AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Towards Understanding Adversarial Robustness of Optical Flow Networks

Simon Schrodi, Tonmoy Saikia, Thomas Brox

CVPR 2022 · Adversarial robustness and security › Black-box and transfer attacks · approach: Robustness benchmarks and defenses · 15 citations (OpenAlex)

Check your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use the deeper Robust FlowNetC encoder with a larger receptive fieldFlowNetC · KITTI 2015 training, 102x102 handcrafted patch at worst locationEPE: 46.12 -> 13.08Improved
Replace attacked correlation features with unattacked features (ablation)FlowNetC · KITTI 2015 training, 102x102 patch attackEPE: 25.95 -> 12.67Improved
Replace attacked conv3<a,b> features with unattacked features (ablation)FlowNetC · KITTI 2015 training, 102x102 patch attackEPE: 25.95 -> 11.31Improved
Replace attacked conv_redir features with unattacked features (ablation)FlowNetC · KITTI 2015 training, 102x102 patch attackEPE: 25.95 -> 28.36Hurt
Replace the RAFT encoder with the original FlowNetC encoder and remove the context encoder (ablation)RAFT · KITTI 2015 training, 102x102 handcrafted patch at worst locationEPE: 8.69 -> 19.12Hurt
Increase receptive field by increasing encoder depth (ablation)FlowNetC variants · KITTI 2015 training, optimized patch attacksEPEImproved
Increase receptive field by increasing convolution dilation (ablation)FlowNetC · KITTI 2015 training, optimized patch attacksEPE: 18.81; 23.4; 22.66; 23.32; 22.3Mixed
Use the wrong input data normalization at inference (ablation)FlowNetC · KITTI 2015 training, 102x102 optimized patchEPE: 11.50 -> 14.52; 52.66 -> 94.51Hurt
Use a handcrafted high-frequency self-similar striped patch instead of an optimized patchFlowNetC · KITTI 2015 training, 102x102 patch at worst locationEPE: 11.50 -> 46.12Hurt
Use untargeted I-FGSM perturbationsRobust FlowNetC, PWC-Net, and RAFT · KITTI 2015 trainingEPEHurt
Use universal adversarial perturbations with small L-infinity normsOptical flow networks · KITTI 2015 trainingEPENo clear effect
Fine-tune with adversarial data augmentation (ablation)RAFT · KITTI 2015 training, unattacked imagesEPE: 5.47 to 5.96, 7.13, 8.96 and 10.15Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy