Confidence Measures in Multiple pronunciations Modeling For Speaker Verification

This paper investigates the use of multiple pronunciations modeling for User-Customized Password Speaker Verification (UCP-SV). The main characteristic of the UCP-SV is that the system does not have any {\it a priori} knowledge about the password used by the speaker. Our aim is to exploit the information about how the speaker pronounces a password in the decision process. This information is extracted automatically by using a speaker-independent speech recognizer. In this paper, we investigate and compare several techniques. Some of them are based on the combination of confidence scores estimated by different models.In this context, we propose a new confidence measure that uses acoustic information extracted during the speaker enrollment and based on {\it log likelihood ratio} measure. These techniques show significant improvement ($15.7%$ relative improvement in terms of equal error rate) compared to a UCP-SV baseline system where the speaker is modeled by only one model (corresponding to one utterance).

Related material