摘要
提出一种基于单张照片的真实感三维人脸表情合成方法，首先自动检测人脸和定位关键点；然后结合三维人脸模型实现人脸的重构；再根据肌肉模型的思想合成表情；最后合成模型的眼睛，增加表情真实感。该方法能够合成出逼真、自然的表情。
Abstract
Presents a 3D facial expression synthesis method based on a single photo, first by automatic face detection and positioning key point, then by reconstruction and combined the face with the 3D face model, then according the idea of muscle model synthesis the expression, final-ly, synthetic model eyes, increased expression of sense of reality. The method can synthesize realistic and natural expression.
NETL
NSTL
维普
万方数据