Text-dependent speaker verification using vector quantization source coding
Abstract
Several vector quantization approaches to the problem of text-dependent speaker verification are described. In each of these approaches, a source codebook is designed to represent a particular speaker saying a particular utterance. Later, this same utterance is spoken by a speaker to be verified and is encoded in the source codebook representing the speaker whose identity was claimed. The speak is accepted if the verification utterance's quantization distortion is less than a prespecified speaker-specific threshold. The best of the approaches achieved a 0.7% false acceptance rate and a 0.6% false rejection rate on a speaker population containing 16 admissible speakers and 111 casual imposters. The approaches are described, and detailed experimental results are presented and discussed.
- Publication:
-
Naval Research Lab. Report
- Pub Date:
- November 1985
- Bibcode:
- 1985nrl..reptS....B
- Keywords:
-
- Coding;
- Speech;
- Speech Recognition;
- Vector Analysis;
- Vectors (Mathematics);
- Demography;
- Distortion;
- Matrix Methods;
- Matrix Theory;
- Security;
- Communications and Radar