Performance study of vector quantization methods (k-means, GMM) for arabic isolated word recognition system based on DHMM

This article present several techniques used in the design of an isolated Arabic words recognition system based on the discrete Hidden Markov Model. We present the results of an experimental study aimed at finding the effect of the quantization methods (k-means and GMM) on the recognition performance. we used in analysis phase Mel Frequency Cepstral Coefficients (MFCC), although experiments are carried out for the choice of the optimal parameters of the system. Good results are obtained using a GMM method.