This paper proposes a newmethod for text-dependent speaker recognition. The scheme is based on learning (what we refer to as) speaker-specific compensators for each speaker in the system. The compensator is essentially a speaker to speaker transformationwhich enables the recognition of the speech of one speaker through a speaker-dependent speech recognition system built for the other. Such a transformation, adequate for our purposes, may be achieved by a simple vector addition in the cepstral domain. This speaker-specific compensator captures the characteristics of the speaker we wish to recognize. For each speaker who is registered into the system, we learn a unique set of compensators. The speaker recognition decision is then based on which compensator achieves best speech recognition scores.

PDF file

In  Proceedings of IEEE TENCON 2003, Conference for convergent technologies for Asia-Pacific Region, Bangalore, India

Publisher  Institute of Electrical and Electronics Engineers, Inc.
© 2007 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Details