Skip to content

Publications

Detail of publication

Citation

Krňoul, Z. and Císař, P. and Železný, M. and Holas, J. : Viseme analysis for speech-driven facial animation for Czech audio-visual speech synthesis . SPECOM 2005 proceedings, p. 227-230, Moscow State Linguistic University, Moscow , 2005.

Abstract

In this paper we present recent advances in the audio-visual speech synthesis for the Czech language. Based on our previous work, we decided to apply the audio-visual speech synthesis also in the mode when the input is not a textual information but a speech signal. At the same time, we decided to carry out more thorough viseme analysis. For both of these purposes, we collected a new audiovisual speech corpus. To be able to parameterize the lip shape with higher precision, we decided to use reflexive markers for interest points and infrared illumination. This approach made the visual data segmentation and the detection of interest points much more robust and thus more precise. Based on this corpus, we performed the viseme analysis for the Czech language and modified the existing facial animation application. Also, we solved the case, when the input for audio-visual speech synthesis is a speech signal and not a text.

Detail of publication

Title: Viseme analysis for speech-driven facial animation for Czech audio-visual speech synthesis
Author: Krňoul, Z. ; Císař, P. ; Železný, M. ; Holas, J.
Language: English
Date of publication: 17 Oct 2005
Year: 2005
Type of publication: Papers in proceedings of reviewed conferences
Title of journal or book: SPECOM 2005 proceedings
Page: 227 - 230
ISBN: 5-7452-0110-X
Publisher: Moscow State Linguistic University
Address: Moscow
Date: 17 Oct 2005 - 19 Oct 2005
/ 2008-06-06 13:44:22 /

Keywords

audio-visual speech synthesis, viseme, artificial neural network, face animation

BibTeX

@INPROCEEDINGS{KrnoulZ_2005_Visemeanalysisfor,
 author = {Kr\v{n}oul, Z. and C\'{i}sa\v{r}, P. and \v{Z}elezn\'{y}, M. and Holas, J.},
 title = {Viseme analysis for speech-driven facial animation for Czech audio-visual speech synthesis},
 year = {2005},
 publisher = {Moscow State Linguistic University},
 journal = {SPECOM 2005 proceedings},
 address = {Moscow },
 pages = {227-230},
 ISBN = {5-7452-0110-X},
 url = {http://www.kky.zcu.cz/en/publications/KrnoulZ_2005_Visemeanalysisfor},
}