mlsanityData to 14 Oct 2026

Distracting Downpour: Adversarial Weather Attacks for Motion Estimation

arXiv 2305.06716 ↗First posted 11 May 2023cs.CVcited by 15

Not on the map. Why?

Abstract (arXiv)

Current adversarial attacks on motion estimation, or optical flow, optimize small per-pixel perturbations, which are unlikely to appear in the real world. In contrast, adverse weather conditions constitute a much more realistic threat scenario. Hence, in this work, we present a novel attack on motion estimation that exploits adversarially optimized particles to mimic weather effects like snowflakes, rain streaks or fog clouds. At the core of our attack framework is a differentiable particle rendering system that integrates particles (i) consistently over multiple time steps (ii) into the 3D space (iii) with a photo-realistic appearance. Through optimization, we obtain adversarial weather that significantly impacts the motion estimation. Surprisingly, methods that previously showed good robustness towards small per-pixel perturbations are particularly vulnerable to adversarial weather. At the same time, augmenting the training with non-optimized weather increases a method's robustness towards weather effects and improves generalizability at almost no additional cost. Our code will be available at https://github.com/cv-stuttgart/DistractingDownpour.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Builds on 34 of its 57 references that are in the map. First posted on arXiv 11 May 2023.

Builds on

34 of 57 references.

References · 34

  1. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  2. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  3. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  4. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,415
  5. 2015FlowNet: Learning Optical Flow with Convolutional Networkscited by 1,059
  6. 2015A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimationcited by 881
  7. 2016FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networkscited by 784
  8. 2016Optical Flow Estimation using a Spatial Pyramid Networkcited by 424
  9. 2019Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Comingcited by 336
  10. 2022FlowFormer: A Transformer Architecture for Optical Flowcited by 275
  11. 2021Learning to Estimate Hidden Motions with Global Motion Aggregationcited by 224
  12. 2023Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereocited by 172
22 more
  1. 2022Shadows can be Dangerous: Stealthy and Effective Physical-world Adversarial Attack by Natural Phenomenoncited by 101
  2. 2020Rain rendering for evaluating and improving robustness to bad weathercited by 57
  3. 2019Physics-Based Rendering for Improving Robustness to Raincited by 49
  4. 2019Attacking Optical Flowcited by 41
  5. 2021When Human Pose Estimation Meets Robustness: Adversarial Algorithms and Benchmarkscited by 41
  6. 2021Closing the Loop: Joint Rain Generation and Removal via Disentangled Image Translationcited by 36
  7. 2021AdvHaze: Adversarial Haze Attackcited by 31
  8. 2023CosPGD: an efficient white-box adversarial attack for pixel-wise prediction taskscited by 27
  9. 2020Optical Flow in Dense Foggy Scenes using Semi-Supervised Learningcited by 20
  10. 2020Stereopagnosia: Fooling Stereo Networks with Adversarial Perturbationscited by 19
  11. 2021Weather GAN: Multi-Domain Weather Translation Using Generative Adversarial Networkscited by 18
  12. 2021Stereoscopic Universal Perturbations across Different Architectures and Datasetscited by 17
  13. 2022A Perturbation-Constrained Adversarial Attack for Evaluating the Robustness of Optical Flowcited by 16
  14. 2021Towards Understanding Adversarial Robustness of Optical Flow Networkscited by 15
  15. 2019Analysis of Deep Networks for Monocular Depth Estimation Through Adversarial Attacks with Proposal of a Defense Methodcited by 10
  16. 2020Adversarial Rain Attack and Defensive Deraining for DNN Perceptioncited by 10
  17. 2022Attacking Motion Estimation with Adversarial Snowcited by 9
  18. 2022fakeWeather: Adversarial Attacks for Deep Neural Networks Emulating Weather Conditions on the Camera Lens of Autonomous Systemscited by 8
  19. 2020A Little Fog for a Large Turncited by 5
  20. 2020Deep Snow: Synthesizing Remote Sensing Imagery with Generative Adversarial Netscited by 5
  21. 2022Scale-free and Task-agnostic Attack: Generating Photo-realistic Adversarial Patterns with Patch Quilting Generatorcited by 5
  22. 2021Consistent Semantic Attacks on Optical Flowcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 4; 2024: 5; 2025: 5; 2026: 1.

15 papers.

Sort:

By area:

  1. 2023A Survey of Robustness and Safety of 2D and 3D Deep Learning Models Against Adversarial Attackscited by 19
  2. 2025Deep Learning and Foundation Models for Weather Prediction: A Surveycited by 16
  3. 2025Controllable Weather Synthesis and Removal with Video Diffusion Modelscited by 11
  4. 2023Adversarial Examples in the Physical World: A Surveycited by 10
  5. 2023Improving Feature Stability during Upsampling -- Spectral Artifacts and the Importance of Spatial Contextcited by 8
  6. 2025RobustSpring: Benchmarking Robustness to Image Corruptions for Optical Flow, Scene Flow and Stereocited by 7
  7. 2023Detection Defenses: An Empty Promise against Adversarial Patch Attacks on Optical Flowcited by 6
  8. 2024Adverse Weather Optical Flow: Cumulative Homogeneous-Heterogeneous Adaptationcited by 4
  9. 2024RenderBender: A Survey on Adversarial Attacks Using Differentiable Renderingcited by 3
  10. 2024Towards Dependable Optical Flow: Benchmarks for Real-World Spatiotemporal Corruptionscited by 3
  11. 2024CosalPure: Learning Concept from Group Images for Robust Co-Saliency Detectioncited by 2
  12. 2024Position: Towards Resilience Against Adversarial Examplescited by 1
3 more
  1. 2025Unveiling Hidden Vulnerabilities in Digital Human Generation via Adversarial Attackscited by 1
  2. 2025LatentStealth: Unnoticeable and Efficient Adversarial Attacks on Expressive Human Pose and Shape Estimationcited by 0
  3. 2026AirflowAttack: Thermal-Airflow Adversarial Perturbations against Infrared Remote-Sensing Vision-Language Modelscited by 0