mlsanityData to 7 Oct 2026

DocPure: Prompt-Free Unified Document Restoration via Degradation-Aware Structure-Guided Wavelet Modulation

arXiv 2608.09536 ↗First posted 10 Aug 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

High-quality document images are pivotal for information archiving and downstream automatic processing. However, they are frequently compromised by diverse degradations during uncontrolled acquisition and transmission. While unified document restoration techniques have been proposed to restore images from multiple degradations, they often struggle with training multiple degradation-specific models, reliance on manual task-specific prompts, or cross-task data pairing. To address these limitations, we propose DocPure, a prompt-free unified framework that achieves degradation-aware document restoration. We design a degradation-aware structure auto-encoder with degradation-informed routing regularization to predict clean structural priors from degraded inputs. The model is prompt-free at inference, and degradation labels are only used as auxiliary supervision for the routing regularization during training. Furthermore, we introduce a structure-guided wavelet interaction mechanism to bridge frequency-domain features and spatial semantics. Within the structure-guided wavelet interaction mechanism, a cross-frequency adaptive modulation utilizes low-frequency sub-bands to modulate high-frequency recovery, ensuring structural consistency. Extensive experiments demonstrate that DocPure achieves strong performance compared with state-of-the-art methods across various tasks, including deblurring, denoising, compression artifact reduction, and deshadowing.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 31 of its 54 references that are in the map. First posted on arXiv 10 Aug 2026.

Builds on

31 of 54 references.

References · 31

  1. 2017Attention Is All You Needcited by 59,932
  2. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  3. 2017Squeeze-and-Excitation Networkscited by 4,763
  4. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,154
  5. 2021Restormer: Efficient Transformer for High-Resolution Image Restorationcited by 1,618
  6. 2017Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentationscited by 614
  7. 2019FUNSD: A Dataset for Form Understanding in Noisy Scanned Documentscited by 345
  8. 2021DocFormer: End-to-End Transformer for Document Understandingcited by 254
  9. 2023Nougat: Neural Optical Understanding for Academic Documentscited by 240
  10. 2018Boundary loss for highly unbalanced segmentationcited by 187
  11. 2022On the Representation Collapse of Sparse Mixture of Expertscited by 156
  12. 2024AdaIR: Adaptive All-in-One Image Restoration via Frequency Mining and Modulationcited by 108
19 more
  1. 2025PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Modelcited by 84
  2. 2020DE-GAN: A Conditional Generative Adversarial Network for Document Enhancementcited by 44
  3. 2021Text Prior Guided Scene Text Image Super-resolutioncited by 37
  4. 2023DocDiff: Document Enhancement via Residual Diffusion Modelscited by 37
  5. 2023High-Resolution Document Shadow Removal via A Large-Scale Real-World Dataset and A Frequency-Aware Shadow Erasing Netcited by 36
  6. 2025MB-TaylorFormer V2: Improved Multi-branch Linear Transformer Expanded by Taylor Formula for Image Restorationcited by 36
  7. 2024DocRes: A Generalist Model Toward Unifying Document Image Restoration Taskscited by 27
  8. 2021Learning from Synthetic Shadows for Shadow Detection and Removalcited by 26
  9. 2018DMCNN: Dual-Domain Multi-Scale Convolutional Neural Network for Compression Artifacts Removalcited by 20
  10. 2019Water-Filling: An Efficient Algorithm for Digitized Document Shadow Removalcited by 17
  11. 2024Diff-Restorer: Unleashing Visual Prompts for Diffusion-based Universal Image Restorationcited by 15
  12. 2024High-Fidelity Document Stain Removal via A Large-Scale Real-World Dataset and A Memory-Augmented Transformercited by 12
  13. 2024Visual Style Prompt Learning Using Diffusion Models for Blind Face Restorationcited by 8
  14. 2021A novel segmentation dataset for signatures on bank checkscited by 4
  15. 2025Uni-DocDiff: A Unified Document Restoration Model Based on Diffusioncited by 4
  16. 2025TextDoctor: Unified Document Image Inpainting via Patch Pyramid Diffusion Modelscited by 2
  17. 2025UHDRes: Ultra-High-Definition Image Restoration via Dual-Domain Decoupled Spectral Modulationcited by 1
  18. 2026Unifying Heterogeneous Degradations: Uncertainty-Aware Diffusion Bridge Model for All-in-One Image Restorationcited by 1
  19. 2026EchoSR: Efficient Context Harnessing for Lightweight Image Super-Resolutioncited by 1

Built on by

Not cited yet.