Presentations -
-
Restoration of Temporal Image Sequence from a Single Image Captured by a Correlation Image Sensor International conference
K. Kawade, A. Wakita, T. Yokota, H. Hontani, and S. Ando
International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP) 2017
Event date： 2017.02 - 2017.03
Language：English Presentation type：Oral presentation (general)
Venue：ポルトガル
-
Automatic detection of brachial artery in ultrasound images for a flow meditated dilation test International conference
K. Sano, H. Masuda, K. Sano, H. Suzuki, T. Koyama, T. Yokota, and H. Hontani
International Forum on Medical Imaging in Asia (IFMIA) 2017
Event date： 2017.01
Language：English Presentation type：Poster presentation
Venue：沖縄
-
アミロイドβ経時変化モデル構築のためのSparse NMFを用いたPET画像解析
永田達也，本谷秀堅， 横田達也，木村裕一，伊藤康一，加藤隆司，岩田香織，中村昭範
MI研究会
Event date： 2017.01
Language：Japanese Presentation type：Oral presentation (general)
-
空間の低ランク性と平滑性を考慮したフーリエ係数最適化によるMR超解像
河村直輝， 横田達也，本谷秀堅
MI研究会
Event date： 2017.01
Language：Japanese Presentation type：Oral presentation (general)
-
低ランク非負行列分解を用いたダイナミックPET同時再構成の初期検討
山田純也，本谷秀堅， 横田達也，坂田宗之，木村裕一
MI研究会
Event date： 2016.11
Language：Japanese Presentation type：Oral presentation (general)
-
低ランク性および平滑性を用いたテンソルデータ補完 Invited
横田達也
医用画像研究会 電子情報通信学会
Event date： 2016.09
Language：Japanese Presentation type：Oral presentation (invited, special)
Venue：東京農工大学
-
時間相関イメージセンサで計測した静止画からのTV-正則化による高精度動画再構成
脇田章裕，川出康平，横田達也，本谷秀堅，安藤繁
コンピュータビジョンとイメージメディア研究会 情報処理学会
Event date： 2016.09
Language：Japanese Presentation type：Oral presentation (general)
Venue：富山大学
-
-
Multilinear rank selection for denoising and dimensionality reduction of multiway data
Namgil Lee, Tatsuya Yokota, Andrzej Cichocki
The Korean Statistical Society Spring Conference 2016 Korean Statistical Society
Event date： 2016.05
Presentation type：Oral presentation (general)
Venue：Deagu, South Korea
-
Tensor Completion via Functional Smooth Component Deflation International conference
T. Yokota, A. Cichocki
The 41st IEEE International Conference on Acoustics, Speech and Signal Processing
Event date： 2016.03
Language：English Presentation type：Poster presentation
Venue：Shanghai, China
-
A fast automatic rank determination algorithm for noisy low-rank matrix completion International conference
T. Yokota, A. Cichocki
Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC 2015)
Event date： 2015.12
Language：English Presentation type：Oral presentation (general)
Venue：Hong-Kong
-
Smooth PARAFAC Decomposition for Tensor Completion International conference
T. Yokota, A. Cichocki
Low-rank Optimization and Applications
Event date： 2015.06
Language：English Presentation type：Poster presentation
Venue：Bonn, Germany
-
Multilinear tensor rank estimation via sparse Tucker decomposition International conference
T. Yokota, A. Cichocki
Joint 7th International Conference on Soft Computing and Intelligent Systems and 15th International Symposium on Advanced Intelligent Systems (SCIS&ISIS2014)
Event date： 2014.12
Language：English Presentation type：Oral presentation (general)
Venue：Kita-Kyushu
-
Linked Tucker2 decomposition for flexible multi-block data analysis International conference
T. Yokota, A. Cichocki
21st International Conference on Neural Information Processing
Event date： 2014.11
Language：English Presentation type：Oral presentation (general)
Venue：Kuching, Malaysia
-
B-Spline Smoothing of Feature Vectors in Nonnegative Matrix Factorization International conference
R. Zdunek, A. Cichocki, T. Yokota
International Conference on Artificial Intelligence and Soft Computing
Event date： 2014.06
Language：English Presentation type：Oral presentation (general)
Venue：Zakopane, Poland
-
多カーネルを用いた二次制約MAPによるパラメータレスの識別器の実現
鷲沢嘉一, 横田達也, 山下幸彦
第28回 信号処理シンポジウム
Event date： 2013.11
Language：Japanese Presentation type：Poster presentation
Venue：海峡メッセ下関 国際貿易ビル
-
Inter-subject Common Spatial Pattern Filter for Motor Imagery based Brain Computer Interface
YOKOTA Tatsuya, YAMASHITA Yukihiko, CICHOCKI Andrzej
The Institute of Electronics, Information and Communication Engineers
Event date： 2013.11
Language：Japanese Presentation type：Poster presentation
Motor imagery based brain computer interface is a technique to control some device (e.g., wheelchair) through motor imagery brain signals. In order to achieve this, supervised learning techniques, which use enough numbers of training samples, are often applied to learn the parameters of spacial filter. However, the recording process of the training samples of motor imagery brain signals requires great time and effort. Our goal is to reduce this load by proposing new parameter learning methods. In this paper, we propose new algorithms to learn the parameters of common spatial pattern filter by using dataset of multiple subject for the motor imagery based brain computer interface, called the "inter-subject common spatial pattern (ISCSP) filter". Use of the ISCSP filter can reduce the load of user to record user's motor imagery brain signals, since it needs to use a few or no samples of user's motor imagery data. In experiments, we show the validities of our methods.
-
Heteroscedastic Gaussian based Correction term for Fisher Discriminant Analysis and Its Kernel Extension International conference
T. Yokota, T. Wakahara, Y. Yamashita
2013 International Joint Conference on Neural Networks
Event date： 2013.08
Language：English Presentation type：Oral presentation (general)
Venue：Dallas, USA
-
Correction Term of Fisher Discriminant Analysis Based on Normal Distribution
YOKOTA Tatsuya, WAKAHARA Toru, SAKANO Hitoshi, YAMASHITA Yukihiko
Technical report of IEICE. PRMU The Institute of Electronics, Information and Communication Engineers
Event date： 2013.03
Language：Japanese Presentation type：Oral presentation (general)
Fisher's linear discriminant is one of the most basic feature extraction methods for pattern recognition. However, even if the distributions of patters is normal but if the correlation matrices of patterns are not the same among categories, it does not provide the best features to classify the patterns. In this paper, we assume that the distributions of patterns are normal and provide two criteria, the one using the distribution of the rival categories as weight and the other based on the Chernoff distance of distributions. We derive corrections terms for the generalized eigenvalue problem for the case of one feature extraction in two class classification. Furthermore, we evaluate its performance by using 2-dimensional normal distributions and MNIST database.
-
Linked PARAFAC/CP Tensor Decomposition and Its Fast Implementation for Multi-block Tensor Analysis International conference
T. Yokota, A. Cichocki, Y. Yamashita
19th International Conference on Neural Information Processing
Event date： 2012.11
Language：English Presentation type：Oral presentation (general)
Venue：Doha, Qatar
