A Critical Analysis of Computer Aided Approaches for Skin Cancer Screening

Skin Cancer is life-threatening when diagnosed at a later stage. Early detection of skin cancers such as melanoma indicates a higher survival rate for the patient. Non-computer aided tools were used in the past such as the visual inspection using tools like the dermoscopy. Commercial tools were later introduced that allowed the examiners to examine the images obtained from the dermoscopy using techniques such as the ABCD rule and 7-point checklist. Deep Learning has proven to be the state-of-the-art for computer vision problems such as image classification. A lot of research has been carried out in the application of deep learning for automating skin cancer screening. This paper presents an analysis of the existing work carried out in the area of automatic skin cancer screening and the different steps involved in building a skin cancer classification tool for skin cancer screening. The limitations of the various existing approaches are explored, and the results of the analysis will be used as part of an ongoing research to design and develop a robust system that will address the identified cons.