Speaker recognition using mfcc and hybrid model of VQ and GMM

Abstract

Speaker recognition is widely used for automatic authentication of speaker's identity based on human biological features. Speaker recognition extracts, characterizes and recognizes the information about speaker identity. For feature extraction and speaker model ing many algorithms are being used. In this paper, we have proposed speaker recognition system based on hybrid approach using Mel Frequency Cepstrum Coefficient (MFCC) as feature extraction and combination of vector quantization (VQ) and Gaussian Mixture Modeling (GMM) for speaker modeling. Our approach is able to recognize speaker for both text dependent and text independent speech and uses relative index as confidence measures in case of contradiction in recognition process by GMM and VQ. Simulation results highlight the efficacy of proposed method compared to earlier work.