研究者詳細 - 中澤　篤志

論文 - 中澤　篤志

分割表示 105 件中 61 - 80 件目／全件表示 >>

Corneal imaging revisited: An overview of corneal reflection analysis and applications 査読

Christian Nitschke, Atsushi Nakazawa, Haruo Takemura

IPSJ Transactions on Computer Vision and Applications 5 1 - 18 2013年7月

記述言語：英語

The cornea of the human eye acts as a mirror that reflects light from a person's environment. These corneal reflections can be extracted from an image of the eye by modeling the eye-camera geometry as a catadioptric imaging system. As a result, one obtains the visual information of the environment and the relation to the observer (view, gaze), which allows for application in a number of fields. The recovered illumination map can be further applied to various computational tasks. This paper provides a comprehensive introduction on corneal imaging, and aims to show the potential of the topic and encourage advancement. It makes a number of contributions, including (1) a combined view on previously unrelated fields, (2) an overview of recent developments, (3) a detailed explanation on anatomic structures, geometric eye and corneal reflection modeling including multiple eye images, (4) a summary of our work and contributions to the field, and (5) a discussion of implications and promising future directions. The idea behind this paper is a geometric framework to solve persisting technical problems and enable non-intrusive interfaces and smart sensors for traditional, ubiquitous and ambient environments. © 2013 Information Processing Society of Japan.

DOI： 10.2197/ipsjtcva.5.1

Corneal Imaging Revisited: An Overview of Corneal Reflection Analysis and Applications

Nitschke Christian, Nakazawa Atsushi, Takemura Haruo

Information and Media Technologies 8 ( 2 ) 389 - 406 2013年

記述言語：英語出版者・発行元：Information and Media Technologies Editorial Board

The cornea of the human eye acts as a mirror that reflects light from a person's environment. These corneal reflections can be extracted from an image of the eye by modeling the eye-camera geometry as a catadioptric imaging system. As a result, one obtains the visual information of the environment and the relation to the observer (view, gaze), which allows for application in a number of fields. The recovered illumination map can be further applied to various computational tasks. This paper provides a comprehensive introduction on corneal imaging, and aims to show the potential of the topic and encourage advancement. It makes a number of contributions, including (1) a combined view on previously unrelated fields, (2) an overview of recent developments, (3) a detailed explanation on anatomic structures, geometric eye and corneal reflection modeling including multiple eye images, (4) a summary of our work and contributions to the field, and (5) a discussion of implications and promising future directions. The idea behind this paper is a geometric framework to solve persisting technical problems and enable non-intrusive interfaces and smart sensors for traditional, ubiquitous and ambient environments.

DOI： 10.11185/imt.8.389

Motion Coherent Tracking Using Multi-label MRF Optimization 査読

David Tsai, Matthew Flagg, Atsushi Nakazawa, James M. Rehg

INTERNATIONAL JOURNAL OF COMPUTER VISION 100 ( 2 ) 190 - 202 2012年11月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SPRINGER

We present a novel off-line algorithm for target segmentation and tracking in video. In our approach, video data is represented by a multi-label Markov Random Field model, and segmentation is accomplished by finding the minimum energy label assignment. We propose a novel energy formulation which incorporates both segmentation and motion estimation in a single framework. Our energy functions enforce motion coherence both within and across frames. We utilize state-of-the-art methods to efficiently optimize over a large number of discrete labels. In addition, we introduce a new ground-truth dataset, called Georgia Tech Segmentation and Tracking Dataset (GT-SegTrack), for the evaluation of segmentation accuracy in video tracking. We compare our method with several recent on-line tracking algorithms and provide quantitative and qualitative performance comparisons.

