SVM Based Speaker Verification using a GMM Supervector Kernel and NAP Variability Compensation

TLDR

Gaussian mixture models with universal backgrounds (UBMs) are the standard for speaker recognition, typically built by MAP adaptation of UBM means. The study investigates using the GMM supervector as input to support vector machines. A GMM supervector is formed by stacking MAP‑adapted UBM means, and this supervector is used to construct an SVM kernel. Latent factor analysis of the GMM supervector effectively compensates variability, and the resulting SVM kernel shows similarities to SVM nuisance attribute projection, with experiments on NIST SRE 2005 confirming the technique’s effectiveness.

Abstract

Gaussian mixture models with universal backgrounds (UBMs) have become the standard method for speaker recognition. Typically, a speaker model is constructed by MAP adaptation of the means of the UBM. A GMM supervector is constructed by stacking the means of the adapted mixture components. A recent discovery is that latent factor analysis of this GMM supervector is an effective method for variability compensation. We consider this GMM supervector in the context of support vector machines. We construct a support vector machine kernel using the GMM supervector. We show similarities based on this kernel between the method of SVM nuisance attribute projection (NAP) and the recent results in latent factor analysis. Experiments on a NIST SRE 2005 corpus demonstrate the effectiveness of the new technique