Fuzzy variants of hard classification rules for speech pattern recognition in Romanian language

Author

Gavat, Inge ; Grigore, Ovidiu ; Zirra, Matei ; Cula, Oana

Author_Institution

Fac. of Electron. & Telecommun., Politehnica Univ. of Bucharest, Romania

fYear

1997

fDate

21-24 Sep 1997

Firstpage

172

Lastpage

176

Abstract

The paper presents results obtained in a vowel recognition task applying unsupervised and supervised fuzzy algorithms and fuzzy neural networks. The vowels, uttered from 10 speakers each in 1000 different contexts are recognized using as features the first three formant frequencies. Feature extraction is presented and two fuzzy algorithms, fuzzy ISODATA and fuzzy k-NN, and the fuzzy multilayer perceptron neural network used for recognition are given. Conclusions about the obtained results with future plans and a reference list close the paper

Keywords

feedforward neural nets; fuzzy neural nets; learning systems; multilayer perceptrons; pattern classification; speech recognition; Romanian language; feature extraction; formant frequencies; fuzzy ISODATA; fuzzy hard classification rules; fuzzy k-NN; fuzzy multilayer perceptron neural network; fuzzy neural networks; speakers; speech pattern recognition; supervised fuzzy algorithms; unsupervised fuzzy algorithms; vowel recognition task; Cepstrum; Clustering algorithms; Electronic mail; Frequency; Fuzzy sets; Histograms; Natural languages; Pattern recognition; Prototypes; Speech processing;

fLanguage

English

Publisher

ieee

Conference_Titel

Fuzzy Information Processing Society, 1997. NAFIPS '97., 1997 Annual Meeting of the North American

Conference_Location

Syracuse, NY

Print_ISBN

0-7803-4078-7

Type

conf

DOI

10.1109/NAFIPS.1997.624031

Filename

624031