises with different variances, the Kalmu filter algorithm cannot restore accurately. The proposed method can estimate the noise variances in the areas of the image by using region segmentation results, and restore it more accurately than the previous Kalman filter algorithm does.
• A Note on an Implementation of a CORDIC ARMA Lattice Synthesis Filter

  SHIRAISHI Shinichi, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  37  -42  1997/12/12  [Not refereed][Not invited]
   

  This paper presents a method for a transformation of a CORDIC ARMA lattice filter into a signal synthesis filter. By using CORDIC processors, the proposed signal synthesis filter can be efficiently implemented without multipliers. Since the obtained signal synthesis filter consists of dedicated CORDIC processors, this filter keeps the advantage of the CORDIC ARMA lattice filter, that is a simp1e structure.
• Image Restoration for Images Contaminated with Pulse Noise Using IFS

  HONDA Hiroyuki, OKA Mikako, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (78)  43  -48  1997/12/12  [Not refereed][Not invited]
   

  One restoration filter for images contaminated with pulse noise is a median filter. This filter often smoothes complex structure in the image. When the density of pulse noise is low, IFS image restoration methods which we previously proposed can keep complexity of structure in the image. However when the density of pulse noise is high, the restorated image is affected the noise. We propose in this paper an improved IFS image restoration method to change
• Image Coding using Adaptive KL Bases for Edge Retention

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (429)  45  -49  1997/12/11  [Not refereed][Not invited]
   

  Image coding using adaptive KL bases that reflect edge properties is investigated. Images are partitioned into blocks of pixels. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Adaptive KL bases that reflect edge features are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. Adaptive KL bases are designed for edge images (high frequency components), and then edge images are coded. The method is compared with a conventional method using the DCT and the effectiveness of using adaptive KL bases is examined.
• Motion Compensation Based on Motion-Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (429)  51  -56  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes motion compensation based on motion-table lookup. Various motion patterns are put into a motion-table. The proposed method selects the motion to minimize prediction error from the motion-table. A pair of motion-tables sets up by distinct methods are compared.
• Image Interpolation using Fractal Dimensions

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (429)  57  -62  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes image interpolation using fractal dimensions. This method calculates fractal dimensions from image. Since fractal dimensions correlate closely with roughness of images, this interpolation can generate images which have lower error rather than images generated by linear and spline interpolations. These interpolations lose high frequency components. And loss of the components cause error of the interpolated images. Fractal interpolations have been used for generation of geographical images. We change the interpolation in order to apply to other images.
• A Method of Finding the Region for Extracting Moving Object with Fuzzy reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (429)  63  -68  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a region segmentation method for extracting moving objects. In the proposed algorithm, after segmentation with a simple region growing algorithm, the centroid of each region is computed. the centroids are used as the initial values of the k-mean clustering. These regions the existent method for extracting moving objects with a fuzzy reasoning. It is verified that. the proposed method can effectively extract moving objects by some experiments.
• A Note on Searching the Shortest Route through Several Designated Points

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (429)  69  -74  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a method of searching the shortest route through several designated points using a genetic algorithm. The proposed method not only searches the shortest route but also selects the, shortest route through several designated points. Current methods, such as the Dijkstra's method, are the algorithms only for routing the shortest route. If the shortest route through several designated points is searched by these methods, its computational cost is O (N^2) where N is the number of the designated points. Therefore, the more the designated points increase, the more the routing demands computational cost. The proposed algorithm can obtain the shortest route through the points by an only search, and is more useful for this kind of routing than the other methods.
• Image Coding using Adaptive KL Bases for Edge Retention

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (427)  45  -49  1997/12/11  [Not refereed][Not invited]
   

  Image coding using adaptive KL bases that reflect edge properties is investigated. Images are partitioned into blocks of pixels. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Adaptive KL bases that reflect edge features are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. Adaptive KL bases are designed for edge images (high frequency components), and then edge images are coded. The method is compared with a conventional method using the DCT and the effectiveness of using adaptive KL bases is examined.
• Motion Compensation Based on Motion-Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (427)  51  -56  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes motion compensation based on motion-table lookup. Various motion patterns are put into a motion-table. The proposed method selects the motion to minimize prediction error from the motion-table. A pair of motion-tables sets up by distinct methods are compared.
• Image Interpolation using Fractal Dimensions

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (427)  57  -62  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes image interpolation using fractal dimensions. This method calculates fractal dimensions from image. Since fractal dimensions correlate closely with roughness of images, this interpolation can generate images which have lower error rather than images generated by linear and spline interpolations. These interpolations lose high frequency components. And loss of the components cause error of the interpolated images. Fractal interpolations have been used for generation of geographical images. We change the interpolation in order to apply to other images.
• A Method of Finding the Region for Extracting Moving Object with Fuzzy reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (427)  63  -68  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a region segmentation method for extracting moving objects. In the proposed algorithm, after segmentation with a simple region growing algorithm, the centroid of each region is computed. the centroids are used as the initial values of the k-mean clustering. These regions the existent method for extracting moving objects with a fuzzy reasoning. It is verified that. the proposed method can effectively extract moving objects by some experiments.
• A Note on Searching the Shortest Route through Several Designated Points

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  97-  (427)  69  -74  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a method of searching the shortest route through several designated points using a genetic algorithm. The proposed method not only searches the shortest route but also selects the, shortest route through several designated points. Current methods, such as the Dijkstra's method, are the algorithms only for routing the shortest route. If the shortest route through several designated points is searched by these methods, its computational cost is O (N^2) where N is the number of the designated points. Therefore, the more the designated points increase, the more the routing demands computational cost. The proposed algorithm can obtain the shortest route through the points by an only search, and is more useful for this kind of routing than the other methods.
• Image Coding using Adaptive KL Bases for Edge Retention.

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (77)  45  -49  1997/12/11  [Not refereed][Not invited]
   

  Image coding using adaptive KL bases that reflect edge properties is investigated. Images are partitioned into blocks of pixels. The pixel blocks are classified into subsets of blocks according to their edge contents and orientations. Adaptive KL bases that reflect edge features are designed for each class. In practice, images are separated into two frequency components, that are low and high frequency components. Adaptive KL bases are designed for edge images (high frequency components), and then edge images are coded. The method is compared with a conventional method using the DCT and the effativeness of using adaptive KL bases is examined.
• Motion Compensation Based on Motion-Table Lookup

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (77)  51  -56  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes motion compensation based on motion-table lookup. Various motion patterns are put into a motion-table. The proposed method selects the motion to minimize prediction error from the motion-table. A pair of motion-tables sets up by distinct methods are compared.
• Image Interpolation using Fractal Dimensions

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (77)  57  -62  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes image interpolation using fractal dimensions. This method calculates fractal dimensions from image. Since fractal dimensions correlate closely with roughness of images, this interpolation can generate images which have lower error rather than images generated ty linear and spline interpolations. These interpolations lose high frequency components. And loss of the components cause error of the interpolated images. Fractal interpolations have been used for generation of geographical images. We change the interpolation in order to apply to other images.
• A Method of Finding the Region for Extracting Moving Object with Fuzzy reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (77)  63  -68  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a region segmentation method for extracting moving objects. In the proposed algorithm, after segmentation with a simple region growing algorithm. the centroid of each region is computed. the centroids are used as the initial values of the k-mean clustering. These regions the existent method for extracting moving objects with a fuzzy reasoning. It is verified that the proposed method can effectively extract moving objects by some experiments.
• A Noto on Searching the Shortest Route through Several Designated Points

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (77)  69  -74  1997/12/11  [Not refereed][Not invited]
   

  This paper proposes a method of searching the shortest route through several designated points using a genetic algorithm. The proposed method not only searches the shortest route but also selects the shortest route through several designated points. Current methods, such as the Dijkstra's method, are the algorithms only for routing the shortest route. If the shortest route through several designated points is searched by these methods, its computational cost is O(N^2) where N is the number of the designated points. Therefore, the more the designated points increase, the more the routing demands computational cost. The proposed algorithm can obtain the shortest route through the points by an only search, and is more useful for this kind of routing than the other methods.
• Image Modeling and Parameter Identification for Image Restoration Using a Kalman Filter

  JO Takashi, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  80-  (11)  2912  -2919  1997/11/25  [Not refereed][Not invited]
   

  より高精度な画像復元を可能とするためにカルマンフィルタによる画像復元アルゴリズムを提案する. カルマンフィルタを用いて画像復元を行う場合, 原画像のモデリングは復元精度を左右する重要な要素であり, 画像の種類や観測画像の劣化の程度によって良好な復元を行うことができるモデルは変わってくる. また, カルマンフィルタを実現する際には, 原画像の相関や雑音の分散などのパラメータが必要とされる. 従ってこれらのパラメータが未知である場合には, 観測画像から推定しなければならない. 以上の問題を解決するために, 画素値を推定するために用いる画素数とそれらの位置を決定する方法および未知パラメータが存在する際にこれらを復元過程に組み込んで推定する方法を提案する.
• Statistical analysis of MFM images

  TAKEKUMA I, KUMAGAI M, YASUI M, OKUMURA Y, AKITA T, MAEDA M, SUEOKA K, HASEYAMA M, MUKASA K  日本応用磁気学会学術講演概要集 = Digest of ... annual conference on magnetics in Japan  21-  1  -1  1997/10/01  [Not refereed][Not invited]
  
• Inherent matrix identities on ARMA lattice filter realization algorithm and their application

  M Haseyama, H Kitajima  IEEE TRANSACTIONS ON SIGNAL PROCESSING  45-  (9)  2395  -2398  1997/09  [Not refereed][Not invited]
   

  This correspondence presents the matrix identities that are inherent in the solution of the normal equations for an ARMA lattice filter. This derivation also makes clear the relationship between the recursive least squares (RLS) method and the ARMA lattice filter realization algorithm. Further, as an application of the matrix identities, a new method for model identification with frequency weighting (MIFW) is presented.
• Inherent matrix identities on ARMA lattice filter realization algorithm and their application

  Miki Haseyama, Hideo Kitajima  IEEE Transactions on Signal Processing  45-  (9)  2395  -2398  1997/09  [Not refereed][Not invited]
   

  This correspondence presents the matrix identities that are inherent in the solution of the normal equations for an ARMA lattice filter. This derivation also makes clear the relationship between the recursive least squares (RLS) method and the ARMA lattice filter realization algorithm. Further, as an application of the matrix identities, a new method for model identification with frequency weighting (MIFW) is presented.
• Image restoration method with multiframe projection filter

  Manabu Kikuchi, Miki Haseyama, Hideo Kitajima  Systems and Computers in Japan  28-  (7)  65  -74  1997/06/30  [Not refereed][Not invited]
   

  In this paper, we propose a multiframe projection filter (MFPF) that restores images using several observed images rather than one observed image by enhancing a conventional projection filter. By superimposing several observed images during image restoration, the MFPF reduces random noise. Use of the MFPF enables high-quality image restoration when several observed images of a specified object are available. Since image selection is crucial to the process of superimposition, we also propose a selection equation. The MFPF differs from the multiframe Wiener filter (MFWF), which also uses several observed images, in that it does not require original image correlation numbers during image restoration. Because the MFPF does not require such numbers, the MFPF offers superior image restoration in cases where the original image cannot be elucidated from observed images. At the end of this paper, we present summary findings from an image restoration experiment carried out using the proposed filter. But first, we will confirm the MFPF's effectiveness in superimposing images and give an example of a situation where, given the absence of an original image, the MFPF provides better image resolution than the MFWF. Finally, combining a method for estimating background noise dispersion and an equation for selecting which images to use for superimposition results in an image restoration system capable of functioning when random noise dispersion is unknown. Following the above scheme will confirm the effectiveness of our approach. © 1997 Scripta Technica, Inc.
• A Note of Implementation of Normalized ARMA Lattice Filter with CORDIC

