What is the role of color symmetry in the detection of melanomas?

Abstract

Several computer aided diagnosis (CAD) systems have been proposed to detect melanomas in dermoscopy images. Most of them rely on the extraction of several types of visual features: color, texture and shape. However, the role of each type of feature is seldom assessed. This paper proposes several features for the analysis of color symmetry in skin lesions and assesses their performance under a wide variety of system configurations. We have obtained very high detection scores (SE=96%, SP=83%) by only using color symmetry features, showing that they play a major role in the analysis of dermoscopy images. © 2013 Springer-Verlag.