Contour extraction from natural scenes is an important and essential task performed by the visual system. Contours delineate and segregate objects from their backgrounds. Psychophysical studies reveal that this seemingly robust mechanism is also susceptible to errors, which we perceive as ‘visual illusions’. The Kaniza Square (see figure) is one such illusion where an object is clearly perceived when there is no continuous square region in the image itself.
The Boundary Contour System (BCS) is a neural model of contour extraction and interpolation. The BCS attempts to explain the formation of all contours, both real and illusory, within a common framework. The BCS model successfully explains many psychophysical observations and in fact predicted the existence of cells in the lower-level visual system that fire for both real and illusory contours.
However, the BCS is difficult to parameterize for complex natural or synthetic images. A comprehensive study of the parametric ranges is now underway.
The Boundary Contour System (BCS) is a neural model of contour extraction and interpolation. The BCS attempts to explain the formation of all contours, both real and illusory, within a common framework. The BCS model successfully explains many psychophysical observations and in fact predicted the existence of cells in the lower-level visual system that fire for both real and illusory contours.
However, the BCS is difficult to parameterize for complex natural or synthetic images. A comprehensive study of the parametric ranges is now underway.
