Text Dependent Speaker Recognition System Using Vector Quantization

Speech is the most important and primary mode of communication among human being and also the most natural and efficient form of exchanging information among humans. Various fields for research in speech processing are Speech Recognition, Speaker Recognition, speech synthesis, speech coding etc. This paper presents a detailed study of text-dependent Speaker Recognition system used to identify an unknown speaker. This recognition system uses vector quantization (VQ) as the modeling technique. The features of the speech signal are extracted using Mel Frequency Cepstum Coefficients (MFCC) followed by the VQ technique. K-means clustering algorithm has been used to obtain the vector quantized codebook. Highest accuracy is obtained using hanning window and mel perceptual feature extraction realized with 35 filter bank. The accuracy also improves as the number of vectors in the VQ codebook is increased from 64 to 100.