Reproduce “A Probabilistic Perspective on Wasserstein-Based Evidential Uncertainty for Out-of-Distribution Segmentation” 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
Arnold Brosch, Abdelrahman Eldesokey, Michael Felsberg, Kira Maag
The paper compares Wasserstein loss orders for evidential segmentation in DeepLabV3+ and SegFormer. Across four datasets, the best-performing order differs: W2 generally performs best for DeepLabV3+, while W3 gives SegFormer the highest AuPRC on all four. In a DeepLabV3+ comparison, W2 reaches 57.9 AuPRC on RoadAnomaly21, versus PGN’s 42.8; gradient- and adversarial-based methods perform better on RoadObstacle21 and Fishyscapes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.