ABSTRACT
Dermoscopy is an inconspicuous diagnostic system for the observation of pigmented skin injuries and one of the greatest momentous imaging procedures for melanoma analysis. At the same time, its indicative exactitude overall relies upon the familiarity of the dermatologists, the optical illumination and evaluation of this type of picture is monotonous, and a few computer-based hold up verdict systems of computerized dermoscopic pictures have been obtainable. This chapter puts forward the segmentation of dermoscopy skin lesion images. For the segmentation process, unsupervised clustering methods, namely K-Means and the enhanced clustering method RKM, have been carried out. The results produced by RKM are analyzed in different scenarios. RKM takes nominal time to segment the skin lesion properly. Despite segregating properly, RKM produces(leads) over segmentation results for most of the images. Moreover, this method produces a result with certain loss of pixels because it is a trimmed process.
