Matlab-Based Intelligent Voiceprint Recognition System

摘要

In this paper, a complete voiceprint recognition based on Matlab was realized, including speech processing and feature extraction at early stage, and model training and recognition at later stage. For speech processing and feature extraction at early stage, Mel Frequency Cepstrum Coefficient (MFCC) was taken as feature parameter. For speaker model method, DTW model was adopted to reflect the voiceprint characteristics of speech, converting voiceprint recognition into speaker speech data evaluation, and breaking up complex speech training and matching into model parameter training and probability calculation. Simulation experiment results show that this system is effective to recognize voiceprint.