DOI： 10.1007/s11263-011-0512-5

Point of Gaze Estimation through Corneal Surface Reflection in an Active Illumination Environment 査読

Atsushi Nakazawa, Christian Nitschke

COMPUTER VISION - ECCV 2012, PT II 7573 ( PART 2 ) 159 - 172 2012年

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：SPRINGER-VERLAG BERLIN

Eye gaze tracking (EGT) is a common problem with many applications in various fields. While recent methods have achieved improvements in accuracy and usability, current techniques still share several limitations. A major issue is the need for external calibration between the gaze camera system and the scene, which commonly restricts to static planar surfaces and leads to parallax errors. To overcome these issues, the paper proposes a novel scheme that uses the corneal imaging technique to directly analyze reflections from a scene illuminated with structured light. This comprises two major contributions: First, an analytic solution is developed for the forward projection problem to obtain the gaze reflection point (GRP), where light from the point of gaze (PoG) in the scene reflects at the corneal surface into an eye image. We also develop a method to compensate for the individual offset between the optical axis and true visual axis. Second, introducing active coded illumination enables robust and accurate matching at the GRP to obtain the PoG in a scene image, which is the first use of this technique in EGT and corneal reflection analysis. For this purpose, we designed a special high-power IR LED-array projector. Experimental evaluation with a prototype system shows that the proposed scheme achieves considerable accuracy and successfully supports depth-varying environments.

DOI： 10.1007/978-3-642-33709-3_12

Display-camera calibration using eye reflections and geometry constraints 査読

Christian Nitschke, Atsushi Nakazawa, Haruo Takemura

COMPUTER VISION AND IMAGE UNDERSTANDING 115 ( 6 ) 835 - 853 2011年6月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：ACADEMIC PRESS INC ELSEVIER SCIENCE

In this paper, we describe a novel method for calibrating display-camera setups from reflections in a user's eyes. Combining both devices creates a capable controlled illumination system that enables a range of interesting vision applications in non-professional environments, including object/face reconstruction and human computer interaction. One major issue barring such systems from average homes is the geometric calibration to obtain the pose of the display which requires special hardware and tedious user interaction. Our proposed approach eliminates this requirement by introducing the novel idea of analyzing screen reflections in the cornea of the human eye, a mirroring device that is always available. We employ a simple shape model to recover pose and reflection characteristics of the eye. Thorough experimental evaluation shows that the basic strategy results in a large error and discusses possible reasons. Based on the findings, a non-linear optimization strategy is developed that exploits geometry constraints within the system to considerably improve the initial estimate. It further allows to automatically resolve an inherent ambiguity that arises in image-based eye pose estimation. The strategy may also be integrated to improve spherical mirror calibration. We describe several comprehensive experimental studies which show that the proposed method performs stably with respect to varying subjects, display poses, eye positions, and gaze directions. The results are feasible and should be sufficient for many applications. In addition, the findings provide general insight on the application of eye reflections for geometric reconstruction. (C) 2011 Elsevier Inc. All rights reserved.

DOI： 10.1016/j.cviu.2011.02.008

サリエンシーディテクション

中澤篤志

映像情報メディア学会誌 : 映像情報メディア = The journal of the Institute of Image Information and Television Engineers 64 ( 12 ) 1830 - 1832 2010年12月

Motion capture systems 査読

Atsushi Nakazawa

Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers 63 ( 9 ) 1224 - 1227 2009年

Task recognition and person identification in cyclic dance sequences with Multi Factor Tensor analysis 査読

Manoj Perera, Takaaki Shiratori, Shunsuke Kudoh, Atsushi Nakazawa, Katsushi Ikeuchi

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E91D ( 5 ) 1531 - 1542 2008年5月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG

In this paper, we present a novel approach to recognize motion styles and identify people using the Multi Factor Tensor (MFT) model. We apply a musical information analysis method in segmenting the motion sequence relevant to the keyposes and the musical rhythm. We define a task model by considering the repeated motion segments, where the motion is decomposed into a person-invariant factor task and a person-dependent factor style. Given the motion data set, we formulate the MFT model, factorize it efficiently in different modes, and use it in recognizing the tasks and the identities of the persons performing the tasks. We capture the motion data of different people for a few cycles, segment it using the musical analysis approach, normalize the segments using a vectorization method, and realize our MFT model. In our experiments, Japanese traditional dance sequences performed by several people are used. Provided with an unknown motion segment which is to be probed and which was performed at a different time in the time space, we first normalize the motion segment and flatten our MFT model appropriately, then recognize the task and the identity of the person. We follow two approaches in conducting our experiments. In one experiment, we recognize the tasks and the styles by maximizing a function in the tensor subdomain, and in the next experiment, we use a function value in the tensorial subdomain with a threshold for recognition. Interestingly, unlike the first experiment, we are capable of recognizing tasks and human identities that were not known beforehand. We conducted various experiments to evaluate the potential of the recognition ability of our proposed approaches, and the results demonstrate the high accuracy of our model.

