MarkSweep: A No-box Removal Attack on AI-Generated Image Watermarking via Noise Intensification and Frequency-aware Denoising

TL;DR

This paper presents MarkSweep, a novel attack that effectively removes AI-generated image watermarks by amplifying noise and using frequency-aware denoising, significantly reducing watermark detection accuracy without affecting image quality.

Contribution

Introduces MarkSweep, a new no-box watermark removal attack leveraging noise amplification and frequency-aware denoising modules, demonstrating high effectiveness against existing AI watermarking schemes.

Findings

Reduces watermark bit accuracy below 67%

Maintains perceptual image quality

Effective against multiple watermarking schemes

Abstract

AI watermarking embeds invisible signals within images to provide provenance information and identify content as AI-generated. In this paper, we introduce MarkSweep, a novel watermark removal attack that effectively erases the embedded watermarks from AI-generated images without degrading visual quality. MarkSweep first amplifies watermark noise in high-frequency regions via edge-aware Gaussian perturbations and injects it into clean images for training a denoising network. This network then integrates two modules, the learnable frequency decomposition module and the frequency-aware fusion module, to suppress amplified noise and eliminate watermark traces. Theoretical analysis and extensive experiments demonstrate that invisible watermarks are highly vulnerable to MarkSweep, which effectively removes embedded watermarks, reducing the bit accuracy of HiDDeN and Stable Signature watermarking schemes to below 67%, while preserving perceptual quality of AI-generated images.