模板化的人体运动合成

为解决现有运动合成方法中控制方式过于复杂的问题,提出一种模板化的运动合成模型,旨在降低运动合成技术的应用门槛.利用稀疏主成分分析(Sparse principal component analysis, SPCA)、Group lasso和Exclusive group lasso对人体运动进行建模,使其对应的每一个低维参数只依赖于少数几个人体关节,构成人体运动的一个内在自由度(Degree of freedom, DOF),并具有直观语义;同时,每个关节被尽量少的低维参数所控制,以减少低维参数对彼此所控制的自由度的交叉影响.实验表明,通过直观地修改低维参数,就能够实时地控制每个参数对应的摆臂幅度、踢腿高度、跳跃距离等运动属性.这种模板学习、模板定制的两步方法,有效地降低了运动合成控制的复杂度,即便非专业人员也可以用其进行艺术创作.

Templated Human Motion Synthesis

Since the existing approaches to control human motion synthesis are too complicated, we propose a templated motion synthesis model to reduce the difficulty of using motion synthetic technology. We use sparse principal component analysis(SPCA), group lasso and exclusive group lasso to model human motions so that each low-dimensional parameter depends on a few human joints which form an intrinsic degree of freedom(DOF) with intuitive meanings. Meanwhile, our approach makes each joint controlled by as few low-dimensional parameters as possible to reduce the interferences between different DOFs. Our experiments demonstrate that users can control the motion features like amplitude of swing arm, kick height and jump distance by modifying the low-dimensional parameters intuitively in real time. This two-step approach of "template learning and template customization" can effectively reduce the complexity of synthesis control, and allows inexperienced users to create a realistic human animation quickly and easily.