DOI： 10.1093/ietisy/e91-d.5.1531

Diskless Linux を用いた情報教育システムの開発とその評価査読

桝田秀夫, 小川剛史, 町田貴史, 中澤篤志, 清川清, 竹村治雄

情報処理学会論文誌 49 ( 3 ) 1239 - 1248 2008年3月

記述言語：日本語出版者・発行元：一般社団法人情報処理学会

教育用計算機システムでは，多数の利用者に対して各種のアプリケーションが動作する一様な計算機環境を，できるだけ低いTCO （Total Cost of Ownership）で実現する必要がある．運用コスト削減のためには，故障時に素早い復旧が可能であることだけでなく，そもそも故障しにくい構成をとることも重要である．本論文では，これまでの運用の中で最も故障の多かったハードディスクをクライアント上では使用しない，ディスクレス構成を用いたLinux ベースの情報教育システムの開発とその評価について述べる．また，予算措置の関係で導入時期や機種が異なる複数種類のパソコンを，一括して運用管理する必要性に対応するための構成上の工夫についても述べる．さらに，Microsoft Office で作られた文書の激増に対して，互換性のあるOpenOffice.org やStarSuite だけではなく，OS としてのMicrosoft Windows を運用せずに，Microsoft Office 自体をCrossOver Office というWINE ベースのエミュレータ上で稼働させることに挑戦している．これらの構成の設計・開発の経緯，ならびに導入後約2 年間の運用実績と得られた知見について報告する．This paper shows the configurations and evaluations of the educational computer system in our university that serves variety types of applications for many numbers of users, reducing Total Cost of Ownership (TCO). As for client operating systems, we developed newly designed diskless Linux. The OS and other applications are loaded from servers through networks; therefore the clients don't have hard drives. This scheme drastically reduces TCO because it is not necessary to replace hard drives which fail most frequently in these types of systems. Also, the software updates of the clients are achieved just by uploading the files in few servers. This also reduces the maintenance cost rather than updating the software stored in local hard drives of many clients. As for the client application software, we installed not only OpenOffice.org or StarSuite, but also Microsoft Office combined with Crossover Office emulator that is highly demanded by end users.

Iterative refinement of range images with anisotropic error distribution 査読

Ryusuke Sagawa, Takeshi Oishi, Atsushi Nakazawa, Ryo Kurazume, Katsushi Ikeuchi

Digitally Archiving Cultural Objects 193 - 205 2008年

記述言語：英語掲載種別：論文集(書籍)内論文出版者・発行元：Springer US

We propose a method which refines the range measurement of range finders by computing correspondences of vertices of multiple range images acquired from various viewpoints. Our method assumes that a range image acquired by a laser range finder has anisotropic error distribution which is parallel to the ray direction. Thus, we find corresponding points of range images along with the ray direction. We iteratively converge range images to minimize the distance of corresponding points. We describe the effectiveness of our method by the presenting the experimental results of artificial and real range data. Also we show that our method refines a 3D shape more accurately as opposed to that achieved by using the Gaussian filter. © 2008 Springer-Verlag US.

