None defined yet.
Vision as Proprioception: Estimating Humanoid Robot Pose from Video
Geometric Action Model for Robot Policy Learning