ABSTRACT
In the real world, the human visual system can distinguish different target region relying on single frame image gray-level information alone. This ability of accurate target segmentation is inseparable with both developed human nervous system and long-term experiential learning. Another important factor is the visible boundary of spatial domain between target and background. If there is no such kind of boundary, human beings can’t distinguish any more; for example, in a dark room, people can’t make a distinction between objects.
