A Review on Early Diagnosis of Skin Cancer Detection Using Deep Learning Techniques

Cancer that is caused on the skin is nothing but the abnormal growth of the cells on the skin which is mainly caused when the skin is exposed to the UV radiation emitted from the sun, for a prolonged period of time. The fourth most common benign illness in the world, according to the Global Burden of Disease project, is a skin disease. The paucity of professional dermatologists and access to formal medical treatment make the diagnosis of dermatological illnesses The early detection or identification of skin cancer can help people in curing and preventing the skin disease which when left uncared can cause serious issues. Skincare identification at the early stage also significantly enhances the likelihood. So that the therapy becomes more effective. An image dataset that contains various images of the skin disease is taken which contains 3500 pictures collected for the purpose of the study. Using a variety of algorithms, including the SVM, CNN, VGG16, Resnet50 and ViT the collected images were assessed for loss and accuracy. In which ResNet50, SVM and VGG16 produced accuracy readings of 84.31, 83.4 and 82.4, respectively, and CNN produced a reading of roughly 97.6. CNN is the deep learning algorithm that is primarily used when comparing Machine Learning and Deep Learning Algorithms. SVM and VGG16 are machine-learning techniques used for classification.