The removal of noise from speech signals has applications ranging from speech enhancement for cellular communications to front ends for speech recognition systems. In this paper, we present a new nonlinear time-domain method called NoiseRegularized Adaptive Filtering.The approach is based on minimum mean-squared estimation using a modified cost function and allows designing both linear and nonlinear filters using only the observed noisy speech.
Cite as: Wan, E.A., Merwe, R.v.d. (1999) Noise-regularized adaptive filtering for speech enhancement. Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 2643-2646, doi: 10.21437/Eurospeech.1999-583
@inproceedings{wan99b_eurospeech, author={Eric A. Wan and Rudolph van der Merwe}, title={{Noise-regularized adaptive filtering for speech enhancement}}, year=1999, booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)}, pages={2643--2646}, doi={10.21437/Eurospeech.1999-583} }