Exploiting principal component analysis in modulation spectrum enhancement for robust speech recognition

In this paper, we present a novel method to improve the noise robustness of speech features based on principal component analysis (PCA). The PCA process is employed to extract a set of basis spectral vectors for the modulation spectra of clean training speech features. The new modulation spectra of the speech features, constructed by mapping the original modulation spectra into the space spanned by these PCA-derived basis vectors, have shown robustness against the noise distortion. The experiments conducted on the Aurora-2 digit string database revealed that the proposed PCA-based approach, together with mean and variance normalization (MVN), can provide average error reduction rates of over 65% and 12% relative as compared with the baseline MFCC system and that using the MVN method alone, respectively.