Faculty Profiles - AMANO Toshiyuki

Conference Activities & Talks -

Division display61 - 80 of about 222／ All the affair displays >>

Controlled illumination for the object recognition with projector camera feedback

Toshiyuki Amano, Kazuki Osamura, Makoto Fujisawa, Jun Miyazaki, Hirokazu Kato

Proceedings of the 12th IAPR Conference on Machine Vision Applications, MVA 2011 2011.12.01

In this paper a novel image capturing method for a solution of shading problem in computer vision is proposed. In the experiments, we performed the evaluation about the conditions of environment lighting and additional lighting and proposed controlled lighting method. Image-processing of self quotient image algorithm, histogram equalization, and luminance equalization were used for pre-processing to remove shading from a captured image for the comparative evaluation. Through the evaluation of robustness assumed appearance based method, the proposed method showed most stable results. In addition to this, the capability of the high dynamic range image reconstruction with approximation by using of projector camera system has been shown.

消費電力を考慮したOLAP問合せ処理

福澤優, 宮崎純, 藤澤誠, 天野敏之, 加藤博一

情報処理学会関西支部支部大会講演論文集(CD-ROM) 2011.09.22

ユーザの嗜好を考慮した情報推薦のためのSkyline演算の拡張

吉武亮, 宮崎純, 藤澤誠, 天野敏之, 加藤博一

情報処理学会関西支部支部大会講演論文集(CD-ROM) 2011.09.22

モデルベースドトラッキングにおける推定可能自由度の変化に対するロバスト性の向上

熊谷謙造, 山本豪志朗, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

日本バーチャルリアリティ学会大会論文集(CD-ROM) 2011.09.20

プロジェクタカメラ系による照明制御を用いた物体認識

天野敏之, 長村一樹, 藤澤誠, 宮崎純, 加藤博一

画像の認識・理解シンポジウム(MIRU2011)論文集 2011.07.20

粒子法におけるサブパーティクルスケールの乱流を考慮した液体表面生成手法

三村豪, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

Visual Computing/グラフィクスとCAD合同シンポジウム予稿集(CD-ROM) 2011.06.24

Development of Augmented Reality Dictionary System and Evaluation of Effectiveness

FUJIMOTO Yuichiro, TSE Chu Cheng, FUJISAWA Makoto, AMANO Toshiyuki, MIYAZAKI Jyun, KATO Hirokazu

IEICE technical report 2011.01.13

Augmented Reality (AR) is a display technology which can overlay virtual objects or information on top of the real world. One effective application of such technology is user 's activities support such as machine operation, maintenance and assembly. In addition, AR can integrate real and virtual world seamlessly even if the information is not associated with a specific location in the real world. Since this type of AR information display can also be achieved using other methods therefore it is necessary to compare these methods and evaluate their effectiveness. In this research, we implemented an AR-Dictionary system as an example of such information display. We also aim to evaluate its effectiveness by comparing it to different types of information display methods.

Development of Augmented Reality Dictionary System and Evaluation of Effectiveness

藤本雄一郎, TSE Chu Cheng, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

電子情報通信学会技術研究報告 2011.01.13

A Improvement of Fluid Surface using SPH for Turbulent Flow

三村豪, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

情報処理学会研究報告(CD-ROM) 2010.10.15

A Method to Display Connelation Between Images by Using a Recursive Photomosaic System

佐藤和男, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

情報処理学会研究報告(CD-ROM) 2010.10.15

A Improvement of Fluid Surface using SPH for Turbulent Flow

MIMURA GO, FUJISAWA MAKOTO, AMANO TOSHIYUKI, MIYAZAKI JUN, KATO HIROKAZU

情報処理学会研究報告. グラフィクスとCAD研究会報告 2010.09.08

乱流まで考慮した現実的な液体アニメーションを高速に生成することは非常に困難である．そこで，本論文では，粒子法においてパーティクルをサブパーティクルに分割することによりシミュレーションでは捉えきれない速度場，粒子配置を再現し，乱流を含む流れの流体表面を生成する手法を提案する．また，その実装を行い，提案手法の有効性を検証する．It is very difficult to make an animation of liquid with turbulence. In this paper, we propose a method to generate fluid surface including turbulent flow by dividing fluid particles into sub-particles and updating positions and velocity field that usual particle method can not capture. We implement this method and verify it.

A Method to Display Connelation Between Images by Using a Recursive Photomosaic System

SATOH KAZUO, FUJISAWA MAKOTO, AMANO TOSHIYUKI, MIYAZAKI JUN, KATO HIROKAZU

情報処理学会研究報告. EC, エンタテインメントコンピューティング = IPSJ SIG technical reports 2010.08.23

フォトモザイクは小さな写真を並べて一つの絵に見立てるモザイク画の一種である．本論文では，再帰的なフォトモザイクと写真に付加されたメタデータを用い，より多くの関連性のある写真をユーザに対して有効に提示させる方法を提案する．Photomosaic consists of a mosaic in which a big picture is constructed using many small pictures. In this paper, we propose a method to effectively display correlated images by using a recursive photomosaic system and by adding metadata to the images.

