Paper Info
Title
Continuous phoneme recognition in Cued Speech for French.
Abstract
Cued Speech is a visual communication mode, which uses hand shapes and lip shapes making all the sounds of spoken language clearly understandable to deaf and hearing-impaired people. Using Cued Speech the problems of lipreading can be overcome resulting thus in understanding of full spoken language by deaf children and adults. In automatic recognition of Cued Speech, lip shape recognition, gesture recognition, and integration of the two modalities are required. Previously, the authors have reported studies on vowel-, consonant, and isolated word recognition in Cued Speech for French. In the current study, continuous phoneme recognition experiments are presented using data from a normal-hearing and a deaf cuer. In the case of the normal-hearing cuer, the obtained phoneme correct was 82.9%, and in the case of the deaf cuer 81.5%. The results showed, that automatic recognition of Cued Speech shows similar performance in both normal-hearing and deaf cuers.
Year
Venue
Keywords
2012
European Signal Processing Conference
Cued Speech,hidden Markov models,fusion,phoneme recognition
Field
DocType
ISSN
Consonant,Computer science,Word recognition,Gesture recognition,Cued speech,Speech recognition,Natural language processing,Visual communication,Artificial intelligence,Vowel,Phoneme recognition,Spoken language
Conference
2076-1465
Citations 
PageRank 
References 
0
0.34
3
Authors
3
Name
Order
Citations
PageRank
Panikos Heracleous16816.27
Denis Beautemps25716.31
Norihiro Hagita32877259.10