arXiv 2511.00477
Investigating Label Bias and Representational Sources of Age-Related Disparities in Medical Segmentation
By Aditya Parikh, Sneha Das, et al.
Published 2025-11-01
Citation lineage
Review the prior work and downstream research connected to this paper.
Algorithmic bias in medical imaging can perpetuate health disparities, yet its causes remain poorly understood in segmentation tasks. While fairness has been extensively studied in classification, segmentation remains underexplored despite its clinical importance. In breast cancer segmentation, models exhibit significant performance disparities against younger patients, commonly attributed to physiological differenc…