A Context-Dependent Sequential Decision for Speaker Verification

摘要

This paper presents an improved method of speaker verification using the sequential probability ratio test (SPRT), which can treat the correlation between successive fea- ture vectors. The hidden Markov model with the mean field ap- proximation enables us to consider the correlation in the SPRT, i.e., using the mean field of previous state, probability compu- tation can be carried out as if input samples were independent each other.