DOI： 10.1007/978-0-387-75807_11

A fast simultaneous alignment of multiple range images 査読

Takeshi Oishi, Atsushi Nakazawa, Ryo Kurazume, Katsushi Ikeuchi

Digitally Archiving Cultural Objects 89 - 107 2008年

記述言語：英語掲載種別：論文集(書籍)内論文出版者・発行元：Springer US

This chapter describes a fast, simultaneous alignment method for a large number of range images. Generally the most time-consuming task in aligning range images is searching corresponding points. The fastest searching method is the "Inverse Calibration" method. However, this method requires pre-computed lookup tables and precise sensor parameters. We propose a fast searching method using "index images", which work as look-up tables and are rapidly created without any sensor parameters by using graphics hardware. To accelerate the computation to estimate rigid transformations, we employed a linear error evaluation method. When the number of range images increases, the computation time for solving the linear equations becomes too long because of the large size of the coefficient matrix. On the other hand, the coefficient matrix has the characteristic of becoming sparser as the number of range images increases. Thus, we applied the Incomplete Cholesky Conjugate Gradient (ICCG) method to solve the equations and found that the ICCG greatly accelerates the matrix operation by pre-conditioning the coefficient matrix. Some experimental results in which a large number of range images are aligned demonstrate the effectiveness of our method. © 2008 Springer-Verlag US.

DOI： 10.1007/978-0-387-75807_6

Breed differentiation among Japanese native chickens by specific skull features determined by direct measurements and computer vision techniques 査読

Y. Ino, T. Oka, K. Nomura, T. Watanabe, S. Kawashima, T. Amano, Y. Hayashi, A. Okabe, Y. Uehara, T. Masuda, J. Takamatsu, A. Nakazawa, K. Ikeuchi, H. Endo, K. Fukuta, F. Akishinonomiya

BRITISH POULTRY SCIENCE 49 ( 3 ) 273 - 281 2008年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：TAYLOR & FRANCIS LTD

1. Inter-breed morphological comparisons were made among 11 breeds of Japanese native chickens (Gifujidori, Hinaidori, Shokoku, Totenko, Tomaru, Satsumadori, Shamo, Koshamo, Koeyoshi, Chabo and Nagoya), White Leghorn, broiler chickens (Chunky) and red junglefowl collected in the Philippines, based on results of direct measurements and analysis by computer vision techniques of the skull.
2. Analysis of direct measurements identified two groups of chicken: a small type that included the Chabo, Koshamo, red junglefowl, Gifujidori and Shokoku and a large type that included the remaining breeds studied. These groupings were made based on size determined both in the first (PC1) and second principal component (PC2). The greatest length of the cranium and condylobasal length greatly contributed to the morphological differences between these two groups.
3. Analysis by computer vision techniques, however, identified three groups: the Bantam group (which includes red junglefowl), Shokoku group and Shamo group. White Leghorn clustered within the Shokoku group while the broiler chicken belonged to the Shamo group. The region around the junction of the neural cranium and the visceral cranium contributed greatly to the morphological differences among breeds, both in the PC1 and PC2.

DOI： 10.1080/00071660802094727

Parallel alignment of a large number of range images 査読

Takeshi Oishi, Atsushi Nakazawa, Ryo Kurazume, Katsushi Ikeuchi, Ryusuke Sagawa

Digitally Archiving Cultural Objects 109 - 126 2008年

記述言語：英語掲載種別：論文集(書籍)内論文出版者・発行元：Springer US

This chapter describes a method for parallel alignment of multiple range images. There are problems of computational time and memory space in aligning a large number of range images simultaneously. We developed a parallel method to address the problems. Searching for corresponding points between two range images is time-consuming and requires considerable memory space when performed independently. However, this process can be preformed in parallel, with each corresponding pair of range images assigned to a node. Because the computation time is approximately proportional to the number of vertices, by assigning the pairs so that the number of vertices computed is equal on each node, the load on each node is effectively distributed. In order to reduce the amount of memory required on each node, a hypergraph that represents the correspondences of range images is created, and heuristic graph partitioning algorithms are applied to determine the optimal assignment of the pairs. Moreover, by rejecting redundant dependencies, it becomes possible to accelerate computation time and reduce the amount of memory required on each node. The method was tested on a 16-processor PC cluster, where it demonstrated high extendibility and improved performance. © 2008 Springer-Verlag US.

