Are Watermarked Images Editable? SafeMark for Watermark-Preserving Text-Guided Image Editing

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wu, Xiaodong, Li, Qi, Li, Xiangman, Zhang, Zelin, Liu, Lingshuang, Ni, Jianbing
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910235304656896
author Wu, Xiaodong
Li, Qi
Li, Xiangman
Zhang, Zelin
Liu, Lingshuang
Ni, Jianbing
author_facet Wu, Xiaodong
Li, Qi
Li, Xiangman
Zhang, Zelin
Liu, Lingshuang
Ni, Jianbing
contents This paper investigates a fundamental yet underexplored question: can watermarked images remain editable without compromising watermark integrity? We propose SafeMark, a framework for watermark-preserving text-guided image manipulation that explicitly integrates watermark integrity into the editing process. Specifically, SafeMark adds a thresholded watermark-decoding loss directly to the diffusion editor's training objective, fine-tuning the editor so that semantically valid edits also preserve the embedded watermark at the final output. This design admits a clean information-theoretic justification: maintaining high bit-accuracy on the edited image lower-bounds the mutual information that the editor channel preserves between watermark and edited output, the quantity that fundamentally controls watermark recoverability. SafeMark is compatible with differentiable diffusion-based editors, and requires no architectural modification. Extensive evaluations across multiple datasets, text-guided editing methods, and post-edit distortion settings demonstrate that SafeMark achieves high watermark bit accuracy across diverse editing settings while maintaining high-quality semantic edits, without sacrificing robustness to common post-edit distortions. These results demonstrate that semantic editability and watermark integrity are fundamentally compatible, enabling trustworthy image provenance in generative editing pipelines.
format Preprint
id arxiv_https___arxiv_org_abs_2605_19511
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Are Watermarked Images Editable? SafeMark for Watermark-Preserving Text-Guided Image Editing
Wu, Xiaodong
Li, Qi
Li, Xiangman
Zhang, Zelin
Liu, Lingshuang
Ni, Jianbing
Computer Vision and Pattern Recognition
This paper investigates a fundamental yet underexplored question: can watermarked images remain editable without compromising watermark integrity? We propose SafeMark, a framework for watermark-preserving text-guided image manipulation that explicitly integrates watermark integrity into the editing process. Specifically, SafeMark adds a thresholded watermark-decoding loss directly to the diffusion editor's training objective, fine-tuning the editor so that semantically valid edits also preserve the embedded watermark at the final output. This design admits a clean information-theoretic justification: maintaining high bit-accuracy on the edited image lower-bounds the mutual information that the editor channel preserves between watermark and edited output, the quantity that fundamentally controls watermark recoverability. SafeMark is compatible with differentiable diffusion-based editors, and requires no architectural modification. Extensive evaluations across multiple datasets, text-guided editing methods, and post-edit distortion settings demonstrate that SafeMark achieves high watermark bit accuracy across diverse editing settings while maintaining high-quality semantic edits, without sacrificing robustness to common post-edit distortions. These results demonstrate that semantic editability and watermark integrity are fundamentally compatible, enabling trustworthy image provenance in generative editing pipelines.
title Are Watermarked Images Editable? SafeMark for Watermark-Preserving Text-Guided Image Editing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.19511