HMM phoneme recognition with supervised training and Viterbi algorithm

Author

Vaich, T. ; Cohen, A.

Author_Institution

Dept. of Electr. & Comput. Eng., Ben-Gurion Univ. of the Negev, Beer-Sheva, Israel

fYear

1995

fDate

7-8 March 1995

Abstract

An HMM continuous Hebrew phoneme recognition system, that requires no manual segmentation for its training was developed. A relatively small Hebrew data base was acquired for training and recognition of phonemes in continuous speech. One of the main problems in phoneme recognition, that of manual segmentation of the training data base, was overcome by a special training algorithm. The Viterbi algorithm was used in the recognition stage, and the evaluation of the results was done with the Levenshtein distance measure. Initial recognition results of Hebrew phonemes for speaker independent, text dependent cases were 69.4% correct phoneme recognition.

Keywords

Viterbi detection; hidden Markov models; learning (artificial intelligence); speech recognition; HMM phoneme recognition; Levenshtein distance measure; Viterbi algorithm; continuous speech; speaker independent text dependent tests; supervised training; Cepstral analysis; Detection algorithms; Hidden Markov models; Linear predictive coding; Speech coding; Speech enhancement; Speech processing; Speech recognition; Testing; Viterbi algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Electronics Engineers in Israel, 1995., Eighteenth Convention of

Conference_Location

Tel Aviv, Israel

Print_ISBN

0-7803-2498-6

Type

conf

DOI

10.1109/EEIS.1995.513820

Filename

513820