An Overview on Skin Cancer Detction Applyting Deep Learning Techniques from Dermoscopic Skin Lesion Images

Abstract

Skin cancer is the most prevalent type of cancer in the world that can be an abnormal growth, an unhealed wound, or an alteration in a mole for which detection of skin cancer becomes confusing in some cases. Almost all of the skin cancers are treatable and do not cause any death unless it spreads to or around the lymphatic nodes or any other organs. Early identification of skin cancer is necessary to ensure proper treatment in time to protect the patient’s precious life. Early detection of skin cancer can be significantly aided by automated methods. In recent times, deep learning models are employed for medical image processing and disease diagnosis. Among the deep learning models, the Convolutional Neural Network has drawn a significant amount of interest because of its great performance and adaptability. Besides CNN, GAN and machine learning techniques associated with deep learning models have left notable efficiency in detecting and classifying various categories of skin cancer. In the paper, a comprehensive analysis has been drawn answering research questions reviewing recent prominent journals related to detection and classification of skin cancer from dermoscopic skin lesion images.

Jahangirnagar University Journal of Information Technology Vol.12(1):2023 p 13-24