The healthcare-related market for Artificial Intelligence (AI) is rapidly developing at an average annual rate of 45.8% (averaged over 5 years from 2019 to 2023) as AI technology is being introduced in the medical industry. A skin cancer AI study is currently being conducted at Kyungpook National University Hospital. While Korea has already researched AI-related skin tumors, our focus is specifically on skin cancer to develop AI for skin cancer differential diagnosis that can assist primary medical institutions in diagnosing and differentiating skin cancer. Unlike previous studies, this research specifically focuses on skin cancer by separating it into regions to account for the uniqueness of each location and enhance detection accuracy through image collection and analysis. We collected images using DSLR, Dermoscopy, and smartphones, and for binary classification, we divided them into benign and cancerous lesions. For 4-class classification, we classified them into benign, basal cell carcinoma, squamous cell carcinoma, and malignant melanoma. We collected a total of 7,327 images and used the ResNet-152 architecture for the convolution layers. We configured the fully connected layers such that the size of the subsequent layer is halved during forward propagation, with the size of the topmost FC layer specified as the number of prediction labels. As of now, we have confirmed an accuracy of 88.6%, and we plan to improve this accuracy by increasing the amount of data, introducing clinical information, and subdividing the classes.