Research map / Adversarial robustness and security / Attacks on text and graphs
Adversarial camera stickers: A physical camera-based attack on deep learning systems
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 a physically realizable 6-dot camera sticker targeting computer mouse | ResNet-50 on ImageNet keyboard images · ImageNet test set, computer keyboard → computer mouse | target predictions: 0% -> 36% | Helped |
| Add a physically realizable 6-dot camera sticker targeting guitar pick | ResNet-50 on ImageNet street sign images · ImageNet test set, street sign → guitar pick | target predictions: 0% -> 34% | Helped |
| Add a printed adversarial camera sticker | ResNet-50 viewing a keyboard · Real camera video, keyboard → mouse | target predictions: 548 | Helped |
| Add a printed adversarial camera sticker | ResNet-50 viewing a street sign · Real camera video, street sign → guitar pick | target predictions: 605 | Helped |
| Increase the sticker from 1 dot to 10 dots (ablation) | Camera sticker targeting computer mouse · ImageNet test set, keyboard → mouse | targeted fooling rate: 27.9% -> 49.6% | Helped |
| Increase the sticker from 1 dot to 3 dots (ablation) | Camera sticker targeting computer mouse · ImageNet test set, keyboard → mouse | targeted fooling rate: 27.9% -> 32.0% | Helped |
| Increase the sticker from 3 dots to 5 dots (ablation) | Camera sticker targeting computer mouse · ImageNet test set, keyboard → mouse | targeted fooling rate: 32.0% -> 34.2% | Helped |
| Increase the sticker from 5 dots to 7 dots (ablation) | Camera sticker targeting computer mouse · ImageNet test set, keyboard → mouse | targeted fooling rate: 34.2% -> 38.2% | Helped |
| Increase the sticker from 7 dots to 10 dots (ablation) | Camera sticker targeting computer mouse · ImageNet test set, keyboard → mouse | targeted fooling rate: 38.2% -> 49.6% | Helped |
Closest papers by meaning
- Invisible Perturbations: Physical Adversarial Examples Exploiting the Rolling Shutter Effect (CVPR 2021 · similarity 0.90)
- Robust Physical-World Attacks on Deep Learning Visual Classification (CVPR 2018 · similarity 0.89)
- Defending Against Physically Realizable Attacks on Image Classification (ICLR 2020 · similarity 0.89)
- Connecting the Digital and Physical World: Improving the Robustness of Adversarial Attacks (AAAI 2019 · similarity 0.89)
- Shadows Can Be Dangerous: Stealthy and Effective Physical-World Adversarial Attack by Natural Phenomenon (CVPR 2022 · similarity 0.88)
- Breaking Barriers in Physical-World Adversarial Examples: Improving Robustness and Transferability via Robust Feature (AAAI 2025 · similarity 0.86)
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.