Abstract
Skin cancer is an excessively common type of cancer. It occurs when mutations appear in the DNA of skin cells. The four main forms of skin cancer are Basal cell carcinoma (BCC), Squamous cell carcinoma (SCC), Merkel cell carcinoma (MCC), and the deadliest form Melanoma. Skin cancer is usually diagnosed late because of the unnoticeable symptoms. Therefore, Reliable automatic detection of skin tumors is needed to help increase the accuracy and efficiency of pathologists. In this paper the DCNN method is used which is designed to perform complex analysis of 2594 images and 2594 of corresponding ground truth (response masks) for training and 1000 images for testing of data using image segmentation and classification for creating model that detect skin cancer in early stages. The models testing produced positive outcome with accuracy 0.95 for classification and 0.895 for segmentation. The results are promising for future enhancement.