mlsanityData to 7 Oct 2026

Disparities in Dermatology AI: Assessments Using Diverse Clinical Images

arXiv 2111.08006 ↗First posted 15 Nov 2021eess.IV cs.CV cs.LGcited by 132

Not on the map. Why?

Abstract (arXiv)

More than 3 billion people lack access to care for skin disease. AI diagnostic tools may aid in early skin cancer detection; however most models have not been assessed on images of diverse skin tones or uncommon diseases. To address this, we curated the Diverse Dermatology Images (DDI) dataset - the first publicly available, pathologically confirmed images featuring diverse skin tones. We show that state-of-the-art dermatology AI models perform substantially worse on DDI, with ROC-AUC dropping 29-40 percent compared to the models' original results. We find that dark skin tones and uncommon diseases, which are well represented in the DDI dataset, lead to performance drop-offs. Additionally, we show that state-of-the-art robust training methods cannot correct for these biases without diverse training data. Our findings identify important weaknesses and biases in dermatology AI that need to be addressed to ensure reliable application to diverse patients and across all disease.

Key facts. Cited by 132 papers in the map (data through 7 Oct 2026). Builds on 10 of its 59 references that are in the map. First posted on arXiv 15 Nov 2021.

Builds on

10 of 59 references.

References · 10

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  3. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  4. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,232
  5. 2016Deep CORAL: Correlation Alignment for Deep Domain Adaptationcited by 1,194
  6. 2019Skin Lesion Analysis Toward Melanoma Detection 2018: A Challenge Hosted by the International Skin Imaging Collaboration (ISIC)cited by 657
  7. 2020How Does Mixup Help With Robustness and Generalization?cited by 193
  8. 2021Evaluating Deep Neural Networks Trained on Clinical Images in Dermatology with the Fitzpatrick 17k Datasetcited by 184
  9. 2019A deep learning system for differential diagnosis of skin diseasescited by 118
  10. 2020TrueImage: A Machine Learning Algorithm to Improve the Quality of Telehealth Photoscited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 7; 2023: 23; 2024: 31; 2025: 34; 2026: 37.

132 papers.

Sort:

By area:

