Comparison of Various Techniques for Speaker Recognition

In this paper, a comparison on different speaker recognition techniques is presented. The techniques we are describing here are Vector Quantization (VQ) by using Linde Bozo Gray (LBG), Hidden Markov Model (HMM) and Gaussian Mixture Model (GMM) using the iterative Expectation-Maximization (EM) algorithm. VQ adds the method of considering a large group of feature vectors of a known user and generating a smaller group of feature vectors that signifies the centroid of spreading, i.e. points set apart, so as to reduce the distance between the points. The GMM can be represented in the form of a summation of the VQ model where the clusters are overlying. HMM is a finite group of states, each of which is united with a probability distribution. These techniques are used with their respective algorithm to compute the accuracy rate of speaker recognition. Based on the theoretical analysis, the GMM is more precise as compared to the other techniques.

[1]  Biing-Hwang Juang,et al.  A vector quantization approach to speaker recognition , 1985, ICASSP '85. IEEE International Conference on Acoustics, Speech, and Signal Processing.

[2]  S. B. Dhonde,et al.  Comparison of Vector Quantization and Gaussian Mixture Model using Effective MFCC Features for Text-independent Speaker Identification , 2016 .

[3]  Huo Chun Bao,et al.  The research of speaker recognition based on GMM and SVM , 2012, 2012 International Conference on System Science and Engineering (ICSSE).

[4]  A. Hussain,et al.  Speaker Verification using Vector Quantization and Hidden Markov Model , 2007, 2007 5th Student Conference on Research and Development.

[5]  Manan Vyas.  A Gaussian Mixture Model Based Speech Recognition System Using Matlab , 2013 .

[6]  John M. Grace,et al.  Speaker Recognition using MFCC and Vector Quantization , 2015 .

[7]  S. Biswas,et al.  Speaker Identification Using Cepstral Based Features and Discrete Hidden Markov Model , 2007, 2007 International Conference on Information and Communication Technology.

[8]  Xiaoqin Zeng,et al.  An improved VQ based algorithm for recognizing speaker-independent isolated words , 2012, 2012 International Conference on Machine Learning and Cybernetics.

[9]  Scott Axelrod,et al.  Subspace constrained Gaussian mixture models for speech recognition , 2005, IEEE Transactions on Speech and Audio Processing.

[10]  Geeta Nijhawan,et al.  Speaker Recognition Using MFCC and Vector Quantisation , 2014 .