Extraction of Primitive Motion and Discovery of Association Rules from Human Motion Data

In past several years, more and more digital multimedia data in the forms of image, video and audio have been captured and archived. This kind of new resource is exiting, yet the sheer volume of data makes any retrieval task overwhelming and its efficient usage impossible. In order to deal with the deficiency, tagging method is required so as to browse the content of multimedia data almost instantly.In this paper, we will focus on tagging human motion data. The motion data have the following features: movements of some body parts have influence on other body parts. We call this dependency motion association rule. Thus, the task of tagging motion data is equal to the task of expressing motion by using motion association rules. Association rules consist of symbols, which uniquely represent basic patterns. We call these basic patterns primitive motions. Primitive motions are extracted from the motion data by using segmentation and clustering processes. Finally, we will discuss some experiments to discover association rules from multi-stream of the motion data.