T his and the following chapter describe how the D ichrom atic Reflection Model and the sensor model can be transform ed into a com putational pro­ cedure for analyzing color images. This chapter describes how the models can be used for color image segm entation. T he goal of segm entation is to identify objects in an image, as delineated by m aterial boundaries. Because m ost color segm entation m ethods in the past have considered object color to be a constant property of an object and color variation to be caused by random cam era noise, they generally segm ent im ages no t only along m aterial boundaries b u t also along other lines exhibiting color variations, such as highlight and shadow boundaries, or object edges w ith significant shading changes. T he D ichrom atic Reflection Model provides an in terpre­ ta tio n scheme th a t relates the physics of light reflection to color changes in the image and the color space. T he approach to color im age segm enta­ tion described here uses the D ichrom atic Reflection Model to distinguish relevant color changes a t m ateria l boundaries from irrelevant ones due to shading or highlights.