  SHIRAISHI Shinichi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  97-  (136)  17  -24  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method for implementation of a normalized ARMA lattice filter with CORDIC. An algorithm CALF is previously reported as a CORDIC processor implementation of an adaptive lattice filter. However, it is difficult to apply the CALF to a normalized ARMA lattice filter because of its problems about rotation angle computed by it and its convergence speed. Therefore, we present a method to implement CORDIC ARMA lattice section which resolves these problems. By using this specified CORDIC algorithm, the proposed lattice can be effectively implemented with VLSI. Moreover, the proposed lattice filter can estimate input, so that it can be applied to unknown-input models. Effectiveness of the proposed method is verified by experiments.
• Image Restoration for Images Contaminated with Pulse Noise

  ICHIOKA Mikako, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  97-  (136)  25  -30  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes with IFS (Iterated Function System) method an image restoration for images contaminated by pulse noise. One of the image restoration methods for images contaminated by pluse noise is the median filter, and another is linear interpolation which is suitable to restore images contaminated by collective pulse noise. However, these image restoration methods often smooth the complex structure part of image. Proposed IFS image restoration methods can keep the complexity of image.
• Extracting Moving Object with Fuzzy Reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  97-  (136)  31  -36  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of extracting a moving object with the fuzzy reasoning. After segmentation of the frame, this proposed method extracts a moving object by merging regions which is contained in an object. The rule to merge regions uses the fuzzy reasoning based on relations which are defined by region's move and location. Moreover, this paper introduces the recursive fuzzy reasoning to extract a moving object with hight accuracy.
• A Note on Deciding Various Solutions by Genetic Algorithm for Routing Applications

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  97-  (136)  37  -42  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of routing with Genetic Algorithm. The proposed method not only selects the shortest route but also decides several semi-shortest routes utilizing the feature of Genetic Algorithm. By using this method, another short route can be presented without re-routing if it is demanded. The existent methods cannot present any other short routes, therefore the proposed algorithm is useful for this kind of routing. Moreover, since the proposed genotype structure has a weighting factor which controls route length, the method can select suitable routes depending on geographical conditions.
• A Note of Implementation of Normalized ARMA Lattice Filter with CORDIC

  SHIRAISHI Shinichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  97-  (140)  17  -24  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method for implementation of a normalized ARMA lattice filter with CORDIC. An algorithm CALF is previously reported as a CORDIC processor implementation of an adaptive lattice filter. However, it is difficult to apply the CALF to a normalized ARMA lattice filter because of its problems about rotation angle computed by it and its convergence speed. Therefore, we present a method to implement CORDIC ARMA lattice section which resolves these problems. By using this specified CORDIC algorithm, the proposed lattice can be effectively implemented with VLSI. Moreover, the proposed lattice filter can estimate input, so that it can be applied to unknown-input models. Effectiveness of the proposed method is verified by experiments.
• Image Restoration for Images Contaminated with Pulse Noise

  ICHIOKA Mikako, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  97-  (140)  25  -30  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes with IFS (Iterated Function System) method an image restoration for images contaminated by pulse noise. One of the image restoration methods for images contaminated by pluse noise is the median filter, and another is linear interpolation which is suitable to restore images contaminated by collective pulse noise. However, these image restoration methods often smooth the complex structure part of image. Proposed IFS image restoration methods can keep the complexity of image.
• Extracting Moving Object with Fuzzy Reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  97-  (140)  31  -36  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of extracting a moving object with the fuzzy reasoning. After segmentation of the frame, this proposed method extracts a moving object by merging regions which is contained in an object. The rule to merge regions uses the fuzzy reasoning based on relations which are defined by region's move and location. Moreover, this paper introduces the recursive fuzzy reasoning to extract a moving object with hight accuracy.
• A Note on Deciding Various Solutions by Genetic Algorithm for Routing Applications

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. DSP  97-  (140)  37  -42  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of routing with Genetic Algorithm. The proposed method not only selects the shortest route but also decides several semi-shortest routes utilizing the feature of Genetic Algorithm. By using this method, another short route can be presented without re-routing if it is demanded. The existent methods cannot present any other short routes, therefore the proposed algorithm is useful for this kind of routing. Moreover, since the proposed genotype structure has a weighting factor which controls route length, the method can select suitable routes depending on geographical conditions.
• A Note of Implementation of Normalized ARMA Lattice Filter with CORDIC

  SHIRAISHI Shinichi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  97-  (138)  17  -24  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method for implementation of a normalized ARMA lattice filter with CORDIC. An algorithm CALF is previously reported as a CORDIC processor implementation of an adaptive lattice filter. However, it is difficult to apply the CALF to a normalized ARMA lattice filter because of its problems about rotation angle computed by it and its convergence speed. Therefore, we present a method to implement CORDIC ARMA lattice section which resolves these problems. By using this specified CORDIC algorithm, the proposed lattice can be effectively implemented with VLSI. Moreover, the proposed lattice filter can estimate input, so that it can be applied to unknown-input models. Effectiveness of the proposed method is verified by experiments.
• Image Restoration for Images Contaminated with Pulse Noise

  ICHIOKA Mikako, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  97-  (138)  25  -30  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes with IFS (Iterated Function System) method an image restoration for images contaminated by pulse noise. One of the image restoration methods for images contaminated by pluse noise is the median filter, and another is linear interpolation which is suitable to restore images contaminated by collective pulse noise. However, these image restoration methods often smooth the complex structure part of image. Proposed IFS image restoration methods can keep the complexity of image.
• Extracting Moving Object with Fuzzy Reasoning

  ARAI Tetsuya, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  97-  (138)  31  -36  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of extracting a moving object with the fuzzy reasoning. After segmentation of the frame, this proposed method extracts a moving object by merging regions which is contained in an object. The rule to merge regions uses the fuzzy reasoning based on relations which are defined by region's move and location. Moreover, this paper introduces the recursive fuzzy reasoning to extract a moving object with hight accuracy.
• A Note on Deciding Various Solutions by Genetic Algorithm for Routing Applications

  INAGAKI Jun, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  97-  (138)  37  -42  1997/06/26  [Not refereed][Not invited]
   

  This paper proposes a method of routing with Genetic Algorithm. The proposed method not only selects the shortest route but also decides several semi-shortest routes utilizing the feature of Genetic Algorithm. By using this method, another short route can be presented without re-routing if it is demanded. The existent methods cannot present any other short routes, therefore the proposed algorithm is useful for this kind of routing. Moreover, since the proposed genotype structure has a weighting factor which controls route length, the method can select suitable routes depending on geographical conditions.
• Motion Parameters for Sequence Image Coding Using a Genetic Algorithm

  ASAMIZU Satoshi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (123)  1  -6  1997/06/20  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for motion-parameter identification. The parameters identified by the conventional method the LBG algorithm have quantization errors which make the decoded images are contaminated. The proposed method decrease the effect of this quantization error by the genetic algorithm. This paper shows the designing method of the genetic algorithm for parameter identification. Some experimental results are included to show the effectiveness of the proposed method.
• Extension of The Collage Theorem

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  97-  (123)  7  -12  1997/06/20  [Not refereed][Not invited]
   

  Fractal image compression using the IFS is based on the collage theorem proposed by M. F. Barnsley. In this conventional method, the errors between the reconstructed image and original image may be greater than the errors between the collage and original image, because the collage theorem does not guarantee the former errors to be smaller than the latter errors. This paper proposes an extended collage theorem. An IFS algorithm based on this theorem determines the parameters after iterations of the contraction mappings. An image reconstructed according to the new theorem has higher quality than one based on the existing collage theorem. The reconstructed image can be gotten by fewer iterations than by the conventional IFS.
• 7)パターン識別を用いた動画像のカット点検出に関する一考察(〔情報ディスプレイ研究会 ヒューマンインフォメーション研究会 マルチメディア情報処理研究会 ネットワーク映像メディア研究会 画像情報システム研究会〕合同)

  槇 宏晃, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (6)  820  -820  1997/06/20  [Not refereed][Not invited]
  
• 1)量子化誤差も考慮したIFSパラメータの量子化法に関する考慮(放送方式研究会)

  本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  342  -343  1997/03/20  [Not refereed][Not invited]
  
• 2)動画像のフレーム間予測誤差信号のベクトル量子化法に関する一考察(放送方式研究会)

  金森 紀文, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  343  -343  1997/03/20  [Not refereed][Not invited]
  
• 3)GAとMRFモデルを用いた画像の領域分割(放送方式研究会)

  宮本 俊洋, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  343  -343  1997/03/20  [Not refereed][Not invited]
  
• 4)定ブロックサイズIFSにおける相似領域検索法に関する考察(放送方式研究会)

  伊藤 進策, 本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  343  -343  1997/03/20  [Not refereed][Not invited]
  
• 5)整数論的アダマール変換を用いた静止画像のロスレス符号化(放送方式研究会)

  星野 亮, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  343  -343  1997/03/20  [Not refereed][Not invited]
  
• 6)遺伝的アルゴリズムを用いたフラクタルブロック符号化に関する一考察(放送方式研究会)

  小林 亜令, 本多 博之, 長谷山 美紀, 北島 秀夫  映像情報メディア学会誌 : 映像情報メディア  51-  (3)  343  -343  1997/03/20  [Not refereed][Not invited]
  
• Analysis of Hadamard Transforms of Integer Signals

  Kitajima Hideo, Hoshino Makoto, Haseyama Miki, Shimono Tetsuo  Proceedings of the IEICE General Conference  1997-  (2)  49  -49  1997/03/06  [Not refereed][Not invited]
   

  This paper analyzes properties of Hadamard transforms of integer signals with intended applications in lossless image coding. It is shown that the vectors holding Hadamard transform coefficients of integer signals form a small subset of integer vectors if they are mapped onto an integer ring modulo the transform size.
• Active Noise Cancellation System with Virtual Reference Singnals.

  AKIHO Masaichi, HASEYAMA Miki, KITAJIMA Hideo  日本音響学会研究発表会講演論文集  1997-  (2)  819  -820  1997/03/01  [Not refereed][Not invited]
  
• A Note Of Cut Detection for Image Sequences Using Pattern Recognition

  MAKI Hiroaki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  96-  (508)  47  -52  1997/02/04  [Not refereed][Not invited]
   

  This paper proposes improved cut detection for image sequences using pattern recognition. A method reported before is based on image segmentation and classification of the regions. This method detects a cut, by the disappearance of classes. But pattern recognition used for the classification is influenced by interframe motion. To cope with this problem, this paper considers features such as motions and region sizes. Effectiveness of the proposed method is verified by experiments.
• A Note Of cut Detection for Image Sequences Using Pattern Recognition

  MAKI Hiroaki, HASEYAMA Miki, KITAJIMA Hideo  ITE Technical Report  21-  (7)  47  -52  1997/02/04  [Not refereed][Not invited]
   

  This paper proposes improved cut detection for image sequences using pattern recognition. A method reported before is based on image segmentation and classification of the regions. This method detects a cut by the disappearance of classes. But pattern recognition used for the classification is influenced by interframe motion. To cope with this problem, this paper considers features such as motions and region sizes. Effectiveness of the proposed method is verified by experiments.
• Removal of Additive White Noise using an Adaptive Wiener Filter.

