Advances on HMM-based Text-dependent Speaker Verification

摘要

This paper presents recent development on text-dependent speaker verification technology in EU project PICASSO, which have improved the SV performance significntly. In the project we adopt HMM appoach for pattern matching. In the paper we describes four different techniques, adaptive variance flooring, multiple use of enrolment sample, generalised competitive measurement for score normalisation and hybrid method for score normalisation to improve hMm method for speakerverification. The evaluation resutls are given for each of these techniques.