Skin Cancer Detection Using Convolutional Neural Network (CNN)

Abstract

Worldwide cancer patients are increasing, Skin Cancer is one of most spreader type of cancer. Early skin cancer identification is crucial and can help prevent some skin cancers, and because of the high expense of treatment, quick growth rate, and mortality rate of melanoma skin cancer. The majority of the time, treating cancer cells requires time and manual detection. Although there is a lot of work in this field, but a few have targeted offering solution within mobile applications which is need by many patients as a primary disease discovery. The proposed solution is to combine both efforts. First the deep learning method convolution neural network (CNN) and image processing strategy for an artificial skin cancer diagnosis system which was used to detect the type of skin cancer, with using of the ISIC2018 dataset, with transfer learning model InceptionV3 which was used for fine-tuning. Secondly Mobile Application has developed with the proposed detection technique to help patients in initial examination for Skin Cancer, and facilitate the process of follow up with medical team. The results shows an accuracy rate was 85.8% in skin cancer detection.