Fast adaptation of GMM-based compact models - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2007

Fast adaptation of GMM-based compact models

Résumé

In this paper, a new strategy for a fast adaptation of acoustic models is proposed for embedded speech recognition. It relies on a general GMM, which represents the whole acoustic space, associated with a set of HMM state-dependent probability functions modeled as transformations of this GMM. The work presented here takes advantage of this architecture to propose a fast and efficient way to adapt the acoustic models. The adaptation is performed only on the general GMM model, using techniques gathered from the speaker recognition domain. It does not require state-dependent adaptation data and it is very efficient in terms of computational cost. We evaluate our approach in the voice-command task, using a car-based corpus. This adaptation method achieved a relative error-rate decrease of about 10% even if few adaptation data are available. The complete system allows a total relative gain of more than 20% compared to a basic HMM-based system.
Fichier non déposé

Dates et versions

hal-01318190 , version 1 (19-05-2016)

Identifiants

  • HAL Id : hal-01318190 , version 1

Citer

Christophe Lévy, Georges Linarès, Jean-François Bonastre. Fast adaptation of GMM-based compact models. INTERSPEECH, Aug 2007, Anvers, Belgium. ⟨hal-01318190⟩

Collections

UNIV-AVIGNON LIA
34 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More