Extraction of Primitive Motions by Using Clustering and Segmentation of Motion-Captured Data

Abstract

Content-based retrieval of multimedia information has been investigated in several research projects. In this paper, we will focus on the automatic indexing method of motion information. Automatic indexing is performed in a pattern matching approach. Reference patterns are necessary for pattern matching, so that we will propose two methods to define primitive motions as reference patterns. One method divides motion data into segmental motion data by detecting the change of motion speed. The other classifies segmental motions such that similar segmental motions are gathered in the same cluster by nearest neighbor algorithm. In order to evaluate the similarity between two segmental motions, we use the dynamic time warping(DTW)method because each segmental motion takes different time length even if one person performed the same motions. Motion data can be converted into symbol sequences which represent primitive motions. Then continuous dynamic programming(CDP)method is used to recognize contents of motion. CDP is one of the extensions of DTW. It makes us possible to recognize a motion even if it is complex. At last, we use a bagging approach to merge the two similar clusters which have same primitive motions.