HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos

ArXi:2605.24934v1 Announce Type: cross Human egocentric video captures rich manipulation nstrations without any robot hardware, yet transferring these skills to robots remains challenging due to the embodiment gap between human and robot in both visual appearance and kinematics. We present HumanEgo, a framework that bridges the embodiment gap by lifting each human nstration to an entity-level representation of hand-object interaction, and