Estimating a rotation matrix R by using higher-order matrices R<sup>N</sup> with application to supervised pose estimation

Toru Tamaki, Bisser Raytchev, Kazufumi Kaneda, Toshiyuki Amano

Proceedings of the 7th IASTED International Conference on Signal Processing, Pattern Recognition and Applications, SPPRA 2010 2010.07.20

In this paper, we show that a more accurate estimation of a 3 × 3 rotation matrix R can be achieved by appropriately decomposing higher-order rotation matrices: R2, R3, and so on. First we discuss an angle estimation of a 2 × 2 rotation matrix inspired by the Electronic Distance Measurement. Then we reformulate the problem for a 3 x 3 rotation matrix: if noise-contaminated measurement matrices R, R2,⋯, Rn are given, find an appropriate rotation matrix R. In the proposed method, the given measurement matrices are first transformed to rotation matrices by using the polar decomposition. Then the rotation angles are obtained by using an eigen decomposition of the rotation matrices. Finally, the ambiguity of the obtained rotation angle is removed. Experimental results show that the use of Rn results in more accurate estimates than when R itself is used.

HANDY : Enhancing Social Presence in Video-chat Systems

DE SOUZA ALMEIDA IGOR, POLO JORDI CARRES, OIKAWA MARINA ATSUMI, HIRAMATSU TATSUYA, AMANO TOSHIYUKI, FUJISAWA MAKOTO, BILLINGHURST MARK, KATO HIROKAZU

情報処理学会研究報告. EC, エンタテインメントコンピューティング = IPSJ SIG technical reports 2010.05.15

既存のビデオ会話システムは，対面でのコミュニケーションを模倣しようとしている．その目的は，自分が遠方にいる話し相手と一緒にいるんだと感じれるようなソーシャルプレセンスの感覚を実現することにある．我々は，ビデオ会話システムにおいて，このソーシャルプレセンスを向上させるために，一方のユーザの存在感を他方のユーザの映像の中に感じさせる方法を提案する．実際にこのアイデアに基づく試作システムを構築し，試験的な評価実験を行ったので，それについて報告する．Current video-chat systems try to copy face-to-face communication. They aim to create a sense of social presence, which is the feeling that one person is present with another person in a mediated environment. We propose an approach to enhance social presence of video-mediated systems by simulating the presence of one user into the second user's video image. A prototype was implemented and a pilot study was conducted.

SNSに基づく災害ボランティア活動支援システムの構築

平松達也, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

ヒューマンインタフェース学会研究報告集(CD-ROM) 2010

共同活動空間におけるメンバーの位置情報履歴を用いたアウェアネス支援

庄司賢太, 藤澤誠, 天野敏之, 宮崎純, 加藤博一

ヒューマンインタフェース学会研究報告集(CD-ROM) 2010

An appearance based fast linear pose estimation

Toshiyuki Amano, Toru Tamaki

Proceedings of the 11th IAPR Conference on Machine Vision Applications, MVA 2009 2009.12.01

In this paper, we describe a high speed and multi degree of freedoms (DOFs) pose estimation method for a 3-D object that called Estimation-by-Completion (EbC) method. The most of employed processes are described with linear calculation, thus, whole procedure for each parameter estimation is expressed by a pair inner production, and it used only an arctangent calculation at the final part of the estimation. The accuracy evaluation by 3DOFs pose estimation that includes rotation around the object's vertical and horizontal translations is shown in experiment. We also describe its calculation cost in discussions.

Linear 3-D object pose estimation with dense sample images -Discussions about limitation of parameter estimation ability by the linear regressions

Toshiyuki Amano, Toru Tamaki

Proceedings of the 11th IAPR Conference on Machine Vision Applications, MVA 2009 2009.12.01

In the image parameter estimation by the linear regression, it has very high degrees of freedom for the decision of regression coefficients, because the dimension of image vector is huge high. In this paper, we discuss its potential by the learning of the dense samples. For the learning process, we employed a sequential regression coefficient calculation algorithm and realize its calculation for dense samples with reasonable computational cost. Through the experimental result, we discuss about the limitation of parameter estimation ability by the linear regression.

Experiments using 100 Objects: Comparison of Pose Representations for Pose Estimations

田中聡子, 原田健吾, 玉木徹, RAYTCHEV Bisser, 金田和文, 天野敏之

電子情報通信学会技術研究報告 2009.10.15

We report a result of an experimental study on properties of pose representations for 3DOF linear pose estimations with 100 CG objects. We use linear regression as a pose estimation method. First, we explain a method of linear pose estimation and two properties of pose representations. Next, we use four pose representations (rotation matrix, ZYX Euler angle, exponential map, and quaternions), for pose estimation experiments, and compare estimation errors. We show that estimation error of rotation matrix is significantly smaller than other representations by using pairwise t-test.

Can R^n estimate a rotation matrix R more accurately than R ? : A method for estimating a rotation matrix R by using R,R^2,R^3,..., obtained by an one-shot measurement

玉木徹, 天野敏之, 金田和文

MIRU 2009 画像の認識・理解シンポジウム論文集 2009.07

In this paper, we show that a more accurate estimation of a 3*3 rotation matrix R can be achieved by appropriately decomposing higher-order rotation matrices: R^2,R^3, and so on. First we discuss an angle estimation of a 2*2 rotation matrix inspired by the Electronic Distance Measurement. Then we reformulate the problem fora 3*3 rotation matrix: if noise-contaminated measurement matrices R,R^2,...,R^n are given, find an appropriate rotation matrix R. In the proposed method, the given measurement matrices are first transformed to rotation matrices by using the polar decomposition. Then the rotation angles are obtained by using an eigen decomposition of the rotation matrices. Finally, the ambiguity of the obtained rotation angle is removed. Numerical simulations and pose estimation experiments show that the use of R^n results in more accurate estimates than when R itself is used.

　 PREV - NEXT 　

<AMANO Toshiyuki>

Conference Activities & Talks -