Diagnosis of skin cancer using convolution neural network

Abstract

Skin cancer is one of the most common skin cancers, and its early detection helps reduce the spread of the disease and reduce the death rate, but visual detection sometimes leads to misdiagnosis due to human errors resulting from visual fatigue, and the wrong diagnosis is costly and stressful for the patient psychologically.
Therefore, it was necessary to develop a working methodology to diagnose skin cancer automatically using an algorithm based on image processing, where a database of skin cancer images belonging to people of different age, gender and color was collected, and the difference in background, distance from the camera and the location of the mole was taken into account, and then the application was applied. Some image processing techniques from primary processing and image segmentation. Based on previous reference studies, we will build an algorithm based on convolutional neural networks to detect skin cancer and classify it into malignant or benign conditions, It achieved an accuracy of up to 96%.