We report our recent work on noise-robust large-vocabulary speech recognition. Three key innovations are developed and evaluated in this work: 1) a new model learning paradigm that comprises a noise-insertion process followed by noise reduction; 2) a noise adaptive training algorithm that integrates noise reduction into probabilistic multi-style system training; and 3) a new algorithm (SPLICE) for noise reduction that makes no assumptions about noise stationarity. Evaluation on a large-vocabulary speech recognition task demonstrates significant and consistent error rate reduction using these techniques. The resulting error rate is shown to be lower than that achieved by the matched-noisy condition for both stationary and nonstationary natural, as well as simulated, noises.
Cite as: Deng, L., Acero, A., Plumpe, M., Huang, X. (2000) Large-vocabulary speech recognition under adverse acoustic environments. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 3, 806-809, doi: 10.21437/ICSLP.2000-657
@inproceedings{deng00d_icslp,
author={Li Deng and Alex Acero and Mike Plumpe and Xuedong Huang},
title={{Large-vocabulary speech recognition under adverse acoustic environments}},
year=2000,
booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
pages={vol. 3, 806-809},
doi={10.21437/ICSLP.2000-657}
}