This lecture presents a paper on how to enable robots to comprehend the world similarly to humans, focusing on semantic understanding and affordance learning. The framework processes RGB-D inputs from multiple cameras to build a detailed map of the environment at the object level.