PHÂN TÍCH HIỆU NĂNG CỦA MÔ HÌNH ẨN MARKOV HỖN HỢP GAUSS TRONG HỆ THỐNG XÁC THỰC ĐỊNH DANH NGƯỜI NÓI

Abstract

This study presents the development and evaluation of a speaker verification system designed
to enhance biometric security. Speech signals are processed using MFCC feature extraction and modeled
with a Gaussian Mixture Hidden Markov Model (GMM-HMM), which effectively captures temporal and
acoustic variations in individual voices. The HTK toolkit is employed for both training and recognition
processes. Experimental results indicate that the proposed model achieves a low Equal Error Rate
(EER) and performs reliably, demonstrating its potential for practical deployment in real-time speaker
authentication systems