The Study Of Speaker Recognition System Based On MFCC

Speaker recognition may be looked as one kind of voice recognition. And it is one of researching hotspots currently. Speaker recognition is the processing of automatically recognition who is speaker by using speaker specific information included in speech signal. It can be classified into speaker identification and speaker verification according to decision modes. With the development of communication and information technology,it is getting more and more attention for its bright future.The main works of this thesis are as follows:(1) This paper implements a speaker recognition system, which uses Mel- Frequency Cepstrum Coefficients (MFCC) that can reflect the human apperception to voice as feature parameters. The System respectively compares the recognition rates of MFCC, combination of difference MFCC and MFCC based on VQ and weighted VQ. The conclusion is that, for the speaker recognition, combination of difference MFCC and MFCC has higher recognition rates than MFCC.(2) We have modified the standard vector quantization (VQ) method and proposed a method with weight based on VQ in speaker recognition.(3)Artificial neural Network is able to simulate human brain to some extend. It provides a new method for speaker recognition. The article introduces the method of probabilistic neural Network (PNN). PNN is a high performance classified neural Network. The simulation shows that the classification accuracy rate of PNN to the training samples is very high, but the classification accuracy rate of PNN to the testing samples is lower.