Role of Spectral Peaks in Autocorrelation Domain for Robust Speech Recognition

This paper presents a new front-end for robust speech recognition. This new front-end scenario focuses on the spectral features of the filtered speech signals in the autocorrelation domain. The autocorrelation domain is well known for its pole preserving and noise separation properties. In this paper, a novel method for robust speech extraction is proposed in the autocorrelation domain. The proposed method is based on a novel rep-resentation of the speech signal corrupted by an additive noise. Initial filtering stage is used to reduce the additive noise when computing the speech features followed by extraction of the autocorrelation spectrum peaks. Robust features based on these peaks are derived by assuming that the corrupting noise is stationary in nature. A task of speaker-independent isolated-word recognition is used to demonstrate the efficiency of these robust features. The cases of white noise and colored noise such as factory, babble and F16 are tested. Experimental results show significant improvement in comparison to the results obtained using traditional front-end methods. Further enhancement has been done by applying cepstral mean normalization (CMN) to the above extracted features