Research map / Adversarial robustness and security / Attacks on text and graphs
Prompt-driven Transferable Adversarial Attack on Person Re-Identification with Attribute-aware Textual Inversion
Yuan Bian, Min Liu, Yunqi Yi, Xueping Wang, Shuai Jiang, Yaonan Wang
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 global CLIP visual feature loss to surrogate feature loss (ablation) | AP-Attack generator · Cross-model&dataset attack | aAP: 45.8 -> 27.9 | Improved |
| add integral text embedding loss instead of global visual feature (ablation) | AP-Attack generator · Cross-model&dataset attack | aAP: 27.9 -> 25.9 | Improved |
| add fine-grained attribute semantic loss instead of integral text embedding (ablation) | AP-Attack generator · Cross-model&dataset attack | aAP: 25.9 -> 21.2 | Improved |
| add fine-grained attribute semantic loss (mDR view) (ablation) | AP-Attack generator · Cross-model&dataset attack | mDR: 40.3 -> 72.4 | Improved |
| use AP-Attack instead of PDCL-Attack | AP-Attack · Cross-model&dataset attack | aAP: 38.7 -> 21.2 | Improved |
| use AP-Attack instead of PDCL-Attack | AP-Attack · Cross-model attack | aAP: 19.0 -> 6.9 | Improved |
| use AP-Attack instead of PDCL-Attack | AP-Attack · Cross-dataset attack | aAP: 6.6 -> 5.7 | Improved |
| compare AP-Attack with multi-dataset MetaAttack | AP-Attack · Cross-dataset attack | aAP: 5.7 -> 3.7 | Hurt |
| use AP-Attack instead of Mis-Ranking against defenses | AP-Attack · Defense evaluation | aAP: 50.2 -> 31.1 | Improved |
| attack unseen model types with AP-Attack | AP-Attack · PASS (Market) mAP | mAP: 92.2 -> 12.1 | Improved |
| attack CLIP-ReID victim with AP-Attack | AP-Attack · CLIP-ReID (Market) mAP | mAP: 89.6 -> 32.7 | Improved |
Closest papers by meaning
- R$^2$TUA: Reconstruction-residual Based Targeted and Untargeted Attack Against Text-Image Person Re-Identification (CVPR 2026 · similarity 0.90)
- Hierarchical Prompt Learning for Image- and Text-Based Person Re-Identification (AAAI 2026 · similarity 0.87)
- Feature-Level Adversarial Attacks and Ranking Disruption for Visible-Infrared Person Re-identification (NeurIPS 2024 · similarity 0.85)
- Unifying Multi-Modal Uncertainty Modeling and Semantic Alignment for Text-to-Image Person Re-identification (AAAI 2024 · similarity 0.84)
- Transferable Adversarial Face Attack with Text Controlled Attribute (AAAI 2025 · similarity 0.84)
- Noisy-Correspondence Learning for Text-to-Image Person Re-identification (CVPR 2024 · similarity 0.84)
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.
