A Systematic Analysis of Skin Cancer Detection Using Machine Learning and Deep Learning Techniques

Abstract

The skin serves as the primary line for protection against oxidative damage by UV rays on the outside of the body. Skin cancer is now the most often reported malignancy in the world, placing a great effect on economy as well as public health. The most frequent kind of cancer in Caucasians, encompassing both melanoma and non-melanoma, is skin cancer. In this proposed taxonomy, a complete overview of recent developments relevant to cancer diagnosis was discussed. Moreover, a review of melanoma cancer detection methods using machine learning and deep learning with image processing techniques was conducted. It also includes comparison of performance metrics, results, and publicly accessible datasets as well. Conclusion drawn from this study was that deep learning, CNN based models provide more accurate results across variety of datasets among all other machine learning and deep learning models and this also leads directions to the other researchers that which areas will be covered in future.