Reproduce “DocPure: Prompt-Free Unified Document Restoration via Degradation-Aware Structure-Guided Wavelet Modulation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lingming Su, Wanglong Lu, Tao Wang, Kaihao Zhang, Nan Zhang, Liyan An, Hanli Zhao
DocPure restores documents across four tasks—deblurring, denoising, compression-artifact reduction, and deshadowing—by predicting a structural map and combining it with frequency features, without a task prompt at inference. Degradation labels guide training; cross-task pairing to the same clean image is not required. On TDD deblurring, PaddleOCR character accuracy is 87.25% versus DocRes’s 83.22%, an OCR-based rather than workflow measure.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.