Presentation 2019-03-15
[Poster Presentation] An Evaluation of Underdetermined Source Separation Based on Multichannel Variational Autoencoder
Shogo Seki, Hirokazu Kameoka, Li Li, Tomoki Toda, Kazuya Takeda,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper deals with a multichannel audio source separation problem under underdetermined conditions. Multichannel Non-negative Matrix Factorization (MNMF) is one of powerful approaches, which adopts the NMF concept for source power spectrogram modeling. This concept is also employed in Independent Low-Rank Matrix Analysis (ILRMA), a special class of the MNMF framework formulated under determined conditions. These methods work reasonably for particular types of sound sources, however, one limitation is that they can fail to work for sources with spectrograms that do not comply with the NMF model. To address this limitation, an extension of ILRMA called the Multichannel Variational Autoencoder (MVAE) method was recently proposed, where a Conditional VAE (CVAE) is used instead of the NMF model for source power spectrogram modeling. This approach has shown to perform impressively in determined source separation tasks thanks to the representation power of DNNs. This paper generalizes MVAE originally formulated under determined mixing conditions so that it can also deal with underdetermined cases. The proposed method was evaluated on an underdetermined source separation task of separating out three sources from two microphone inputs. Experimental results revealed that the generalized MVAE method achieved better performance than the MNMF method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Underdetermined source separation / Multichannel variational autoencoder / Multichannel non-negative matrix factorization
Paper # EA2018-154,SIP2018-160,SP2018-116
Date of Issue 2019-03-07 (EA, SIP, SP)

Conference Information
Committee EA / SIP / SP
Conference Date 2019/3/14(2days)
Place (in Japanese) (See Japanese page)
Place (in English) i+Land nagasaki (Nagasaki-shi)
Topics (in Japanese) (See Japanese page)
Topics (in English) Engineering/Electro Acoustics, Signal Processing, Speech, and Related Topics
Chair Suehiro Shimauchi(Kanazawa Inst. of Tech.) / Shogo Muramatsu(Niigata Univ.) / Yoichi Yamashita(Ritsumeikan Univ.)
Vice Chair Kenichi Furuya(Oita Univ.) / Kanji Watanabe(Akita Pref. Univ.) / Naoyuki Aikawa(TUS) / Kazunori Hayashi(Osaka City Univ) / Akinobu Ri(Nagoya Inst. of Tech.)
Secretary Kenichi Furuya(Shizuoka Inst. of Science and Tech.) / Kanji Watanabe(NHK) / Naoyuki Aikawa(Takushoku Univ.) / Kazunori Hayashi(Hiroshima Univ.) / Akinobu Ri(Kyoto Univ.)
Assistant Keisuke Imoto(Ritsumeikan Univ.) / Daisuke Morikawa(Toyama Pref Univ.) / Katsumi Konishi(Hosei Univ.) / hyihsin(Takushoku Univ.) / Tomoki Koriyama(Tokyo Inst. of Tech.) / Satoshi Kobashikawa(NTT)

Paper Information
Registration To Technical Committee on Engineering Acoustics / Technical Committee on Signal Processing / Technical Committee on Speech
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) [Poster Presentation] An Evaluation of Underdetermined Source Separation Based on Multichannel Variational Autoencoder
Sub Title (in English)
Keyword(1) Underdetermined source separation
Keyword(2) Multichannel variational autoencoder
Keyword(3) Multichannel non-negative matrix factorization
1st Author's Name Shogo Seki
1st Author's Affiliation Nagoya University(Nagoya Univ.)
2nd Author's Name Hirokazu Kameoka
2nd Author's Affiliation Nippon Telegraph and Telephone Corporation(NTT)
3rd Author's Name Li Li
3rd Author's Affiliation University of Tsukuba(Univ. Tsukuba)
4th Author's Name Tomoki Toda
4th Author's Affiliation Nagoya University(Nagoya Univ.)
5th Author's Name Kazuya Takeda
5th Author's Affiliation Nagoya University(Nagoya Univ.)
Date 2019-03-15
Paper # EA2018-154,SIP2018-160,SP2018-116
Volume (vol) vol.118
Number (no) EA-495,SIP-496,SP-497
Page pp.pp.323-328(EA), pp.323-328(SIP), pp.323-328(SP),
#Pages 6
Date of Issue 2019-03-07 (EA, SIP, SP)