  塚原正人, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Consideration on image interpolation considering fractal dimension.

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Utilization of color image for object extraction using fuzzy reasoning.

  新井哲也, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Advanced Motion Compensation Based on Motion-Table Lookup.

  浅水仁, 吉田仁, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Consideration on Kalman filter image restoration using regional division.

  熊谷政輝, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Utilization of audio data for high-precision cut point detection.

  薄田隆博, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Consideration on image restoration using IFS.

  一岡美佳子, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Improvement of evaluation formula in cut point detection method using regional division.

  槙宏晃, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Speeding up of IFS parameter determination using genetic algorithm.

  小林亜令, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Reversible encoding and quasi-reverse encoding of static image using the number theory Hadamard transformation.

  星野亮, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Determination method of multiple solution candidates in GA path planning.

  稲垣潤, 長谷山美紀, 北島秀夫, 利部直  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Consideration on implementation method of the CORDIC ARMA lattice filter.

  白石真一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Movement compensation prediction using three-dimensional model in moving image encoding.

  小池理一郎, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• A Complex ARX Model for Shape Classification.

  長谷山美紀, 秋保政一, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• Corresponding point search using genetic algorithm.

  武藤剛史, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1997-  1997
• ARMA direct form filter designing method to reconstruct FIR digital filters based upon genetic algorithm and simulated annealing

  Masaichi Akiho, Miki Haseyama, Hideo Kitajima  Journal of the Acoustical Society of Japan (E) (English translation of Nippon Onkyo Gakkaishi)  18-  (4)  153  -161  1997  [Not refereed][Not invited]
   

  In this paper, the method for reconstructing FIR digital filter into ARMA direct form digital filter based upon adaptive ARMA Four Line Lattice filter with the phase estimation is proposed. The phase estimation is performed by the combination of GA (Genetic Algorithm) and SA (Simulated Annealing) in addition to the spectrum estimation. In order to reduce the calculation cost that will be significant issue for the system implementation of digital filters into the products, it will be necessary to design ARMA digital filter. Introducing the phase estimation method, it contributes the possibility to provide optimum ARMA digital filters that is not only relatively lower coefficient sensitivity and better robustness, but also it will maintain all its desired characteristics that include the phase performance. Maintaining the phase characteristics is the unforgettable factor for the fields of the inverse filtering problem for any opportunities. We conduct the original cost function that will cooperate with GA and SA searching in order to realize the phase estimation. In this paper, experimental results are also discussed to confirm the performance of proposed method.
• A realization method of an ARMAX lattice filter

  Miki Haseyama, Hideo Kitajima  IEEE Transactions on Signal Processing  45-  (2)  471  -476  1997  [Not refereed][Not invited]
   

  This correspondence proposes a realization method of an ARMAX lattice filter for frequency-weighting ARMAX model identification. The proposed lattice filter uses an exponentially weighted sliding window for the same application as the extended least squares (ELS) achieves. Based on the proposed structure, the algorithm can perform the frequency-weighting model identification more easily than the ELS. Further, applied to the ARMAX model identification, the proposed algorithm requires fewer multiplications than the ELS does. © 1997 IEEE.
• Extension of the collage theorem

  H Honda, M Haseyama, H Kitajima, S Matsumoto  INTERNATIONAL CONFERENCE ON IMAGE PROCESSING - PROCEEDINGS, VOL II  2-  306  -309  1997  [Not refereed][Not invited]
   

  Fractal image compression using the IFS is based on the collage theorem proposed by M. F. Barnsley. In this conventional method, the errors between the reconstructed image and original image may be greater than the errors between the collage and original image because the collage theorem does not guarantee the former errors to be smaller than the latter errors. This paper proposes an extended collage theorem. An IFS algorithm based on this theorem determines the parameters after iterations of the contraction mappings. An image reconstructed according to the new theorem has higher quality than one based on the existing collage theorem. The reconstructed image can be gotten by fewer iterations than by the conventional IFS.
• Residue-based classification of Hadamard transforms of integer signals

  H Kitajima, M Hoshino, M Haseyama  DSP 97: 1997 13TH INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING PROCEEDINGS, VOLS 1 AND 2  1-  357  -359  1997  [Not refereed][Not invited]
   

  This paper discusses, the classification of Hadamard transform coefficients. It is shown that the coefficient vectors can be classified into a small number of residue classes for a modulus equal to an integer power of 2. It is suggested that the result can be used to define an efficient product code for lossless coding of digital images.
• The Selection of IFS Parameters with Effects of Their Quantization-Errors Being Considered

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  96-  (402)  1  -6  1996/12/13  [Not refereed][Not invited]
   

  In conventional fractal image coding using by IFS (Iterated Function Systems), the IFS Parameters are quantified and coded after their analog values are determined. But, quantization errors influence the reconstructed image. In this paper, we propose a quantization method to reduce the influence. In this method, the IFS parameters are selected from predetermined, quantized values.
• Vector Quantization of Interframe Prediction Residues for Motion Images

  KANAMORI Toshiya, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  96-  (402)  7  -12  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes, for image sequence coding, vector quantization of interframe prediction residues with thresholding. Interframe prediction residues are divided with thresholding into two groups, and we individually construct vector quantizers for them. Interframe prediction residues to be quantized are also divided with the thresholding, and they are quantized. In motion compensation, we use motion-table lookup. Various motion patterns are put into a motion-table. Each block of pixels is predicted through a motion that minimizes interframe prediction error. By combining the motion compensation using motion-table lookup and vector quantization of interframe prediction residues with thresholding, we resize efficient coding of image sequence.
• Picture Segmentation Using a Genetic Algorithm and an MRF Model

  MIYAMOTO Toshihiro, HASEYAMA Miki, KITAJIMA Hideo  CS  96-  (402)  13  -18  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes a method of picture segmentation using a Genetic algorithm (GA). One of the picture segmentation methods is derived with a Markov random field (MRF). It anodizes the texture of a picture and segments it into some regions. However, if the picture has gradation regions, the method can not segment it accurately. Further, a total number of the regions need to be known before execution. The proposed method estimates the parameters and can segment a picture of which the brightness is gradually changed. In this paper, the performance of the proposed method is verified by some simulations.
• A Note of a Method of Similar Region Search on Fixed Block Size IFS

  ITOH Shinsaku, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  96-  (402)  19  -24  1996/12/13  [Not refereed][Not invited]
   

  In this paper, we propose a improvement of lame quality for the fixed block size IFS with the bit rate kept low. In the case of a fixed block size, especially when the block size is large, we cannot represent, detains of images well while we can restrain the bit rate. In the proposed method, we improve image quality, by searching similar regions closely for a block whose variance is large. Search is rough for a block whose variance is small. The adaptive process is based on the fact that MSE of each block of the reconstructed image and the variance of the corresponding block of an original image is closely correlated.
• Lossless Image Coding Using a Number-Theoretic Hadamard Transform

  HOSHINO Makoto, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  96-  (402)  25  -30  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes losses image coding using a number-theoretic Hadamard transform. The number-theoretic Hadamard transform is derived from the Hadamard transform computed by modulo arithmetic. The transform effectively decreases the bit rate because all the coefficients are integer less than the modulus. Losses image coding is performed with the transform. Statistics of the least significant bits and gray-scale remapping are used to decrease the bit rate. This paper also proposes quasilossless coding by revising image.
• A Note of Fractal Block Coding Using a Genetic Algorithm

  KOBAYASHI Arei, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Communication systems  96-  (402)  31  -36  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for scaling-parameter identification. The conventional method quantize the parameters are identified by using the RMS metric and effect of quantization error increase when image is reconstructed. The genetic algorithm is constructed to cope with the problem. This paper shows the setting method of the genetic algorithm for parameter identification. Results of some experiments are included to show the effectiveness of the proposed method.
• The Selection of IFS Parameters with Effects of Their Quantization-Errors Being Considered

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  1  -6  1996/12/13  [Not refereed][Not invited]
   

  In conventional fractal image coding using by IFS (Iterated Function Systems), the IFS parameters are quantized and coded after their analog values are determined. But, quantization errors influence the reconstructed image. In this paper, we propose a quantization method to reduce the influence. In this method, the IFS paramters are selected from predetermined, quantized values.
• Vector Quantization of Interframe Prediction Residues for Motion Images

  KANAMORI Toshiya, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  7  -12  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes, for image sequence coding, vector quantization of interframe prediction residues with thresholding. Interframe prediction residues are divided with thresholding into two groups, and we individually construct vector quantizers for them. Interframe prediction residues to be quantized are also divided with the thresholding, and they are quantized. In motion compensation, we use motion-table lookup. Various motion patterns are put into a motion-table. Each block of pixels is predicted through a motion that minimizes interframe prediction error. By combining the motion compensation using motion-table lookup and vector quantization of interframe prediction residues with thresholding, we realize efficient coding of image sequence.
• Picture Segmentation Using a Genetic Algorithm and an MRF Model

  MIYAMOTO Toshihiro, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  13  -18  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes a method of picture segmentation using a Genetic algorithm (GA). One of the picture segmentation methods is derived with a Markov random field (MRF). It analyzes the texture of a picture and segments it into some regions. However, if the picture has gradation regions, the method can not segment it accurately. Further, a total number of the regions need to be known before execution. The proposed method estimates the parameters and can segment a picture of which the brightness is gradually changed. In this paper, the performance of the proposed method is verified by some simulations.
• A Note of a Method of Similar Region Search on Fixed Block Size IFS

  ITOH Shinsaku, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  19  -24  1996/12/13  [Not refereed][Not invited]
   

  In this paper, we propose a improvement of image quality for the fixed block size IFS with the bit rate kept low. In the case of a fixed block size, especially when the block size is large, we cannot represent details of images well while we can restrain the bit rate. In the proposed method, we improve image quality by searching similar regions closely for a block whose variance is large. Search is rough for a block whose variance is small. The adaptive process is based on the fact that MSE of each block of the reconstructed image and the variance of the corresponding block of an original image is closely correlated.
• Lossless Image Coding Using a Number-Theoretic Hadamard Transform

  HOSHINO Makoto, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  25  -30  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes lossless image coding using a number-theoretic Hadamard transform. The number-theoretic Hadamard transform is derived from the Hadamard transform computed by modulo arithmetic. The transform effectively decreases the bit rate because all the coefficients are integer less than the modulus. Lossless image coding is performed with the transform. Statistics of the least significant bits and gray-scale remapping are used to decrease the bit rate. This paper also proposes quasilossless coding by revising image.
• A Note of Fractal Block Coding Using a Genetic Algorithm

  KOBAYASHI Arei, HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (70)  31  -36  1996/12/13  [Not refereed][Not invited]
   

