Detecting Melanoma Fairly: Skin Tone Detection and Debiasing for Skin Lesion Classification

TL;DR

This paper introduces a skin tone detection algorithm and applies bias unlearning techniques to mitigate skin tone bias in melanoma classification, improving fairness and generalization across skin tones.

Contribution

It presents a novel automated skin tone labeling method and demonstrates its effectiveness in reducing bias in melanoma detection models.

Findings

Skin tone detection outperforms existing methods.

Bias unlearning reduces performance disparity.

Improved generalization across skin tones.

Abstract

Convolutional Neural Networks have demonstrated human-level performance in the classification of melanoma and other skin lesions, but evident performance disparities between differing skin tones should be addressed before widespread deployment. In this work, we propose an efficient yet effective algorithm for automatically labelling the skin tone of lesion images, and use this to annotate the benchmark ISIC dataset. We subsequently use these automated labels as the target for two leading bias unlearning techniques towards mitigating skin tone bias. Our experimental results provide evidence that our skin tone detection algorithm outperforms existing solutions and that unlearning skin tone may improve generalisation and can reduce the performance disparity between melanoma detection in lighter and darker skin tones.