Delving Aleatoric Uncertainty in Medical Image Segmentation via Vision Foundation Models

ArXi:2604.10963v1 Announce Type: new Medical image segmentation s clinical workflows by precisely delineating anatomical structures and lesions. However, medical image datasets medical image datasets suffer from acquisition noise and annotation ambiguity, causing pervasive data uncertainty that substantially undermines model robustness. Existing research focuses primarily on model architectural improvements and predictive reliability estimation, while systematic exploration of the intrinsic data uncertainty remains insufficient.