Fractional Differentiation Based Image Enhancement for Automatic Detection of Malignant Melanoma

Abstract

Abstract Recent improvements in artificial intelligence and computer vision make it possible to automatically detect some of the abnormalities in medical images. Skin lesion is one of them and the treatments are greatly aided by the quick and precise diagnosis of these conditions. The identification and delineation of boundaries inside skin lesions has showed promise when using the basic image processing approach of edge detection, to be used in automatic detection. We investigate the use of fractional differentiation in this work to improve edge-detection in order to enhance the images. We propose a framework based on fractional differential filters for edge detection in skin lesion images. The derived images are then used to enhance the original images. Obtained images then undergo a classification process based on deep learning models. A well-studied dataset of HAM10000 is used in the experiments. The results show that these filters are effective in avoiding noise and detecting intricate edge details that can be informative during the recognition processes.