  This paper proposes the use of a genetic algorithm for scaling-parameter identification. The conventional method quantize the parameters are identified by using the RMS metric and effect of quantization error increase when image is reconstructed. The genetic algorithm is constructed to cope with the problem. This paper shows the setting method of the genetic algorithm for parameter identification. Results of some experiments are included to show the effectiveness of the proposed method.
• A method quantizing filter coefficients with genetic algorithm and simulated annealing

  Miki Haseyama, Yoshihiro Aketa, Hideo Kitajima  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E79-A-  1130  -1134  1996/12/01  [Not refereed][Not invited]
   

  In this paper, a quantization method which can keep the phase and gain characteristics of a reference filter is proposed. The proposed method uses a genetic algorithm and a simulated annealing algorithm. The objective function used in this method is described with two kinds of weighting functions for identifying the phase and gain characteristics respectively. Therefore, the quantization accuracy on the gain characteristic is independent of the accuracy on the phase characteristic. Further, the proposed algorithm can be applied to any types of filters, because the chromosome expresses only their coefficients values. The efficiency of the proposed algorithm is verified by some experiments.
• A map matching method with the innovation of the Kalman filtering

  Takashi Jo, Miki Haseyama, Hideo Kitajima  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  E79-A-  1853  -1855  1996/12/01  [Not refereed][Not invited]
   

  This letter proposes a map-matching method for automotive navigation systems. The proposed method utilizes the innovation of the Kalman filter algorithm and can achieve more accurate positioning than the correlation method which is generally used for the navigation systems. In this letter, the performance of the proposed algorithm is verified by some simulations.
• An Image Restoration Method with Multiframe Projection Filter

  KIKUCHI Manabu, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  79-  (9)  1516  -1524  1996/09/25  [Not refereed][Not invited]
   

  本論文では,射影フィルタを拡張し,1枚の観測画像だけではなく,相互に関連のある複数枚の観測画像を用いて画像復元を行うマルチフレーム射影フィルタ(MFPF)を提案する.MFPFは,観測画像の重ね合せによって,雑音を低減する効果を利用した画像復元フィルタである.本フィルタを用いれば,特定の物体を観測する場合に撮影された複数枚の観測画像により,高精度な画像復元が可能となる.更に,本文では,効果的に重ね合わせる観測画像を選択するために,判定式を提案している.MFPFは,複数枚の観測画像を利用するウィーナーフィルタ(MFWF)と異なり,画像復元を行う際に原画像の相関関数を必要としない.そのため,原画像の相関関数が未知であり,観測画像から原画像の相関関数を高精度に推定できない場合は,MFPFの復元精度がMFWFよりも良好となる.本文の最後に,提案フィルタによる画像復元実験の概要を説明する.まず,観測画像の重ね合せの効果を確認し,次に原画像の相関関数が未知の場合に,MFPFがMFWFよりも復元精度が良好となる復元例を示す.最後に,雑音の分散が未知の場合の画像復元システム例を,既存の雑音の分散推定方法と重ね合わせる観測画像の選択を行う判定式を用いて実現する.以上から,本手法の有効性が確認される.
• Another Look at the Collage Theorem

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo, MATSUMOTO Syuichi  Proceedings of the Society Conference of IEICE  1996-  173  -173  1996/09/18  [Not refereed][Not invited]
   

  反復関数システム(Iterated Function System:IFS)においては無限回の反復写像の後に得られるattractorを用いてパラメータ決定を行うのが理想的ではあるが,計算時間が膨大になるためにCollage定理に基づき1回の縮小写像により得られるCollageを用いパラメータ決定を行う場合がほとんどである.しかし,Collage定理は得られる再生画像が最良のものであることを保証していない.したがって,他により良い再生画像が得られるようなパラメータが存在していると考えられる.本文では,このようなCollage定理の問題点について述べ,さらにCollage定理の拡張を行い,IFSにおける影響について調べる.
• Image representation using adaptive KL basis

  TSUKAHARA Masato, HASEYAMA Miki, KITAJIMA Hideo  Proceedings of the Society Conference of IEICE  1996-  222  -222  1996/09/18  [Not refereed][Not invited]
   

  KL変換は,MSEを最小にするという観点などから最も効率の良い変換であるが,その基底を求める計算の複雑さなどから,実用性に欠けると言われている.しかし,画像を小ブロックに分割し,その局所的性質に応じた基底を用意できることから変換効率などの向上が期待できる.なお,DCTは相関が非常に高いところでKLTと同等の性能を有するが,固定基底であり,ブロックの局所的性質に応じた基底とは言えない.そこで,本文ではブロックの性質に応じた基底を用いた場合の有効性を,DCTと比較して調べる。
• Ficture Segmentation Using a Genetic Algorithm Based on an MRF Texture Model

  MIYAMOTO Toshihiro, HASEYAMA Miki, KITAJIMA Hideo  Proceedings of the Society Conference of IEICE  1996-  381  -381  1996/09/18  [Not refereed][Not invited]
   

  画像を領域に分割する方法が、これまでいくつか提案されている。領域分割は間値処理によって領域を抽出している。間値を各部位ごとに求め、さらに滑らかに画素値の変化する領域にも対応する遺伝的アルゴリズム(Genetic Algorithm:以下GA)を用いた画像の領域分割手法が提案されている。しかしこれらの手法はテクスチュアを考慮に入れていないため、徴小領域が多く発生していた。本文では、テクスチュアに対応するために2次多項式及びマルコフ確立場(Markov Random Field;以下MRF)モデルによるテクスチュア表現に基づく遺伝的アルゴリズムを用いた画像の領城分割法を提案する。従来のMRFモデルを用いた領域分割法においては領域内で滑らかに画素値の変化する領域には対応できず、また領域数が既知である事が要求されるが、本手法はGAと組み合わせる事によりその問題を解決している。
• Image representation through gray-scale iterated function systems

  H Honda, M Haseyama, H Kitajima  SYSTEMS AND COMPUTERS IN JAPAN  27-  (9)  55  -62  1996/08  [Not refereed][Not invited]
   

  Image representation using the random iterated function system (Random IFS) has a disadvantage in that it requires many iterations. This paper proposes image representation through a gray-scale iterated function system. This method requires fewer iterations and it earn obtain better images than the random IFS. The saving in iterations is realized first by the use of a deterministic method, whereas in the random IFS, each mapping is defined as a probablistic event, and it needs many iterations to satisfy the law of large numbers. In the proposed method, the reconstructed image converges faster than in the conventional method, and it is shown to yield better reconstructed images. Second, statistical characteristics of the IFS parameters are investigated to be used for further accelerating the IFS parameter search. The characteristics are shown to be stable among different images. Simulation results are included to demonstrate the effectiveness of the accelerated algorithm.
• A method quantizing filter coefficients with genetic algorithm and simulated annealing

  M Haseyama, Y Aketa, H Kitajima  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E79A-  (8)  1130  -1134  1996/08  [Not refereed][Not invited]
   

  In this paper, a quantization method which can keep the phase and gain characteristics of a reference filter is proposed. The proposed method uses a genetic algorithm and a simulated annealing algorithm. The objective function used in this method is described with two kinds of weighting functions for identifying the phase and gain characteristics respectively. Therefore, the quantization accuracy on the gain characteristic is independent of the accuracy on the phase characteristic. Further, the proposed algorithm can be applied to any types of filters, because the chromosome expresses only their coefficients values. The efficiency of the proposed algorithm is verified by some experiments.
• A method for quantizing coefficients of a filter with genetic algorithm

  Y Aketa, M Haseyama, H Kitajima, N Nagai  ELECTRONICS AND COMMUNICATIONS IN JAPAN PART III-FUNDAMENTAL ELECTRONIC SCIENCE  79-  (4)  1  -10  1996/04  [Not refereed][Not invited]
   

  This paper proposes a new method to quantize coefficients of filters where frequency-characteristic deterioration is suppressed. When a digital filter is realized, quantization errors cannot be avoided because operation circuits and registers have finite word lengths. Generally, filter coefficients are quantized by popular methods, such as rounding and omitting, etc. The pro; posed method obtains the quantized coefficients by search. The Genetic Algorithm is used for the search because it needs only objective function information. Further, by using the proposed method, the filter coefficients can be quantized with the frequency weight. Therefore, deterioration of the coefficients quantized by the proposed method is suppressed in the specified frequency region. The effectiveness of this method is shown by some simulations that quantize the four-line lattice and the direct form filter coefficients.
• A Method of Map-Matching Using the Innovation in Kalman Filter

  JO Takashi, HASEYAMA Miki, KITAJIMA Hideo, TAKASEKI Shigeru  Proceedings of the IEICE General Conference  1996-  163  -163  1996/03/11  [Not refereed][Not invited]
   

  自動車用ナビゲーションシステムにおいて、車両の現在位置を正確に検出することは必要不可欠である。現在、自立航法、および衛星航法を用いて車両位置の検出が行われているが、これらの方法を用いることにより得られた車両位置には誤差が含まれているために、車両位置の表示が道路から外れるという問題が生じる。そこで、車両位置を道路上に修正するためにマップマッチング処理が必要とされる。マップマッチング法としては、走行軌跡と候補となる道路との相関を用いるパターンマッチング法などが提案されている。しかし、パターンマッチング法では相関が高くなるような範囲が広いために真の軌跡を特定することは困難である。そこで本文では、より高精度に走行軌跡を求めるためにカルマンフィルタのイノベーションを用いたマップマッチング法を提案する。
• Model-Based Fully Automatic Extraction of the Lip Contour from TV Face Images

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  95-  (504)  75  -80  1996/02/02  [Not refereed][Not invited]
   

  Automatic lipreading requires an automatic lip contour extraction tool. In this paper, based on an active contour model, a Lip contour extraction method is presented which can automatically extract the lip contour from TV face images. It needs neither a special artificial condition i.e., smearing the lips with a black lipstick, nor initializing the start points by human assistance. It accepts one person's TV shot images as its input, and extracts the person's lip contour fully automatically. Experimental results show the effectiveness of the proposed method.
• Model-Based Fully Automatic Extraction of the Lip Contour from TV Face Images

  ZHONG Wen, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  20-  (5)  75  -80  1996/02/02  [Not refereed][Not invited]
   

  Automatic lipreading requires an automatic lip contour extraction tool. In this paper. based on an active contour model, a lip contour extraction method is presented, which can automatically extract the lip contour from TV face images. It needs neither a special artificial condition i.e., smearing the lips with a black lipstick, nor initializing the start points by human assistance. It accepts one person's TV shot images as its input, and extracts the person's lip contour fully automatically. Experimental results show the effectiveness of the proposed method.
• Image Representation through Gray-Scale Iterated Function Systems

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  The Transactions of the Institute of Electronics,Information and Communication Engineers.  79-  (1)  18  -25  1996/01/25  [Not refereed][Not invited]
   

  縮小写像による画像表現法のうち, ランダム反復関数システム (Random IFS) による方法には非常に多くの反復回数を必要とするという欠点がある. そこで本論文では, 局所方式の新たな画像表現の手法として, 濃淡画像対応反復関数システムによる画像表現法を提案する. 本手法は従来法に比べ少ない反復回数でIFSパラメータが決定可能であり, かつ従来手法よりも誤差の少ない再生画像が得られる. 本提案手法が従来手法より高速かつ良好なパラメータ決定および再生が可能であるのは以下の理由による. (1)ランダム反復関数システムでは, 写像が確率事象として定義され, 大数の法則が成立するように多数回の試行を必要とする. これに対し本手法では決定論的に写像を行うので, ランダム反復関数システムに比べて反復写像の収束が速くなる. また, 写像方法の違いにより精度が向上する. (2)本手法によって抽出される画像パラメータの画像中における統計的性質を明らかにする. この統計的性質は異なる画像間でも安定した形となることが実験的に明らかにされる. この性質を利用して, 画像のIFSパラメータ決定の高速化を図る. 最後に実験で, その高速化の検証を行っている.
• Reversible encoding of static image using Hadamard transformation similar to theory of numbers.

