Melanoma is the leading type of skin cancer that causes death in a large number of people yearly. The key to survival from this disease is early detection. Currently, the medical community uses microscopy, a very complicated procedure, to detect Melanoma. This research plans to work on the disadvantage of the microscopy procedure by creating a deep neural network that will be able to diagnose skin lesions digitally with high accuracy. The proposed approach is expected to provide a cost-effective tool available to communities with limited access to medical care. Currently, the network has shown that it could be able to increase the accuracy and systematically analyze the trade-offs between hits, correct rejections, misses, and false alarms. Presenting that, the network can further be improved to an accuracy higher than that of microscopy, to become a better option for the detection of skin cancer.