BDBComp
Parceria:
SBC
Desenvolvimento de um Sistema de Reconhecimento Automatico ´ de Voz Cont´?nua com Grande Vocabulário para o Português Brasileiro

Ênio SilvaLuiz BaptistaHelane FernandesAldebaro Klautau

This work presents some steps towards the development of a large vocabulary continuousspeech recognition system for Brazilian Portuguese, with a vocabulary of more than 60 thousand words. The paper discusses the creation of a phonetic dictionary, with pronunciation(s) of each word, the improvement of a N-gram language model through the adoption of a class-based scheme,and acoustic modeling using triphones clustered through decision trees. Some preliminary results for the Spoltech corpus are also presented.

http://www.lbd.dcc.ufmg.br/colecoes/til/2005/0025.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD