Research map / Adversarial robustness and security / Black-box and transfer attacks
Adversarial Sticker: A Stealthy Attack Method in the Physical World
Xingxing Wei, Ying Guo, Jie Yu
Read the paperCheck 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 changed | Where | Measured | Effect |
|---|---|---|---|
| add adaptive adjustment strategy to DE (ablation) | DE · LFW dodging, FaceNet | fooling rate: 29.03% -> 41.96% | Improved |
| add region-based offspring generation to DE (ablation) | DE · LFW dodging, FaceNet | fooling rate: 29.03% -> 54.81% | Improved |
| combine adaptive adjustment and region-based generation (RHDE) (ablation) | DE · LFW dodging, FaceNet | fooling rate: 29.03% -> 76.26% | Improved |
| combine both strategies (RHDE) (ablation) | DE · LFW dodging, FaceNet | queries: 1107 -> 478 | Improved |
| combine both strategies (RHDE) (ablation) | DE · LFW dodging, SphereFace | fooling rate: 21.88% -> 64.08% | Improved |
| combine both strategies (RHDE) (ablation) | DE · LFW dodging, CosFace | fooling rate: 31.83% -> 69.82% | Improved |
| use RHDE instead of exhaustive search | RHDE · LFW dodging, FaceNet | fooling rate: 79.52% -> 76.26% | No clear effect |
| model 3D deformation when solving sticker parameters (ablation) | adversarial sticker · physical frames, FaceNet subject A | attack success rate: 59.78% -> 98.46% | Improved |
| model 3D deformation when solving sticker parameters (ablation) | adversarial sticker · physical frames, SphereFace subject B | attack success rate: 33.33% -> 83.45% | Improved |
| adversarially train with same sticker (sticker 2) | target FaceNet · LFW dodging | fooling rate: 76.26% -> 60.35% | Hurt |
| adversarially train with a different sticker (sticker 1) | target FaceNet · LFW dodging | fooling rate: 76.26% -> 74.18% | No clear effect |
| use RHDE sticker instead of adv-hat | ours vs adv-hat · LFW black-box, FaceNet | fooling rate: 28.85% -> 76.26% | Improved |
Closest papers by meaning
- Adversarial camera stickers: A physical camera-based attack on deep learning systems (ICML 2019 · similarity 0.90)
- Shadows Can Be Dangerous: Stealthy and Effective Physical-World Adversarial Attack by Natural Phenomenon (CVPR 2022 · similarity 0.89)
- Infrared Adversarial Car Stickers (CVPR 2024 · similarity 0.87)
- Robust Physical-World Attacks on Deep Learning Visual Classification (CVPR 2018 · similarity 0.86)
- Simultaneously Optimizing Perturbations and Positions for Black-Box Adversarial Patch Attacks (TPAMI 2022 · similarity 0.86)
- SlowFormer: Adversarial Attack on Compute and Energy Consumption of Efficient Vision Transformers (CVPR 2024 · similarity 0.85)
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.
