Published Paper Details:

The depleting ozone capacity causes more Ultra Violet (UV) rays towards the earth and can cause skin cancer. World Health Organization (WHO) states that people with skin cancer continues to increase from year to year in each decade. There are about 2-3 million non-melanoma skin cancer patients and 2 million patients with the disease melanoma skin cancer. Therefore it is necessary to detect skin cancer early to reduce the number of skin cancer sufferers. Early detection of skin cancer can be done by doctors by looking directly based on the ABCD feature, but this method is less efficient because visual response can be different for different persons, so Computer-Aided Diagnosis (CAD) was made to help parties in detecting skin cancer. The process carried out is pre-processing by using an image or color augmentation filtering, using noise removal filtering, brightness, contrast enhancement using histogram equalization and resizing. Furthermore, the classification process is carried out using Convolutional Neural Network (CNN) to classify skin cancer. The study was conducted on 2808 benign skin cancer data and 2394 non-cancer malignant data totaling 5202 data. Furthermore, the data will be broken down with the distribution and divided into 80% training data and 20% testing data. The best results obtained by the proposed model using maximum of 1024 hidden layers with an accuracy value of 94.85%, and a loss of 22.96%.