FOCUSED COLOR INTERSECTION WITH EFFICIENT SEARCHING FOR OBJECT EXTRACTION

We propose focused color intersection with efficient searching for ide ntifying and extracting the objects in a complex scene based on color similarity. The method matches the models against different parts of a scene, called focus regions, using normalized color histogram interse ction. The best matching focus region is determined by an efficient se arch strategy employing upper bound pruning. This search strategy, cal led active search, concentrates its effort on parts of the scene havin g high similarity with the object. Consequently, it achieves a large r eduction in computational effort without sacrificing accuracy. An effi cient algorithm for evaluating the color histogram intersection betwee n a model and a focus region is also given. Experiments conducted demo nstrate that multiple known objects in complex scenes can be extracted by this process. The method is stable against scale changes, two-dime nsional rotation, moderate changes in shape and partial occlusion. (C) 1997 Pattern Recognition Society. Published by Elsevier Science Ltd.