Rule Induction Based-On Coevolutionary Algorithms for Image Annotation

This paper describes our experiments in the field of evolutionary algorithms for rule extraction applied to automating image annotation and classification problems. Presented approach is based on classical evolutionary algorithm with binary representation of 'if-then' rules. We want to show that some search space reduction techniques make possible to get problem's solution. Paper shows that the gap between classification and image annotation problem can be bridged easily. Some experiments with EA parametrization in image annotation problem are presented. There are presented first results on ECCV 2002 dataset in image annotation problem.