Deep neural networks achieve superior performance in semantic segmentation, but are limited to a predefined set of classes, which leads to failures when they encounter unknown objects in open-world scenarios. Recognizing and segmenting these out-of-distribution (OOD) objects is crucial for safety-critical applications such as automated driving. In this work, we present an evidence segmentation framework using a Wasserstein loss, which captures distributional distances while respecting the probability simplex geometry. Combined with Kullback-Leibler regularization and Dice structural consistency terms, our approach leads to improved OOD segmentation performance compared to uncertainty-based approaches.
@article{arxiv.2512.11373,
title = {Out-of-Distribution Segmentation via Wasserstein-Based Evidential Uncertainty},
author = {Arnold Brosch and Abdelrahman Eldesokey and Michael Felsberg and Kira Maag},
journal= {arXiv preprint arXiv:2512.11373},
year = {2025}
}