Using Diphones in Large Vocabulary Word Recognition

Auteurs-es

  • C. Vicenzi
  • D. Sciarra

Résumé

In this paper we present a large vocabulary,speaker dependent, isolated word recognition system with diphones as basic units, so that the training session is much faster and useful for any application. The system, tested on a vocabulary of 910 words on one speaker, gave a word recognition rate of 78%, slightly lower than an Itakura recognizer with whole word templates (WRR= 85%)

Fichiers supplémentaires

Publié-e

2022-12-03

Comment citer

1.
Vicenzi C, Sciarra D. Using Diphones in Large Vocabulary Word Recognition. Canadian Acoustics [Internet]. 3 déc. 2022 [cité 11 juill. 2024];14(3 bis):47-8. Disponible à: https://jcaa.caa-aca.ca/index.php/jcaa/article/view/3516

Numéro

Rubrique

Actes du congrès de la Semaine canadienne d'acoustique