Typically, when AI is presented with a scene of cluttered objects, the system will try to identify each item based only on its visible parts.