Skip to content

FeatMark: Feature-level Watermark Protection against Mimicry Attacks with Diffusion Models

Sep 2026 · arXiv.org · Vol abs/2609.30980 · 0 citations · 91 references
Computer Science

TL;DR

FeatMark is introduced, a watermarking framework that shifts from pixel-level, energy-starved perturbations to inconspicuous semantic features: small, scene-consistent micro- features that remain natural to humans while providing a stronger, machine-verifiable provenance signal.

Abstract

Text-to-image diffusion models enable data-efficient"mimicry"attacks, wherein adversaries fine-tune the model on a handful of public photos to synthesize convincing forgeries of a target individual. A common countermeasure is to embed imperceptible, low-energy watermarks, yet recent studies show these signatures are brittle: modest post-processing or lightweight adversarial perturbations readily suppress detection, exposing a fundamental tension between imperceptibility and robustness. We introduce FeatMark, a watermarking framework that shifts from pixel-level, energy-starved perturbations to inconspicuous semantic features: small, scene-consistent micro-features that remain natural to humans while providing a stronger, machine-verifiable provenance signal. FeatMark builds domain-specific feature banks that encode each watermark as a compact concept program, pairing open-vocabulary semantic cues with reliable edit regions and instruction templates. It then automatically selects features that are both feasible and executable and injects them through modular, mask-guided concept editing, yielding highly localized, scene-consistent micro-edits that are difficult to perceive. We conduct extensive experiments across VGGFace2, CelebA-HQ, and WikiArt, evaluating against 10 strong watermark removal/purification attacks (including regeneration-style purification) and several bespoke adaptive attacks tailored to FeatMark, to assess perceptual fidelity, watermark detection accuracy, and robustness. We further demonstrate FeatMark's extensibility to video mimicry attacks. The results show FeatMark remains virtually impervious, withstanding all evaluated attacks with negligible bit-accuracy and fidelity degradation.

View source

Similar papers

Open access Sep 2026

FakeMark: gradient-guided false watermark claims via robust feature fusion

FakeMark is presented, a gradient-guided false-claim attack for image classifiers that uses a white-box surrogate but never queries or accesses the victim model during attack construction, to motivate provenance-aware, multi-factor ownership protocols.

Yu-Tong Wu, Wen-Yue Li, He-Wang Nie et al. · 0 citations

An Experimental Study on Attacks and Vulnerabilities of Text Watermarks

Results show that the copy–paste attack degrades watermark detectability when using very sparingly and only short sentences are used, while deletion becomes effective only at extreme levels on short texts, and underscore the need for future watermark designs to prioritise resilience to semantic and cross-lingual transf...

Jakob Elias Albrecht, A. Jakoby, Benno Maria et al. · 0 citations
Conference Open access Sep 2026

Latents-Inv:Robust Semantic Watermark via Dual-Path Mutual Information Redundancy for Diffusion Models

A dual-path network is proposed to encode watermark information into both the generated image and the owner’s secret key, which achieves superior robustness against various adversarial attacks while maintaining high visual quality across diverse generative models.

Cong-Rong Li, Ling-Yun Yu, Pei-Qi Jiang et al. · 0 citations
Sep 2026

Semantic Watermarking for Malicious Image Manipulation Detection

The proliferation of high-fidelity generative editing models has made it possible to inject violent or sexual content into otherwise ordinary images while preserving visual plausibility, with concrete consequences for public discourse and vulnerable populations. We propose a robust semantic watermarking framework that...

Yoonseo Kim, S. Baek, Jun-Yong Park · 0 citations
Preprint Sep 2026

MarkSec: Capability-Aware Evaluation of Adversarial Attacks Against LLM Watermarks

Results show that apparent attack winners depend on text-quality constraints, attack generality, and capability assumptions, and that apparent attack winners depend on text-quality constraints, attack generality, and capability assumptions.

Kai-Rong Li, Zhi-Kun Zhang, Xiao-Nan Ren et al. · 1 citation
Sep 2026

DeMark: A Query-Free Black-Box Attack for Quality-Preserving Audio Watermark Removal

Audio watermarking protects digital speech by embedding imperceptible signals for ownership verification and misuse tracing. However, the security of learning-based watermarking remains insufficiently understood under realistic adversarial removal, where attackers cannot access or query the watermark encoder, decoder,...

Wei-Kang Ding, Bin-Hao Ma, Han-Qing Guo et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.