Research map / Image generation / Autoregressive and latent generation
Attack-Resistant Watermarking for AIGC Image Forensics via Diffusion-based Semantic Deflection
Qingyu Liu, Yitao Zhang, Zhongjie Ba, Chao Shuai, Peng Cheng, Tianhang Zheng, Zhibo Wang
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 PAI's multi-dimensional latent-space verification instead of scalar-threshold verification (Gaussian Shading) | PAI vs Gaussian Shading · Spoofing attacks (average) | ownership ACC (%): 15.77 -> 96.33 | Improved |
| Use PAI instead of Stable Signature under removal attacks | PAI vs Stable Signature · Removal attacks (average) | ownership ACC (%): 43.95 -> 99.00 | Improved |
| Use PAI instead of EditGuard on full-image semantic editing (Stable Inpainting) | PAI vs EditGuard · Tamper localization (Stable Inpainting) | F1 (%): 34.00 -> 80.00 | Improved |
| Use PAI instead of EditGuard on partial edits (PS) | PAI vs EditGuard · Tamper localization (PS) | F1 (%): 86.07 -> 66.24 | Hurt |
| Use PAI instead of EditGuard on average tamper localization | PAI vs EditGuard · Tamper localization (average) | AUC (%): 75.77 -> 89.77 | Improved |
| Use PAI instead of Gaussian Shading on image quality (FID) on CelebA-HQ | PAI vs Gaussian Shading · CelebA-HQ (clean) | FID: 23.72 -> 21.37 | Improved |
| Randomly tamper the saved timestamp on watermarked images (ablation) | PAI · Adaptive attack (timestamp tampering) | ownership verification accuracy: remained 100% | No clear effect |
| Embedded baselines under unseen degradation such as blur versus inherent schemes | EditGuard · Image degradation (blur) | ownership verification accuracy | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- EditGuard — EditGuard: Versatile Image Watermarking for Tamper Localization and Copyright Protection (CVPR 2024) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- StableGuard: Towards Unified Copyright Protection and Tamper Localization in Latent Diffusion Models (NeurIPS 2025 · similarity 0.91)
- EditGuard: Versatile Image Watermarking for Tamper Localization and Copyright Protection (CVPR 2024 · similarity 0.89)
- Black-Box Forgery Attacks on Semantic Watermarks for Diffusion Models (CVPR 2025 · similarity 0.89)
- SleeperMark: Towards Robust Watermark against Fine-Tuning Text-to-image Diffusion Models (CVPR 2025 · similarity 0.88)
- Robustness of AI-Image Detectors: Fundamental Limits and Practical Attacks (ICLR 2024 · similarity 0.88)
- Invisible Image Watermarks Are Provably Removable Using Generative AI (NeurIPS 2024 · similarity 0.88)
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.
