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

2026-03-13

AI Summary

This work proposes TRACE, a novel document watermarking framework that overcomes the limitations of existing methods—such as reliance on edge features or predefined codebooks, poor noise robustness, and limited generalization—by leveraging the structural stability of characters. For the first time, TRACE adaptively embeds watermarks into local character regions using a diffusion model without requiring a predefined codebook. The approach integrates adaptive initialization, guided diffusion, and a masked region replacement mechanism, synergistically optimized through a Moving Probability Estimator (MPE), Target Point Estimator (TPE), and Mask Drawing Model (MDM). Experimental results demonstrate that TRACE achieves over 5 dB higher PSNR and a 5% improvement in extraction accuracy after cross-media transmission, significantly outperforming state-of-the-art methods while exhibiting strong cross-lingual and cross-font generalization capabilities.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

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.

Problem

• document watermarking
• robustness
• generalizability
• character encoding
• cross-media transmission

Innovation

• structure-aware watermarking
• diffusion models
• character encoding
• adaptive initialization
• masked region replacement