We model noise as a stationary component plus a time varying residual. The stationary part is estimated off-line and compensated using Log-Add noise compensation. The time varying residual is estimated and compensated using a sequential EM algorithm. The residual noise compensation proceeds in parallel with the recognition process. Experimental results demonstrate that the proposed algorithm improves the recognition performance not only in highly nonstationary noise but also in slow-varying noise, compared with Log-Add noise compensation alone.
Cite as: Yao, K., Shi, B.E., Nakamura, S., Cao, Z. (2000) Residual noise compensation by a sequential EM algorithm for robust speech recognition in nonstationary noise. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 1, 770-773, doi: 10.21437/ICSLP.2000-191
@inproceedings{yao00_icslp,
author={Kaisheng Yao and Bertram E. Shi and Satoshi Nakamura and Zhigang Cao},
title={{Residual noise compensation by a sequential EM algorithm for robust speech recognition in nonstationary noise}},
year=2000,
booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
pages={vol. 1, 770-773},
doi={10.21437/ICSLP.2000-191}
}