Publication Details

Classifiers in speech recognition systems based on TRAPS

GRÉZL, F. Classifiers in speech recognition systems based on TRAPS. Proceedings of 8th Conference STUDENT EEICT 2002. Brno: Faculty of Electrical Engineering and Communication BUT, 2002. p. 74-77. ISBN: 80-214-2116-9.
Czech title
Kalifikatory zalozene na TRAPS v systemech pro rozpoznavani reci
Type
conference paper
Language
English
Authors
Grézl František, Ing., Ph.D. (DCGM)
Keywords

speech recognition, temporal patterns, linear discriminant analysis

Abstract

Comparison of different configuration of neural network classifiers in speech recognition based on TRAPS. Linear Discriminant Analysis (LDA) was used for reducing the dimensionality of temporal vectors.

Annotation

This article deals with feature extraction for speech recognition using temporal patterns (TRAPS). TRAP-systems use neural network classifiers. Four different configuration of classifiers are compared in this work. Linear Discriminant Analysis (LDA) was used for reducing the dimensionality of temporal vectors. Accuracies on the output of classifiers are compared. Global recognition accuracies are shown.

Published
2002
Pages
74–77
Proceedings
Proceedings of 8th Conference STUDENT EEICT 2002
ISBN
80-214-2116-9
Publisher
Faculty of Electrical Engineering and Communication BUT
Place
Brno
BibTeX 
@inproceedings{BUT10284,
  author="František {Grézl}",
  title="Classifiers in speech recognition systems based on TRAPS",
  booktitle="Proceedings of 8th Conference STUDENT EEICT 2002",
  year="2002",
  pages="74--77",
  publisher="Faculty of Electrical Engineering and Communication BUT",
  address="Brno",
  isbn="80-214-2116-9"
}
Back to top