Relistic 3D Facial Expression Synthesis

In people's daily life, facial expression is one of the most important ways to communication and frankly give voice to person's motion such as joy, anger, sadness and so on. In recent years, as computer technology has developed rapidly, multimedia technology has entered a golden period of development. Because people have increasingly high desire of visual interaction of man-machine interface, which makes facial modeling and expression animation technology play more and more important role in the man-machine interaction. Three-dimensional facial expression animation technology has important applied value and theoretic meaning, which is widely applied to domains such as computer game and entertainment, filmmaking, man-machine interaction, advertising production, and so on. The ultimate goal of facial modeling and expression animation is to establish a highly realistic, self-adaptive and automatic real-time facial expression synthesis system. Through the researchers efforts made in recent years, three-dimensional facial expression animation technology has made great progress and development.To create more realistic three-dimensional expression animation, the research is focus on the three-dimensional expression animation technology based on MPEG-4. The main research contents are shown below:Based on information statistics method parameter to determine the human face of animation(1) Faical animation parameters determination based on information statisticsOn the basis of review of three-dimensional facial expression animation technology and analysis of Beijing University of Technology three-dimensional facial database, this paper proposes a method based on information statistics to determine the facial animation parameters of three-dimensional facial expression animation. Through experiment, we get the facial animation parameters of the six basic expressions respectively.(2) Facial region division and grid deformation method based on feature pointsFrom the analysis of facial action, this paper proposes a facial region division method to improve the independence of facial action. Based on discussion of existing animation, we also propose an interpolation algorithm with self-adaptive weight, which is based on feature points. This interpolation algorithm can avoid the tensile distortion of synthesized expression near the boundary of different FAP influencing area. Through the experiment, this system generates realistic animation of six kinds of basic expression, and the generated expression has a strong sense of reality. It also confirms the method proposed in this paper could reproduce human faceexpression really and naturally.(3) Simulation the facial movement character in the process of facial expression changingFrom the analysis of rule in the process of facial expression changing, this paper proposed a method to simulate the facial movement character based on feature points. This method decomposes the facial holistic movement character into feature points's movement character, which reduce the complex degrees of problem. Through analyzing the movement law of feature points along time axis, and simulating this law in the process of facial expression animation, finally we can realize real and natural facial expression animation along time axis.