Machine learning of melanocytic skin lesion images

Abstract

We use some machine learning methods to build classifiers of pigmented skin lesion images. We take advantage of natural induction methods based on the attributional calculus (AQ21) and MLP and SVM supervised methods to discover patterns in the melanocytic skin lesion images. This methodology can be treated as a non-invasive approach to early diagnosis of melanoma. Our feature set is composed of wavelet-based multi-resolution filters of the dermatoscopic images. Our classifiers show good efficiency and may potentially be important diagnostic aids.