A New Color Constancy Model for Machine Vision

Both physiological and psychological evidences suggest that the humanvisual system analyze images in neural subsystems tuned to differentattributes of the stimulus. Color module and lightness module are suchsubsystems. Under this general result, a new physical model oftrichromatic system has been developed to deal with the color constancyof computer vision. A normal color image is split into two images: thegray scale image and the equal lightness color image for the twomodules. Relatively, a two-dimensional descriptor is applied to describethe property of surface reflectance in the equal lightness color image.This description of surface spectral reflectance has the property ofcolor constancy. Image segmentation experiments based on colorproperty of object show that the presented model is effective.