Awesome-Medical-Image-Analysis-Fairness

A collection of papers in fairness of medical image analysis

For more details, please refer to our recent survey on fairness in medical image analysis

Notes: We may miss some relevant papers in the list. If you have any suggestions or would like to add some papers, please submit a pull request or email me at [email protected]. Your contribution is much appreciated!

---

Fairness Papers in Medical Image Analysis

Fairness Evaluation

Unfairness Existence

1. Fairness-related performance and explainability effects in deep learning models for brain image analysis.JMI, 2022. (paper)
2. How fair is your graph? exploring fairness concerns in neuroimaging studies. ML4HC, 2022. (paper)
3. Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations. Nature medicine,2021. (paper)
4. Fairness in cardiac magnetic resonance imaging: assessing sex and racial bias in deep learning-based segmentation. Frontiers in Cardiovascular Medicine, 2022. (paper)

Fairness & Data Distribution

1. Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis. PNAS, 2020. (paper)
2. Algorithmic encoding of protected characteristics and its implications on disparities across subgroups. ArXiv, 2021. (paper)
3. Feature robustness and sex differences in medical imaging: a case study in mri-based alzheimer’s disease detection. ArXiv, 2022. (paper)
4. Fairness of classifiers across skin tones in dermatology. MICCAI, 2020. (paper)

Fairness Benchmarking

1. Improving the Fairness of Chest X-ray Classifiers. PCHIL, 2022. (paper)
2. Medfair: Benchmarking fairness for medical imaging. ICLR, 2023. (paper, code)
3. Detecting Melanoma Fairly: Skin Tone Detection and Debiasing for Skin Lesion Classification. MICCAI-DART, 2021. (paper, code)

Fairness & Uncertainty

1. Evaluating subgroup disparity using epistemic uncertainty in mammography. ArXiv, 2021. (paper)

Fairness & Model Selection

1. Model selection’s disparate impact in real-world deep learning applications. ArXiv, 2021. (paper)

Fairness Dataset

1. Harvard Glaucoma Fairness: A Retinal Nerve Disease Dataset for Fairness Learning and Fair Identity Normalization. ArXiv, 2023. (paper, code, dataset)

---

Unfairness Mitigation via Pre-processing Methods

Data Resampling

1. Fairness in cardiac MR image analysis: An investigation of bias due to data imbalance in deep learning based segmentation. MICCAI, 2021. (paper)
2. Detecting and preventing shortcut learning for fair medical ai using shortcut testing (short). ArXiv, 2022. (paper)

Data Synthesis

1. AI fairness via domain adaptation. ArXiv, 2021. (paper)
2. CIRCLe: Color Invariant Representation Learning for Unbiased Classification of Skin Lesions. ECCV-ISIC Workshop, 2022. (paper, code)

Data Aggregation

1. CheXclusion: Fairness gaps in deep chest X-ray classifiers. Proceedings of the Pacific Symposium, 2021. (paper)
2. Radfusion: Benchmarking performance and fairness for multimodal pulmonary embolism detection from ct and ehr. ArXiv, 2021. (paper)

Data Embellishment

1. Improving fairness in image classification via sketching. NeurIPS Workshop, 2022. (paper, code)
2. Fairness of classifiers across skin tones in dermatology. MICCAI, 2020. (paper)

---

Unfairness Mitigation via In-processing Methods

Adversarial Learning

Training confounder-free deep learning models for medical applications. Nat. Com, 2020. (paper)

1. Representation learning with statistical independence to mitigate bias. WACV, 2021. (paper)
2. Risk of training diagnostic algorithms on data with demographic bias. IAELMIC, 2020. Springer (paper)
3. Estimating and improving fairness with adversarial learning. ArXiv, 2021. (paper)
4. Technical challenges for training fair neural networks. ArXiv, 2021. (paper)

Disentanglement Learning

1. On the fairness of privacy-preserving representations in medical applications. MICCAI-DART, 2020. (paper)
2. On fairness of medical image classification with multiple sensitive attributes via learning orthogonal representations. IPMI, 2023. (paper)
3. CIRCLe: Color Invariant Representation Learning for Unbiased Classification of Skin Lesions. ECCV-ISIC Workshop, 2022. (paper, code)

Contrastive Learning

1. FairDisCo: Fairer AI in Dermatology via Disentanglement Contrastive Learning. ECCVW, 2022. (paper, code)

Universal Learning

1. FairAdaBN: Mitigating unfairness with adaptive batch normalization and its application to dermatological disease classification. ArXiv, 2023. (paper)

Federated Learning

1. On the fairness of swarm learning in skin lesion classification. MICCAI Workshop, 2021. (paper)

Unfairness Mitigation via Post-processing Methods

Network Pruning

1. Fairprune: Achieving fairness through pruning for dermatological disease diagnosis. MICCAI, 2022. (paper)
2. Debiasing Deep Chest X-Ray Classifiers using Intra- and Post-processing Methods. ML4HC, 2022. (paper, code)

---

Medical Datasets with Sensitive Attributes

MISC