Laplacian Segmentation Networks Improve Epistemic Uncertainty Quantification

Kilian Zepf, Selma Wanna, Marco Miani, Juston Moore, Jes Frellsen +3 more
3/23/2023
cs.CVcs.LG

No official code found. Best implementation: kilianzepf/laplacian_segmentation (7 stars, last commit 2024-12-11).

Abstract

Image segmentation relies heavily on neural networks which are known to be overconfident, especially when making predictions on out-of-distribution (OOD) images. This is a common scenario in the medical domain due to variations in equipment, acquisition sites, or image corruptions. This work addresses the challenge of OOD detection by proposing Laplacian Segmentation Networks (LSN): methods which jointly model epistemic (model) and aleatoric (data) uncertainty for OOD detection. In doing so, we propose the first Laplace approximation of the weight posterior that scales to large neural networks with skip connections that have high-dimensional outputs. We demonstrate on three datasets that the LSN-modeled parameter distributions, in combination with suitable uncertainty measures, gives superior OOD detection.

View on arXivView PDF

Code Implementations(1)

kilianzepf/laplacian_segmentationCommunity100%

Official Repo for the Paper "Laplacian Segmentation Networks Improve Epistemic Quantification"

70PythonJul 1, 20241 years ago

Similar Papers

Out-of-Distribution Segmentation via Wasserstein-Based Evidential Uncertainty

84% match2025

A Simple and Effective Method for Uncertainty Quantification and OOD Detection

84% match2025

Out-of-Distribution Detection in LiDAR Semantic Segmentation Using Epistemic Uncertainty from Hierarchical GMMs

83% match2025

Uncertainty Estimation and Out-of-Distribution Detection for LiDAR Scene Semantic Segmentation

82% match2024

Uncertainty-Aware Out-of-Distribution Detection with Gaussian Processes

82% match2024

A Metacognitive Approach to Out-of-Distribution Detection for Segmentation

82% match2023

Cite this paper

@article{zepf2023laplacian,
  title  = {Laplacian Segmentation Networks Improve Epistemic Uncertainty Quantification},
  author = {Kilian Zepf and Selma Wanna and Marco Miani and Juston Moore and Jes Frellsen and Søren Hauberg and Frederik Warburg and Aasa Feragen},
  year   = {2023},
  eprint = {2303.13123},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2303.13123v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2303.13123.svg)](https://codeofpaper.com/paper/2303.13123)

Discussion