Análise comparativa entre as implicações Lukasiewicz, Dienes-Rescher e Mandani aplicadas ao reconhecimento de voz

Washington SilvaGinalber Serra

The concept of fuzzy sets and fuzzy logic can be applied widely in systems modeling, classification and pattern recognition. This paper will approach systems based on fuzzy rules, where the relationship between the variables to be modeled are represented using IF-THEN rules. The approach of this paper proposes a pattern recognition system using the Lukasiewicz, Dienes-Rescher and Mamdani implications, where it intends to evaluate the performance implications for applications in Speech Recognition Systems. This work also makes the preparation pattern (speech) to be recognized by pre-processing with mel-cepstral coefficients. In addition to the pre-processing, will use the discrete cosine transform (DCT) as the standard classification step.

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

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: