AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Removing Box-Free Watermarks for Image-to-Image Models via Query-Based Reverse Engineering

Haonan An, Guang Hua, Hangcheng Cao, Zhengru Fang, Guowen Xu, Susanto Rahardja, Yuguang Fang

AAAI 2026 · Adversarial robustness and security › Black-box and transfer attacks · approach: Data-free substitute training · 2 citations (OpenAlex)

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 changedWhereMeasuredEffect
train a forward surrogate of HNet instead of an inverse of HNetquery-based watermark removal · Deraining, victim VZhangPSNR (dB): 33.98 -> 34.69Improved
train a forward surrogate of HNet instead of an inverse of HNetquery-based watermark removal · Image generation, victim VWuPSNR (dB): 31.75 -> 32.05Improved
train a forward surrogate of HNet instead of an inverse of HNetquery-based watermark removal · Image generation, victim VZhangMS-SSIM: 0.991 -> 0.990No clear effect
query with identity-transformation images that bypass GNet (rather than images from domain A) (ablation)HNet inverse attack · Image generationimage qualityImproved
query with identity-transformation images that bypass GNet (rather than images from domain A) (ablation)Forward HNet attack · Image generationwatermark removalImproved
apply Forward HNet attack versus generic JPEG compressionVWu deraining · Deraining, victim VWuPSNR (dB): 26.53 -> 33.75 (JPEG-20% vs Forward HNet)Improved
apply Forward HNet attack versus WEvade-B-QVWu deraining · Deraining, victim VWuremoval success rate: 0.380 -> 1.000Improved
apply Forward HNet attack versus AWGN-30dB noiseVWu image generation · Image generation, victim VWuremoval success rate: 0.000 -> 1.000Improved
apply attack versus Regeneration-VAEVZhang image generation · Image generation, victim VZhangPSNR (dB): 32.54 -> 32.60No clear effect
approximate delta of ai by delta of bi (additive equivalence)Forward HNet attack · Deraining, victim VZhangcorrelation: 0.635 (deraining VZhang) vs 0.894 (image generation VZhang)Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy