Speaker Recognition Research Based On GMM Speaker Clustering Technology

Speaker recognition is a biometrics that the identifier of a person can be recognized via his voice. It is applied to man-machine interface, ensure public security, military affairs, judicature, and so on. Speech signals pre-procession, point detection, feature extraction were discussed. The system on hierarchical speaker recognition based on GMM speaker clustering technology was set up. And then, the model of wavelet neural network was presented. The speaker recognition system using wavelet neural network as the classifier is constructed in this paper. The experiment results show that the proposed methods have faster training speed and higher recognition rate than the system based on BP neural network. The system on hierarchical speaker recognition has not only improved the speed of training, but also increased the running speed of recognition greatly, with the rate of recognition. The system based on hierarchical speaker recognition is useful and effective.