DOI： 10.1007/978-0-387-75807_7

The Great Buddha project: Digitally archiving, restoring, and analyzing cultural heritage objects 査読

Katsushi Ikeuchi, Takeshi Oishi, Jun Takamatsu, Ryusuke Sagawa, Atsushi Nakazawa, Ryo Kurazume, Ko Nishino, Mawo Kamakura, Yasuhide Okamoto

INTERNATIONAL JOURNAL OF COMPUTER VISION 75 ( 1 ) 189 - 208 2007年10月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SPRINGER

This paper presents an overview of our research project on digital preservation of cultural heritage objects and digital restoration of the original appearance of these objects. As an example of these objects, this project focuses on the preservation and restoration of the Great Buddhas. These are relatively large objects existing outdoors and providing various technical challenges. Geometric models of the great Buddhas are digitally achieved through a pipeline, consisting of acquiring data, aligning multiple range images, and merging these images. We have developed two alignment algorithms: a rapid simultaneous algorithm, based on graphics hardware, for quick data checking on site, and a parallel alignment algorithm, based on a PC cluster, for precise adjustment at the university. We have also designed a parallel voxel-based merging algorithm for connecting all aligned range images. On the geometric models created, we aligned texture images acquired from color cameras. We also developed two texture mapping methods. In an attempt to restore the original appearance of historical objects, we have synthesized several buildings and statues using scanned data and a literature survey with advice from experts.

DOI： 10.1007/s11263-007-0039-y

Learning from observation paradigm: Leg task models for enabling a biped humanoid robot to imitate human dances 査読

Shin'ichiro Nakaoka, Atsushi Nakazawa, Fumio Kanehiro, Kenji Kaneko, Mitsuharu Morisawa, Hirohisa Hirukawa, Katsushi Ikeuchi

INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH 26 ( 8 ) 829 - 844 2007年8月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SAGE PUBLICATIONS LTD

This paper proposes a framework that achieves the Learning from Observation paradigm for learning dance motions. The framework enables a humanoid robot to imitate dance motions captured from human demonstrations. This study especially focuses on leg motions to achieve a novel attempt in which a biped-ope robot imitates not only upper body motions but also leg motions including steps. Body differences between the robot and the original dancer make the problem difficult because the differences prevent the robot front straight forward-v following the original motions and they also change dynamic body balance. We propose leg task models, which play a key role in solving the problem. Low-level tasks in leg motion are modelled so that they clearly provide essential information required,for keeping dynamic stability and important motion characteristics. The models divide the problem of adapting motions into the problem of recognizing a sequence of the tasks and the problem of executing the task sequence. We have developed a method for recognizing the tasks from captured motion data and a method for generating the motions of the tasks that can be executed by existing robots including HRP-2. HRP-2 successfully performed the generated motions, which imitated a traditional folk dance performed by human dancers.

DOI： 10.1177/0278364907079430

Real-time space carving using graphics hardware 査読

Christian Nitschke, Atsushi Nakazawa, Haruo Takemura

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E90D ( 8 ) 1175 - 1184 2007年8月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG

Reconstruction of real-world scenes from a set of multiple images is a topic in computer vision and 3D computer graphics with many interesting applications. Attempts have been made to real-time reconstruction on PC cluster systems. While these provide enough performance, they are expensive and less flexible. Approaches that use a GPU hardware-acceleration on single workstations achieve real-time framerates for novel-view synthesis, but do not provide an explicit volumetric representation. This work shows our efforts in developing a GPU hardware-accelerated framework for providing a photo-consistent reconstruction of a dynamic 3D scene. High performance is achieved by employing a shape from silhouette technique in advance. Since the entire processing is done on a single PC, the framework can be applied in mobile environments, enabling a wide range of further applications. We explain our approach using programmable vertex and fragment processors and compare it to highly optimized CPU implementations. We show that the new approach can outperform the latter by more than one magnitude and give an outlook for interesting future enhancements.

