Irregularity and Asymmetry Analysis of Skin Lesions Based on Multi-Scale Local Fractal Distributions

To tackle the problems of detecting structural irregularity and asymmetry of skin lesion boundaries in the early CAD diagnosis of melanomas, a multiscaled local fractal algorithm is proposed in the paper where sequent removing textural irregularity from skin lesion boundaries is performed and features of boundary irregularity under multiscale are extracted. Meanwhile, a novel method of structural irregularity measurements, as a key diagnostic factor, is given and boundary asymmetry measurements using local fractals are also presented. Experiments show that features extracted by the proposed scheme have advantages over global fractal and geometric measures in enlarging inter-differences between tumor and benign lesions using multiscaled contour irregular information. The local fractal based asymmetry measures obviously strengthen asymmetric degrees of melanoma lesions to provide multi-source irregularity descriptors and enhance the capability of discriminating skin tumors on characteristics of boundary irregularity and asymmetry.