Genetic optimization mapping applied to medical image segmentation

Abstract

A number of important optimization problems have been classified as mapping applied towards segmentation of important features. The segmentation of important features can be formulated as configurational mapping problems by representing mapping configurations as solutions to problems of interest. One example of such configuration mapping is found in image segmentation where an image can be represented as unique subsets of a complete image and then evolved through mapping to become a segment of specific interest within an image. An effective segmentation mapping algorithm must determine the specific image subsets of an image field that best exhibit an a priori set of quantitative and qualitative characteristics. In this paper, a Genetic Optimization Mapping Algorithm is used to produce a population of sub-images, characteristic of specific image subsets of interest that were tested via a quantitative objective function, ranked using a linear fitness scheme, and modified using a genetic Crossover operator. The mapping algorithm is found to converge, within fifty to one hundred generations of maps, to a good fit to the targeted mapping configuration in a very robust and efficient manner.