In this paper, a noise robust speech recognition method is proposed, by combining temporal domain singular value decomposition (SVD) based speech enhancement and Gaussian mixture model (GMM) based speech estimation. The critical neck of the GMM based approach is the noise estimation problem. For this noise estimation problem, we investigated the adaptive noise estimation in the GMM based approach. Furthermore, in order to obtain higher recognition accuracy, we employed a temporal domain SVD based speech enhancement method as the pre-processing module of the GMM based approach. In addition, to reduce more influence of the noise included in the noisy speech, we introduce an adaptive over subtraction factor into the temporal domain SVD based speech enhancement. In evaluation on the AURORA2 tasks, our method showed the significant improvement in the recognition accuracy at all the noise conditions.
展开▼