The problem of Personalized Face Animation is addressed in this paper by proposing the ShowFace streaming structure. This structure is based on most widely accepted industry standards in multimedia presentation like MPEG-4 and SMIL, extending them by defining image transformations required for certain facial movements, and also by providing a higher level Face Modeling Language (FML) for modeling and control purposes. It defines a comprehensive framework for face animation consisting of components for parsing the input script, generating and splitting the audio and video "behaviors", creating the required images and sounds, and eventually displaying or writing the data to files. This component-based design and scripted behavior make the framework suitable for many purposes including web-based applications. Various examples illustrate the efficiency of our method.
PERSONALIZED FACE ANIMATION IN SHOWFACE SYSTEM
