基于人体运动规律的运动融合技术

Motion fusion technique based on characteristics of human motion

介绍一种不需要人工干预的运动融合方法,提出基于关节运动规律的运动周期判别方法。通过分析运动捕获数据,计算双膝与臀部节点连线夹角的变化确定运动周期,然后再进行时空变形、插值与约束重建,从而生成高质量的运动融合动画。实验结果表明,该算法能够准确计算出运动周期,并且使约束后的融合动作更加真实。

This paper proposed a motion fusion method without manual intervention, and suggested a way to identify the motion cycle based on joint movement. The cycle was calculated through the analysis of the motion capture data and the angle between the knees and the hips node. Then high-quality motion blending animation came into being after the procedures of time and space warping, interpolation, and constraints reconstruction. The experimental result shows that the proposed algorithm can calculate the motion cycle accurately and make the fusion motion under constraints more realistic.