Speaker Recognition Based on Gaussian Mixture Models Using Variational Bayesian Method

摘要

This paper presents a speaker identification system based on Gaussian Mixture Models (GMM) using the variational Bayesian method. Maximum Likelihood (ML) and Maximum A Posterior (MAP) are well-known methods for estimating GMM parameters. However, the overtraining problem occurs with insufficient data due to a point estimate of model parameters. The Bayesian approach estimates a posterior distribution of model param eters and achieves a robust prediction. To solve complicated integral calculations in the Bayesian approach, the variational Bayesian method has been proposed. This paper investigates the performance of the Bayesian approach in large speaker identification tasks.%ガウス混合モデル（Gaussian Mixture Model；GMM）に基づく話者認識は，これまで尤度最大化（Maximum Likelihood；ML）基準や事後確率最大化（Maximum a Posterior；MAP）基準が用いられてきた．しかし，ML基準は．モデルパラメータを確定的変数として点推定するため，学習データが十分に得られない場合，モデルの推定精度が低化する可能性がある．この問題に対し．近年，変分ベイズ法が提案され，様々なモデルにべイズ学習が適用可能となった.本研究では，GMMの学習に変分ベイズ法を適用し，ベイズ基準による話者認識の有効性について検討する．