HAM10000 is a dataset of 10,000 training images for detecting pigmented skin lesions. The authors collected dermoscopic images from different populations, acquired and stored in different ways.

This benchmark dataset can be used to compare machine learning diagnoses with human experts. Cases include a representative set of all important diagnostic categories in the field of pigmented lesions. More than 50% of lesions have been confirmed by pathology, while the ground truth for the remaining cases is follow-up, expert consensus, or in vivo confocal microscopy confirmation.