Geometric Analysis of Skin Lesion for Skin Cancer Using Image Processing

Abstract

This study focuses on the geometric features of skin lesions for detecting and classifying skin cancer. Geometric features of the skin lesion are extracted following the asymmetry, border, and diameter parameters of the ABCD-Rule of Dermoscopy. In particular, area, perimeter, circularity index, greatest and shortest diameter, irregularity index and equivalent diameter are the parameters loaded in the dataset for classification. Three classifications of skin lesion are considered in this study such as malignant melanoma, benign melanoma, and unknown. Classification of skin lesion images is done through k-Nearest Neighbors (k-NN) algorithm and shows an accuracy of 90.0% in the functionality testing.