Removal of artifacts from dermatoscopic images

Abstract

Dermoscopy is the primary tool used for pigmented skin lesion diagnosis. Despite the use of this relative new clinical method dermoscopy based, diagnose is still subjective and the diagnosis detection accuracy is about 75-80%. In this paper we present several enhancement pre-processing techniques applied on dermatoscopic images, such as black frame removal and hair removal in an automatically manner. We have tested our algorithms on 45 dermoscopic images and compared the automated enhancement methods results with other existing methods.