DOI： 10.1093/ietisy/e90-d.8.1175

Human pose estimation from volume data and topological graph database 査読

Hidenori Tanaka, Atsushi Nakazawa, Haruo Takemura

COMPUTER VISION - ACCV 2007, PT I, PROCEEDINGS 4843 ( PART 1 ) 618 - + 2007年

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：SPRINGER-VERLAG BERLIN

This paper proposes a novel volume-based motion capture method using a bottom-up analysis of volume data and an example topology database of the human body. By using a two-step graph matching algorithm with many example topological graphs corresponding to postures that a human body can take, the proposed method does not require any initial parameters or iterative convergence processes, and it can solve the changing topology problem of the human body. First, three-dimensional curved lines (skeleton) are extracted from the captured volume data using the thinning process. The skeleton is then converted into an attributed graph. By using a graph matching algorithm with a large amount of example data, we can identify the body parts from each curved line in the skeleton. The proposed method is evaluated using several video sequences of a single person and multiple people, and we can confirm the validity of our approach.

階層化3次元データを扱う対話的なコンテンツマネジメントシステム

木村洋介, 間下以大, 中澤篤志, 町田貴史, 清川清, 竹村治雄

ヒューマンインタフェースシンポジウム 2006年9月

記述言語：日本語掲載種別：研究論文（その他学術会議資料等）

本研究では、三次元データを扱うコンテンツマネジメントシステム(CMS)を提案する。システムに三次元データ等のコンテンツを登録し、それらを用いたウェブサイトの構築を可能とする。三次元データは単にHTMLファイルに組み込むだけでなく、他のコンテンツとの表示の連携を定義することが可能であり、ユーザによりわかりやすく三次元データを提示することが出来る。三次元データの描画にはQSplatをJavaアプレット化したシステムを用いる。これにより、端末の処理速度に合わせた解像度で、また端末のプラットフォームに依存せずに描画することが可能である。

モバイル端末向け注釈付加機能付き大規模三次元データ高速描画用Javaアプレット

木村洋介, 中澤篤志, 町田貴史, 清川清, 竹村治雄

情報処理学会マルチメディア，分散，協調とモバイルシンポジウム(DICOMO)論文集 2006年7月

記述言語：日本語掲載種別：研究論文（その他学術会議資料等）

近年，モバイル端末の普及に伴って，インターネットなどのネットワーク上でデータを閲覧する機会が増えている．しかし，モバイル環境でネットワーク上の大規模三次元データを閲覧する際には，通信速度及び描画能力が制限される，プラットフォームが統一されていないなどの問題がある．そこで，本研究では，大規模三次元データの高速描画方式であるQSplatをJavaアプレットを用いて実装することで，大規模な三次元データを端末の表示環境に合わせて高速に描画可能な，マルチプラットフォームシステムを提案する．これにより，システムの実行にあたってウェブブラウザを利用することができるため，専用アプリケーションが不要となる．また，本システムを他のシステムへ統合することも容易になる．さらに，ユーザの三次元データへの理解を補助するため，三次元データに対して階層的な意味構造を持つ注釈情報を付加し，画像と共に表示する機能を実装した．

脚タスクモデルを用いた2足歩行ヒューマノイドロボットによる人の舞踊動作の再現

中岡慎一郎, 中澤篤志, 金広文男, 金子健二, 森澤光晴, 比留川博久, 池内克史

日本ロボット学会誌 = Journal of Robotics Society of Japan 24 ( 3 ) 388 - 399 2006年4月

記述言語：日本語出版者・発行元：The Robotics Society of Japan

In this paper, we propose a method that enables a biped humanoid robot to reproduce human dance motions with its whole body. Our method is based on the paradigm ofLearning from Observation. In this study, a robot uses its own legs to support the body during a dance performance. We proposeleg task models, which can solve the problems caused by severe constraints in adapting human motions to the legs of a robot. First, elements of the leg task models are recognized from motion data captured from human performances. Then motion data of a robot is regenerated from the recognized elements so that the motion is stably executable on the robot. Our method was verified by experiments on a humanoid robotHRP-2using a traditional folk dance. HRP-2 successfully performed dance motions that were automatically reproduced from motion data captured from human dance performances.

DOI： 10.7210/jrsj.24.388

その他リンク： https://jlc.jst.go.jp/DN/JALC/00278397673?from=CiNii

　前のページ - 次のページ　

1
2
3
4
5
6