A Model of Figure/Ground Separation Based on Correlated Neural Activity in the Visual System

Humans and animals can recognize objects even in viewing situations where the objects are seen against a patterned background and when several objects are adjoining or intersecting each other. An extreme example illustrating the figure/ground separation power of the human visual system is the well-known picture of a dog by R.C. JAMES (Fig.1 ). When one looks at the picture, it seems as if the regions belonging to the object “dog” were somehow labeled to the extent that even contiguous black or white areas appear to be divided into sections that belong to the object and into others that belong to the background.