mlsanityData to 14 Oct 2026

Are Synthetic Corruptions A Reliable Proxy For Real-World Corruptions?

arXiv 2505.04835 ↗First posted 7 May 2025cs.CVcited by 5

Not on the map. Why?

Abstract (arXiv)

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

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 45 of its 62 references that are in the map. First posted on arXiv 7 May 2025.

Builds on

45 of 62 references.

References · 45

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  4. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  5. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  6. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
  7. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,223
  8. 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
  9. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  10. 2016Pyramid Scene Parsing Networkcited by 2,479
  11. 2021BEiT: BERT Pre-Training of Image Transformerscited by 2,375
  12. 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
33 more
  1. 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,332
  2. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,161
  3. 2018Unified Perceptual Parsing for Scene Understandingcited by 1,271
  4. 2016Semantic Understanding of Scenes through the ADE20K Datasetcited by 1,069
  5. 2019AugMix: A Simple Data Processing Method to Improve Robustness and Uncertaintycited by 901
  6. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 661
  7. 2022InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutionscited by 482
  8. 2019Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Comingcited by 336
  9. 2021SegmentMeIfYouCan: A Benchmark for Anomaly Segmentationcited by 150
  10. 2014Hypercolumns for Object Segmentation and Fine-grained Localizationcited by 128
  11. 2017On the Robustness of Semantic Segmentation Models to Adversarial Attackscited by 126
  12. 20223D Common Corruptions and Data Augmentationcited by 119
  13. 2019Benchmarking the Robustness of Semantic Segmentation Modelscited by 111
  14. 2022SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustnesscited by 74
  15. 2024Improving 2D Feature Representations by 3D-Aware Fine-Tuningcited by 71
  16. 2023Using DUCK-Net for Polyp Image Segmentationcited by 31
  17. 2022FrequencyLowCut Pooling -- Plug & Play against Catastrophic Overfittingcited by 28
  18. 2023CosPGD: an efficient white-box adversarial attack for pixel-wise prediction taskscited by 27
  19. 2022Robust Models are less Over-Confidentcited by 25
  20. 2023Robustifying Token Attention for Vision Transformerscited by 19
  21. 2024Adversarial Supervision Makes Layout-to-Image Diffusion Models Thrivecited by 18
  22. 2022Intra-Source Style Augmentation for Improved Domain Generalizationcited by 17
  23. 2022A Perturbation-Constrained Adversarial Attack for Evaluating the Robustness of Optical Flowcited by 16
  24. 2023Classification robustness to common optical aberrationscited by 13
  25. 2023On the unreasonable vulnerability of transformers for image restoration -- and an easy fixcited by 11
  26. 2022Attacking Motion Estimation with Adversarial Snowcited by 9
  27. 2023Intra- & Extra-Source Exemplar-Based Style Synthesis for Improved Domain Generalizationcited by 8
  28. 2023Differentiable Sensor Layouts for End-to-End Learning of Task-Specific Camera Parameterscited by 6
  29. 2023Detection Defenses: An Empty Promise against Adversarial Patch Attacks on Optical Flowcited by 6
  30. 2024Beware of Aliases -- Signal Preservation is Crucial for Robust Image Restorationcited by 6
  31. 2024FAIR-TAT: Improving Model Fairness Using Targeted Adversarial Trainingcited by 5
  32. 2024How Do Training Methods Influence the Utilization of Vision Models?cited by 3
  33. 2024Towards Class-wise Robustness Analysiscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 3.

5 papers.

  1. 2025RobustSpring: Benchmarking Robustness to Image Corruptions for Optical Flow, Scene Flow and Stereocited by 7
  2. 2026Images as Tables: In-Context Learning with TabPFN for Low-Data Detection of AI-Generated Imagescited by 1
  3. 2025DispBench: Benchmarking Disparity Estimation to Synthetic Corruptionscited by 0
  4. 2026Few-Shot Synthetic Data Generation with Diffusion Models for Downstream Vision Taskscited by 0
  5. 2026Failure or Drift? Evaluating Monocular SLAM under Synthetic and Real-World Corruptionscited by 0