Research map / Image generation / Text-to-image diffusion
Foreground and Text-lines Aware Document Image Rectification
Heng Li, Xiangping Wu, Qingcai Chen, Qianjin Xiang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use both foreground and text-line attention modules instead of one (ablation) | document rectification transformer · DocUNet Benchmark | MS-SSIM: 0.4785 -> 0.4978 | Improved |
| use both attention modules instead of one (ablation) | document rectification transformer · DocUNet Benchmark | CER: 0.1889 -> 0.1705 | Improved |
| use both attention modules instead of one (ablation) | document rectification transformer · DocUNet Benchmark | ED: 748.46 -> 697.52 | Improved |
| compare proposed method with DocGeoNet on OCR | Ours vs DocGeoNet · DocUNet Benchmark | CER: 0.1821 -> 0.1705 | Improved |
| compare proposed method with DocGeoNet on OCR | Ours vs DocGeoNet · DIR300 | ED: 664.96 -> 652.80 | Improved |
| compare proposed method with DocGeoNet on image-distortion metrics | Ours vs DocGeoNet · DIR300 | MS-SSIM: 0.6380 -> 0.6070 | Hurt |
| higher text content ratio in image | proposed method · DocUNet Benchmark | LD: 10.74 -> 6.34 (LD, low to high text ratio); MS-SSIM 0.54 -> 0.44 | Mixed |
Closest papers by meaning
- Can You Read Me Now? Content Aware Rectification using Angle Supervision (ECCV 2020 · similarity 0.88)
- Exploiting Vector Fields for Geometric Rectification of Distorted Document Images (ECCV 2018 · similarity 0.88)
- ForCenNet: Foreground-Centric Network for Document Image Rectification (ICCV 2025 · similarity 0.87)
- Geometric Representation Learning for Document Image Rectification (ECCV 2022 · similarity 0.87)
- Prune and Repaint: Content-Aware Image Retargeting for any Ratio (NeurIPS 2024 · similarity 0.80)
- Predicting the Original Appearance of Damaged Historical Documents (AAAI 2025 · similarity 0.80)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
