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

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Cao, Jie, Zhang, Zelin, Li, Qi, Ni, Jianbing
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918342858637312
author Cao, Jie
Zhang, Zelin
Li, Qi
Ni, Jianbing
author_facet Cao, Jie
Zhang, Zelin
Li, Qi
Ni, Jianbing
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2602_15364
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MarkSweep: A No-box Removal Attack on AI-Generated Image Watermarking via Noise Intensification and Frequency-aware Denoising
Cao, Jie
Zhang, Zelin
Li, Qi
Ni, Jianbing
Cryptography and Security
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.
title MarkSweep: A No-box Removal Attack on AI-Generated Image Watermarking via Noise Intensification and Frequency-aware Denoising
topic Cryptography and Security
url https://arxiv.org/abs/2602.15364