Speaker-independent isolated word recognition using multiple hidden Markov models

Author

Zhang, Y. ; deSilva, C.J.S. ; Togneri, R. ; Alder, M. ; Attikiouzel, Y.

Author_Institution

Centre for Intelligent Inf. Process. Syst., Western Australia Univ., Nedlands, WA, Australia

Volume

141

Issue

3

fYear

1994

fDate

6/1/1994 12:00:00 AM

Firstpage

197

Lastpage

202

Abstract

A multi-HMM speaker-independent isolated word recognition system is described. In this system, three vector quantisation methods, the LBG algorithm, the EM algorithm, and a new MGC algorithm, are used for the classification of the speech space. These quantisations of the speech space are then used to produce three HMMs for each word in the vocabulary. In the recognition step, the Viterbi algorithm is used in the three subrecognisers. The log probabilities of the observation sequences matching-the models are multiplied by the weights determined by the recognition accuracies of individual subrecognisers and summed to give the log probability that the utterance is of a particular word in the vocabulary. This multi-HMM system results in a reduction of about 50% in the error rate in comparison with the single model system

Keywords

hidden Markov models; probability; speech coding; speech recognition; vector quantisation; EM algorithm; LBG algorithm; MGC algorithm; Viterbi algorithm; error rate; hidden Markov models; log probabilities; observation sequences; recognition accuracies; speaker-independent isolated word recognition; speech space classification; vector quantisation; vocabulary;

fLanguage

English

Journal_Title

Vision, Image and Signal Processing, IEE Proceedings -

Publisher

iet

ISSN

1350-245X

Type

jour

DOI

10.1049/ip-vis:19941142

Filename

296563