Using Eigenvoice Coefficients as Features in Speaker Recognition

Eigenvoice speaker adaptation has been shown to be effective in recent years. In this paper, we propose to use eigenvoice coefficients as features for speaker recognition. We use a simplified version of probabilistic subspace adaptation (PSA) to estimate eigenvoice coefficients, and the coefficients are concatenated to construct supervectors of support vector machines. This approach significantly reduces the dimension of feature vector, and leads to a great reduction of training time cost. We then design a simple and effective feature normalization method, which uses eigenvalues for variance normalization. Our approach is evaluated on the SRE2008 NIST evaluation and exhibits better performance than the conventional eigenGMM approach.