OmniGuard: Hybrid Manipulation Localization via Augmented Versatile Deep Image Watermarking

TL;DR

OmniGuard is a hybrid watermarking method that enhances image copyright protection and tamper localization by combining proactive embedding with passive extraction, achieving superior robustness and flexibility against AI-generated edits.

Contribution

It introduces a novel augmented versatile watermarking framework with a degradation-aware tamper localization network and flexible watermark selection, surpassing existing methods in fidelity and robustness.

Findings

Outperforms state-of-the-art EditGuard by 4.25dB PSNR

Achieves 20.7% higher F1-Score under noisy conditions

Improves average bit accuracy by 14.8%

Abstract

With the rapid growth of generative AI and its widespread application in image editing, new risks have emerged regarding the authenticity and integrity of digital content. Existing versatile watermarking approaches suffer from trade-offs between tamper localization precision and visual quality. Constrained by the limited flexibility of previous framework, their localized watermark must remain fixed across all images. Under AIGC-editing, their copyright extraction accuracy is also unsatisfactory. To address these challenges, we propose OmniGuard, a novel augmented versatile watermarking approach that integrates proactive embedding with passive, blind extraction for robust copyright protection and tamper localization. OmniGuard employs a hybrid forensic framework that enables flexible localization watermark selection and introduces a degradation-aware tamper extraction network for precise localization under challenging conditions. Additionally, a lightweight AIGC-editing simulation layer is designed to enhance robustness across global and local editing. Extensive experiments show that OmniGuard achieves superior fidelity, robustness, and flexibility. Compared to the recent state-of-the-art approach EditGuard, our method outperforms it by 4.25dB in PSNR of the container image, 20.7% in F1-Score under noisy conditions, and 14.8% in average bit accuracy.