Interfering noise severely degrades the performance of a speech recognition system. The Parallel Model Combination (PMC) technique is one of the most efficient techniques for dealing with such noise. Another method is to use features local in the frequency domain. Recently, we proposed Mel-Frequency Discrete Wavelet Coefficients (MFDWCs) as speech features local in frequency domain. In this paper, we discuss using PMC along with MFDWC features to take advantage of both noise compensation and local features (MFDWCs) to decrease the effect of noise on recognition performance. In addition we discuss the effect of increasing the number of the noise model mixture component on the performance of the Mel-Frequency Cepstral Coefficients (MFCCs) and MFDWCs. We evaluate the performance of MFDWCs using the NOISEX-92 database for various noise types and noise levels. We also compare the performance of these versus MFCCs and both using PMC for dealing with additive noise.
Cite as: Tufekci, Z., Gowdy, J.N., Gurbuz, S., Patterson, E. (2001) Applying parallel model compensation with mel-frequency discrete wavelet coefficients for noise-robust speech recognition. Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001), 873-876, doi: 10.21437/Eurospeech.2001-266
@inproceedings{tufekci01_eurospeech,
author={Zekeriya Tufekci and John N. Gowdy and Sabri Gurbuz and E. Patterson},
title={{Applying parallel model compensation with mel-frequency discrete wavelet coefficients for noise-robust speech recognition}},
year=2001,
booktitle={Proc. 7th European Conference on Speech Communication and Technology (Eurospeech 2001)},
pages={873--876},
doi={10.21437/Eurospeech.2001-266}
}