Presentation | 2019-10-28 FastMNMF based on multivariant complex Student's t distribution for blind source separation Keigo Kamo, Yuki Kubo, Norihiro Takamune, Daichi Kitamura, Hiroshi Saruwatari, Yu Takahashi, Kazunobu Kondo, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | FastMNMF is a blind source separation technique, which is an accelerated algorithm of multichannel nonnegative matrix factorization under the assumption of jointly diagonalizable spatial covariance matrices. Source signals' spectrograms were assumed to follow a multivariate complex Gaussian distribution in FastMNMF. In this paper, we propose the model extension of FastMNMF to a multivariate complex Student's textit{t} distribution. We derive a new parameter update rule using the auxiliary-function-based method, and show that the proposed method outperforms the conventional Gaussian-model-based FastMNMF. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | blind source separation / spatial covariance model / multivariate complex Student's t distriburion |
Paper # | EA2019-40 |
Date of Issue | 2019-10-21 (EA) |
Conference Information | |
Committee | EA / ASJ-H |
---|---|
Conference Date | 2019/10/28(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | NHK Science&Technology Research Lab. |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Engineering/Electro Acoustics, Psychological and Physiological Acoustics, High-Reality Audio, Spatial Audio, and Related Topics |
Chair | Kenichi Furuya(Oita Univ.) / 小澤 賢司(山梨大) |
Vice Chair | Suehiro Shimauchi(Kanazawa Inst. of Tech.) / Shigeto Takeoka(Shizuoka Inst. of Science and Tech.) / 鵜木 祐史(北陸先端大) |
Secretary | Suehiro Shimauchi(NHK) / Shigeto Takeoka(Univ. of Tokyo) / 鵜木 祐史(NTT) |
Assistant | Keisuke Imoto(Ritsumeikan Univ.) / Daisuke Morikawa(Toyama Pref Univ.) |
Paper Information | |
Registration To | Technical Committee on Engineering Acoustics / Auditory Research Meeting |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | FastMNMF based on multivariant complex Student's t distribution for blind source separation |
Sub Title (in English) | |
Keyword(1) | blind source separation |
Keyword(2) | spatial covariance model |
Keyword(3) | multivariate complex Student's t distriburion |
1st Author's Name | Keigo Kamo |
1st Author's Affiliation | The University of Tokyo(UTokyo) |
2nd Author's Name | Yuki Kubo |
2nd Author's Affiliation | The University of Tokyo(UTokyo) |
3rd Author's Name | Norihiro Takamune |
3rd Author's Affiliation | The University of Tokyo(UTokyo) |
4th Author's Name | Daichi Kitamura |
4th Author's Affiliation | National Institute of Technology, Kagawa Collage(Kagawa NCIT) |
5th Author's Name | Hiroshi Saruwatari |
5th Author's Affiliation | The University of Tokyo(UTokyo) |
6th Author's Name | Yu Takahashi |
6th Author's Affiliation | Yamaha Corporation(Yamaha) |
7th Author's Name | Kazunobu Kondo |
7th Author's Affiliation | Yamaha Corporation(Yamaha) |
Date | 2019-10-28 |
Paper # | EA2019-40 |
Volume (vol) | vol.119 |
Number (no) | EA-253 |
Page | pp.pp.23-29(EA), |
#Pages | 7 |
Date of Issue | 2019-10-21 (EA) |