| Presentation | 2008-12-09 Two-channel input speech recognition using sparsness-based blind source separation Kenta NISHIKI, Yousuke IZUMI, Shinji WATANABE, Takuya NISHIMOTO, Nobutaka ONO, Shigeki SAGAYAMA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | This paper discusses a two-channel input speech recognition using a sparsness-based blind source separation. The target speech is extracted from observed signals under diffusive noises (e.g. reverberation) by the source separation technique where a time-frequency mask is dynamically designed for speech separation using the EM algorithm. Cepstral Mean Normalization is exploited to reduce a remaining distortions or a newly introduced distortions in separated speech features. In a connected digit recognition task with multiple noise sources, the proposed method drastically improved the word accuracy in anechoic and reverberant environments. The proposed method achieved higher performance especially in a reverberant environment than conventional methods. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | sparsness / 2-channel blind source separation / reverberation / speech recognition |
| Paper # | NLC2008-24,SP2008-79 |
| 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) | Two-channel input speech recognition using sparsness-based blind source separation |
| Sub Title (in English) | |
| Keyword(1) | sparsness |
| Keyword(2) | 2-channel blind source separation |
| Keyword(3) | reverberation |
| Keyword(4) | speech recognition |
| 1st Author's Name | Kenta NISHIKI |
| 1st Author's Affiliation | Department of Information Physics and Computing, University of Tokyo() |
| 2nd Author's Name | Yousuke IZUMI |
| 2nd Author's Affiliation | Department of Information Physics and Computing, University of Tokyo |
| 3rd Author's Name | Shinji WATANABE |
| 3rd Author's Affiliation | NTT Communication Science Laboratories |
| 4th Author's Name | Takuya NISHIMOTO |
| 4th Author's Affiliation | Department of Information Physics and Computing, University of Tokyo |
| 5th Author's Name | Nobutaka ONO |
| 5th Author's Affiliation | Department of Information Physics and Computing, University of Tokyo |
| 6th Author's Name | Shigeki SAGAYAMA |
| 6th Author's Affiliation | Department of Information Physics and Computing, University of Tokyo |
| Date | 2008-12-09 |
| Paper # | NLC2008-24,SP2008-79 |
| Volume (vol) | vol.108 |
| Number (no) | 337 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |