AI Paper Atlas

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

ICCV 2021 · Graph learning › Knowledge and temporal graphs · approach: Heterogeneous graph learning · 88 citations (OpenAlex)

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 changedWhereMeasuredEffect
add GCN to the logical location branch (ablation)TGRNet logical location branch · TableGraph-24KA_all: 0.697 -> 0.788Improved
add ordinal regression loss on top of GCN (ablation)TGRNet logical location branch · TableGraph-24KA_all: 0.788 -> 0.824Improved
add focal loss on top of GCN and ordinal regression (ablation)TGRNet logical location branch · TableGraph-24KA_all: 0.824 -> 0.832Improved
add GCN (ablation)TGRNet logical location branch · TableGraph-24KF_beta=0.5: 0.850 -> 0.876Improved
add ordinal regression loss (ablation)TGRNet logical location branch · TableGraph-24KF_beta=0.5: 0.876 -> 0.882Improved
add focal loss (ablation)TGRNet logical location branch · TableGraph-24KF_beta=0.5: 0.882 -> 0.890Improved
add GCN (ablation)TGRNet logical location branch · TableGraph-24Kcell spatial Hmean: 0.899 -> 0.901No clear effect
remove 20% of cells (80% kept) on CMDD, TGRNet vs ReS2TIMTGRNet · CMDD robustnessA_all: 0.857 vs 0.705Improved
use TGRNet instead of ReS2TIM on 100% cellsTGRNet · CMDDA_all: 0.995 vs 0.999No clear effect
use TGRNet instead of ReS2TIM on 100% cellsTGRNet · ICDAR13-TableA_all: 0.174 -> 0.334Improved
use TGRNet instead of ReS2TIM for logical locationTGRNet · ICDAR19-cTDaR TrackB1A_all: 0.138 -> 0.267Improved
use TGRNet instead of ReS2TIM on adjacency metricTGRNet · ICDAR19-cTDaR TrackB1WAF: 0.481 -> 0.283Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy