基于多分辨率模型的快速运动过渡生成算法  

A Fast Motion Transition Algorithm Based on Multi-resolution Model

作　　者：杨春玲[1] 董传良[1] 

机构地区：[1]上海交通大学,上海200030

出　　处：《计算机仿真》2007年第1期186-187,195,共3页Computer Simulation

摘　　要：运动捕获技术可以记录人体关节运动的细节,是当前最有前景的计算机动画技术。然而,运动数据的重用性一直是个难点,为此,多种运动编辑手段被提出。运动过渡是一种常见的编辑技术,它可以将输入的两端运动序列拼接,形成新的运动序列。其中,过渡点选择的合理与否直接影响着结果运动的质量。在两运动间选择过渡点,需要对输入运动的每一对帧之间分别计算帧间的距离,其计算复杂度是O(n2)的,通过引入多分辨率模型,文中将该复杂度降低到O(n log2n),同时试验结果表明,此方法并未损害到结果运动的质量。Motion capture technique, which can record the details of human joint motions, is a promising computer animation technique. However, how to reuse the captured motion data has been an unsolved problem. As a result, many editing means are proposed. Motion transition, which can concatenates two input motions into a new motion, is a widely used editing technique in human animation field. In motion transition, the selection of transition point influences the quality of result motions greatly. The complexity of transition selection algorithm between two motions is O （n2）. In this paper, a multi -resolution model is introduced to reduce the complexity to O（nlog2n）. Moreover, the experiment results show that our algorithm does not damage the quality of the result motions.

关 键 词：计算机动画 运动捕获 运动过渡 多分辨模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]