Beyond Invisibility: Learning Robust Visible Watermarks for Stronger Copyright Protection

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liu, Tianci, Yang, Tong, Zhang, Quan, Lei, Qi
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908391494909952
author Liu, Tianci
Yang, Tong
Zhang, Quan
Lei, Qi
author_facet Liu, Tianci
Yang, Tong
Zhang, Quan
Lei, Qi
contents As AI advances, copyrighted content faces growing risk of unauthorized use, whether through model training or direct misuse. Building upon invisible adversarial perturbation, recent works developed copyright protections against specific AI techniques such as unauthorized personalization through DreamBooth that are misused. However, these methods offer only short-term security, as they require retraining whenever the underlying model architectures change. To establish long-term protection aiming at better robustness, we go beyond invisible perturbation, and propose a universal approach that embeds \textit{visible} watermarks that are \textit{hard-to-remove} into images. Grounded in a new probabilistic and inverse problem-based formulation, our framework maximizes the discrepancy between the \textit{optimal} reconstruction and the original content. We develop an effective and efficient approximation algorithm to circumvent a intractable bi-level optimization. Experimental results demonstrate superiority of our approach across diverse scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02665
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond Invisibility: Learning Robust Visible Watermarks for Stronger Copyright Protection
Liu, Tianci
Yang, Tong
Zhang, Quan
Lei, Qi
Machine Learning
As AI advances, copyrighted content faces growing risk of unauthorized use, whether through model training or direct misuse. Building upon invisible adversarial perturbation, recent works developed copyright protections against specific AI techniques such as unauthorized personalization through DreamBooth that are misused. However, these methods offer only short-term security, as they require retraining whenever the underlying model architectures change. To establish long-term protection aiming at better robustness, we go beyond invisible perturbation, and propose a universal approach that embeds \textit{visible} watermarks that are \textit{hard-to-remove} into images. Grounded in a new probabilistic and inverse problem-based formulation, our framework maximizes the discrepancy between the \textit{optimal} reconstruction and the original content. We develop an effective and efficient approximation algorithm to circumvent a intractable bi-level optimization. Experimental results demonstrate superiority of our approach across diverse scenarios.
title Beyond Invisibility: Learning Robust Visible Watermarks for Stronger Copyright Protection
topic Machine Learning
url https://arxiv.org/abs/2506.02665