Comparative Study of GMM, DTW, and Ann on Thai Speaker Identification System

摘要

This paper proposes a new investigation on Gaussian mixture model (GMM) by comparing it with some preliminary experiemtns on multilayered perceptron network (MLP) with backpropagation learning algorithm (BKP) and dynamic time warping (DTW) techniques o nthai test-dependent speaker identification system. Three major identification engines are conducted on 50 speakers with isolated digits 0-9. Training and testing utterances were recorded over a five week duration. Furhtermore, three well-known speech features, namely linear predictive coding derived cepstrum (LPCC), postfiltered ceptrum (PFL), and Mel frequency cepstral coefficient (MFCC) were evaluted. From our previous experimetns, the MFCC has given the highest identification rate on DTW and MLP. Therefore, GMM with MFCC feature was experimented and attained 87.54