Physical Table Structure Recognition on ICDAR partial 2013 (test)
99.1PrecisionTabStruct-Net
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TabStruct-NetTrain Dataset=SciTSR, Setup=S-B (image + ground truth boxes)2020.10 | 99.1 | 99.3 | 99.2 | |
| Tabstruct-NetTraining Dataset=ICDAR-2013-partial, #Images=0.124K, Exp. Setup=S-B2020.10 | 99.1 | 98.9 | 99 | |
| Tabstruct-NetTraining Dataset=SCITSR + ICDAR-2013-partial, #Images=12.124K, Exp. Setup=S-B2020.10 | 99.1 | 99.3 | 99.2 | |
| DGCNNTraining Dataset=SCITSR + ICDAR-2013-partial, #Images=12.124K, Exp. Setup=S-B2020.10 | 98.6 | 99 | 98.8 | |
| Bi-directional GRUTraining Dataset=ICDAR-2013-partial, #Images=0.124K, Exp. Setup=S-A2020.10 | 96.9 | 90.1 | 93.4 | |
| DeepDESRTTraining Dataset=ICDAR-2013-partial, #Images=0.124K, Exp. Setup=S-A2020.10 | 95.9 | 87.4 | 91.4 | |
| TableNetTraining Dataset=Marmot extended, #Images=1.016K, Exp. Setup=S-B2020.10 | 93.1 | 90 | 91.5 | |
| TabStruct-NetTrain Dataset=SciTSR, Setup=S-A (image only)2020.10 | 93 | 90.8 | 91.9 | |
| Tabstruct-NetTraining Dataset=SCITSR + ICDAR-2013-partial, #Images=12.124K, Exp. Setup=S-A2020.10 | 93 | 90.8 | 91.9 | |
| Tabstruct-NetTraining Dataset=ICDAR-2013-partial, #Images=0.124K, Exp. Setup=S-A2020.10 | 92.8 | 90.3 | 91.5 | |
| SPLERGETraining Dataset=ICDAR-2013-partial, #Images=0.124K, Exp. Setup=S-A2020.10 | 91.7 | 91.1 | 91.4 | |
| GraphTSRTraining Dataset=SCITSR + ICDAR-2013-partial, #Images=12.124K, Exp. Setup=S-B2020.10 | 85.4 | 89.1 | 87.2 |