基于最长轨迹投影的3D空间手写字符维数约简

Dimensionality Reduction Based on the Longest Trajectory Projections for 3D Space Handwritten Characters

提出一种基于最长轨迹投影的3D空间手写字符维数约简算法。首先,获取运动指尖的3D坐标,依次连接坐标点生成3D运动轨迹;将3D轨迹上所有的点分别投影到XOY、XOZ、YOZ平面形成2D轨迹;分别计算3个平面内的2D轨迹上相邻点的长度和,选择长度和最大的平面作为最佳投影平面。实验结果表明,所提算法可以得到固定方向的2D图像,不需要方向调整算法,就能够使3D空间手写字符的识别率达到96. 2%。

In order to overcome the above shortcomings,a 3D space handwritten characters dimensionality reduction method based on the longest trajectory projections is proposed in this paper.Firstly,the 3D coordinates of moving fingertip are got and connected to the 3D fingertip coordinates in turn to generate the 3D trajectory.Next,all points on 3D trajectory are projected to XOY,XOZ,YOZ planes and the 2D trajectories are formed.Finally,the length sum of the adjacent points on the 2D trajectory is calculated in the three planes.And the plane that has the maximum 2D projection trajectory length sum is selected as the optimal projection plane.Experimental results have shown that the proposed algorithm can obtain 2D images with fixed direction,and the recognition rate of 3D space handwritten characters can reach to 96.2%without the direction adjustment.