We address the problem of robust lip tracking, visual speech feature extraction, and sensor integration for audio-visual speech recognition applications. An appearance based model of the articulators, which represents linguistically important features, is ...
Springer Verlag1998