摘要
描述一种3维人脸模型的单视频驱动方法。该方法在传统肌肉模型的基础上,根据嘴部运动特性采用机构学原理建立嘴部运动控制模型,通过视频图像序列跟踪得到特征点的运动规律曲线,进而驱动眼部、嘴部及面部其他部分的网格点运动,产生具有真实感的面部表情动作。仿真结果表明,采用此方法可以得到逼真的人脸表情模拟动画。
A method to drive 3D facial model using single video was presented. Based on the movement characteristics, the mouth was controlled by using a mechanism model. The motion trace of the feature points was acquired by tracking video image sequence. Then we drive the grid nodes on the facial model to obtain a realistic facial animation. The results show that this method could simulate the facial movements and expressions very well.
出处
《中国图象图形学报》
CSCD
北大核心
2011年第3期449-453,共5页
Journal of Image and Graphics
基金
国家高技术研究发展计划(863)项目(2007AA01Z324)
关键词
机构学
肌肉模型
情感表达
人脸仿真
theory of mechanism
muscle model
expressing feelings
face emulation