Suprasegmental features and continuous speech recognition

Author

Dumouchel, P.

Author_Institution

INRS-Telecommun., Quebec Univ., Verdun, Que., Canada

Volume

ii

fYear

1994

fDate

19-22 Apr 1994

Abstract

We first propose to model microprosody by means of a Bayesian classifier assuming multivariate Gaussian distributions on suprasegmental features. Second, we normalize the suprasegmental features by using dynamic parameters extracted from diphones. Third, we examine three different types of covariance matrices and show that a full covariance matrix per diphone gives the best results. Finally, the insertion of the microprosodic model into the INRS large vocabulary speech recognition improves the word recognition slightly from 48% to 52%

Keywords

Bayes methods; Gaussian distribution; Gaussian processes; covariance matrices; feature extraction; pattern classification; speech recognition; Bayesian classifier; INRS large vocabulary speech recognition; continuous speech recognition; covariance matrices; diphones; dynamic parameters; feature extraction; microprosodic model; microprosody; multivariate Gaussian distributions; suprasegmental features; word recognition; Delay; Disk recording; Face; Filters; Frequency; Hidden Markov models; Speech processing; Speech recognition; Stress; Viterbi algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1994. ICASSP-94., 1994 IEEE International Conference on

Conference_Location

Adelaide, SA

ISSN

1520-6149

Print_ISBN

0-7803-1775-0

Type

conf

DOI

10.1109/ICASSP.1994.389690

Filename

389690