TRACE: Structure-Aware Character Encoding for Robust and Generalizable Document Watermarking
On the map: Diffusion models → Robustness, privacy and adversarial aspects
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 17 of its 35 references that are in the map. First posted on arXiv 13 Mar 2026.
Builds on
17 of 35 references.
References · 17
- 2020Denoising Diffusion Probabilistic Modelscited by 20,851
- 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
- 2018OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fieldscited by 1,161
- 2023A Watermark for Large Language Modelscited by 667
- 2023DragDiffusion: Harnessing Diffusion Models for Interactive Point-based Image Editingcited by 293
- 2019Character Region Awareness for Text Detectioncited by 273
- 2024UltraEdit: Instruction-based Fine-Grained Image Editing at Scalecited by 266
- 2019StegaStamp: Invisible Hyperlinks in Physical Photographscited by 256
- 2022SINE: SINgle Image Editing with Text-to-Image Diffusion Modelscited by 159
- 2020Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hidingcited by 121
- 2022CATER: Intellectual Property Protection on Text Generation APIs via Conditional Watermarkscited by 68
- 2023REMARK-LLM: A Robust and Efficient Watermarking Framework for Generative Large Language Modelscited by 64
5 more
- 2024Drag Your Noise: Interactive Point-based Editing via Diffusion Semantic Propagationcited by 50
- 2024Watermark Anything with Localized Messagescited by 49
- 2021Tracing Text Provenance via Context-Aware Lexical Substitutioncited by 40
- 2024InstantDrag: Improving Interactivity in Drag-based Image Editingcited by 30
- 2025CoreMark: Toward Robust and Universal Text Watermarking Techniquecited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Global Sketch-Based Watermarking for Diffusion Language Models
- 2025Dynamic watermarks in images generated by diffusion models
- 2026ShapeMark: Robust and Diversity-Preserving Watermarking for Diffusion Models
- 2026When Denoising Becomes Unsigning: Theoretical and Empirical Analysis of Watermark Fragility Under Diffusion-Based Image Editing
- 2024SleeperMark: Towards Robust Watermark against Fine-Tuning Text-to-image Diffusion Models
- 2025PT-Mark: Invisible Watermarking for Text-to-image Diffusion Models via Semantic-aware Pivotal Tuning
- 2026TANGO: Watermarking Masked Diffusion Language Models in Token Pairs
- 2025SWA-LDM: Toward Stealthy Watermarks for Latent Diffusion Models