Eurographics/ACM SIGGRAPH Symposium on Computer Animation (2005) K. Anjyo, P. Faloutsos (Editors) Motion Modeling for On-Line Locomotion Synthesis Taesoo Kwon and Sung Yong Shin † Korea Advanced Institute of Science and Technology Abstract In this paper, we propose an example-based approach to on-line locomotion synthesis. Our approach consists of two parts: motion analysis and motion synthesis. In the motion analysis part, an unlabeled motion sequence is first decomposed into motion segments, exploiting the behavior of the COM (center of mass) trajectory of the performer. Those motion segments are subsequently classified into groups of motion segments such that the same group of motion segments share an identical footstep pattern. Finally, we construct a hierarchical motion transition graph by representing these groups and their connectivity to other groups as nodes and edges, respectively. The coarse level of this graph models locomotive motions and their transitions, and the fine level mainly captures the cyclic nature of locomotive motions. In the motion synthesis part, given a stream of motion specifications in an on-line manner, the motion transition graph is traversed while blending the motion segments to synthesize a motion at a node, one by one, guided by the motion specifications. Our main contributions are the motion labeling scheme and a new motion model, embodied by the hierarchical motion transition graph, which together enable not only artifact-free motion blending but also seamless motion transition. Categories and Subject Descriptors (according to ACM CCS): I.3.7 [Computer Graphics]: Animation 1. Introduction Demand for on-line, real-time motion synthesis has ever been increasing, propelled mainly by computer games and VR applications. Recent technological advancement in mo- tion capture and reuse increases the potential for on-the-fly motion synthesis to replace playback of prerecorded mo- tions according to scenarios. There have been two main re- search streams in real-time motion synthesis: motion blend- ing [GR96, PSS02, PSS04, RCB98, WH97, KG04] and mo- tion rearrangement [AF02, AFO03, GJH00, KGP02, LCR ∗ 02, LWS02, PB02, TH00]. The former increases efficiency and controllability, whereas the latter retains naturalness embed- ded in captured motions. A hybrid approach has recently been proposed to combine the idea of motion blending and that of motion rearrange- ment. In this approach, motion transition graphs are instru- mental in combining these ideas, as demonstrated in locomo- tive motion generation [PSS02, PSS04] and rhythmic motion synthesis [KPS03]. The origin of these graphs is traced back to “verb graphs" in [RCB98], which were used mainly for † syshin@jupiter.kaist.ac.kr motion style control. Later, Park et al. [PSS02, PSS04] en- hanced them for on-line motion blending. A node in a motion transition graph represents a group of basic motions of an identical structure, and an edge rep- resents the transition from a blended motion to a blended motion (possibly including self-transition). Given a stream of motion specifications, the graph is traversed from node to node, while blending motions at nodes and making transi- tions at edges. To facilitate the hybrid approach, the major premise is the availability of labeled motion data satisfying the following two properties. • The group of motions at each node should have an identi- cal structure for motion blending. • The group of motions at a node should transit seamlessly to that of motions at a node connected by an edge (possi- bly a self-edge). A straightforward way to prepare these motion data would be to direct a performer to repeat motions with an iden- tical structure but different aspects. Aside from their ac- curacy, however, the resulting captured motions would be quite unnatural. A better approach would be to let the c  The Eurographics Association 2005.