Skin cancer is a frequent and fatal tumor in the world. Melanoma is the primary source of skin cancer deaths, although constitutes only a low fraction of all skin cancer types. Dermoscopy was introduced to aid professionals and raise the diagnostic rate by accurately detecting the condition in its early stages. A new research study has demonstrated that deep learning is particularly significant for detecting skin cancer and clinical diagnosis. In this case, we recommend applying two deep learning (DL) techniques to identify both benign and malignant skin cancers. The first approach involves acquiring features from CNN pre-trained networks and employing machine learning (ML) categorization to produce predictions. The other approach involves fine-tuning and data augmentation. The paper evaluates two approaches for performing classification and time efficiency. The research study uses a public database with photos of melanoma and benign skin lesions. Research findings show that pre-trained frameworks operate better with regard to accuracy and processing time.