To make an automatic speech recognition system robust with respect 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 investigate several noise estimation approaches, such as moving average, long-term average, long-term Fourier analysis, etc. We will then introduce a sub-band based scheme to remove the noise effect from corrupted speech to make recognition system immune to additive noise. We will report on experiments on TI digits database and NOISEX database to justify the proposed approach.
Cite as: Chen, J., Paliwal, K.K., Nakamura, S. (2001) Sub-band based additive noise removal for robust speech recognition. Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001), 571-574, doi: 10.21437/Eurospeech.2001-152
@inproceedings{chen01d_eurospeech,
author={Jingdong Chen and Kuldip K. Paliwal and Satoshi Nakamura},
title={{Sub-band based additive noise removal for robust speech recognition}},
year=2001,
booktitle={Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001)},
pages={571--574},
doi={10.21437/Eurospeech.2001-152}
}