AI Paper Atlas

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

Finding a needle in a haystack: A Black-Box Approach to Invisible Watermark Detection

Minzhou Pan, Zhenting Wang, Xin Dong, Vikash Sehwag, Lingjuan Lyu, Xue Lin

ECCV 2024 · Adversarial robustness and security › Black-box and transfer attacks · approach: Data-free substitute training

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
use asymmetric loss (softmax on clean data, linear on detection data) instead of symmetric linear loss (ablation)WMD · ImageNet DctDwtSvdAUC: 0.786 -> 0.968Improved
use asymmetric loss instead of symmetric BCE loss (ablation)WMD · ImageNet DctDwtSvdAUC: 0.757 -> 0.968Improved
use exponential loss on detection data (maximize side) (ablation)WMD · ImageNet DctDwtSvdAUC: 0.968 -> 0.513Hurt
add iterative pruning with model reinitialization to the detection dataset (ablation)WMD · ImageNetAUCImproved
raise the pruning rate to a high value (ablation)WMD · ImageNetAUCHurt
lower the pruning rate from 50% to 1% (ablation)WMD · ImageNetcompute overhead: approximately 66.2 times more overheadMixed
apply multiple watermark types at once instead of a single typeWMD · ImageNet LSBAUC: 0.852 -> 0.732Hurt
replace the clean reference set with a mismatched-domain dataset (CelebA) (ablation)WMD · Detection ImageNetAUC: 0.968 -> 0.612Hurt
replace 80% of the clean set with synthesized images from detection-set captions (ablation)WMD · ImageNet DctDwtSvdAUC: 0.968 -> 0.684Hurt
replace 20% of the clean set with synthesized images (case 1) (ablation)WMD · ImageNet DctDwtSvdAUC: 0.968 -> 0.966No clear effect
use WMD detection output to build a pure watermarked set for PGD removalblack-box watermark removal · Tree-RingAUC after removal: AUC 0.504 after removalImproved

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