Combinaison de différents jeux de param etres acoustiques pour la reconnaissance de la parole - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2008

Combinaison de différents jeux de param etres acoustiques pour la reconnaissance de la parole

Loic Barrault
Driss Matrouf
Georges Linarès

Résumé

With the purpose of improving Automatic Speech Recognition (ASR) systems performance, many different approaches on combining them have been studied. In this paper, a combination of state a posteriori probabilities given by different feature sets is proposed. In order to perform a coherent combination of state posterior probabilities, the acoustic models trained on different feature sets must have the same topo-logy (i.e. same set of states). For this purpose, a fast and efficient twin model training protocol is proposed. Then, two different strategies for combining probabilities are presented : the linear and the log linear interpolation. By using log linear interpolation, a relative Word Error Rate (WER) reduction of about 15% on MEDIA and 14% on ESTER corpora have been respectively observed.
Fichier non déposé

Dates et versions

hal-01312834 , version 1 (09-05-2016)

Identifiants

  • HAL Id : hal-01312834 , version 1

Citer

Loic Barrault, Driss Matrouf, Renato de Mori, Georges Linarès. Combinaison de différents jeux de param etres acoustiques pour la reconnaissance de la parole. Les Journées d’Etude sur la Parole (JEP), Jun 2008, Avignon, France. ⟨hal-01312834⟩

Collections

UNIV-AVIGNON LIA
81 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More