Presentation | 2023-03-01 Analysis of Noisy-target Training for DNN-based speech enhancement and investigation towards its practical use Takuya Fujimura, Tomoki Toda, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Deep neural network (DNN)-based speech enhancement usually uses a clean speech as a training target. However, it is hard to collect large amounts of clean speech because its recording is very costly. To relax this limitation, we proposed Noisy-target Training (NyTT) that utilizes noisy speech as a training target. It has been experimentally shown that NyTT can train a DNN without clean speech. However, sufficient investigations have not been conducted to clarify the reason why NyTT works, its detailed property, and the effectiveness of utilizing large amounts of noisy speech. In this paper, we conduct various analyses to deepen our understanding of NyTT. Based on the property of NyTT, we also propose a refined method that performs higher-quality speech enhancement. Furthermore, we investigate whether using a huge amount of noisy speech is effective for improving speech enhancement performance. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Single channel speech enhancement / Deep Neural Network / Unsupervised learning / Behavior analysis |
Paper # | EA2022-112,SIP2022-156,SP2022-76 |
Date of Issue | 2023-02-21 (EA, SIP, SP) |
Conference Information | |
Committee | SP / IPSJ-SLP / EA / SIP |
---|---|
Conference Date | 2023/2/28(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Tomoki Toda(Nagoya Univ.) / Tomoki Toda(Nagoya Univ.) / Kenichi Furuya(Oita Univ.) / Toshihisa Tanaka(Tokyo Univ. Agri.&Tech.) |
Vice Chair | / / Tatsuya Kako(NTT) / Junki Ono(Tokyo Metropolitan Univ.) / Koichi Ichige(Yokohama National Univ.) / Takayuki Nakachi(Ryukyu Univ.) |
Secretary | (NTT) / (Univ. of Electro-Comm.) / Tatsuya Kako(NTT) / Junki Ono(Univ. of Electro-Comm.) / Koichi Ichige(NTT) / Takayuki Nakachi(RitsumeikanUniv.) |
Assistant | Ryo Aihara(Mitsubishi Electric) / Daisuke Saito(Univ. of Tokyo) / Ryo Aihara(Mitsubishi Electric) / Daisuke Saito(Univ. of Tokyo) / Masato Nakayama(Osaka Sangyo Univ.) / Kouhei Yatabe(Tuat) / Taichi Yoshida(UEC) / Shoko Imaizumi(Chiba Univ.) |
Paper Information | |
Registration To | Technical Committee on Speech / Special Interest Group on Spoken Language Processing / Technical Committee on Engineering Acoustics / Technical Committee on Signal Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Analysis of Noisy-target Training for DNN-based speech enhancement and investigation towards its practical use |
Sub Title (in English) | |
Keyword(1) | Single channel speech enhancement |
Keyword(2) | Deep Neural Network |
Keyword(3) | Unsupervised learning |
Keyword(4) | Behavior analysis |
1st Author's Name | Takuya Fujimura |
1st Author's Affiliation | Nagoya University(Nagoya Univ.) |
2nd Author's Name | Tomoki Toda |
2nd Author's Affiliation | Nagoya University(Nagoya Univ.) |
Date | 2023-03-01 |
Paper # | EA2022-112,SIP2022-156,SP2022-76 |
Volume (vol) | vol.122 |
Number (no) | EA-387,SIP-388,SP-389 |
Page | pp.pp.221-226(EA), pp.221-226(SIP), pp.221-226(SP), |
#Pages | 6 |
Date of Issue | 2023-02-21 (EA, SIP, SP) |