  星野亮, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Consideration on speeding-up of fractal block coding using genetic algorithm.

  小林亜令, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Consideration on picture quality improvement in constant block size IFS.

  伊藤進策, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Expression in the region by MRF model and regional division of image using genetic algorithm.

  宮本俊洋, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Image restoration method using IFS.

  一岡美佳子, 本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Quantization of the IFS Parameters.

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• A method for reconstruction of filter forms with the GA and SA.

  浅水仁, 秋保政一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Consideration on modeling of image for high-precision image restoration using Kalman filter.

  城崇, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Effect of threshold value processing in quantization of prediction error signal between frames of moving image.

  金森紀文, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Search of path using genetic algorithm.

  稲垣潤, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Cut point detection method for moving image using region division.

  槙宏晃, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Consideration on object extraction noticing movement.

  新井哲也, 長谷山美紀, 北島秀夫, 下野哲雄  電気関係学会北海道支部連合大会講演論文集  1996-  1996
• Cut detection for image sequences through image segmentation

  Akira Mizuno, Miki Haseyama, Hideo Kitajima  Terebijon Gakkaishi  50-  (1)  79  -86  1996  [Not refereed][Not invited]
   

  This paper proposes a new cut-detection algorithm for image sequences. The new method is based on image segmentation and classification of the regions resulting from the segmentation. Each region is assigned to a class that terminates at a shot boundary. This technique is shown to be effective for gradual shot transitions, where conventional cut-detection algorithms are known to fail. It is also shown that the threshold with which the outputs of the cut detector are compared can be selected easily as the outputs are close to unity at cuts, while low elsewhere. The effectiveness of the proposed algorithm is demonstrated by experiments.
• Introduction of Number-Theoretic Transforms for Lossless Image Coding

  Kitajima Hideo, Haseyama Miki, Shimono Tetsuo  Proceedings of the IEICE General Conference  29  -29  1996  [Not refereed][Not invited]
   

  Orthogonal transforms such as the KLT, DCT, etc. are widely used in image data compression. The major virture of such transforms is energy compaction in the transform domain. Recently promising results have been reported in the use of orthogonal transforms in lossless coding, where special attention has been paid to possible arithmetic inaccuracies arising from finite register length. A more fundamental solution to the problem is the use of a finite integer ring. Orthogonal transforms over finite integer rings were investigated in efforts to efficiently compute convolution. In spite of their elegance and efficiency, they are in history due to their dynamic-range restrictions and the advent of VLSI-implemented floating-point arithmetic operations. In this paper we consider the use of transforms over rings of integers for possible image data compression, recalling them from the archive. The purpose of this paper is not to show any conclusive evidence but possibilities.
• Error in Reconstructed Images Caused by Quantization of IFS Parameters

  Honda Hiroyuki, Haseyama Miki, Kitajima Hideo, Matsumoto Syuichi  Proceedings of the IEICE General Conference  74  -74  1996  [Not refereed][Not invited]
   

  IFS(反復関数システム)を用いたフラクタル画像符号化に関する研究が近年盛んに行なわれている。この際に問題となるのは決定されるパラメータの量子化誤差が実際に画像を再生する際に大きな影響を及ぼすことである。このため、量子化の際には最適なビット配分を用いることが重要となってくる。そこで、本文では、我々が以前提案した濃淡画像対応IFSを用い、このパラメータの量子化誤差を考慮する代わりに、パラメータに乱数によりノイズを付加し、その再生画像への影響を実験により調べた結果を報告する。
• A Method for Quantize Coefficients of a Filter with Genetic Algorithm

  AKETA Yshihiro, HASEYAMA Miki, KITAJIMA Hideo, NAGAI Nobuo  The Transactions of the Institute of Electronics,Information and Communication Engineers. A  78-  (12)  1542  -1549  1995/12/25  [Not refereed][Not invited]
   

  本論文では周波数特性の劣化を抑えたフィルタ係数の量子化手法を提案する.ディジタルフィルタを実際に実現する場合,演算回路やレジスタの語長が有限なため,量子化誤差は避けることができない.従来,各係数は切捨てや丸めなどの一様な手法で量子化されている.提案手法は,量子化後表現可能な値を範囲として探索を行い,周波数特性の劣化を抑えた有限語長のフィルタを得る.探索アルゴリズムとして,広域探索に有効とされる遺伝的アルゴリズムを用いている.なぜならば,本手法では遺伝的アルゴリズムが評価関数の値以外に,評価関数の微分のような派生的な情報を必要としない性質をもつためである.更に,評価関数に周波数重みを考慮し周波数特性の劣化を抑えたい領域を指定することを可能とした.最後に提案手法を用いてARMA4線ラティスフィルタおよび直接形ARMAフィルタを量子化し,その有効性を確認する.
• 6)可変ブロックサイズ動き補償アルゴリズムにおけるパラメータ設定法についての考察([情報ディスプレイ研究会 視聴覚技術研究会 画像処理・コンピュータビジョン研究会 画像通信システム研究会 画像応用研究会]合同)

  吉田 仁, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (7)  954  -954  1995/07/20  [Not refereed][Not invited]
  
• A Study of Quantize Filter Coefficients with a Genetic Algorithm and a Simulated Annealing Algorithm

  AKETA Yoshihiro, HASEYAMA Miki, KITAJIMA Hideo, NAGAI Nobuo  IEICE technical report. Circuits and systems  95-  (105)  97  -102  1995/06/22  [Not refereed][Not invited]
   

  When a digital filter is used for synthesizing a signal, the change of the frequency characteristics caused by coefficient-quantization errors is a serious problem. A method for a quantization of ARMA lattice filter coefficients has been proposed. The method can suppress the deterioration of the frequency characteristic. In this paper, a genetic algorithm in combination with a simulated annealing algorithm is used for a improvement of the search ability. The improvement of a search ability is confirmed by the simulation of filters coefficients quantization.
• A Note of Picture Segmentation Using a Genetic Algorithm

  MIYAMOTO Toshihiro, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  95-  (104)  45  -50  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes a method of picture segmentation using a Genetic algorithm. The methods of the picture segmentation which have been proposed apply the same thresholds to a whole picture, and make fragmentation and overmerging of regions. To cope with these problems, the proposed method optimizes the thresholds in each part by using the Genetic algorithm. This paper shows the setting method of several parameters of the Genetic algorithm for the picture segmentation. Some simulations are carried out and their results shows the effectiveness of the proposed method.
• Novel Use of the Innovation in Kalman Filter Based Image Restoration

  JO Takashi, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  95-  (104)  59  -64  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes a method of estimating the variance of the observation noise, which is genarally unknown when the Kalman filter is apllied to image restoration. The proposed method divides the degraded image into blocks to cope with a practical situation where the spatial distribution of the noise is not uniform.
• The Proposal with Image Restoration by Multiframe Projection Filter

  KIKUCHI Manabu, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Circuits and systems  95-  (104)  65  -70  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes multiframe projection filter (MFPF) that uses some images having large correlation with each frames on estimating the original images from the degraded images. Projection filter is better than Wiener filter when the auto-correlation function is unknown, because projection filter is capable of estimating the original image in relationless to the auto-correlation function. We derivate multiframe projection filter and prove that multiframe projection filter is more effective by simulating image restoration under the various conditions.
• A Note of Picture Segmentation Using a Genetic Algorithm

  MIYAMOTO Toshihiro, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  95-  (107)  45  -50  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes a method of picture segmentation using a Genetic algorithm. The methods of the picture segmentation which have been proposed apply the same thresholds to a whole picture, and make fragmentation and overmerging of regions. To cope with these problems, the proposed method optimizes the thresholds in each part by using the Genetic algorithm. This paper shows the setting method of several parameters of the Genetic algorithm for the picture segmentation. Some simulations are carried out and their results shows the effectiveness of the proposed method.
• Novel Use of the Innovation in Kalman Filter Based Image Restoration

  JO Takashi, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  95-  (107)  59  -64  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes a method of estimating the variance of the observation noise, which is genarally unknown when the Kalman filter is apllied to image restoration. The proposed method divides the degraded image into blocks to cope with a practical situation where the spatial distribution of the noise is not uniform.
• The Proposal with Image Restoration by Multiframe Projection Filter

  KIKUCHI Manabu, HASEYAMA Miki, KITAJIMA Hideo  Technical report of IEICE. VLD  95-  (107)  65  -70  1995/06/21  [Not refereed][Not invited]
   

  This paper proposes multiframe projection filter (MFPF) that uses some images having large correlation with each frames on estimating the original images from the degraded images. Projection filter is better than Wiener filter when the auto-correlation function is unknown, because projection filter is capable of estimating the original image in relationless to the auto-correlation function. We derivate multiframe projection filter and prove that multiframe projection filter is more effective by simulating image restoration under the various conditions.
• The Use of Contraction Mappings with Triangular Ranges in Fractal Image Representation

  HONDA Hiroyuki, HASEYAMA Miki, KITAJIMA Hideo  IEICE technical report. Image engineering  95-  (52)  1  -6  1995/05/19  [Not refereed][Not invited]
   

  In fractal image representation, rectangular mapping ranges are conventionally used. But, with the mapping, block distortion error occures in reconstructed images. In this paper, the use of triangular ranges is investigated to solve the problem. Images reconstructed by triangular-range mappings and rectangular-range mappings and their combination are compared. And it is by using gray-scale mapping based on least squares method.
• Feature-Point-Based Image Representation

  KITAJIMA Hideki, HASEYAMA Miki, SHIMONO Tetsuo  Proceedings of the IEICE General Conference  1995-  (2)  23  -23  1995/03/27  

  This paper proposes feature-point-based image representation.Smooth windows centered on arbitrarily specified feature points are used to extract image contents associated with them.The image contents can be combined with background information to reconstruct the image.
• Threshold Selection for Variable Block Size Motion Compensation

  YOSHIDA Hitoshi, HASEYAMA Miki, KITAJIMA Hideo  Proceedings of the IEICE General Conference  1995-  (2)  42  -42  1995/03/27  [Not refereed][Not invited]
   

  文献で提案されている動き補償アルゴリズム(以下VBSMPMCアルゴリズムと呼ぶ)を用いる場合、動き補償に与えるビット数と予測誤差の符号化に与えるビット数の設定によって圧縮効率が変化する。従って、実際にユーザーがこのアルゴリズムを用いる場合、ビット部分を適切に決定することが必要となる。そこで本文では、このアルゴリズムの閾値選択の実験結果を述べる。
• Parameter Optimization for Image-Sequence Coding Using Variable-Block-Size Motion Compensation

  YOSHIDA Hitoshi, HASEYAMA Miki, KITAJIMA Hideo  ITEJ Technical Report  19-  (6)  31  -36  1995/02/02  [Not refereed][Not invited]
   

  This paper reviews earlier work concerning variable-block-size motion compensation and proposes a plan for optimum bit allocation between motion-related information and motion-compensated interframe prediction residues. Detection of the parameter which detect optimum bit allocation is confirmed by video coding simulation.
• 2)連続画像の表現におけるKL基底の問題点(画像通信システム研究会)

  塚原 正人, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (1)  112  -112  1995/01/20  [Not refereed][Not invited]
  
• 8)テレビジョン画像のシーン境界の検出法に関する一考案(画像通信システム研究会)

  水野 哲, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (1)  113  -113  1995/01/20  [Not refereed][Not invited]
  
• 10)領域分割符号化に関する検討 : 領域作成法と領域内信号の符号化について(画像通信システム研究会)

  泉 洋平, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (1)  113  -113  1995/01/20  [Not refereed][Not invited]
  
• 14)可変ブロックサイズ動き補償画像符号化におけるビット配分に関する考察(画像通信システム研究会)

  吉田 仁, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (1)  113  -114  1995/01/20  [Not refereed][Not invited]
  
• 4)画像生成における縮小写像の実験的検討(画像通信システム研究会)

  本多 博之, 長谷山 美紀, 北島 秀夫  テレビジョン学会誌  49-  (1)  114  -114  1995/01/20  [Not refereed][Not invited]
  
• Regional division of image using genetic algorithm based on regional expression by quadratic approximation.

