BCN20000: Dermoscopic Lesions in the Wild

TL;DR

This paper introduces the BCN20000 dataset of nearly 20,000 dermoscopic skin lesion images collected over six years, aimed at advancing automatic classification of diverse and challenging skin cancer images.

Contribution

The paper presents a large, diverse dataset for dermoscopic image analysis, facilitating research on unconstrained skin cancer classification tasks.

Findings

Dataset includes images from various difficult lesion types.

Supports development of robust skin cancer classification algorithms.

Aims to improve diagnostic accuracy in challenging cases.

Abstract

This article summarizes the BCN20000 dataset, composed of 19424 dermoscopic images of skin lesions captured from 2010 to 2016 in the facilities of the Hospital Cl\'inic in Barcelona. With this dataset, we aim to study the problem of unconstrained classification of dermoscopic images of skin cancer, including lesions found in hard-to-diagnose locations (nails and mucosa), large lesions which do not fit in the aperture of the dermoscopy device, and hypo-pigmented lesions. The BCN20000 will be provided to the participants of the ISIC Challenge 2019, where they will be asked to train algorithms to classify dermoscopic images of skin cancer automatically.