A New Algorithm for Melanoma Detection

Abstract

Melanoma is considered as one of the most dangerous skin tumors and most deadly, however the early detection of this disease helps in treating it successfully. This research presents an automated detection system based on RGB technique for the processing of pigmented lesion colored image and isolating it from the other elements, and then adding HSL technique to support and improve the detection of the lesion, which we have achieved since the color range of the pigmented lesion was identified. Using RGB and HSL techniques together has helped in distinguishing between Lentigo and Melanoma cancerous diseases despite their color resemblance, also between vascular lesion and Lentigo. What increases the system reliability is that it doesn’t confuse these diseases