CONF GenoMoreMayo97b/IDIAP Text dependent speaker verification using binary classifiers Genoud, Dominique Moreira, Miguel Mayoraz, Eddy EXTERNAL http://publications.idiap.ch/attachments/reports/1997/rr97-08.pdf PUBLIC http://publications.idiap.ch/index.php/publications/showcite/genomoremayo97a Related documents IEEE - Proceedings of the 1998 IEEE International Conference on Acoustics, Speech, and Signal Processing --- ICASSP'98 I 129-132 1998 IEEE May 1998 IDIAP-RR 97-08 This paper describes how a speaker verification task can be advantageously decomposed into a series of binary classification problems, i.e. each problem discriminating between two classes only. Each binary classifier is specific to one speaker, one anti-speaker and one word. Decision trees dealing classifiers. The set of classifiers is then pruned to eliminate the less relevant ones. Diverse pruning methods are experimented, and it is shown that when the speaker verification decision is performed with an a priori threshold, some of them give better results than a reference HMM system. REPORT GenoMoreMayo97a/IDIAP Text dependent speaker verification using binary classifiers Genoud, Dominique Moreira, Miguel Mayoraz, Eddy EXTERNAL http://publications.idiap.ch/attachments/reports/1997/rr97-08.pdf PUBLIC Idiap-RR-08-1997 1997 IDIAP in the Proceedings of ICASSP'98 This paper describes how a speaker verification task can be advantageously decomposed into a series of binary classification problems, i.e. each problem discriminating between two classes only. Each binary classifier is specific to one speaker, one anti-speaker and one word. Decision trees dealing classifiers. The set of classifiers is then pruned to eliminate the less relevant ones. Diverse pruning methods are experimented, and it is shown that when the speaker verification decision is performed with an a priori threshold, some of them give better results than a reference HMM system.