Program

EPIMI will be held as a joint workshop with FAIMI and BRIDGE involving keynote presentations from thought leaders both inside and outside of the Medical Image Computing and Computer-Assisted Interventions community as well as oral and poster presentations from selected author submissions. 

The combined FAIMI-BRIDGE-EPIMI workshop will take place at the Palais de la Musique et des Congrès in Strasbourg France on September 27, 2026.

The combined programme is as follows:

10:30 -10:35 Opening Remarks

10:35 -11:15 Keynote 1: Aasa Feragen, Technical University of Denmark - "Algorithmic fairness is harder than you think"

11:15 - 12:05 Oral Session 1: Understanding nuanced disparities

• Intersectional Disentangling of Temporal and Acquisition Bias in Fetal Ultrasound

• Understanding Sources of Demographic Predictability in Brain MRI via Disentangling Anatomy and Contrast

• On Evaluating Subgroup Discovery in Medical Image Classification

• From Fairness Findings to Fairness Claims: An Evidence Classification Scheme for Clinical AI

12:05 - 12:30 Poster Session Highlight Talks

12:30 -13:30 Lunch

13:30 -14:10 Keynote 2: Florence Xini Doo, University of Maryland - Title TBD

14:10 - 15:00 Oral Session 2: Issues in bridging research and deployment

• Equal Accuracy, Unequal Agreement: Auditing Clinician–AI Agreement Fairness in Lung Nodule CT

• False Confidence: Automated Labels Confound Fairness Audits in Cervical Spine Segmentation

• When Oracle Conditioning Misleads Deployment: Conditioning-Availability Bias in Echocardiographic Segmentation

• Rethinking Explainability for Clinical Trust: A Task-Specific Communicative Layer

15:00 -16:00 Poster Session / Break

16:00 -16:50 Oral Session 3: Issues with foundation models and federated learning

• Gradient Erasure and Contributory Injustice in Federated Medical Imaging AI

• Look What the Probes Dragged In! Real-World Chest X-ray Shortcuts in MedCLIP

• Subgroup performance analysis of adaptation strategies for chest X-ray foundation models

• Foundational values for foundation models

16:50 -17:50 Round-table Discussion with experts from industry, academia, and regulatory bodies

17:50 -18:00 Awards and Closing

The poster session will include the following papers:

Poster Theme 1: Demographic Fairness & Intersectional Disparities

1. Contrast-Induced Class Overlap as a Fairness Bottleneck in Dermatological AI: Evidence from HAM10000

2. When Fairness Transfer Backfires: Dark-Skin Inversion in Dermatology Foundation Models and a Minimal In-Distribution

3. Subgroup performance analysis of adaptation strategies for chest X-ray foundation models

4. Beyond Predictive Fairness: Quantifying Attribution Consistency Across Demographic Groups in Diabetic Retinopathy Screening

5. Understanding Sources of Demographic Predictability in Brain MRI via Disentangling Anatomy and Contrast

6. An Intersectional Fairness-Aware Framework for Alzheimer’s Disease Detection Using Multimodal Data

7. Intersectional Disentangling of Temporal and Acquisition Bias in Fetal Ultrasound

8. Investigating Sex and Ethnicity Bias in Deep Learning-based Echocardiography Image Segmentation

Poster Theme 2: Model Robustness & Deployment Risks

9. When Oracle Conditioning Misleads Deployment: Conditioning-Availability Bias in Echocardiographic Segmentation

10. Marginal Coverage Hides a Reduced-Ejection-Fraction Reliability Gap: Severity-Conditional Conformal Intervals on Public Echocardiography

11. Loss-Conditioned Utility-Fairness Boundary Modeling for Medical Imaging

12. Multiple Shortcut Pathways in Mammography-Based Breast Cancer Risk Prediction

13. Look What the Probes Dragged In! Real-World Chest X-ray Shortcuts in MedCLIP

14. Counterfactual Stress Testing for Image Classification Models

15. Hallucinations and constraints : Regulating surgical workflow recognition beyond accuracy

Poster Theme 3: Human-AI Interaction & Values

16. Foundational values for foundation models

17. Gradient Erasure and Contributory Injustice in Federated Medical Imaging AI

18. What is AI to Us: Exploring Patient Values in Integrating AI for Epilepsy Management

19. Futures Before Failures: Design Fictions for Anticipatory Fairness in Surgical AI

Poster Theme 4: Auditing Frameworks

20. Rethinking Explainability for Clinical Trust: A Task-Specific Communicative Layer

21. On Evaluating Subgroup Discovery in Medical Image Classification

22. From Fairness Findings to Fairness Claims: An Evidence Classification Scheme for Clinical AI

23. Equal Accuracy, Unequal Agreement: Auditing Clinician–AI Agreement Fairness in Lung Nodule CT

24. False Confidence: Automated Labels Confound Fairness Audits in Cervical Spine Segmentation