|
naistar (NAIST Academic Repository) >
学術リポジトリ naistar / NAIST Academic Repository naistar >
国際会議発表論文 / Proceedings >
情報科学研究科 / Graduate School of Information Science >
Please use this identifier to cite or link to this item:
http://hdl.handle.net/10061/8106
|
| Title: | Spectral Subtraction in Noisy Environments Applied to Speaker Adaptation Based on HMM Sufficient Statistics |
| Authors: | Shingo Yamade
Kanako Matsunami
Akira Baba
Akinobu Lee
Hiroshi Saruwatari
Kiyohiro Shikano |
| Issue Date: | Sep-2002 |
| Start page: | 1045 |
| End page: | 1048 |
| Abstract: | Noise and speaker adaptation techniques are essential to realize robust speech recognition in real noisy environments . In this paper, we applied spectral subtraction to an unsupervised speaker adaptation algorithm in noisy environments. The adaptation algorithm consists of the following five steps. (1) Spectral subtraction is carried out for noise added database. (2) Noise matched acoustic models are trained by using noise added speech database. (3) HMM sufficient statistics for each speaker are calculated from noise added speech database, and stored. (4) According to one arbitrary utterance, speakers close to a test speaker are selected by using speaker GMMs. (5) Speaker adapted acoustic models are constructed from HMM sufficient statistics of the selected speakers. We evaluated our unsupervised speaker adaptation algorithm in noisy environments in the 20k dictation task. The recognition experiments show that our speaker adapted acoustic model can achieve 82% word accuracy in 20dB SNR, which is about 6% higher than that of the noise matched models trained by Forward-Backward algorithm. |
| Description: | ICSLP2002: the 7th International Conference on Spoken Language Processing , September 16-20, 2002, Denver, Colorado, USA. |
| URI: | http://hdl.handle.net/10061/8106 |
| Rights: | Copyright 2002 ISCA |
| Text Version: | Publisher |
| Appears in Collections: | 情報科学研究科 / Graduate School of Information Science
|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
|
