Facial Expression Generation in 3D Space

摘要

Facial expression is one of the most powerful resources for people to coordinate conversation and communicate emotions and other mental, social, and physiological cues. Because of human's high sensitivity for facial expressions in many areas such as game, movie, avatars and social robot tries to mimic these facial expressions to their artificial human models to increase realistic feature. This paper proposes a novel methodology of combining ratio mapping technique and difference feature adding mechanism to generate facial expressions for mesh models in 3D space. A 3D point cloud data is taken as the input and preprocessing techniques such as spike removing and down sampling are applied to remove noise points and unwanted data points to improve the smoothness of construction model. Then a 2.5D mesh reconstruction algorithm is applied to generate a 3D face mesh. In the final stage proposed techniques applied to the generated face meshes.