Publication Details

Robust heteroscedastic linear discriminant analysis and LCRC posterior features in meeting data recognition

KARAFIÁT, M.; GRÉZL, F.; SCHWARZ, P.; BURGET, L.; ČERNOCKÝ, J. Robust heteroscedastic linear discriminant analysis and LCRC posterior features in meeting data recognition. Proc. 3nd Joint Workshop on Multimodal Interaction and Related Machine Learning Algorithms (MLMI 2006). Lecture Notes in Computer Science. Berlin: Springer Verlag, 2006. p. 275-284. ISBN: 3-540-69267-3.
Czech title
Robustní heteroscedastická lineární dikriminační analýza a LCRC posteriorní příznaky v rozpoznávání meetingových dat
Type
conference paper
Language
English
Authors
URL
Keywords

speech recognition, feature extraction, acoustic modeling

Abstract

The paper deals with robust heteroscedastic linear discriminant analysis and LCRC posterior features in meeting data recognition

Annotation

This paper investigates into feature extraction for meeting recognition. Three robust variants of popular HLDA transform are investigated. Influence of adding posterior features to PLP feature stream is studied. The experimental results are obtained on two data-sets: CTS (continuous telephone speech) and meeting data from NIST RT'05 evaluations. Silence-reduced HLDA and LCRC phoneme-state posterior features are found to be suitable for both recognition tasks.

Published
2006
Pages
275–284
Proceedings
Proc. 3nd Joint Workshop on Multimodal Interaction and Related Machine Learning Algorithms (MLMI 2006)
Series
Lecture Notes in Computer Science
Volume
4299
ISBN
3-540-69267-3
Publisher
Springer Verlag
Place
Berlin
BibTeX
@inproceedings{BUT33392,
  author="Martin {Karafiát} and František {Grézl} and Petr {Schwarz} and Lukáš {Burget} and Jan {Černocký}",
  title="Robust heteroscedastic linear discriminant analysis and LCRC posterior features in meeting data recognition",
  booktitle="Proc. 3nd Joint Workshop on Multimodal Interaction and Related Machine Learning Algorithms (MLMI 2006)",
  year="2006",
  series="Lecture Notes in Computer Science",
  volume="4299",
  pages="275--284",
  publisher="Springer Verlag",
  address="Berlin",
  isbn="3-540-69267-3",
  url="http://www.fit.vutbr.cz/~cernocky/publi/2006/karafiat_mlmi2006_honza.pdf"
}
Back to top