| On model quality and evaluation in speaker verification |
Title: On model quality and evaluation in speaker verification
Author(s): Johan Koolwaaij, Lou Boves, H. Jongebloed & E.A. den Os
Reference: Proceedings of the International Conference on Acoustics, Speech, and Signal Processing (ICASSP'00), Istanbul, Turkey, pp. 3759-3762
Keywords: Speaker Recognition
There is a PostScript version (413666 bytes) available.
There is a PDF version (180762 bytes) available.
This paper addresses differences between operational speaker verification (SV)
systems and laboratory experiments in terms of performance and methods for measuring
performance. It is concluded that operational SV systems need an indication of the
quality of newly enrolled speaker models, to decide whether
to re-enrol or request more enrolment material. We have investigated the
impact of ASR errors on modelquality. While attempting to design measures for the
quality of speaker models we have developed a novel method for assigning weights
to the contribution of modelsin accordance with their discriminative ability.
Error processing SSI file