Research map / Adversarial robustness and security / Attacks on text and graphs
Perceptual-Sensitive GAN for Generating Adversarial Patches
Aishan Liu, Xianglong Liu, Jiaxin Fan, Yuqing Ma, Anlan Zhang, Huiyuan Xie, Dacheng Tao
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 |
|---|---|---|---|
| PS-GAN generated adversarial patch vs seed patch | VY classifier · GTSRB | classification accuracy (lower=stronger attack): 85.6% -> 12.5% | Improved |
| strong patch distortion constraint (PS-GAN vs PS-GAN*) (ablation) | PS-GAN · GTSRB | classification accuracy (lower=stronger attack): 1.5% -> 12.5% | Hurt |
| PS-GAN* vs LaVAN / GoogleAp | adversarial patch attack · GTSRB | classification accuracy (lower=stronger attack): 3.1% -> 1.5% | Improved |
| feed-forward generator instead of per-patch optimization | adversarial patch generation · GTSRB | time per patch: 61.2s -> 0.106s | Improved |
| apply PS-GAN printed patches to real traffic sign | traffic sign classifier · physical world | classification accuracy: 86.7% -> 17.2% | Improved |
Closest papers by meaning
- CamoPatch: An Evolutionary Strategy for Generating Camoflauged Adversarial Patches (NeurIPS 2023 · similarity 0.88)
- Naturalistic Physical Adversarial Patch for Object Detectors (ICCV 2021 · similarity 0.87)
- PAD: Patch-Agnostic Defense against Adversarial Patch Attacks (CVPR 2024 · similarity 0.87)
- Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation (AAAI 2024 · similarity 0.86)
- A Unified, Resilient, and Explainable Adversarial Patch Detector (CVPR 2025 · similarity 0.86)
- AIM: Additional Image Guided Generation of Transferable Adversarial Attacks (AAAI 2025 · 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.
