Facing the Machine: Data-Driven Expressive Facial Animation Synthesis and Editing: 03.30.06
Computer facial animation is one of the most challenging problems in computer graphics field. In this talk, I will present our eFASE system that automatically generates full expressive facial animation including visual speech, eye motion, and head motion while providing high-level intuitive user controls. Based on a large expressive facial motion database composed of captured motions of an actress, our system concatenates captured motion data in accordance with animator-established constraints and goals using a constrained dynamic programming search, given novel phoneme-aligned speech input and emotion specifications. To model incidental facial gestures, I will present a texture-synthesis based technique for synthesizing realistic eye motion and a trained Hidden Markov Models (HMM) based approach for modeling audio-driven head motion.
I will also describe my latest research efforts in blendshape animation:
reduction of interference among blendshape basis and the direct mapping of facial motion capture data onto blendshape faces. With these problems resolved, the painstaking effort of repeatedly tuning the weights of blendshapes can be significantly saved.
Speaker Bio:
Zhigang Deng is a Ph.D. candidate (expected 05/2006) in the Department of Computer Science and a Graduate Research Assistant at the Integrated Media System Center (NSF Engineering Research Center for Multimedia and Internet Research) at University of Southern California. He has received his M.S. in Computer Science and B.S. in Mathematics from Peking University and Xiamen University respectively. His research interests include Computer Graphics, Computer Animation, Interactive Media and Gaming, Information Visualization, and Human Computer Interaction. He is a student member of ACM, ACM SIGGRAPH, and IEEE Computer Society.
