Skip to content

Publications

Detail of publication

Citation

Psutka Josef V. and Müller, L. and Šmídl, L. and Psutka, J. : Feature space reduction and decorrelation in a large number of speech recognition experiments . Signal and Image Processing, , p. 158-161, ACTA Press, Anaheim, 2007.

Abstract

The paper studies the influence of significant space reduction and decorrelation techniques on the performance of an automatic speech recognition (ASR) system. A baseline PLP-based ASR system, which works with zero-gram language model, was trained using speech of one thousand speakers. The Linear Discriminant Analysis (LDA), Heteroscedastic LDA (HLDA) and Smoothed HLDA (SHLDA) techniques were applied and compared in many experiments in which an optimum dimension of a feature space was searched for.

Detail of publication

Title: Feature space reduction and decorrelation in a large number of speech recognition experiments
Author: Psutka Josef V. ; Müller, L. ; Šmídl, L. ; Psutka, J.
Language: English
Date of publication: 20 Sep 2007
Year: 2007
Type of publication: Papers in proceedings of reviewed conferences
Title of journal or book: Signal and Image Processing
Edition:
Page: 158 - 161
ISBN: 978-0-88986-675-1
Publisher: ACTA Press
Address: Anaheim
Date: 20 Sep 2007 - 22 Sep 2007
/ 2011-06-09 13:00:42 /

Keywords

automatic speech recognition, feature space reduction, PLP-based parameterization

BibTeX

@INPROCEEDINGS{PsutkaJosefV_2007_Featurespace,
 author = {Psutka Josef V. and M\"{u}ller, L. and \v{S}m\'{i}dl, L. and Psutka, J.},
 title = {Feature space reduction and decorrelation in a large number of speech recognition experiments},
 year = {2007},
 publisher = {ACTA Press},
 journal = {Signal and Image Processing},
 address = {Anaheim},
 pages = {158-161},
 ISBN = {978-0-88986-675-1},
 url = {http://www.kky.zcu.cz/en/publications/PsutkaJosefV_2007_Featurespace},
}