表演驱动的矢量化二维表情动画

Performance-Driven Two-Dimensional Vector Expression Animation

为了方便快捷生成生动的卡通表情动画,提出将表情数据重定向到卡通形象上的算法.该算法由用户在卡通形象上标记出与表情数据相对应拓扑结构的控制点,以此控制点集合与表情数据首帧控制点集合做比对;使用得到的比例关系将后续表情数据重定向到卡通形象上,从而形成针对卡通形象的表情数据.实验结果表明,运用文中算法可以制作出较为生动形象的卡通表情动画.

To produce vivid expression animation quickly and conveniently,we present a new algorithm to retarget the motion capture data onto the cartoon face.The algorithm allows users mark the cartoon face with the same topology as the expression data,and compares the first frame of the expression data with the marked control points.The resulting ratio is used to retarget the rest of the expression data onto the cartoon face,yielding the right expression data.The experimental results show that our approach is practical to generate vivid expression animation.