A facial image synthesis system for human-machine interface

摘要

The authors introduce a 'face' interface which is a user-friendly human-machine interface with multi-media and can realize a face-to-face communication environment between the operator and the machine. In this system, a human face appears on the display of the machine and it can talk to the operator with a natural voice. This paper describes the face motion and expression synthesis schemes which can be applied to this 'face' interface. The authors express a human head with a 3D model. The surface model is built by texture mapping with a 2D real image. All the motions and expressions are synthesized and controlled automatically by the movement of some feature points on the model. This 'face' interface is one of the applications of the model based image coding and media conversion schemes.