Reproduce “Delving Aleatoric Uncertainty in Medical Image Segmentation via Vision Foundation Models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ruiyang Li, Fang Liu, Licheng Jiao, Xinglin Xie, Jiayao Hao, Shuo Li, Xu Liu, Jingyi Yang and 3 more
The authors use a pretrained medical vision model to estimate aleatoric uncertainty in segmentation and guide filtering or training. AUV derives from class-specific feature diversity; DUO combines noise denoising and semantic-scale loss adjustment. Across five datasets and three backbones, DUO improves every Dice and mIoU result. For nnU-Net on LiTS, Dice rises from 82.96% to 84.13%, a test-set result, not clinical validation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.