Research map / Graph learning / Knowledge and temporal graphs
TGRNet: A Table Graph Reconstruction Network for Table Structure Recognition
Wenyuan Xue, Baosheng Yu, Wen Wang, Dacheng Tao, Qingyong Li
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 |
|---|---|---|---|
| add GCN to the logical location branch (ablation) | TGRNet logical location branch · TableGraph-24K | A_all: 0.697 -> 0.788 | Improved |
| add ordinal regression loss on top of GCN (ablation) | TGRNet logical location branch · TableGraph-24K | A_all: 0.788 -> 0.824 | Improved |
| add focal loss on top of GCN and ordinal regression (ablation) | TGRNet logical location branch · TableGraph-24K | A_all: 0.824 -> 0.832 | Improved |
| add GCN (ablation) | TGRNet logical location branch · TableGraph-24K | F_beta=0.5: 0.850 -> 0.876 | Improved |
| add ordinal regression loss (ablation) | TGRNet logical location branch · TableGraph-24K | F_beta=0.5: 0.876 -> 0.882 | Improved |
| add focal loss (ablation) | TGRNet logical location branch · TableGraph-24K | F_beta=0.5: 0.882 -> 0.890 | Improved |
| add GCN (ablation) | TGRNet logical location branch · TableGraph-24K | cell spatial Hmean: 0.899 -> 0.901 | No clear effect |
| remove 20% of cells (80% kept) on CMDD, TGRNet vs ReS2TIM | TGRNet · CMDD robustness | A_all: 0.857 vs 0.705 | Improved |
| use TGRNet instead of ReS2TIM on 100% cells | TGRNet · CMDD | A_all: 0.995 vs 0.999 | No clear effect |
| use TGRNet instead of ReS2TIM on 100% cells | TGRNet · ICDAR13-Table | A_all: 0.174 -> 0.334 | Improved |
| use TGRNet instead of ReS2TIM for logical location | TGRNet · ICDAR19-cTDaR TrackB1 | A_all: 0.138 -> 0.267 | Improved |
| use TGRNet instead of ReS2TIM on adjacency metric | TGRNet · ICDAR19-cTDaR TrackB1 | WAF: 0.481 -> 0.283 | Hurt |
Closest papers by meaning
- TDeLTA: A Light-Weight and Robust Table Detection Method Based on Learning Text Arrangement (AAAI 2024 · similarity 0.88)
- Grab What You Need: Rethinking Complex Table Structure Recognition with Flexible Components Deliberation (AAAI 2024 · similarity 0.88)
- TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition (CVPR 2026 · similarity 0.87)
- HeGTa: Leveraging Heterogeneous Graph-enhanced Large Language Models for Few-shot Complex Table Understanding (AAAI 2025 · similarity 0.84)
- Polynomial-based Self-Attention for Table Representation Learning (ICML 2024 · similarity 0.84)
- TabGLM: Tabular Graph Language Model for Learning Transferable Representations Through Multi-Modal Consistency Minimization (AAAI 2025 · similarity 0.84)
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.
