Optimal linear feature transformations for semi-continuous hidden Markov models

Author

Schukat-Talamazzini, E. Günter ; Hornegger, Joachim ; Niemann, Heinrich

Author_Institution

Friedrich-Alexander Univ., Erlangen, Germany

Volume

1

fYear

1995

fDate

9-12 May 1995

Firstpage

369

Abstract

Linear discriminant or Karhunen-Loeve transforms are established techniques for mapping features into a lower dimensional subspace. This paper introduces a uniform statistical framework, where the computation of the optimal feature reduction is formalized as a maximum-likelihood estimation problem. The experimental evaluation of this suggested extension of linear selection methods shows a slight improvement of the recognition accuracy

Keywords

feature extraction; hidden Markov models; maximum likelihood estimation; optimisation; speech recognition; statistical analysis; transforms; Karhunen-Loeve transforms; experimental evaluation; linear discriminant transforms; linear selection methods; lower dimensional subspace; maximum-likelihood estimation; optimal linear feature transformations; recognition accuracy; semi-continuous hidden Markov models; uniform statistical framework; Continuous production; Feature extraction; Hidden Markov models; Karhunen-Loeve transforms; Labeling; Maximum likelihood estimation; Parameter estimation; Speech processing; Speech recognition; Vectors; Yttrium;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1995. ICASSP-95., 1995 International Conference on

Conference_Location

Detroit, MI

ISSN

1520-6149

Print_ISBN

0-7803-2431-5

Type

conf

DOI

10.1109/ICASSP.1995.479598

Filename

479598