This paper describes new method for the speaker adaptation of HMM parameters in environments with background noise. This method is based on Bayesian estimation, and calculates the a posteriori distribution of clean-speech HMM parameters from their a priori distribution by using noisy speech observations. The advantage of the method is that the distribution of the noise can be taken into account in adapting clean-speech HMMs to a target speakers speech without noise. The results of the experiments using noninformative prior show that the recognition performance in a noise-free environment was improved by this method even when the SNR of the noisy speech data used for the adaptation was -6 dB.
Cite as: Iwahashi, N., Kawasaki, A. (2000) Speaker adaptation in noisy environments based on parameter estimation using uncertain data. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 4, 528-531, doi: 10.21437/ICSLP.2000-865
@inproceedings{iwahashi00b_icslp,
author={Naoto Iwahashi and Akihiko Kawasaki},
title={{Speaker adaptation in noisy environments based on parameter estimation using uncertain data}},
year=2000,
booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
pages={vol. 4, 528-531},
doi={10.21437/ICSLP.2000-865}
}