基于MPEG-4和PCA的人脸动画合成方法

三维人脸动画技术融合了多通道交互技术和多功能感知技术,在虚拟现实、虚拟主持人、虚拟会议、辅助教学、医疗研究、电影制作、游戏娱乐等很多领域有广泛的应用。文章分析了基于MPEG-4标准的人脸动画合成方法,并针对其在实际过程中产生的两个问题,包括人脸模式太多和局部调整带来的影响,分别采用聚类和主成分分析(PCA)的方法进行改进。实验表明基于PCA来解决局部调整人脸模式的方法可以部分实现动画过程中表情的表达,从而极大地提高了人脸动画合成的可操作性。

A Facial Animation Synthesis Approach Based on MPEG-4 and PCA

Combined with Multi-Modal Interaction and Multi-functional Perception technologies,the technique of three-dimensional human facial animation can be widely applied in the fields such as virtual reality,virtual presenter,virtual meeting,network education,medical treatment researches,movie effects,games and so on.This paper has investigated the facial animation synthesis approach based on MEPG-4standard,and then describes how to use the clustering and prin-cipal component analysis(PCA for short)algorithms to partially resolve the two practical issues in the model application process,including the problem of the overmany facial patterns and the influence of local adjustment.Experiments show our PCA-based approach for locally adjusting facial patterns can remarkably boost the interoperability of facial anima-tion synthesis process.