Classification of Images using Color, CBIR Distance Measures and Genetic Programming - An evolutionary Experiment
MetadataShow full item record
In this thesis a novel approach to image classification is presented. The thesis explores the use ofcolor feature vectors and CBIR ? retrieval methods in combination with Genetic Programming toachieve a classification system able to build classes based on training sets, and determine if animage is a part of a specific class or not.A test bench has been built, with methods for extracting color features, both segmented andwhole, from images. CBIR distance-algorithms have been implemented, and the algorithms usedare histogram Euclidian distance, histogram intersection distance and histogram quadraticdistance. The genetic program consists of a function set for adjusting weights which corresponds tothe extracted feature vectors. Fitness of the individual genomes is measured by using the CBIRdistance algorithms, seeking to minimize the distance between the individual images in the trainingset. A classification routine is proposed, utilizing the feature vectors from the image in question,and weights generated in the genetic program in order to determine if the image belongs to thetrained class.A test?set of images is used to determine the accuracy of the method. The results shows thatit is possible to classify images using this method, but that it requires further exploration to make itcapable of good results.