Research map / Image generation / Image restoration
Parsing Table Structures in the Wild
Rujiao Long, Wen Wang, Nan Xue, Feiyu Gao, Zhibo Yang, Yongpan Wang, Gui-Song Xia
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 quadrilateral (polygon) cell regression instead of horizontal boxes (ablation) | CenterNet · WTW test | TEDS: 58.7 -> 70.1 | Improved |
| use quadrilateral cell regression instead of horizontal boxes (ablation) | CenterNet · WTW test | physical structure F1: 73.1 -> 75.4 | Improved |
| use quadrilateral cell regression instead of horizontal boxes (ablation) | CenterNet · WTW test | adjacency relation F1: 84.8 -> 91.1 | Improved |
| add the cycle-pairing module (ablation) | CenterNet (polygon box) · WTW test | TEDS: 70.1 -> 74.3 | Improved |
| add the cycle-pairing module (ablation) | CenterNet (polygon box) · WTW test | physical structure F1: 75.4 -> 78.2 | Improved |
| add the cycle-pairing module (ablation) | CenterNet (polygon box) · WTW test | adjacency relation F1: 91.1 -> 92.2 | Improved |
| add the pairing loss (ablation) | Cycle-CenterNet · WTW test | TEDS: 74.3 -> 83.3 | Improved |
| add the pairing loss (ablation) | Cycle-CenterNet · WTW test | physical structure F1: 78.2 -> 78.3 | No clear effect |
| use the anchor-free CenterNet detector instead of anchor-based Cascade-RCNN | cell detector baseline · WTW test | TEDS: 53.2 -> 58.7 | Improved |
| use Cycle-CenterNet with pairing loss instead of vanilla CenterNet | table structure parsing · WTW test | TEDS: 58.7 -> 83.3 | Improved |
| use Cycle-CenterNet with pairing loss instead of vanilla CenterNet | table structure parsing · WTW test | adjacency relation F1: 84.8 -> 92.4 | Improved |
| train on WTW only and evaluate wired tables of ICDAR-2013 versus mixed wired and wireless tables | Cycle-CenterNet · WTW test | adjacency F1 (IoU 0.6): 91.7 -> 98.0 | Mixed |
Closest papers by meaning
- Grab What You Need: Rethinking Complex Table Structure Recognition with Flexible Components Deliberation (AAAI 2024 · similarity 0.87)
- Multimodal Tabular Reasoning with Privileged Structured Information (NeurIPS 2025 · similarity 0.83)
- Towards Real-World Document Parsing via Realistic Scene Synthesis and Document-Aware Training (CVPR 2026 · similarity 0.83)
- TDeLTA: A Light-Weight and Robust Table Detection Method Based on Learning Text Arrangement (AAAI 2024 · similarity 0.83)
- TableDART: Dynamic Adaptive Multi-Modal Routing for Table Understanding (ICLR 2026 · similarity 0.83)
- TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition (CVPR 2026 · similarity 0.83)
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.
