Implementación de un Sistema de Detección Temprana de Melanoma Humano usando el software Matlab

Abstract

The present project proposes a software developed in Matlab, which provides an preliminary classification of skin lesions as melanoma, lunar or suspicious condition. Many of the criteria analyzed in a diagnostic detection are given through visualization. Hence the idea of concentrating efforts on digital image processing. The developed system extracts from the dermatoscopic images relevant parameters such as: edge, symmetry and color. These parameters are quantified through the ABCD Rule and Method Menzies.The operation of the software is evaluated on a database of dermatoscopic images obtained through the American Cancer Society and the Skin Center (CEPI UIO). These images are a compendium of clinical cases that indiscriminately cover the conditions under study. The values achieved by the software in both sensitivity and specificity allowed the project to be endorsed by the medical staff of the dermatoscopy area of the Solon Espinoza Ayala Oncology Hospital (SOLCA) and the Skin Center (CEPI UIO).