Multi-Layer Channel Normalization for Frequency-Dynamic Feature Extraction
Dong Wang · 2003
Despite the steady progress made in the area of speech recognition and a high number of practical applications, it is widely acknowledged that recognition technology today is not at the desired level.One main obstacle is what said 搑obustness?This paper focus on one popular idea in antagonizing speech system vulnerability-channel normalization,and presents a new normalization algorithm MLCN (multi-layer channel normalization),which exploits the recursive compensation progress in two domains (spectral domain and cepstral domain) to depress the noise and channel distortion, so that the more robust speech representation for the following processing is achieved.A new frequency-dynamic feature extraction algorithm is also proposed due to the introduction of MLCN,which allows dynamic information integrated in the final feature vectors.Experimental results of the gallina system demonstrate the validity of the new algorithm.