Research map / Adversarial robustness and security / Attacks on text and graphs
Unified Adversarial Patch for Visible-Infrared Cross-Modal Attacks in the Physical World
Xingxing Wei, Yao Huang, Yitong Sun, 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 |
|---|---|---|---|
| use optimized shape instead of basic shapes (ablation) | unified patch · LLVIP YOLOv3 | cross-modal ASR: 5.23% -> 73.33% | Improved |
| use score-aware iterative fitness instead of direct sum (ablation) | unified patch · LLVIP YOLOv3 | cross-modal ASR: 50.83% -> 73.33% | Improved |
| jointly optimize position with shape (ablation) | unified patch · LLVIP YOLOv3 | cross-modal ASR: 73.33% -> 83.33% | Improved |
| use proposed shape modeling instead of hotcold block nine-square-grid (ablation) | unified patch · LLVIP YOLOv3 | cross-modal ASR: 30.83% -> 83.33% | Improved |
| increase patch radius r from 10 (ablation) | unified patch · LLVIP YOLOv3 | ASR: 66.67% at r=10 | Improved |
| increase population from 20 to 30 then to 60 (ablation) | differential evolution · LLVIP YOLOv3 | ASR: +6.66% then +2.50% | Mixed |
| increase lambda from 2 to 3 (ablation) | fitness function · LLVIP YOLOv3 | ASR | Hurt |
| increase differential weight beta from 0.6 to 0.8 (ablation) | differential evolution · LLVIP YOLOv3 | ASR | Hurt |
| add multi-angle affine enhancement (ablation) | unified patch · multi-angle physical | ASR: average ASR increase of 30% | Improved |
| apply 5 pixel translation error to patch (ablation) | unified patch · LLVIP YOLOv3 | ASR: 73.33% -> 55.00% | Hurt |
| apply spatial smoothing defense | unified patch · LLVIP YOLOv3 | ASR: 73.33% -> 64.17% | Hurt |
| apply adversarial training defense | unified patch · LLVIP YOLOv3 | ASR: 73.33% -> 50.00% | Hurt |
Closest papers by meaning
- Unified Adversarial Patch for Cross-Modal Attacks in the Physical World (ICCV 2023 · similarity 0.98)
- Naturalistic Physical Adversarial Patch for Object Detectors (ICCV 2021 · similarity 0.89)
- Revisiting Adversarial Patch Defenses on Object Detectors: Unified Evaluation, Large-Scale Dataset, and New Insights (ICCV 2025 · similarity 0.89)
- PAD: Patch-Agnostic Defense against Adversarial Patch Attacks (CVPR 2024 · similarity 0.88)
- Revisiting Adversarial Patches for Designing Camera-Agnostic Attacks against Person Detection (NeurIPS 2024 · similarity 0.86)
- DAP: A Dynamic Adversarial Patch for Evading Person Detectors (CVPR 2024 · 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.
