B inary images, as discussed in the preceding chapter, consist of pixels and groups of pix-els selected on the basis of some property. The selection may be performed by thresh-olding brightness values, perhaps using several color channels, or after processing to extract texture or other information, or based on inclusion within a delineated boundary. The goal of binarization is to separate features (the designation used in this text for groups of selected, contiguous pixels) representing objects or structures from the background, so that counting, measurement, or matching operations can be performed.