  宮本俊洋, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Consideration on higher precision of image restoration by multi- frame projection filter.

  菊池学, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Quantization of Filter Coefficients with GA Considering a Phase Characteristic.

  明田吉浩, 長谷山美紀, 北島秀夫, 永井信夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Consideration on high precision image restoration using Kalman filter.

  城崇, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Consideration on evaluation of movement compensation in moving image encoding.

  吉田仁, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Analysis of time varying signal using fuzzy clustering.

  江村真史, 長谷山美紀, 北島秀夫, 永井信夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• Statistical Properties of Scaling Parameters in Fractal Block Coding of Images.

  伊藤進策, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• IFS based on generalized collage theorem.

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1995-  1995
• A Method to Quantize ARMA Lattice Filter Coefficients with a Genetic Algorithm.

  明田吉浩, 長谷山美紀, 北島秀夫, 永井信夫  電子情報通信学会技術研究報告  94-  (522(SP94 91-97))  1995
• A lattice filter for ARMAX model identification.

  長谷山美紀, 北島秀夫  電子情報通信学会大会講演論文集  1995-  (Sogo Pt 1)  1995
• A METHOD FOR REALIZATION OF AN ARMAX LATTICE FILTER

  M HASEYAMA, Y AKETA, H KITAJIMA  1995 INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING - CONFERENCE PROCEEDINGS, VOLS 1-5  3-  1705  -1708  1995  [Not refereed][Not invited]
   

  This paper proposes a method for realization of an ARMAX lattice filter. ARMAX (Autoregressive Moving Average model with Exogenous Variable) model identification is significant because the ARMAX model is a standard tool in the control field, and it can be performed by the proposed algorithm. One of the recursive least-square methods for the ARMAX model identification is the ELS (Extended Least Squares). Applied to the ARMAX model identification, the ELS uses o(N2) multiplications, where N = AR order + MA order + X order. When the proposed realization method of the ARMAX lattice filter is used, o(M) multiplications are needed for the ARMAX model identification, where M = max{AR order, MA order, X order}.
• Realization method of an ARMAX lattice filter

  Miki Haseyama, Tohru Hirohku, Hideo Kitajima  Proceedings - IEEE International Symposium on Circuits and Systems  1-  365  -368  1995/01/01  [Not refereed][Not invited]
   

  In this paper, a method for realization of an ARMAX lattice filter is proposed. Since the ARMAX model is a standard tool in control, it is significant to solve ARMAX model identification problems and the proposed algorithm can solve it. One of the recursive least-square methods for the ARMAX model identification is the ELS (Extended Least Squares). If the ELS is applied to the ARMAX model identification, it uses o(N2) multiplications, where N Δ/= AR order + MA order + X order. By using the proposed realization method, o(N) multiplications are needed for the ARMAX model identification.
• The Use of the Complex Number Theoretic Transform for Fast SAR Image Reconstruction

  MATSUYAMA Yoshiyuki, HASEYAMA Miki, KITAJIMA Hideo, SHIRAKAWA Tomoaki, OGAWA Yoshihiko  The Transactions of the Institute of Electronics,Information and Communication Engineers.  77-  (10)  2054  -2062  1994/10/25  [Not refereed][Not invited]
   

  本論文では,SAR(合成開口レーダ)の画像再生における計算量を軽減するために,数論変換の適用を提案する.SAR画像再生処理において重要な部分を占める相関計算に対して,量子化器が最適化されている場合,画像再生に必要な量子化ビット数は3ビットまで下げられることを示す.また,overlap-add法を導入しレーダ信号の長い系列を分割することにより,複素数論変換のダイナミックレンジに関する問題を克服することが可能であることを示す.メルセンヌ素数q=2^<13>-1とq=2^<17>-1を法とする変換が,それぞれ3ビット,4ビットに量子化して長さをN=4096に分割したレーダ信号の系列に対して適合していることを示す.
• Preliminary Research of Contraction Mapping in Image Representation

  Honda Hiroyuki, Haseyama Miki, Kitajima Hideo  ITEJ Technical Report  18-  (49)  21  -26  1994/09/22  [Not refereed][Not invited]
   

  There are two grayscale image coding methods by fractal. One, global scheme. The other, local scheme. Now, the main current method of local scheme is random iteration system (Random IFS). But, this method has a weak point that it needs iteration time. In this paper, a new image coding method based on collage theorem and gray-scale iterated function system, is proposed. The statistical properties between reconstructed image and origainal image are investigated. And the efficacy of this method is shown.
• Problems with the Karhunen-Loeve basis used in continuous-image expression

  Tsukahara Masato, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. OFC  94-  (245)  9  -14  1994/09/21  [Not refereed][Not invited]
   

  The Karhunen-Lodve expansion is one of the methods for expressing continuous-image as discrete value.In this study,we investigate an influence of using truncated coefficients of KL expansion by one dimensional signal.We explain experimentally and statistically that errors,caused by truncated coefficients,tend to be large in the neighborhood of the both ends of signal.
• Consideration to study on the cut-detection methods for image sequences

  Mizuno Akira, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. OFC  94-  (245)  49  -54  1994/09/21  [Not refereed][Not invited]
   

  This paper proposes a new algorithm which compares contents in the frames.This algorithm divides original images into the regions, and tracks separated regions.
• Region-Growing Based Image Coding -Region-Growing and Texture- Information Coding-

  Izumi Yohei, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. OFC  94-  (245)  61  -66  1994/09/21  [Not refereed][Not invited]
   

  This paper shows that region sizes have influence for the human visual system to recognize subjects in the image,and proposes an algorithm that makes use of the region sizes to get the satisfying segmented image for the human visual system.Simulation is carried out by approximating texture information using a two-dimensional polynomial function. Texture-information coding is also investigated with an easy way of expression.
• Problems with The Karhunen-Loeve Basis used in Continuous-Image Expression

  Tsukahara Masato, Haseyama Miki, Kitajima Hideo  ITEJ Technical Report  18-  (48)  9  -14  1994/09/21  [Not refereed][Not invited]
   

  The Karhunen-Loeve expansion is one of the methods for expressing continuous-image as discrete value. In this study, we investigate an influence of using truncated coefficients of kL expansion by one dimensional signal. We explain experimentally and statistically that errors, caused by truncated coefficients, tend to be large in the neighborhood of the both ends of signal.
• Consideration to Study on the Cut-Detection Methods for Image Sequences

  Mizuno Akira, Haseyama Miki, Kitajima Hideo  ITEJ Technical Report  18-  (48)  49  -54  1994/09/21  [Not refereed][Not invited]
   

  This paper proposes a new algorithm which compares contents in the frames. This algorithm divides original images into the regions, and tracks separated regions.
• Region-Growing Based Image Coding : Region-Growing and Texture-Information Coding

  Izumi Yohei, Haseyama Miki, Kitajima Hideo  ITEJ Technical Report  18-  (48)  61  -66  1994/09/21  [Not refereed][Not invited]
   

  This paper shows that region sizes have influence for the human visual system to recognize subjects in the image, and proposes an algorithm that makes use of the region sizes to get the satisfying segmented image for the human visual system. Simulation is carried out by approximating texture information using a two-dimensional polynomial function. Texture-information coding is also investigated with an easy way of expression.
• Bit Allocation in Image-Sequence Coding Using Variable Block Size Motion Compensation

  Yoshida Hitoshi, Haseyama Miki, Kitajima Hideo  ITEJ Technical Report  18-  (48)  91  -96  1994/09/21  [Not refereed][Not invited]
   

  This paper reviews earlier work concerning variable-block-size motion compensation and proposes a plan for optimum bit allocation between motion-related information and motion-compensated interframe prediction residues. The block-size is generally fixed in motion compensation for image sequence. But fixed-block-size motion compensation uses the same block size in a region which has complicated motions as in a region with calm motions. Motions cannot be compensated effectively and the method invites blocking effects.
• Motion Copensation for Image Sequence Coding Using Motion Parameters with Variable Block Size

  KIMURA Yoshikatsu, HASEYAWA Miki, KITAJIMA Hideo, SHIRAKAWA Tomoaki, OGAWA Yoshihiko  The Transactions of the Institute of Electronics,Information and Communication Engineers.  77-  (7)  1273  -1281  1994/07/25  [Not refereed][Not invited]
   

  動画像の符号化において,従来動き補償の方式として用いられているブロックマッチング法は,対象ブロックの動きが平行移動であるという前提で動き探索を行うものである.そのため,平行移動以外の動きには対応できないという問題があった.この問題を解決するために,本論文では,動きを1次式モデルで表現し,動き補償を行う方法を提案する.画素ごとに検出した動きベクトルをもとに,動きモデルにおけるパラメータをブロックごとに決定し,これを伝送する.受信側では,この動きパラメータから画素ごとに動きを決定するため,平行移動以外の動きにも対応することができる.更に,ブロックサイズを可変にすることによりブロック数の低減を図る.動きパラメータの符号化においては,ベクトル量子化を施す.これらによるブロック当りのパラメータ数の増加に伴う動き情報に要するビット数増加の抑制をねらう.シミュレーションでは,実際に動画像に対してフレーム間動き補償を行い,動き補償予測誤差を示す.更に,予測誤差信号の符号化も行い,復元画像のSNRとビットレートを示して,総合的に符号化特性を評価することにより本方式の有効性を確認する.
• An analysis of sensitivity for ARMA four-line lattice filter

  Aketa Yoshihiro, Haseyama Miki, Kitajima Hideo, Nagai Nobuo  IEICE technical report. Circuits and systems  94-  (117)  25  -30  1994/06/24  [Not refereed][Not invited]
   

