A gesture recognition architecture for sign language
Annelies Braffort · 1996
This paper presents a gesture recognition architecture dedicated to Sign Languages. Sign Language gestures include five co-occurring parameters, which convey complementary independent information. Some signs belong to a predefined lexicon which can be learned by the recognition system, but some other signs may be created during the discourses, depending on the context. The proposed recognition system is able to recognise both kinds of signs, by using specific classification tools, and a virtual scene for context storage. It is based on a study of French Sign Language (LSF).