Skin Melanoma Diagnosis Using Machine Learning and Deep Learning with Optimization Techniques: Survey

¹ Dep. Computer Science, College of Computer Science and Information Technology, Diwaniyah, Iraq
² Dep. Computer Science, College of computer science and information technology, Diwaniyah, Iraq

^* Corresponding Author: it.mast.23.6@qu.edu.iq

Abstract

Skin cancer is regarded as one of the most perilous forms of cancer and is recognized as a leading contributor to mortality worldwide. The likelihood of fatalities can be diminished significantly if skin cancer is identified at an early stage. Among the various types of skin cancer, melanoma stands out due to its remarkably high fatality rates. This is primarily attributed to its propensity to metastasize to other bodily regions if not promptly detected and treated. The process of diagnosing melanoma is notably intricate, even for seasoned dermatologists, primarily due to the extensive morphological diversity observed in patients’ moles. Consequently, the automated diagnosis of melanoma presents a formidable challenge that necessitates the development of proficient computational techniques capable of facilitating diagnosis, thereby assisting dermatologists in their decision-making process. In this study, we meticulously examined the most recent scientific papers on melanoma diagnosis, specifically focusing on applying deep learning and machine learning techniques in conjunction with optimization techniques.