Large Margin GMM for discriminative speaker verifi cation

Abstract : Gaussian mixture models (GMM), trained using the generative cri- terion of maximum likelihood estimation, have been the most popular ap- proach in speaker recognition during the last decades. This approach is also widely used in many other classi cation tasks and applications. Generative learning in not however the optimal way to address classi cation problems. In this paper we rst present a new algorithm for discriminative learning of diagonal GMM under a large margin criterion. This algorithm has the ma- jor advantage of being highly e cient, which allow fast discriminative GMM training using large scale databases. We then evaluate its performances on a full NIST speaker veri cation task using NIST-SRE'2006 data. In particular, we use the popular Symmetrical Factor Analysis (SFA) for session variability compensation. The results show that our system outperforms the state-of-the- art approaches of GMM-SFA and the SVM-based one, GSL-NAP. Relative reductions of the Equal Error Rate of about 9.33% and 14.88% are respec- tively achieved over these systems.
Type de document :
Article dans une revue
Multimedia Tools and Applications, Springer Verlag, 2012
Liste complète des métadonnées

Littérature citée [20 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-00647983
Contributeur : Khalid Daoudi <>
Soumis le : dimanche 4 décembre 2011 - 15:32:21
Dernière modification le : mercredi 12 septembre 2018 - 17:46:02
Document(s) archivé(s) le : dimanche 4 décembre 2016 - 23:28:19

Fichier

JMTA.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00647983, version 1

Citation

Reda Jourani, Khalid Daoudi, Régine André-Obrecht, Driss Aboutajdine. Large Margin GMM for discriminative speaker verifi cation. Multimedia Tools and Applications, Springer Verlag, 2012. 〈hal-00647983〉

Partager

Métriques

Consultations de la notice

331

Téléchargements de fichiers

149