A theory on optimal construction of dynamic features of speech for HMM-based speech recognition

Author

Deng, Li

Author_Institution

Dept. of Electr. & Comput. Eng., Waterloo Univ., Ont., Canada

fYear

1994

fDate

13-16 Apr 1994

Firstpage

351

Abstract

The construction of dynamic (delta) features of speech, which has been in the past confined to the pre-processing domain in hidden Markov modelling (HMM), is generalized and formulated as an integrated speech modeling problem. This generalization allows to utilize state-dependent weights to transform static speech features into dynamic ones. The author describes a rigorous theoretical framework that naturally incorporates the generalized dynamic-parameter technique, and presents a maximum-likelihood based algorithm for integrated optimization of the conventional HMM parameters and of the time-varying weighting functions that define the dynamic features of speech

Keywords

hidden Markov models; maximum likelihood estimation; optimisation; parameter estimation; speech recognition; HMM-based speech recognition; delta features; dynamic features; generalized dynamic-parameter technique; hidden Markov modelling; integrated optimization; integrated speech modeling problem; maximum-likelihood based algorithm; optimal construction; speech; state-dependent weights; static speech features; time-varying weighting functions; Hidden Markov models; Speech recognition; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Speech, Image Processing and Neural Networks, 1994. Proceedings, ISSIPNN '94., 1994 International Symposium on

Print_ISBN

0-7803-1865-X

Type

conf

DOI

10.1109/SIPNN.1994.344894

Filename

344894