CONICET | Buscador de Institutos y Recursos Humanos

? Melanoma is the most malignant type of skin cancer due to its ability to metastasize. Since early detection is the best way to prevent it, many detection techniques have been developed, such as the ABCD rule, the Menzies method and the 7-point checklist. Digital image processing is nowadays a powerful and useful tool for the analysis of biomedical images, resulting in great help in the diagnosis of many diseases. Consequently, by using this tool it is possible to develop software capable of automatically recognizing different patterns of pigmented skin lesions. In this paper we propose an algorithm based on the morphological analysis of an image of a pigmented skin lesion in order to characterize and quantify its malignity according to the ABCD rule developed by Stolz in 1994. The proposed method includes image enhancement techniques, the segmentation of the injured area together with morphological algorithms to obtain the ABCD characteristics, and the calculation of the Total Dermoscopy Score (TDS), which will let us classify the lesion as benign, suspicious or malign. The results obtained show that the proposed algorithm is reliable and could help the medical professional in the diagnosis of melanoma.