Title: | Gender-dependent acoustic models fusion developed for automatic subtitling of Parliament meetings broadcasted by the Czech TV |
Other Titles: | Fúze akustických modelů závislých na pohlaví řečníka při automatickém titulkování TV přenosů jednání Parlamentu ČR |
Authors: | Vaněk, Jan Psutka, Josef V. |
Citation: | VANĚK, Jan; PSUTKA, Josef V. Gender-dependent acoustic models fusion developed for automatic subtitling of Parliament meetings broadcasted by the Czech TV. In: Text, speech and dialogue. Berlin: Springer, 2010, p. 431-438. (Lecture notes in computer science; 6231). ISBN 3-642-15759-9. |
Issue Date: | 2010 |
Publisher: | Springer |
Document type: | článek article |
URI: | http://hdl.handle.net/11025/16958 http://www.kky.zcu.cz/cs/publications/VanekJan_2010_Gender-dependent |
ISBN: | 3-642-15759-9 |
Keywords: | automatické titulkování;akustické modelování;genderové modelování;rozpoznávání řeči v reálném čase |
Keywords in different language: | automatic subtitling;acoustics modeling;gender modeling;speech recognition in real time |
Abstract: | Článek popisuje využití fúze akustických modelů závislých na pohlaví řečníka v úloze on-line automatického titulkování TV přenosů jednání Parlamentu ČR. Titulkování je prováděno přímo zpracováním audio zvukové stopy. |
Abstract in different language: | 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. |
Rights: | © Jan Vaněk - Josef V. Psutka |
Appears in Collections: | Články / Articles (KIV) Články / Articles (KKY) |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
VanekJan_2010_Gender-dependent.pdf | Plný text | 58,89 kB | Adobe PDF | View/Open |
Please use this identifier to cite or link to this item:
http://hdl.handle.net/11025/16958
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.