基于KPCA降维的运动捕捉数据自动分割

针对运动捕捉获得的数据其维度通常较高,在进行运动片段分割时候需要进行大量计算,人工方法可以对运动捕获数据进行分割,准确率比较高,但是效率偏低的问题.提出KPCA结合PPCA对捕获数据进行降维处理的方法,将56维数据映射到2维空间;以及构造特征函数获得投射误差的导数,分析几何特征的变化,探测运动捕捉数据的分割点的方法,实现运动捕捉数据的自动分割.

Automatic segmentation of motion capture data based on KPCA

The dimension of motion capture data is generally higher during the movement split fragments requires a lot of computing time, artificial methods of motion capture data segmentation, the accuracy rate is relatively high, but low efficiency. PCA and KPCA are two commonly technique in dimension reduction. PCA is a linear dimension reduction techniques, and KPCA formulate the feature function is use in non-linear data of motion capture function. It can simplify the calculation. In this paper, we present the automatic segmentation technique of the motion capture data, in which the motion data is automatic segmented through KPCA combined with PPCA to reduce the dimension and project the 56 dimensional data in 2 dimensional space; formulate the feature function to obtain the derivative of projection error, detect the segmentation point of data through analyzing the change of geometric features to realize the automatic segmentation. Quanzhou chest-clapping dance is one of the national intangible cultural heritages of China, and the motion capture technique is an important method of the national intangible cultural heritages to realize its protection and inheritance. It is indicated in the experiment that the motion capture technique has certain feasibility.