TRACE: Structure-Aware Character Encoding for Robust and Generalizable Document Watermarking

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Meng, Jiale, Zhang, Jie, Hu, Runyi, Lu, Zhe-Ming, Zhang, Tianwei, Li, Yiming
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915859308478464
author Meng, Jiale
Zhang, Jie
Hu, Runyi
Lu, Zhe-Ming
Zhang, Tianwei
Li, Yiming
author_facet Meng, Jiale
Zhang, Jie
Hu, Runyi
Lu, Zhe-Ming
Zhang, Tianwei
Li, Yiming
contents We propose TRACE, a structure-aware framework leveraging diffusion models for localized character encoding to embed data. Unlike existing methods that rely on edge features or pre-defined codebooks, TRACE exploits character structures that provide inherent resistance to noise interference due to their stability and unified representation across diverse characters. Our framework comprises three key components: (1) adaptive diffusion initialization that automatically identifies handle points, target points, and editing regions through specialized algorithms including movement probability estimator (MPE), target point estimation (TPE) and mask drawing model (MDM), (2) guided diffusion encoding for precise movement of selected point, and (3) masked region replacement with a specialized loss function to minimize feature alterations after the diffusion process. Comprehensive experiments demonstrate \name{}'s superior performance over state-of-the-art methods, achieving more than 5 dB improvement in PSNR and 5\% higher extraction accuracy following cross-media transmission. \name{} achieves broad generalizability across multiple languages and fonts, making it particularly suitable for practical document security applications.
format Preprint
id arxiv_https___arxiv_org_abs_2603_12873
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TRACE: Structure-Aware Character Encoding for Robust and Generalizable Document Watermarking
Meng, Jiale
Zhang, Jie
Hu, Runyi
Lu, Zhe-Ming
Zhang, Tianwei
Li, Yiming
Computer Vision and Pattern Recognition
We propose TRACE, a structure-aware framework leveraging diffusion models for localized character encoding to embed data. Unlike existing methods that rely on edge features or pre-defined codebooks, TRACE exploits character structures that provide inherent resistance to noise interference due to their stability and unified representation across diverse characters. Our framework comprises three key components: (1) adaptive diffusion initialization that automatically identifies handle points, target points, and editing regions through specialized algorithms including movement probability estimator (MPE), target point estimation (TPE) and mask drawing model (MDM), (2) guided diffusion encoding for precise movement of selected point, and (3) masked region replacement with a specialized loss function to minimize feature alterations after the diffusion process. Comprehensive experiments demonstrate \name{}'s superior performance over state-of-the-art methods, achieving more than 5 dB improvement in PSNR and 5\% higher extraction accuracy following cross-media transmission. \name{} achieves broad generalizability across multiple languages and fonts, making it particularly suitable for practical document security applications.
title TRACE: Structure-Aware Character Encoding for Robust and Generalizable Document Watermarking
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.12873