Melanoma, or skin cancer, is usually detected visually from dermoscopic pictures, which is a time-consuming and difficult job for the dermatologist. Existing systems either utilize classic machine learning models that concentrate on hand-picked acceptable features, or deep learning-based approaches that learn features from full pictures. Melanoma, or skin cancer, is usually detected visually from dermoscopic pictures, which is a time-consuming and difficult job for the dermatologist. Due to many artifacts such as poor contrast, diverse noise, presence of hair, fiber, and air bubbles, etc., such a visual examination with the naked eye for skin malignancies is tough and onerous. This paper provides a robust and automated system for Skin Lesion Classification (SLC), in which image augmentation, Deep Convolutional Neural Network (DCNN), and transfer learning are all combined. Our lesion classification experiments show that the suggested technique can effectively classify skin cancer with a high degree of accuracy, and that it can also identify skin lesions for melanoma detection