Psoriasis image analysis using color lacunarity

Texture analysis of color images is one of the challenging and complex problems to solve in order to perform operations like segmentation and classification in digital images. There are several approaches of textural features extraction among which the fractal-based approach is one of the most popular, due to the exhibited properties of invariance to scale and rotation. The fractal dimension and the lacunarity are the two fundamental measures used to describe a fractal set. We present a segmentation technique, which uses the lacunarity and furthermore, lacunarity-derived scalar features, as local texture descriptors. We use this approach for the analysis and segmentation of medical color images of psoriasis skin lesions, in the context of a dermatological application. We comment the results generated on a set of psoriasis images and we conclude the paper.