Presentation 2012-03-16
Speech Recognition under Noisy Environment using Bidirectional OM-LSA Noise Suppression
Yasunari OBUCHI, Masahito TOGAMI, Ryu TAKEDA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a new noise suppression method, which is an extention of optimally-modified log-spectrum amplitude (OM-LSA) noise suppression with the bi-directional processing on the time axis. We also show the evaluation results which confirm the effectiveness of the proposed method. The proposed method works better than standard OM-LSA in the case where the estimation error of OM-LSA accumulates over time. A simple look-ahead mechanism is also implemented in each uni-directional estimation, and makes the estimation more robust. The evaluation experiments are carried out using our proprietary database and CENSREC-2, and the results approved the effectiveness of the proposed method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Noise suppression / OM-LSA / Bidirectional / Speech recognition / CENSREC-2
Paper # EA2011-126
Date of Issue

Conference Information
Committee EA
Conference Date 2012/3/9(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Engineering Acoustics (EA)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Speech Recognition under Noisy Environment using Bidirectional OM-LSA Noise Suppression
Sub Title (in English)
Keyword(1) Noise suppression
Keyword(2) OM-LSA
Keyword(3) Bidirectional
Keyword(4) Speech recognition
Keyword(5) CENSREC-2
1st Author's Name Yasunari OBUCHI
1st Author's Affiliation Central Research Laboratory()
2nd Author's Name Masahito TOGAMI
2nd Author's Affiliation Central Research Laboratory
3rd Author's Name Ryu TAKEDA
3rd Author's Affiliation Central Research Laboratory
Date 2012-03-16
Paper # EA2011-126
Volume (vol) vol.111
Number (no) 490
Page pp.pp.-
#Pages 6
Date of Issue