# Rel-Zero: Harnessing Patch-Pair Invariance for Robust Zero-Watermarking Against AI Editing

Source: [Rel-Zero: Harnessing Patch-Pair Invariance for Robust Zero-Watermarking Against AI Editing](<https://www.sota2.com/research/paper/69bb83240636e5bf4afc1b90>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-17T05:00:21.724Z

Authors: Pengzhen Chen, Yanwei Liu, Xiaoyan Gu, Xiaojun Chen, Wu Liu, Weiping Wang

Published: 2026-03-18T09:40:00.000Z

## Abstract

Recent advancements in diffusion-based image editing pose a significant threat to the authenticity of digital visual content. Traditional embedding-based watermarking methods often introduce perceptible perturbations to maintain robustness, inevitably compromising visual fidelity. Meanwhile, existing zero-watermarking approaches, typically relying on global image features, struggle to withstand sophisticated manipulations. In this work, we uncover a key observation: while individual image patches undergo substantial alterations during AI-based editing, the relational distance between patch pairs remains relatively invariant. Leveraging this property, we propose Relational Zero-Watermarking (Rel-Zero), a novel framework that requires no modification to the original image but derives a unique zero-watermark from these editing-invariant patch relations. By grounding the watermark in intrinsic structural consistency rather than absolute appearance, Rel-Zero provides a non-invasive yet resilient mechanism for content authentication. Extensive experiments demonstrate that Rel-Zero achieves substantially improved robustness across diverse editing models and manipulations compared to prior zero-watermarking approaches.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Watermarking Robustness Detection on ControlNet-Inpainting](<https://www.sota2.com/research/sota/watermarking-robustness-detection-on-controlnet-inpainting.md>): Watermarking Robustness Detection; ControlNet-Inpainting; TPR@0.1%FPR: 97.43.

- [Watermark Extraction on Watermark Robustness and Fidelity Regeneration, AI Editing, Distortions 1.0](<https://www.sota2.com/research/sota/watermark-extraction-on-watermark-robustness-and-fidelity-regeneration-ai-editing-distortions-1-0.md>): Watermark Extraction; Watermark Robustness and Fidelity Regeneration, AI Editing, Distortions 1.0; SSIM: 100.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.17531>)

[PDF](<https://arxiv.org/pdf/2603.17531.pdf>)
