Title:	Fast speaker adaptation in automatic online subtitling
Other Titles:	Rychlá adaptace v úloze online automatického titulkování
Authors:	Pražák, Aleš Zajíc, Zbyněk Machlica, Lukáš Psutka, Josef V.
Citation:	PRAŽÁK, Aleš; ZAJÍC, Zbyněk; MACHLICA, Lukáš; PSUTKA, Josef V. Rychlá adaptace v úloze online automatického titulkováníí­. In: Milano: International Conference on Signal Processing and Multimedia Applications SIGMAP 2009, July 7-10, Milano, Italy. SIGMAP, 2009, p. 126-130.
Issue Date:	2009
Publisher:	SIGMAP
Document type:	článek article
URI:	http://www.kky.zcu.cz/cs/publications/AlesPrazak_2009_FastSpeaker http://hdl.handle.net/11025/16934
Keywords:	ASR;online titulkování;MAP;fMLLR
Keywords in different language:	ASR;online subtitling;MAP;fMLLR
Abstract in different language:	This paper deals with speaker adaptation techniques well suited for the task of online subtitling. Two methods are briefly discussed, namely MAP adaptation and fMLLR. The main emphasis is laid on the description of improvements involved in the process of adaptation subject to the time requirements. Since the adaptation data are gathered continuously, simple modifications of the accumulated statistics have to be carried out in order to make the adaptation more accurate. Another proposed improvement efficiently employs the combination of fMLLR and MAP. In the case of online adaptation no prior transcriptions of the data are available. They are handled by a recognition system, thus it is suitable to assign a well-applied confidence measure to each of the transcriptions. We have performed experiments focused on the trade-off between the adaptation speed and the amount of adaptation data. We were able to gain a relative reduction of WER 16.2 %.
Rights:	© Alešˇ Pražák - Zbyněk Zajíc - Lukáš Machlica - Josef V. Psutka
Appears in Collections:	Články / Articles (KKY) Články / Articles (NTIS)

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.