Synthesis Of Chinese Sign Language Coarticulation Based On Key Frame Animation

The viewpoint that sign language is the primary language of deaf people hasbeen accepted. However, in today’s society most information is based on soundlanguage, which greatly prevents deaf people from receiving information. The maingoal of sign language animation synthesis is to create an information platform fordeaf people to communicate without obstacle through an accurate, high-efficient signlanguage expression generating method. Along with the development of3D virtualperson technology, sign language animation based on virtual person has become themainstream of sign language synthesis research.In fluent sign language expression, signs are not mutually independent.Appearance of a sign in context will be impacted by its neighbors. This impactmakes shape, location, and speed of a sign conspicuously different in comparisonwith the situation when it appears alone. This phenomenon is called coarticulation ofsign language.The incorporation of coarticultion in sign language animation can improve thereality of the synthesized animation, but existing sign language synthesis researchhas not considered it. This dissertation aims at this problem and presents acoarticulation animation synthesis scheme, trying to enha nce the effect of thesynthesized animation.According to the different information transmitting ability of palm joints, aweighted Euclidean distance is used to judge the similarity between hand shapes.Motion displacement map is used to emphasize the similarity or dissimilaritybetween hand shapes. The impact of a non-neutral space sign on a neutral space signis also considered. Finally, a key-frame interpolation method which makes the jerkduring virtual character’s motion minimal is presented. Experimental results showthat the presented scheme makes the virtualcharacter’s behavior close to real human.Reality and naturalness is improved.