| Presentation | 2008-12-09 Noisy speech recognition using integrated method of statistical model-based voice activity detection and noise suppression Masakiyo FUJIMOTO, Kentaro ISHIZUKA, Tomohiro NAKATANI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | This paper addresses robust front-end processing for automatic speech recognition in noise. The proposed method integrates voice activity detection (VAD) and noise suppression, and consists of three core techniques, i.e., (1) statistical model sharing, (2) Wiener filter design by using speech/non-speech probabilities, and (3) VAD improvement by using enhance speech. In addition, the proposed method can perform sequential processing without frame delay. In an evaluation, the proposed method significantly improves accuracy of concatenated speech recognition without frame delay. In addition, we investigate to combine cepstrum mean normalization and sequential acoustic model adaptation with the proposed method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | integrated front-end processing / voice activity detection / noise suppression / sequential processing / speech recognition |
| Paper # | NLC2008-26,SP2008-81 |
| Date of Issue |
| Conference Information | |
| Committee | NLC |
|---|---|
| Conference Date | 2008/12/2(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 | Natural Language Understanding and Models of Communication (NLC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Noisy speech recognition using integrated method of statistical model-based voice activity detection and noise suppression |
| Sub Title (in English) | |
| Keyword(1) | integrated front-end processing |
| Keyword(2) | voice activity detection |
| Keyword(3) | noise suppression |
| Keyword(4) | sequential processing |
| Keyword(5) | speech recognition |
| 1st Author's Name | Masakiyo FUJIMOTO |
| 1st Author's Affiliation | NTT Communicaition Science Laboratories, NTT Corp.() |
| 2nd Author's Name | Kentaro ISHIZUKA |
| 2nd Author's Affiliation | NTT Communicaition Science Laboratories, NTT Corp. |
| 3rd Author's Name | Tomohiro NAKATANI |
| 3rd Author's Affiliation | NTT Communicaition Science Laboratories, NTT Corp. |
| Date | 2008-12-09 |
| Paper # | NLC2008-26,SP2008-81 |
| Volume (vol) | vol.108 |
| Number (no) | 337 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |