Evolutionary Methods in Clinical Diagnostics

This paper presents two methods for the segmentation of medical images based on evolutionary algorithms. Such image segmentation is a useful first step in the efficient and speedy analysis of medical images and provides a basis for the next stages of the image analysis process. In particular, the processing of images obtained from the inside of the human body allows for more effective diagnostics of lesions of human tissues. The first approach developed is based on evolutionary algorithms which divide the image into a given number of areas based on the intensity of RGB colours. The second segmentation method is based on the detection of edges in an image. An evolutionary algorithm was also used for this segmentation, based on the search for threshold values in the grey level in the image. We also compare the above segmentation methods with the numerical k-means segmentation method.