3D motion retrieval with motion index tree

摘要: With the development of Motion capture techniques, more and 3D motion libraries become available. In this paper, we present a novel content-based retrieval algorithm. We partition library construct index tree based on hierarchical description. The serves as classifier to determine sub-library that contains promising similar motions query sample. Nearest Neighbor rule-based dynamic clustering algorithm is adopted tree. similarity between sample in calculated through elastic match. To improve efficiency calculation, an adaptive clustering-based key-frame extraction adopted. experiment demonstrates effectiveness