基于自由形状变形的三维人脸表情控制

Surface-Oriented Free-Form Deformation Based on 3D Facial Expression Control

人脸表情控制是生物特征识别研究的重要内容,本文提出了一种基于MPEG-4中FAP、FAT标准进行三维人脸表情合成的方法。首先对人脸进行关键点定义和区域分割,然后使用面向表面的自由形状变形方法(SOFFD)生成人脸表情。在特定人脸表情生成过程中,使用了基表情的比例合成方法。实验表明,该方法可以有效地合成各种真实的人脸表情。

Abstract： 3D facial expression control is very important in biometrics. This paper gives a method of 3D facial expression control based on FAP ＆ FAT from the MPEG-4 standard. First,we find the key points and divide the face into several segments, and then generate 3D facial expressions based on Surface Oriented Free-Form Deformation. An expression proportion synthetic method is used for generating the specific face expressions. Finally, the experimental results show that the method presented in this paper can synthesize face expressions efficiently.