Early Detection of Melanoma Skin Cancer Using Image Processing and Deep Learning

Abstract

AbstractThe skin occupies the largest portion of the human body, and thus, the risk of exposure to UV rays is high, and DNA damage can easily occur. Because humans benefit from sunlight, it is impossible to avoid sun exposure, so it is impossible to completely avoid diseases such as melanoma and skin cancer. Melanoma skin cancer is one of the most common cancers worldwide, with approximately 300,000 new cases reported in year 2018. However, this is likely to be an underestimated because the number of people diagnosed per year with different type of skin cancer is projected to increase over the next 20 years. In this paper, we propose an early detection mechanism for melanoma skin cancer using image processing and deep learning techniques. The image processing is used for image segmentation such as threshold, edge detection, and geometry-based feature extraction for melanoma features asymmetry, border, color, diameter, and evolving (ABCDE). Whereas, the deep learning is used trained the deep learning model to predict low/high risk of melanoma skin cancer. The results of our proposed solution are clearly showing that it is highly accurate to detect the melanoma skin cancer using our e-health application.KeywordsImage segmentationDeep learningHealthcareMelanomaSkin cancer