Skin Lesions Segmentation by Artificial Learning and Immune System

Medical image diagnosis increases the efficiency of health specialists. In order to further improve image diagnosis researcher has proposed various image processing models. This paper has proposed a model that works on skin lesions detection. Segmentation of the infected region of the image was done by the use of an artificial immune genetic algorithm. It was found that the use of artificial neural networks for the initial population generation has increased the cluster center selection of the genetic algorithm. The proposed model does not need prior information on an input image. Artificial affinity estimation for cloning and mutation reduces the iteration of the model. The experiment was done on reading skin lesion dataset. Results demonstrate that the suggested model, when compared to earlier existing skin segmentation models, has enhanced the work precision, recall, and accuracy values factors.