Detecting Melanoma Fairly: Skin Tone Detection and Debiasing for Skin Lesion Classification
This work addresses fairness issues in skin lesion classification for medical AI applications, though it is incremental as it builds on existing bias mitigation methods.
The authors tackled performance disparities in melanoma detection across skin tones by developing an automated skin tone detection algorithm and applying bias unlearning techniques, resulting in improved generalization and reduced performance gaps between lighter and darker skin tones.
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.