ABSTRACT

The ability of the human brain to extract objects from a cluttered visual scene is called segmentation. Visual segmentation may be based on the coherence of features, such as luminance, color, motion, or depth, that define a given object. Alternatively, objects may be defined by the contrast of their features with those of the surrounding objects. Natural objects are usually defined by both-coherence of their internal features and contrast with the surrounding objects. Their borders are usually defined by more than one visual attribute. In everyday life, figures are automatically, effortlessly extracted from a visual scene. This remarkable ease is presumably based on the mutual facilitation of binding and segmentation along multiple stimulus dimensions.