| Presentation | 2000/12/14 LONG-TERM EFFECT REMOVAL FOR NOISY SPEECH RECOGNITION J. Chen, K.K. Paliwal, T. Matsui, K. Yao, K.P. Markov, S. Nakamura, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Noise speech recognition is of great interests in speech research recently. To make an automatic speech recognition system robust to noise, we will probably have to solve two problems. One is the detection and identification of noise. Another is the consideration of noise effect during recognition process. In this paper, we will address a new method to estimate the noise effect using a long-term Fourier analysis. We will then discuss how to remove the noise effect from corrupted speech to make recognition system immune to uncertainties. The rationale behind our noise estimation and removal approach can be described as follows. Speech signal is a nonstationary stochastic process. Much phonetic information in speech is encoded inthe changes of the speech spectrum over time. Relatively less phonetic information is encapsulated in the long-term speech spectrum. Noise, however can be treated as a stationary process. Long-term spectrum will provide a good estimate of noise. Hence the subtraction of long-term effect from short-term spectra will keep the discrimination information which is necessary for speech recognition, and meanwhile remove the noise effect. We will report on experiments on DARPA speech in noise environments evaluation (SPINE) database to demonstrate the properties of the proposed approach. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | speech recognition / noise subtraction / long-term power spectrum / noise estimation |
| Paper # | NLC2000-29,SP2000-77 |
| 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 | ENG |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | LONG-TERM EFFECT REMOVAL FOR NOISY SPEECH RECOGNITION |
| Sub Title (in English) | |
| Keyword(1) | speech recognition |
| Keyword(2) | noise subtraction |
| Keyword(3) | long-term power spectrum |
| Keyword(4) | noise estimation |
| 1st Author's Name | J. Chen |
| 1st Author's Affiliation | School of Microelectronic Engineering, Griffith University:ATR Spoken Language Translation Research Laboratories() |
| 2nd Author's Name | K.K. Paliwal |
| 2nd Author's Affiliation | School of Microelectronic Engineering, Griffith University:ATR Spoken Language Translation Research Laboratories |
| 3rd Author's Name | T. Matsui |
| 3rd Author's Affiliation | ATR Spoken Language Translation Research Laboratories |
| 4th Author's Name | K. Yao |
| 4th Author's Affiliation | ATR Spoken Language Translation Research Laboratories |
| 5th Author's Name | K.P. Markov |
| 5th Author's Affiliation | ATR Spoken Language Translation Research Laboratories |
| 6th Author's Name | S. Nakamura |
| 6th Author's Affiliation | ATR Spoken Language Translation Research Laboratories |
| Date | 2000/12/14 |
| Paper # | NLC2000-29,SP2000-77 |
| Volume (vol) | vol.100 |
| Number (no) | 522 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |