Visual objects in the real world are seen in contextual scenes. These contexts are usually coherent in terms of their physical and semantic content, and they usually occur in typical configurations.