Speaker normalized spectral subband parameters for noise robust speech recognition

This paper proposes speaker normalized spectral subband centroids (SSCs) as supplementary features in noise environment speech recognition. SSCs are computed as frequency centroids for each subband from the power spectrum of the speech signal. Since the conventional SSCs depend on the formant frequencies of a speaker, we introduce a speaker normalization technique into SSC computation to reduce the speaker variability. Experimental results on spontaneous speech recognition show that the speaker normalized SSCs are more useful as supplementary features for improving the recognition performance than the conventional SSCs