Presentation 2000/12/14
Voice Activity Detection Using Non-speech Models and HMM Composition
Narimasa WATANABE, Takeshi YAMADA, Futoshi ASANO, Nobuhiko KITAWAKI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper proposes a new voice activity detection (VAD) methodusing non-speech models and HMM composition. The proposed method predicts a non-speech sound that overlaps with speech by a Viterbi algorithm with a speech model and non-speech models. Then, a Viterbi algorithm with composed models of the speech model and the predicted non-speech sound models is applied again. As a result, the overlap segments are detected. Furthermore, the non-speech sound that overlaps with speech and its SNR are identified. To evaluate the performance of the proposed method, preliminary experiments and conducted. These results showed that the VAD accuracy of the proposed method is improved by a maximum of 40 % compared to that of the conventional method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Voice activity detection / Viterbi alignment / non-speech models / HMM composition
Paper # NLC2000-36,SP2000-84
Date of Issue

Conference Information
Committee SP
Conference Date 2000/12/14(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 Speech (SP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Voice Activity Detection Using Non-speech Models and HMM Composition
Sub Title (in English)
Keyword(1) Voice activity detection
Keyword(2) Viterbi alignment
Keyword(3) non-speech models
Keyword(4) HMM composition
1st Author's Name Narimasa WATANABE
1st Author's Affiliation Master's Program in Sciences and Engineering, University of Tsukuba()
2nd Author's Name Takeshi YAMADA
2nd Author's Affiliation Institute of Information Science and Electronics, University of Tsukuba
3rd Author's Name Futoshi ASANO
3rd Author's Affiliation Electrotechnical Laboratory
4th Author's Name Nobuhiko KITAWAKI
4th Author's Affiliation Institute of Information Sciences and Electronics, University of Tsukuba
Date 2000/12/14
Paper # NLC2000-36,SP2000-84
Volume (vol) vol.100
Number (no) 522
Page pp.pp.-
#Pages 6
Date of Issue