基于仿生关节链球坐标系的虚拟手势生成

Virtual gesture generation based on biomimetic articulated chain-sphericity coordinate

为解决传统手势获取手段外设昂贵、使用复杂等问题,提出基于仿生关节链球坐标系的虚拟手势生成方法。根据手部特征设计虚拟手骨结构,对其中运动节点建立不同的链球坐标系,改进原有的坐标系转换算法,可通过调整角度参数产生任意静态手势。利用同源连续性原理将静态手势转化为高维空间点,仿照二维点运动轨迹规律进行采样,产生动态手势序列。结合提出的链球坐标系角度判断方法,通过真实姿态模拟,验证算法可以准确地生成指定手势,是方便高效的虚拟手势生成方法。

To reduce the cost of traditional gesture acquisition methods and to improve the simulation efficiency,a virtual gesture generation method based on biomimetic articulated chain-sphericity coordinate was proposed.Based on the improved algorithm of coordinate system transformation,various gestures were generated by changing the coordinate parameters of key nodes in virtual hand skeleton.According to the principle of homology continuity,the static gesture was a point in high dimensional space which could generate dynamic gesture sequence by sampling in some direction.With the new judgment method of angles in biomimetic articulated chain-sphericity coordinate,through real posture simulation,it is verified to be a convenient and efficient way to ge-nerate virtual gestures based on the new coordinate.