Unseen handset mismatch compensation based on a priori knowledge interpolation for robust speaker recognition

Unseen handset mismatch is the major source of performance degradation for speaker recognition in telecommunication environment since handset distortions are tightly coupled with speaker characteristics. In this paper, a soft-decision unseen handset characteristics estimation method based on a priori knowledge interpolation is proposed to decouple the characteristics of the unseen handset and speaker for robust speaker recognition. Experimental results on HTIMIT database showed that the proposed method improved the speaker recognition rate for both seen and unseen handsets.