Event
HumiTransfer transfers human-human interaction videos to humanoid robots
1 reports · 1 sources
Nghia Le and other authors propose the HumiTransfer framework, which converts in-the-wild human-human interaction videos into physically feasible human-humanoid interaction motions. The method first reconstructs 3D human motion with foot-ground contact, then transfers one party to the target humanoid embodiment through interaction modeling, and uses post-retargeting optimization to ensure balance, contact, and collision avoidance. The authors state that they have applied HumiTransfer to 1,200 in-the-wild videos, generating approximately 500,000 frames of interaction data for the 26-DoF Unitree R1 Edu and the 23-DoF HighTorque Mini-hi humanoid robots, forming the HumiMimic dataset; the authors also state that real-robot deployment confirmed the physical executability of the transferred motions in real environments.
Report timeline
- HumiTransfer transfers human-human interactions from in-the-wild videos to humanoid robots
The authors propose HumiTransfer, a framework that converts in-the-wild human-human interaction videos into physically feasible human-humanoid interaction motions. The method first reconstructs 3D human motion with foot-ground contact, then transfers one participant to the target humanoid embodiment through interaction modeling, and uses post-retargeting optimization to ensure balance, contact, and collision avoidance.
arXiv cs.RO ·