ABSTRACT
In recent decades, the incidence of malignant melanoma as a deadly skin cancer has increased worldwide. With its high medical costs and death rates, this cancer has prioritized the need for early diagnosis. Computer-based detection systems can improve the diagnosis rate of melanoma by 5%–30% compared to the naked eye and reduce human error. Although much effort has been made to advance the detection of skin cancers, there are still serious concerns about it. This chapter introduces automatic skin cancer diagnosis and an overview of methods in each step toward detection. A novel algorithm in feature selection and classification stages of automatic skin cancer diagnosis is designed and implemented to identify malignant and benign lesions. A smart algorithm is proposed based on inertia-based particle swarm optimization (IPSO) and the self-advising SVM (SA-SVM). This algorithm optimizes the feature selection stage. Additionally, SA-SVM, known as a new classifier in skin cancer detection systems, is employed along with the proposed algorithm. The statistical and performance measurement analyses of algorithms are presented to prove the superiority of the proposed algorithms.
