Text-Independent Speaker Verification Using Artificially Generated GMMs for Cohorts

Yuuji MUKAI; Hideki NODA; Michiharu NIIMI; Takashi OSANAI

doi:10.1093/ietisy/e91-d.10.2536

Regular Section

Text-Independent Speaker Verification Using Artificially Generated GMMs for Cohorts

Yuuji MUKAI, Hideki NODA, Michiharu NIIMI, Takashi OSANAI

Author information

Keywords: speaker verification, Gaussian mixture model, artificial cohort, score normalization, background model

JOURNAL FREE ACCESS

2008 Volume E91.D Issue 10 Pages 2536-2539

DOI https://doi.org/10.1093/ietisy/e91-d.10.2536

Details

Abstract

This paper presents a text-independent speaker verification method using Gaussian mixture models (GMMs), where only utterances of enrolled speakers are required. Artificial cohorts are used instead of those from speaker databases, and GMMs for artificial cohorts are generated by changing model parameters of the GMM for a claimed speaker. Equal error rates by the proposed method are about 60% less than those by a conventional method which also uses only utterances of enrolled speakers.

Corresponding author