mlsanityData to 14 Oct 2026

Debiasing Skin Lesion Datasets and Models? Not So Fast

arXiv 2004.11457 ↗First posted 23 Apr 2020cs.CVcited by 43

Not on the map. Why?

Abstract (arXiv)

Data-driven models are now deployed in a plethora of real-world applications - including automated diagnosis - but models learned from data risk learning biases from that same data. When models learn spurious correlations not found in real-world situations, their deployment for critical tasks, such as medical decisions, can be catastrophic. In this work we address this issue for skin-lesion classification models, with two objectives: finding out what are the spurious correlations exploited by biased networks, and debiasing the models by removing such spurious correlations from them. We perform a systematic integrated analysis of 7 visual artifacts (which are possible sources of biases exploitable by networks), employ a state-of-the-art technique to prevent the models from learning spurious correlations, and propose datasets to test models for the presence of bias. We find out that, despite interesting results that point to promising future research, current debiasing methods are not ready to solve the bias issue for skin-lesion models.

Key facts. Cited by 43 papers in the map (data through 14 Oct 2026). First posted on arXiv 23 Apr 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 12; 2022: 7; 2023: 9; 2024: 9; 2025: 3; 2026: 3.

43 papers.

Sort:

By area:

  1. 2021Evaluating Deep Neural Networks Trained on Clinical Images in Dermatology with the Fitzpatrick 17k Datasetcited by 185
  2. 2021Salient ImageNet: How to discover spurious features in Deep Learning?cited by 114
  3. 2021Leveraging Sparse Linear Layers for Debuggable Deep Networkscited by 82
  4. 2021Fair Conformal Predictors for Applications in Medical Imagingcited by 74
  5. 2023Discover and Cure: Concept-aware Mitigation of Spurious Correlationcited by 71
  6. 2021Editing a classifier by rewriting its prediction rulescited by 65
  7. 2024A Multimodal Automated Interpretability Agentcited by 63
  8. 2022Spurious Features Everywhere -- Large-Scale Detection of Harmful Spurious Features in ImageNetcited by 41
  9. 2022Addressing Fairness Issues in Deep Learning-Based Medical Image Analysis: A Systematic Reviewcited by 28
  10. 2021GAN-Based Data Augmentation and Anonymization for Skin-Lesion Analysis: A Critical Reviewcited by 25
  11. 2021Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysiscited by 22
  12. 2021Ethics Sheets for AI Taskscited by 18
31 more
  1. 2021Skin Deep Unlearning: Artefact and Instrument Debiasing in the Context of Melanoma Classificationcited by 18
  2. 2022Core Risk Minimization using Salient ImageNetcited by 15
  3. 2024Backdoor Secrets Unveiled: Identifying Backdoor Data with Optimized Scaled Prediction Consistencycited by 13
  4. 2022Epistemic Uncertainty-Weighted Loss for Visual Bias Mitigationcited by 11
  5. 2022Artifact-Based Domain Generalization of Skin Lesion Modelscited by 11
  6. 2023Interpreting and Controlling Vision Foundation Models via Text Explanationscited by 11
  7. 2021An Evaluation of Self-Supervised Pre-Training for Skin-Lesion Analysiscited by 10
  8. 2024Controllable Prompt Tuning For Balancing Group Distributional Robustnesscited by 10
  9. 2023EPVT: Environment-aware Prompt Vision Transformer for Domain Generalization in Skin Lesion Recognitioncited by 9
  10. 2023Preemptively Pruning Clever-Hans Strategies in Deep Neural Networkscited by 9
  11. 2023Even Small Correlation and Diversity Shifts Pose Dataset-Bias Issuescited by 8
  12. 2025In the Picture: Medical Imaging Datasets, Artifacts, and their Living Reviewcited by 8
  13. 2024Copycats: the many lives of a publicly available medical imaging datasetcited by 6
  14. 2024BMFT: Achieving Fairness via Bias-based Weight Masking Fine-tuningcited by 6
  15. 2024PatchAlign:Fair and Accurate Skin Disease Image Classification by Alignment with Clinical Labelscited by 5
  16. 2023Data AUDIT: Identifying Attribute Utility- and Detectability-Induced Bias in Task Modelscited by 4
  17. 2025Subgroup Performance Analysis in Hidden Stratificationscited by 3
  18. 2021Content Preserving Image Translation with Texture Co-occurrence and Spatial Self-Similarity for Texture Debiasing and Domain Adaptationcited by 2
  19. 2023Tackling Shortcut Learning in Deep Neural Networks: An Iterative Approach with Interpretable Modelscited by 2
  20. 2024APPLE: Adversarial Privacy-aware Perturbations on Latent Embedding for Unfairness Mitigationcited by 2
  21. 2024Dataset Distribution Impacts Model Fairness: Single vs. Multi-Task Learningcited by 2
  22. 2025Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Datacited by 2
  23. 2022Equivariant Disentangled Transformation for Domain Generalization under Combination Shiftcited by 1
  24. 2023Test-Time Selection for Robust Skin Lesion Analysiscited by 1
  25. 2023Targeted Data Augmentation for bias mitigationcited by 1
  26. 2021Mixing-AdaSIN: Constructing a De-biased Dataset using Adaptive Structural Instance Normalization and Texture Mixingcited by 0
  27. 2022Intrinsic Bias Identification on Medical Image Datasetscited by 0
  28. 2024Are Explanations Helpful? A Comparative Analysis of Explainability Methods in Skin Lesion Classifierscited by 0
  29. 2026A Causal Framework for Mitigating Data Shifts in Healthcarecited by 0
  30. 2026Exploring the Impact of Skin Color on Skin Lesion Segmentationcited by 0
  31. 2026Effect of Demographic Bias on Skin Lesion Classificationcited by 0