  1. 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Modelscited by 63
  2. 2023SkinCon: A skin disease dataset densely annotated by domain experts for fine-grained model debugging and analysiscited by 44
  3. 2024The Clever Hans Mirage: A Comprehensive Survey on Spurious Correlations in Machine Learningcited by 39
  4. 2022Towards Transparency in Dermatology Image Datasets with Skin Tone Annotations by Experts, Crowds, and an Algorithmcited by 35
  5. 2023Towards Trustable Skin Cancer Diagnosis via Rewriting Model's Decisioncited by 27
  6. 2022Improving dermatology classifiers across populations using images generated by large diffusion modelscited by 25
  7. 2023Beyond Confidence: Reliable Models Should Also Consider Atypicalitycited by 25
  8. 2022A Survey on Deep Learning for Skin Lesion Segmentationcited by 22
  9. 2023WBCAtt: A White Blood Cell Dataset Annotated with Detailed Morphological Attributescited by 22
  10. 2023Augmenting medical image classifiers with synthetic data from latent diffusion modelscited by 21
  11. 2024Self-eXplainable AI for Medical Image Analysis: A Survey and New Outlookscited by 18
  12. 2022CIRCLe: Color Invariant Representation Learning for Unbiased Classification of Skin Lesionscited by 17
38 more
  1. 2023Beyond Skin Tone: A Multidimensional Measure of Apparent Skin Colorcited by 17
  2. 2024Integrating Clinical Knowledge into Concept Bottleneck Modelscited by 17
  3. 2024The Data Addition Dilemmacited by 15
  4. 2024Slicing Through Bias: Explaining Performance Gaps in Medical Image Analysis using Slice Discovery Methodscited by 14
  5. 2025MM-Skin: Enhancing Dermatology Vision-Language Model with an Image-Text Dataset Derived from Textbookscited by 14
  6. 2023Are Sex-based Physiological Differences the Cause of Gender Bias for Chest X-ray Diagnosis?cited by 13
  7. 2023Skin Deep: Investigating Subjectivity in Skin Tone Annotations for Computer Vision Benchmark Datasetscited by 11
  8. 2023Dividing and Conquering a BlackBox to a Mixture of Interpretable Models: Route, Interpret, Repeatcited by 10
  9. 2023Ecosystem-level Analysis of Deployed Machine Learning Reveals Homogeneous Outcomescited by 10
  10. 2024What's documented in AI? Systematic Analysis of 32K AI Model Cardscited by 10
  11. 2024DERM12345: A Large, Multisource Dermatoscopic Skin Lesion Dataset with 38 Subclassescited by 10
  12. 2024Benchmark Data Repositories for Better Benchmarkingcited by 10
  13. 2023Achieving Reliable and Fair Skin Lesion Diagnosis via Unsupervised Domain Adaptationcited by 9
  14. 2024From Majority to Minority: A Diffusion-based Augmentation for Underrepresented Groups in Skin Lesion Analysiscited by 9
  15. 2025Doctor Approved: Generating Medically Accurate Skin Disease Images through AI-Expert Feedbackcited by 9
  16. 2023Intrinsic Self-Supervision for Data Quality Auditscited by 8
  17. 2025Are generative models fair? A study of racial bias in dermatological image generationcited by 7
  18. 2023China and the U.S. produce more impactful AI research when collaborating togethercited by 6
  19. 2023Towards Reliable Dermatology Evaluation Benchmarkscited by 6
  20. 2024A Generalist Model for Diverse Text-Guided Medical Image Synthesiscited by 6
  21. 2024BMFT: Achieving Fairness via Bias-based Weight Masking Fine-tuningcited by 6
  22. 2024Towards Multi-dimensional Explanation Alignment for Medical Classificationcited by 6
  23. 2022Identifying the Context Shift between Test Benchmarks and Production Datacited by 5
  24. 2024The Case for Globalizing Fairness: A Mixed Methods Study on Colonialism, AI, and Health in Africacited by 5
  25. 2024SkinCaRe: A Multimodal Dermatology Dataset Annotated with Medical Caption and Chain-of-Thought Reasoningcited by 5
  26. 2024Intersectional Unfairness Discoverycited by 5
  27. 2024What is Fair? Defining Fairness in Machine Learning for Healthcited by 5
  28. 2024PASSION for Dermatology: Bridging the Diversity Gap with Pigmented Skin Images from Sub-Saharan Africacited by 5
  29. 2022Fairness via In-Processing in the Over-parameterized Regime: A Cautionary Talecited by 4
  30. 2023FEDD -- Fair, Efficient, and Diverse Diffusion-based Lesion Segmentation and Malignancy Classificationcited by 4
  31. 2023(Predictable) Performance Bias in Unsupervised Anomaly Detectioncited by 4
  32. 2023Towards objective and systematic evaluation of bias in artificial intelligence for medical imagingcited by 4
  33. 2024SUDO: a framework for evaluating clinical artificial intelligence systems without ground-truth annotationscited by 4
  34. 2024Enhancing Skin Disease Diagnosis: Interpretable Visual Concept Discovery with SAMcited by 4
  35. 2025DermDiff: Generative Diffusion Model for Mitigating Racial Biases in Dermatology Diagnosiscited by 4
  36. 2025Concept-Based Unsupervised Domain Adaptationcited by 4
  37. 2025Learning Concept-Driven Logical Rules for Interpretable and Generalizable Medical Image Classificationcited by 4
  38. 2026Unsupervised Causal Prototypical Networks for De-biased Interpretable Dermoscopy Diagnosiscited by 4

50 most cited of 132 All 132 (JSON)