Automated Diagnosis of Skin Cancer for Healthcare: Highlights and Procedures

This chapter outlines the framework of the automated dermoscopy‐based skin cancer diagnostic systems by highlighting the main approaches that are commonly followed throughout the different processes of the system. It introduces the dermoscopic‐based features which are derived from clinical rules, such as the ABCD rule, Menzies method and the 7‐point checklist to diagnose the malignancy of a lesion. The ABCD rule differentiates between malignant melanomas and the benign nevus through main criteria which are the asymmetry (A), the border irregularity (B), the color variations (C), and the diameter or differential structures (D). Subsequently, the chapter discusses the feature selection process, along with some of its techniques. The different classifiers that were adopted in the context of dermoscopy‐based skin cancer classification are reported. Finally, the chapter discusses the evaluation criteria of the classifier’s performance.