||Approach to the characterization of uncertain lesions, to
detect characteristic profiles of benign and malignant lesions .
Based on suitable multi-value descriptors (scalar, interval, and histogram data) extracted by dermoscopic images, it consists in
selecting through discriminant analysis the most discriminating features and detecting through dynamic clustering the characteristic profiles.
For the automatic segmentation of skin lesions in dermoscopic images, we proposed the SDI algorithm 
and the SDI+ algorithm ,
extensively tested on the lesion segmentation dataset made available for the
ISIC 2017 challenge on Skin Lesion Analysis Towards Melanoma Detection.