The thesis presents and implements a novel Speaker Recognition System on PDA.

Prevention of impostors entering speaker recognition systems;

Application of the improved VQ arithmetic to speaker recognition;

Application of vector quantification in speaker recognition;

Speaker identification using modified radial basis function networks by genetic algorithm;

New codebook design method based on hybrid immune algorithm for text-independent speaker identification;

Research and experiment of speaker identification algorithms based on artificial neural network;

Speaker verification based on Cohort Similarity

Through analyzing some features that be used usually in speaker identification system nowadays,in order to improve the rate of identification,this paper puts forward a method that combining efficiently more features such as MFCC and LPCC and their one ranks coefficients and keynote period and so on to do speaker verification under Matlab 6.

Most stateoftheart speaker verification systems need a speaker independent(SI) background model,and use some adaptive method to train the background model.