Skip to content

Publications

Detail of publication

Citation

Vaněk Jan and Psutka Josef V. : Gender-dependent acoustic models fusion developed for automatic subtitling of Parliament meetings broadcasted by the Czech TV . Lecture Notes in Computer Science, vol. 2010, p. 431-438, Springer, Berlin, 2010.

Download PDF

PDF

Abstract

Gender-dependent (male/female) acoustic models are more acoustically homogeneous and therefore give better recognition performance than single gender-independent model. This paper deals with a problem how to use these gender-based acoustic models in a real-time LVCSR (Large Vocabulary Continuous Speech Recognition) system that is for more than one year used by the Czech TV for automatic subtitling of Parliament meetings that are broadcasted on the channel \v{C}T24. Frequent changes of speakers and the direct connection of the LVCSR system to the TV audio stream require switching/fusion of models automatically and as soon as possible. The paper presents various techniques based on using the output probabilities for quick selection of a better model or their combinations. The best proposed method achieved over 11\% relative WER reduction in comparision with the GI model.

Detail of publication

Title: Gender-dependent acoustic models fusion developed for automatic subtitling of Parliament meetings broadcasted by the Czech TV
Author: Vaněk Jan ; Psutka Josef V.
Language: English
Date of publication: 8 Sep 2010
Year: 2010
Type of publication: Papers in journals
Title of journal or book: Lecture Notes in Computer Science
Číslo vydání: 2010
Page: 431 - 438
ISSN: 0302-9743
Publisher: Springer
Address: Berlin
2011-03-15 16:21:58 / 2012-02-17 11:33:08 / 1

Keywords

Gender-dependent AM, automatic subtitling, LVCSR system

BibTeX

@ARTICLE{VanekJan_2010_Gender-dependent,
 author = {Van\v{e}k Jan and Psutka Josef V.},
 title = {Gender-dependent acoustic models fusion developed for automatic subtitling of Parliament meetings broadcasted by the Czech TV},
 year = {2010},
 publisher = {Springer},
 journal = {Lecture Notes in Computer Science},
 address = {Berlin},
 volume = {2010},
 pages = {431-438},
 ISSN = {0302-9743},
 url = {http://www.kky.zcu.cz/en/publications/VanekJan_2010_Gender-dependent},
}