基于样例生成面部表情基系统的设计与实现

Design and Implementation of Expression Radical Generation System Based on Examples

摘要: 在面部表情生成方面已经有了很多研究，其中有一种是将多个面部表情基通过线性组合来得到目标表情，这也就要求有一种可以自动生成面部表情基的系统来提供所需要的表情基。本文针对人脸结构的特征，将人脸模型转化为三角网格模型，再根据给定的样例表情基生成目标表情基。研究内容主要包括两个主要方面，即将源三角网格的形变表征为仿射变换和通过仿射变换的一致性约束网格形变保证网格形变的连续性。最后仿真结果表明基于样例生成面部表情基系统最终可以成功生成目标人脸模型的表情基并且具有较强的真实感。

Abstract: There are quite a few factors that determine the authenticity of the characters together, and facial expressions undoubtedly occupy a large proportion of them. And, there have been many studies on the generation of facial expressions. One of them is to obtain a target expression by linearly combining multiple facial expressions. That requires a system that can automatically generate facial expression to provide desired expressions. This paper focuses on a sample-based facial expression generation system. It mainly talks about using the affine transformation to represent the deformation of the source triangle mesh, and constrain the deformation mesh through the consistency of affine transformation. The final simulation result shows that the sample-based facial expression generation system can effectively imitate person’s facial expression.