The IBM system for the NIST-2002 cellular speaker verification evaluation

Author

Ramaswamy, Ganesh N. ; Navratil, Jirí ; Chaudhari, Upendra V. ; Zilca, Ran D.

Author_Institution

IBM T. J. Watson Res. Center, Yorktown Heights, NY, USA

Volume

2

fYear

2003

fDate

6-10 April 2003

Abstract

This paper presents an overview of the architecture and algorithms implemented in IBM´s text-independent speaker verification system developed for the 2002 NIST speaker recognition evaluation, particularly for the 1-speaker detection task using cellular test data. We describe individual components including a Gaussianization front-end, celluar-codec post-processing, modeling, discriminative optimization and scoring steps. A combination of multiple, data-perturbed systems using a discriminative objective so as to achieve optimum performance for a low false alarm operating region obtained the top performance in the NIST 2002 1-speaker detection task.

Keywords

cellular radio; optimisation; speaker recognition; speech codecs; Gaussianization front-end; IBM system; NIST 2002 1-speaker detection task; NIST-2002 cellular speaker verification evaluation; algorithms; architecture; celluar-codec post-processing; cellular test data; data-perturbed systems; discriminative objective; discriminative optimization; low false alarm operating region; scoring steps; text-independent speaker verification system; Codecs; Data mining; Feature extraction; Gaussian channels; Gaussian processes; NIST; Radio access networks; Speaker recognition; Speech; System testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). 2003 IEEE International Conference on

ISSN

1520-6149

Print_ISBN

0-7803-7663-3

Type

conf

DOI

10.1109/ICASSP.2003.1202294

Filename

1202294