Automatic construction of 3D human face models based on 2D images

In order to achieve low bit-rate video coding, model-based coding systems have attracted great interests in visual telecommunications, e.g., videophone and teleconferencing where the human faces are the major part in the scenes. The main idea of this approach is to construct a 3D model for human face. Only the moving parts on the face are analyzed and the motion parameters are transmitted, finally the original facial expressions could be synthesized by deforming the face model using the facial motion parameters. We propose an approach to the automatic construction of 3D human face models using a generic face model and several 2D face images. A template matching based algorithm is developed to automatically extract all necessary facial features from the front and side profile face images. Then the generic face model is fitted to these feature points by geometric transforms. Finally, texture mapping is performed to achieve realistic results.