Lightweight Feature Watermark for Robust Image Attribution in Latent Diffusion Models
IEEE Signal Processing Letters · 2026 · 33 · 1501-1505
DOI: 10.1109/lsp.2026.3679279 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Zhangyi Shen und 3 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben. Gefördert von: Zhejiang Provincial Natural Science Foundation of China, National Natural Science Foundation of China, Sino-French Joint Laboratory for Digital Media Forensics of Zhejiang Province.
Abstract (Originalsprache)
Latent diffusion models (LDMs) have become the dominant paradigm for AI image synthesis, while also posing unprecedented challenges to AI image attribution. To mitigate these risks, digital watermarking has emerged as a crucial technique for embedding traceable identity information into generated content. However, existing watermarking techniques for LDMs often suffer from degraded image quality, limited robustness, and substantial computational and storage overhead. To address these limitations, we propose Feature Watermarking, a novel watermarking paradigm in which pre-trained, randomly sampled feature vectors are leveraged to enable robust attribution capabilities. Subsequently, we employ a dedicated encoder-decoder architecture to generate initial noise and reconstruct features, respectively. Image attribution is performed by measuring the similarity between the original and reconstructed features. Experimental results demonstrate that the proposed method matches or outperforms existing state-of-the-art approaches in image quality and attribution accuracy, while significantly reducing resource consumption.
Förderung
- Zhejiang Provincial Natural Science Foundation of China
- National Natural Science Foundation of China
- Sino-French Joint Laboratory for Digital Media Forensics of Zhejiang Province
Zitieren
APA
Shen, Z., Zhuang, Y., Zhu, Y., & Yao, Y. (2026). Lightweight Feature Watermark for Robust Image Attribution in Latent Diffusion Models. IEEE Signal Processing Letters, 33, 1501-1505. https://doi.org/10.1109/lsp.2026.3679279
Vancouver
Shen Z, Zhuang Y, Zhu Y, Yao Y. Lightweight Feature Watermark for Robust Image Attribution in Latent Diffusion Models. IEEE Signal Processing Letters. 2026;33:1501-1505. doi:10.1109/lsp.2026.3679279
BibTeX
@article{Shen2026,
author = {Shen, Zhangyi and Zhuang, Yuzhong and Zhu, Yani and Yao, Ye},
title = {Lightweight Feature Watermark for Robust Image Attribution in Latent Diffusion Models},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {1501--1505},
doi = {10.1109/lsp.2026.3679279}
}
EUVIMED