A Relational Approach to Content-based Analysis of Motion Capture Data Meinard M¨ uller and Tido R¨ oder Universit¨at Bonn, Institut f¨ ur Informatik III R¨omerstr. 164, 53117 Bonn, Germany {meinard, roedert}@cs.uni-bonn.de Abstract. Motion capture or mocap systems allow for tracking and recording of human motions at high spatial and temporal resolutions. The resulting 3D mocap data is used for motion analysis in fields such as sports sciences, biomechanics, or computer vision, and in particular for motion synthesis in data-driven computer animation. In view of a rapidly growing corpus of motion data, automatic retrieval, annotation, and classification of such data has become an important research field. Since logically similar motions may exhibit significant spatio-temporal variations, the notion of similarity is of crucial importance in comparing motion data streams. After reviewing various aspects of motion simi- larity, we discuss as the main contribution of this paper a relational approach to content-based motion analysis, which exploits the existence of an explicitly given kinematic model underlying the 3D mocap data. Considering suitable combinations of boolean relations between specified body points allows for capturing the motion content while disregarding motion details. Finally, we sketch how such relational features can be used for automatic and efficient segmentation, indexing, retrieval, clas- sification, and annotation of mocap data. 1 Introduction Historically, the idea of motion capturing originates from the field of gait anal- ysis, where locomotion patterns of humans and animals were investigated us- ing arrays of analog photographic cameras, see Chapter ??. With technological progress, motion capture data or simply mocap data became popular in computer animation to create realistic motions for both films and video games. Here, the motions are performed by live actors, captured by a digital mocap system, and finally mapped to an animated character. However, the lifecycle of a motion clip in the production of animations is very short. Typically, a motion clip is captured, incorporated in a single 3D scene, and then never used again. For efficiency and cost reasons, the reuse of mocap data as well as methods for mod- ifying and adapting existing motion clips are gaining in importance. Applying editing, morphing, and blending techniques for the creation of new, realistic mo- tions from prerecorded motion clips has become an active field of research [3, 13, 17, 18, 30, 39]. Such techniques depend on motion capture databases covering a