DocumentCode :
2393656
Title :
Telephone based speaker recognition using multiple binary classifier and Gaussian mixture models
Author :
Castellano, Pierre J. ; Slomka, S. ; Sridharan, Sridha
Author_Institution :
Signal Process. Res. Centre, Queensland Univ. of Technol., Brisbane, Qld., Australia
Volume :
2
fYear :
1997
fDate :
21-24 Apr 1997
Firstpage :
1075
Abstract :
The present study evaluates multiple binary classifier model (MBCM) and Gaussian mixture model (GMM) solutions for both automatic speaker verification (ASV) and automatic speaker identification (ASI) problems involving text-independent telephone speech from the King speech database. The MBCM´s accuracy is enhanced by selectively removing those classifiers within the model which perform worst (pruning). An unpruned MBCM outperforms a GMM for ASV and speakers taken from within the same dialectic region (San Diego, CA). Once pruned, the MBCM is found to be 2.6 times more accurate than the GMM. For closed set ASI, based on the same data, the MBCM is roughly twice as accurate as the GMM but only after pruning
Keywords :
Gaussian distribution; pattern classification; speaker recognition; speech processing; telephony; Gaussian mixture model; King speech database; automatic speaker identification; automatic speaker verification; multiple binary classifier model; pruning; telephone based speaker recognition; text-independent telephone speech; Automatic speech recognition; Data mining; Databases; Signal processing; Speaker recognition; Speech analysis; Speech processing; Telephony; Testing; Training data;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1997. ICASSP-97., 1997 IEEE International Conference on
Conference_Location :
Munich
ISSN :
1520-6149
Print_ISBN :
0-8186-7919-0
Type :
conf
DOI :
10.1109/ICASSP.1997.596127
Filename :
596127
Link To Document :
بازگشت