Bridging Perception and Action: Spatially-Grounded Mid-Level Representations for Robot Generalization

Open in new window