  An ARMA lattice filter which can synthesize signals after the signal analysis eeewithout transforming its structure has been proposed.The sensitivity analysis is important when the ARMA lattice filter is used for synthesizing a gignal,but it is not easy to theoretically analyze its sensitivity characteristic.In this paper,it is considered that the ARMA lattice filter has lower sensitivity than a direct form ARMA filter by some experiments.In the experiments,when the word length of lattice coefficients is varied,the movement of the transfer function′s pole and zero is se arched.Its sensitivity is compared with the direct form filter′s o ne by using a popular measure.After analyzing the sensitivity of each lattice filter′s coefficients,the differences among coefficie nts′ sensitivities are found.
• An Arma order selection method with fuzzy reasoning for time varying model

  Emura Masafumi, Haseyama Miki, Kitajima Hideo, Nagai Nobuo  IEICE technical report. Circuits and systems  94-  (117)  31  -36  1994/06/24  [Not refereed][Not invited]
   

  In order to identify a reference model with an ARMA model,its ARMA order needs be determined.Many order selection methods,such as AIC,etc,have been proposed,however,most of the methods do not use knowoledge of the reference model effectively.Because much knowledge is expressed vaguely and it is difficult to express ambiguous information.If such knowledge can be applied to the order selection method,it is expected to achieve a higher accurate model identification.One of the methods which can express subjective and vague information is the fuzzy reasoning method.The fuzzy reasoning method can easily express the vague knowledge by fuzzy propositions.Therefore,first,this paper introduce the ARMA order selection method with the fuzzy reasoning which has been already proposed.Second,a method with the fuzzy clustering is shown,which can detect the time varying part of the reference model.Finally,an ARMA order selection method with the recursive fuzzy reasoning is proposed which determines its forgetting factor with the result of the fuzzy clustering.
• AN ARMA ORDER SELECTION METHOD WITH FUZZY THEOREM

  M HASEYAMA, H KITAJIMA, M EMURA, N NAGAI  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E77A-  (6)  937  -943  1994/06  [Not refereed][Not invited]
   

  In this paper, an ARMA order selection method is proposed with a fuzzy reasoning method. In order to identify the reference model with the ARMA model, we need to determine its ARMA order. A less or more ARMA order, other than a suitable order causes problems such as; lack of spectral information, increasing calculation cost, etc. Therefore, ARMA order selection is significant for a high accurate ARMA model identification. The proposed method attempts to select an ARMA order of a time-varying model with the following procedures: (1) Suppose the parameters of the reference model change slowly, by introducing recursive fuzzy reasoning method, the estimated order is selected. (2) By introducing a fuzzy c-mean clustering method, the period of the time during which the reference model is changing is detected and the forgetting factor of the recursive fuzzy reasoning method is set. Further, membership functions used in our algorithm are original, which are realized by experiments. In this paper, experiments are documented in order to validate the performance of the proposed method.
• RELATION BETWEEN RLS AND ARMA LATTICE FILTER REALIZATION-ALGORITHM AND ITS APPLICATION

  M HASEYAMA, N NAGAI, H KITAJIMA  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E77A-  (5)  839  -846  1994/05  [Not refereed][Not invited]
   

  In this paper, the relationship between the recursive least square (RLS) method with a U-D decomposition algorithm and ARMA lattice filter realization algorithm is presented. Both the RLS method and the lattice filter realization algorithm are used for the same applications, such as model identification, etc, therefore, it is expected that the lattice filter algorithm is in some ways related to the RLS. Though some of the proposed lattice filter algorithms have been derived by the RLS method, they do not express the relationship between RLS and ARMA lattice filter realization algorithm. In order to describe the relation clearly, a new structure of ARMA lattice fitter is proposed. Further, based on the relationship, a method of model identification with frequency weighting (MIFW), which is different from a previous method, is derived. The new MIFW method modifies the lattice parameters which are acquired without a frequency weighting and obtain the parameters of an ARMA model, which is identified with frequency weighting. The proposed MIFW method has the following restrictions: (1) The used frequency weighting is FIR filter with a low order. (2) By using the parameters of the ARMA lattice filter with ARMA (N,M) order and the frequency weighting with L order, the new ARMA parameter with the frequency weighting is with ARMA(N-L,M-L) order. By using the proposed MIFW method, the ARMA parameters estimated with the frequency weighting can be obtained without starting the computation again.
• RELATION BETWEEN RLS AND ARMA LATTICE FILTER REALIZATION-ALGORITHM AND ITS APPLICATION

  M HASEYAMA, N NAGAI, H KITAJIMA  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E77A-  (5)  839  -846  1994/05  [Not refereed][Not invited]
   

  In this paper, the relationship between the recursive least square (RLS) method with a U-D decomposition algorithm and ARMA lattice filter realization algorithm is presented. Both the RLS method and the lattice filter realization algorithm are used for the same applications, such as model identification, etc, therefore, it is expected that the lattice filter algorithm is in some ways related to the RLS. Though some of the proposed lattice filter algorithms have been derived by the RLS method, they do not express the relationship between RLS and ARMA lattice filter realization algorithm. In order to describe the relation clearly, a new structure of ARMA lattice fitter is proposed. Further, based on the relationship, a method of model identification with frequency weighting (MIFW), which is different from a previous method, is derived. The new MIFW method modifies the lattice parameters which are acquired without a frequency weighting and obtain the parameters of an ARMA model, which is identified with frequency weighting. The proposed MIFW method has the following restrictions: (1) The used frequency weighting is FIR filter with a low order. (2) By using the parameters of the ARMA lattice filter with ARMA (N,M) order and the frequency weighting with L order, the new ARMA parameter with the frequency weighting is with ARMA(N-L,M-L) order. By using the proposed MIFW method, the ARMA parameters estimated with the frequency weighting can be obtained without starting the computation again.
• 9)流れ画像復元におけるカルマンフィルタのパラメータ設定についての考察([情報ディスプレイ研究会 視聴覚技術研究会 画像処理・コンピュータビジョン研究会 画像通信システム研究会 画像応用研究会]合同)

  近藤 慶一, 長谷山 美紀, 北島 秀夫, 白川 智昭, 小川 吉彦  テレビジョン学会誌  48-  (4)  531  -531  1994/04/20  [Not refereed][Not invited]
  
• 6)画像データ圧縮用直交変換の決定論的性質の検討([情報ディスプレイ研究会 視聴覚技術研究会 画像処理・コンピュータビジョン研究会 画像通信システム研究会 画像応用研究会]合同)

  北島 秀夫, 長谷山 美紀  テレビジョン学会誌  48-  (4)  532  -532  1994/04/20  [Not refereed][Not invited]
  
• Deterministic Properties of Orthogonal Transforms for Image Data Compression

  Kitajima Hideo, Haseyama Miki  IEICE technical report. Image engineering  93-  (451)  39  -44  1994/01/28  [Not refereed][Not invited]
   

  This paper proposes new ways to look at orthogonal transforms developed for image-data compression.The impulse response is introduced to evaluate low-pass filtering resulting from the truncation of transform coefficients in the first stage of data compression.Fourier analysis of the basis vectors is proposed for testing noise shaping capabilities,where effects of quantization noise are assumed to be monitored not in the transform sequency domain but in the traditional Fourier frequency domain.Sample results involving the DCT and SCT are included to show that they are quite varied in the new light despite that they look alike when compared in terms of traditional performance measures such as residual correlation and energy packing efficiency.
• Deterministic Properties of Orthogonal Transforms for Image Data Compression

  Kitajima Hideo, Haseyama Miki  ITEJ Technical Report  18-  (7)  39  -44  1994/01/28  [Not refereed][Not invited]
   

  This paper proposes new ways to look at orthogonal transforms developed for image-data compression. The impulse response is introduced to evaluate low-pass filtering resulting from the truncation of transform coefficients in the first stage of data compression. Fourier analysis of the basis vectors is proposed for testing noise shaping capabilities, where effects of quantization noise are assumed to be monitored not in the transform sequency domain but in the traditional Fourier frequency domain. Sample results involving the DCT and SCT are included to show that they are quite varied in the new light despite that they look alike when compared in terms of traditional performance measures such as residual correlation and energy packing efficiency.
• Kalman Filter Design for the Restoration of Motion-Blurred Images

  Kondo Keiichi, Haseyama Miki, Kitajima Hideo, Shirakawa Tomoaki, Ogawa Yoshihiko  IEICE technical report. Image engineering  93-  (450)  61  -66  1994/01/27  [Not refereed][Not invited]
   

  Kalman filter design used for the restoration of motion-blurred images assumes the knowledge of signal statics and the blurring system.This paper discusses the estimation of parameters required in Kalman filter design for the restoration of motion-blurred images.It is shown that the innovation of the Kalman filter can be used to improve the estimations.
• Kalman Filter Design for the Restoration of Motion-Blurred Images

  Kondo Keiichi, Haseyama Miki, Kitajima Hideo, Shirakawa Tomoaki, Ogawa Yoshihiko  ITEJ Technical Report  18-  (6)  61  -66  1994/01/27  [Not refereed][Not invited]
   

  Kalman filter design used for the restoration of motion-blurred images assumes the knowledge of signal statics and the blurring system. This paper discusses the estimation of parameters required in Kalman filter design for the restoration of motion-blurred images. It is shown that the innovation of the Kalman filter can be used to improve the estimations.
• 可変ブロックサイズ動き補償画像符号化法利用のための一考察

  吉田仁, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• A Method for Coarse of an ARMA Lattice Filter with GA.

  明田吉浩, 長谷山美紀, 北島秀夫, 永井信夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• カルマンフィルタのイノベーションによるマルチフレーム画像復元

  近藤慶一, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• ファジィクラスタリングを用いた次数推定法について

  江村真史, 長谷山美紀, 北島秀夫, 永井信夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• Study on Cut-Detection Methods of Image Sequences.

  水野哲, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• ノイズを含んだ画像における勾配法を用いた3次元領域拡張

  菊池学, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• Image Coding by Gray-Scale Iterated Function System.

  本多博之, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• 画像符号化のための自然な領域分割法

  泉洋平, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• Time Distribution of Truncated Errors in Continuous-Signal KL Expansion.

  塚原正人, 長谷山美紀, 北島秀夫  電気関係学会北海道支部連合大会講演論文集  1994-  1994
• A model identification with frequency weighting by ARMA lattice filter.

  長谷山美紀, 北島秀夫, 永井信夫  電子情報通信学会大会講演論文集  1994-  (Shunki Pt 1)  1994
• A realization method for an ARMA lattice filter with reduced number of coefficients

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki  Electronics and Communications in Japan (Part III: Fundamental Electronic Science)  77-  (8)  14  -27  1994  [Not refereed][Not invited]
   

  This paper proposes realization methods of two kinds of low sensitivity ARMA lattice filters which can perform frequency‐weighting model identification with input estimation and low sensitivity of their filters. Using the proposed realization method, the filter with the fewer coefficients than the previous filter can be obtained. Since the calculation costs for the realization of the filter are reduced, the proposed realization methods are significant. Furthermore, assuming that the ARMA lattice filters have low sensitivity, however, this paper shows that the proposed ARMA lattice filters have low sensitivity by simulations, which is important for practical implementation of filters. Copyright © 1994 Wiley Periodicals, Inc., A Wiley Company
• A Designing Method of a Kalman Filter for the Restoration of Motion-Blurred Images.

