Skin cancer parameterisation algorithm based on epiluminiscence image processing

In this paper a digital system for enhancing the diagnosis of skin cancer is presented. The algorithm is based on the standard ABCD dermatologic protocol, and on the measurements of contours and surfaces of some specific areas into the mole image captured by epiluminiscence techniques. Automatic algorithms of digital image processing have been developed in order to detect the appropriate segmentation of the images of suspected moles. This allows one to calculate some quantitative features based on the external contour using the Isodata Algorithm. In addition, other features are found based on the internal contours segmented using color heterogeneity criteria. These small internal moles help the diagnosis of the skin cancer to extract information about propagation and irregularities. The database used consists of 65 images already catalogued by dermatologists and the results are successful according to the assessment of medical experts.