Análise de reconhecedor de fala dependente de locutor quando usado para locutores não treinados

The present paper proposes to develop an evaluation of a speech recognition tool to analyze the accuracy rate variation against the parameters variations for both vocal features and peculiar parameters controlled by the respective tool. The focus in this study is the case where different speaker gro...

ver descrição completa

Autor principal: Pavelski, Guilherme Antonio
Formato: Trabalho de Conclusão de Curso (Graduação)
Idioma: Português
Publicado em: Universidade Tecnológica Federal do Paraná 2020
Assuntos:
Acesso em linha: http://repositorio.utfpr.edu.br/jspui/handle/1/9824
Tags: Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
Resumo: The present paper proposes to develop an evaluation of a speech recognition tool to analyze the accuracy rate variation against the parameters variations for both vocal features and peculiar parameters controlled by the respective tool. The focus in this study is the case where different speaker groups are used in order to training and testing the speech recognition system. This work discuss the use of speech recognition based on Hidden Markov Models and sweep line algorithms to search the best results. A tutorial for the tool was also developed in order to detail the steps of the recognition system preparation.