Selectionist relaxation: genetic algorithms applied to image segmentation

Abstract

This paper describes an unsupervised image segmentation method based on a fine-grained distributed genetic algorithm. Unlike other proposed applications of genetic algorithms to this problem, the method does not require the definition of an objective fitness function evaluating candidate segmentation results. The output segmentation instead emerges as a by-product of the evolution of a population of chromosomes that are mapped onto the image and that locally adapt to its features. A sketchy analysis of the algorithm is proposed, according to which the optimal GA parameters can be predicted. The predictions are experimentally tested on artificial data. Results obtained on natural data are reported and compared with the output of a standard region segmentation method.