Effect of Color Enhancement on Early Detection of Skin Cancer using Convolutional Neural Network

The prevalence of skin cancer is significantly increasing each year due to the damage of the ozone layer in the atmosphere that makes more ultraviolet radiation passing through. Knowing this situation, it is important to develop a simple image processing technique that can be used in the early detection of skin cancer. The skin cancer detection becomes highly active research since 2016 due to the ISIC has released a large skin cancer image dataset. Several types of research propose hand-crafted image processing with machine learning, but the technique is a little bit complicated. The aim of this study is to investigate the effect of simple image processing technique, contrast enhancement using CLAHE and MSRCR as contrast enhancements with CNN. The results show that compares to MSRCR, CLAHE is more suitable to be used in color image enhancement for early detection of skin cancer using CNN. But, the original and CLAHE-enhanced dataset give the same accuracy in the training and validation. The main contribution of this study is that the image contrast enhancement is not required for the skin cancer screening purpose.