NIST Speaker Recognition Evaluation -- Overview, Methodology, Systems, Results, Perspective

Published

June 2, 2000

Author(s)

G R. Doddington, Mark A. Przybocki, Alvin F. Martin, D A. Reynolds

Abstract

This paper, based on three presentations at La Reconnaissance du Locuteur et ses Applications Commerciales et Criminalistiques (RLA2C) Workshop in Avignon, discusses the evaluation of speaker recognition systems from several perspectives. A general discussion of the speaker recognition task and the challenges and issues involved in its evaluation is offered. The NIST evaluations in this area, and specifically the 1998 evaluation, its objectives, protocols, test data, and participants, are described. The algorithms used by the systems that were developed for this evaluation are summarized and compared and contrasted. Overall performance results of this evaluation without identification of specific systems, are presented by means of Detection error trade-off (DET) curves. These show the performance tradeoff of missed detections and false alarms for each system, and the effects on performance of training condition, test segment duration, the speaker's sex, and the match or mismatch of training and test handsets. Several factors that were found to have an impact on performance, including pitch frequency, handset type, and noise, are discussed and DET curves showing their effects are presented. The paper concludes with some perspective on the history of this technology and where it may be going.

Citation

Speech Communication

Pub Type

Journals

Keywords

DET curve, NIST evaluations, performance evaluation, speaker identification, speaker recognition, speaker verification

Citation

Doddington, G. , Przybocki, M. , Martin, A. and Reynolds, D. (2000), NIST Speaker Recognition Evaluation -- Overview, Methodology, Systems, Results, Perspective, Speech Communication (Accessed May 17, 2024)