  Kondo Keiichi, Haseyama Miki, Kitajima Hideo, Shirakawa Tomoaki, Ogawa Yoshihiko  The Journal of The Institute of Image Information and Television Engineers  48-  (8)  1058  -1063  1994  [Not refereed][Not invited]
   

  本論文では, 流れ画像の復元におけるカルマンフィルタの設計パラメータの中で, 流れの長さが復元精度を向上するために重要であることを示す.さらに, イノベーションの性質に着目し, 流れの長さの精度向上を可能とする一手法を示す.
• Bit Allocation in Image-Sequence Coding Using Variable Block Size Motion Compensation

  Yoshida Hitoshi, Haseyama Miki  Technical report of IEICE. OFC  91  -96  1994  [Not refereed][Not invited]
   

  This paper reviewes earlier work cosecering variable-block-size motion compensation and proposes a plan for optimum bit allocation between motion-related information and motion-compensated interframe prediction residues.The block-size is generally fixed in motion compensation for image sequence.But fixed-block-size motion compensation uses the same block size in a region which has complicated motions as in a region with calm motions.Motions cannot be compensated effectively and the method invites blocking effects.
• Preliminary reserch of contraction mapping in image representation

  Honda Hiroyuki, Haseyama Miki, Kitajima Hideo  Technical report of IEICE. OFC  94-  (246)  21  -26  1994  [Not refereed][Not invited]
   

  There are two grayscale image coding methods by fractal.One, global scheme.The other,local scheme.Now,the main current method of local scheme is random iteration system(Random IFS).But,this method has a weak point that it needs many iteration time.In this paper,a new image coding method based on collage theorem and gray- scale iterated function system,is proposed.The statistical properties between reconstructed image and origainal image are investigated.And the efficacy of this method is shown.
• A Realization Method for ARMA Lattice Filter with Reduced Number of Coefficients

  HASEYAMA Miki, NAGAI Nobuo, MIKI Nobuhiro  The Transactions of the Institute of Electronics,Information and Communication Engineers. A  76-  (12)  1714  -1724  1993/12/01  [Not refereed][Not invited]
   

  入力信号が未知のモデルについて,周波数重みを付けたモデル同定を可能とするARMA4線格子形フィルタが既に提案され,その応用が重要であることが述べられているが,乗算器個数が多いことが欠点とされていた.そこで本論文では,このような同定法が可能な乗算器個数の少ないARMA格子形フィルタの実現法を提案する.乗算器個数の削減は,直接計算量の削減につながり,特に適応処理を行う場合には重要である.本論文では,2種類のフィルタの実現法を提案している.提案されている2種類のARMA格子形フィルタは,実現に用いられる予測誤差の違いにより異なった性質をもつ.本論文では,提案する二つのフィルタおよび乗算器個数の多い従来のフィルタについて,各々を用いる場合にどのような点が異なるかを示す.最後に,本論文で提案するフィルタおよび従来のフィルタは,得られたラティス構造を保ったまま,信号合成フィルタとして用いることができることから,そのような応用を行った場合の係数感度についての考察を実験により行う.実験より,乗算器個数の削減による感度特性の劣化が生じていないことを確認する.
• Considerations of relations between RLS method and ARMA latice filter algorithm

  Haseyama Miki, Nagai Nobuo, Miki Nobuhiro  IEICE technical report. Circuits and systems  93-  (33)  35  -40  1993/05/20  [Not refereed][Not invited]
   

  Recursive Least Square(:RLS)method is one of recursive algorithms for identification of linear systems.The RLS is easily applied to Least Square(:LS)estimation problems and its estimated solutions are stably obtained,therefore,it is often used.A realization algorithm of ARMA lattice filter is another recursive algorithm.The RLS and ARMA lattice algorithms are applied to similar problems,but they are different algorithms.In this paper, relations between them are considered.Especially,the lattice algorithm is compared with RLS method with U-D decomposition. Furthermore,we propose a model identification method with frequency weighting of low-order FIR filter by using the above derivation.
• Solution Normal Equation on ARMA Lattice Filter Realization Algorithm.

  長谷山美紀, 永井信夫, 三木信弘  電気関係学会北海道支部連合大会講演論文集  1993-  1993
• Modification of a normal equation on ARMA lattice filter algorithm.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会大会講演論文集  1993-  (Shuki Pt 1)  1993
• An Adaptive ARMA Four-Line Lattice Filter for Spectral Estimation with Frequency Weighting

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki  IEEE Transactions on Signal Processing  41-  (6)  2193  -2207  1993  [Not refereed][Not invited]
   

  This paper proposes a method for designing an adaptive four-line lattice filter which can perform “frequency-weighting spectral estimation.” Frequency-weighting spectral estimation is a method by which we can obtain more accurate spectral estimation for some frequency bands than for others. By using a suitable “frequency-weighting function,” denoted as an ARMA model, we can obtain an estimated spectrum by arbitrarily weighting some frequency bands more heavily than others. If the frequency-weighting function has the property of a lowpass filter, we can estimate the spectrum of the reference model accurately with a reduced ARMA order in the low frequency band. Therefore, the frequency-weighting method is useful for accurately estimating models with a reduced ARMA order in the specified frequency band. Further, using the proposed algorithm, we can estimate spectra of time-varying models with an exponentially weighted sliding window, and we can also estimate the input signal of the reference model by assumption. In this paper, we introduce the order-update and the time-update recursive formulas and also the frequency-weighting method for the proposed filter. In order to verify the proposed algorithm, we present a number of experiments of spectral estimation for a time-varying and unknown-input model. © 1993 IEEE
• A Note on Sensitivity of Coefficients for ARMA Lattice Filter.

  長谷山美紀, 永井信夫, 鈴木正清, 三木信弘  電子情報通信学会大会講演論文集  1992-  (Shunki Pt 1)  1992
• A note on estimation of ARMA orders with fuzzy reasoning.

  津島峰生, 長谷山美紀, 三木信弘, 永井信夫  電子情報通信学会大会講演論文集  1992-  (Shunki Pt 1)  1992
• A method of realizing an adaptive normalized ARMA lattice filter

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki  Electronics and Communications in Japan (Part III: Fundamental Electronic Science)  75-  (4)  71  -78  1992  [Not refereed][Not invited]
   

  This paper proposes a method of realizing an adaptive normalized ARMA lattice filter. This filter was proposed in a previous paper and can be realized with an arbitrary ARMA order and with all the lattice coefficients less than one. Furthermore, these coefficients are equal, so the structure is simplified. The adaptive realization method proposed in this paper retains these features. In addition, we perform spectral estimation of a time‐varying signal with frequency weighting and model identification of an unknown input system. In our paper we describe an experiment with a real speech signal to prove that our algorithm can be used for frequency‐weighted spectral estimation. The use of real data to perform spectral estimation has not been attempted in other references. Copyright © 1992 Wiley Periodicals, Inc., A Wiley Company
• Realization Method of Adaptive Normalized ARMA Lattice Filter.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会論文誌 A  74-  (10)  p1483  -1489  1991/10  [Not refereed][Not invited]
  
• A note on Realization of ARMA Lattice Filter.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会技術研究報告  91-  (54(CAS91 1-13))  1991
• Designing method for ARMA four‐line lattice filter with sliding rectangular window

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki  Electronics and Communications in Japan (Part III: Fundamental Electronic Science)  74-  (9)  11  -21  1991  [Not refereed][Not invited]
   

  This paper derives a design method for an ARMA 4‐line lattice filter using a sliding rectangular window. The adaptive ARMA 4‐line lattice filter already proposed uses a forgetting factor, which is one of the weighting functions to estimate coefficients of a time‐varying system in which system coefficients vary with sufficient smoothness. Therefore, the effect of past observed signal over the estimated coefficients decreases exponentially. The filter presented here is realized using a rectangular window because the concern is over the effects of past observation signal rather than window length. Using this filter, an input signal is estimated when designing furthermore, a system can be identified in which an arbitrary section in the frequency domain is weighted. Thus, by not only analyzing voice signal which is considered a model whose input signal is unknown and EEG data but also by weighting the frequency domain, for example, the holmant in the low‐frequency domain can be estimated in low degree with very high precision and a specific wave (such as a α activities) in EEG data can also be detected. Moreover, in this paper the algorithm is verified by model experiments. Copyright © 1991 Wiley Periodicals, Inc., A Wiley Company
• Designing method for ARMA four-line lattice filter with sliding rectangular window.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会論文誌 A  74-  (1)  p9  -18  1991/01  [Not refereed][Not invited]
  
• A study on cascading elementary sections of adaptive ARMA 4-line lattice filter.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会全国大会講演論文集  1990-  (Spring Pt.1)  1990
• Considerations for adaptive ARMA four-line lattice filter. Spectral estimation with sliding rectangular window.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会技術研究報告  90-  (179(SP90 39-43))  1990
• Model identification weighted in the frequency domain with ARMA 4-lines lattice filter

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki  Electronics and Communications in Japan, Part III: Fundamental Electronic Science (English translation of Denshi Tsushin Gakkai Ronbunshi)  73-  1  -8  1990/01/01  [Not refereed][Not invited]
   

  This paper presents a frequency-weighted model identification method which employs an ARMA four-lines lattice filter developed by the authors. This lattice filter is used to estimate the ARMA identification weighting in the optional frequency region. Using this method, the frequency-weighted characteristics of a model with an unknown input can be extracted by employing the ARMA four-lines lattice filter. The ARMA coefficients are approximated as a reduced-order model for frequency-weighted model identification using an unknown input system. Finally, time-update and order-update recursions are derived and used to estimate the input of an unknown input system.
• Normalized ARMA lattice filter and its terminal conditions

  Miki Haseyama, Nobuo Nagai, Nobuhiro Miki, Yoshikazu Miyanaga  Electronics and Communications in Japan (Part III: Fundamental Electronic Science)  72-  (2)  1  -13  1989  [Not refereed][Not invited]
   

  In this paper it is shown that covariant data of a synthesized signal corresponds to the covariant data of a real signal. This occurs when an ARMA lattice synthesis filter is used and the minimum mean square value of the estimation error is zero. Hence, a new terminal condition of the ARMA lattice filter is derived so that the covariant data of the synthesized signal will always correspond to the covariant data of a real signal. To this end, a design method for the normalized ARMA lattice filter is presented in this paper, and normalized estimation errors are derived. Finally, the starting condition and design for the normalized ARMA lattice filter is given. Copyright © 1989 Wiley Periodicals, Inc., A Wiley Company
• A model identification weighted in the frequency domain with ARMA 4-lines lattice filter.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会論文誌 A  72-  (1)  p80  -86  1989/01  [Not refereed][Not invited]
  
• A method to reduce and increace to arbitrary orders on ARMA lattice filter.

  長谷山美紀, 永井信夫, 三木信弘  電子情報通信学会論文誌 A  71-  (6)  p1242  -1248  1988/06  [Not refereed][Not invited]
  
• Normalized ARMA lattice filter and its terminal conditions.

  長谷山美紀, 宮永喜一, 永井信夫, 三木信弘  電子情報通信学会論文誌 A  70-  (8)  p1115  -1125  1987/08  [Not refereed][Not invited]
  
• Consideration on design for ARMA lattice filter with optimum order.

  長谷山美紀, 永井信夫, 三木信弘, 宮永喜一  電子情報通信学会技術研究報告  87-  (148)  1987