Early diagnosis of skin cancer based on segmentation and measurement of vascularization and pigmentation in Nevoscope images

Abstract

This work presents techniques developed for automated image segmentation and classification of skin lesions as malignant or benign based on the ground truth. For each skin lesion two images are obtained, one in each of two different modalities of epiluminescence microscopy (ELM): side-transillumination which highlights the subsurface vasculature and surface pigmentation, and cross polarization, which only highlights the details of skin surface pigmentation. The automated procedure consists of three steps: i) Segmentation of images, using three segmentation methods; ii) Selection of the most accurate segmentation results based on a weighted scoring technique; and iii) classification of the lesion as malignant or benign by verifying the presence of a ring of hypervascularity around the lesion in the side transillumination images. The segmentation results were validated against manual segmentation by an expert and the malignancy results were validated against the result from pathology.