Abstract

In most of the numerical methods for facial expressions in the literatures, the number of the key controls for animators or robot programmers are too complex and trivial, because what they want are less and critical controls. This study explored varied types of human smiles and extracted proper key factors affecting the smiles. These key factors then were converted into a set of control points which could serve to satisfy the needs for creation of facial expression for 3D animators and be further applied to the face simulation for robots in the future. First, hundreds of human smile pictures were collected and analyzed to identify the key factors for face expression. Then, the factors were converted into a set of control points and sizing parameters calculated proportionally. Finally, two different faces were constructed for validating the parameters via the process of simulating smiles of the same type as the original one.

Key words: 3D animation, facial expression, numerical, robot, smile, parameter.