Title :
Comparison of parametric representations for hidden Markov models and multilayer perceptron recognizers
Author :
Martins, José A. ; Violaro, Fábio
Author_Institution :
CPqD-Telebras, Campinas, Brazil
Abstract :
This paper presents a comparison among different parametric representations of speech signals employed in speech recognition. In order to make this comparison, isolated word recognizers were built using hidden Markov models (HMM) and multilayer perceptrons (MLP). All the implemented recognizers were speaker-independent and the same database was used for their evaluation. The database consisted of a 50-word vocabulary spoken in Brazilian Portuguese
Keywords :
hidden Markov models; multilayer perceptrons; parameter estimation; signal representation; speech recognition; Brazilian Portuguese database; HMM; MLP; hidden Markov models; isolated word recognizers; multilayer perceptron; parametric representations; speaker-independent recognizer; speech recognition; speech signals; Cepstral analysis; Cepstrum; Filter bank; Fourier transforms; Frequency; Hidden Markov models; Linear predictive coding; Multilayer perceptrons; Nonlinear filters; Speech recognition;
Conference_Titel :
Telecommunications Symposium, 1998. ITS '98 Proceedings. SBT/IEEE International
Conference_Location :
Sao Paulo
Print_ISBN :
0-7803-5030-8
DOI :
10.1109/ITS.1998.713106
