Learning physics-based motion style with nonlinear inverse optimization
ACM SIGGRAPH 2005 Papers
Style translation for human motion
ACM SIGGRAPH 2005 Papers
| Hi-index | 0.00 |
There is a great demand for motion style deformation in motion editing. Style deformation is a process of transforming a human motion from one emotion to another emotion. The existing commercial animation systems (e.g. Autodesk MotionBuilder, 3ds MAX, and Maya), provide a powerful tools for animator to edit various types of motions but requires skills and knowledge to familiarize and master the animation system. Thus, it will take much time. The previous motion style deformation methods [Terasaki and Oshita 2006][Hsu et al. 2005][Liu et al. 2005] are focusing on automatic approach to learn a specific style from existing motion data and then deforming an input motion by applying the learned motion style. To use these methods, the user must prepare motion data with the specific style that the user wants to realize. Moreover, since the entire process is automatic, the user does not have freedom to tune the specific style.