The authors propose TopX, a flow-matching transformer that reads a skeleton’s tree, rest pose, and joint names, and outputs joint rotations and a root trajectory.
Text-to-motion model animates unseen skeletons from their skeleton file alone
TopX generates joint rotations and root trajectories for any kinematic tree without training clips, fine-tuning, or retargeting of the target character.
Top University
Tengjiao Sun · Pengcheng Fang · Xiaoyu Zhan · Dongjie Fu · Xiaohao Cai · Hansung Kim
University of Southampton · Nanjing University · MOGO AI
Research Digest··3 min read
Sun et al. present TopX, a 73M-parameter flow-matching transformer that generates motion for any kinematic tree directly from its skeleton file, tree topology, rest pose, and joint names, with no clips of the character and no per-rig normalisation.
The Daily Edition
The front page in your inbox each weekday morning: the lead, the picks, and why each one earned its place.