Saved in:
Bibliographic Details
Main Authors: Agnihotri, Shashank, Schader, David, Sharei, Nico, Kaçar, Mehmet Ege, Keuper, Margret
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.04835
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916725943959552
author Agnihotri, Shashank
Schader, David
Sharei, Nico
Kaçar, Mehmet Ege
Keuper, Margret
author_facet Agnihotri, Shashank
Schader, David
Sharei, Nico
Kaçar, Mehmet Ege
Keuper, Margret
contents Deep learning (DL) models are widely used in real-world applications but remain vulnerable to distribution shifts, especially due to weather and lighting changes. Collecting diverse real-world data for testing the robustness of DL models is resource-intensive, making synthetic corruptions an attractive alternative for robustness testing. However, are synthetic corruptions a reliable proxy for real-world corruptions? To answer this, we conduct the largest benchmarking study on semantic segmentation models, comparing performance on real-world corruptions and synthetic corruptions datasets. Our results reveal a strong correlation in mean performance, supporting the use of synthetic corruptions for robustness evaluation. We further analyze corruption-specific correlations, providing key insights to understand when synthetic corruptions succeed in representing real-world corruptions. Open-source Code: https://github.com/shashankskagnihotri/benchmarking_robustness/tree/segmentation_david/semantic_segmentation
format Preprint
id arxiv_https___arxiv_org_abs_2505_04835
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Are Synthetic Corruptions A Reliable Proxy For Real-World Corruptions?
Agnihotri, Shashank
Schader, David
Sharei, Nico
Kaçar, Mehmet Ege
Keuper, Margret
Computer Vision and Pattern Recognition
Deep learning (DL) models are widely used in real-world applications but remain vulnerable to distribution shifts, especially due to weather and lighting changes. Collecting diverse real-world data for testing the robustness of DL models is resource-intensive, making synthetic corruptions an attractive alternative for robustness testing. However, are synthetic corruptions a reliable proxy for real-world corruptions? To answer this, we conduct the largest benchmarking study on semantic segmentation models, comparing performance on real-world corruptions and synthetic corruptions datasets. Our results reveal a strong correlation in mean performance, supporting the use of synthetic corruptions for robustness evaluation. We further analyze corruption-specific correlations, providing key insights to understand when synthetic corruptions succeed in representing real-world corruptions. Open-source Code: https://github.com/shashankskagnihotri/benchmarking_robustness/tree/segmentation_david/semantic_segmentation
title Are Synthetic Corruptions A Reliable Proxy For Real-World Corruptions?
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.04835