Table Detection on ICDAR cTDaR 2019 (modern)
98IoU@0.6DiT-L
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DiT-LBackbone=DiT-L2022.03 | 98 | 97.56 | 96.23 | 91.57 | 95.5 | |
| DiT-L (Cascade)Backbone=DiT-L, Detection Framework=Cascade R-CNN2022.03 | 97.89 | 97.22 | 97 | 93.88 | 96.29 | |
| DiT-B (Cascade)Backbone=DiT-B, Detection Framework=Cascade R-CNN2022.03 | 97.33 | 96.89 | 96.67 | 93.33 | 95.85 | |
| 1st place in cTDaR2022.03 | 96.86 | 95.74 | 95.07 | 89.69 | 93.97 | |
| MAE-BBackbone=MAE-B2022.03 | 96.47 | 95.58 | 94.48 | 90.07 | 93.81 | |
| ResNeXt-101-32x8d (Cascade)Backbone=ResNeXt-101-32x8d, Detection Framework=Cascade R-CNN2022.03 | 96.41 | 95.52 | 95.07 | 92.38 | 94.63 | |
| ResNeXt-101-32x8dBackbone=ResNeXt-101-32x8d2022.03 | 96.3 | 95.63 | 95.18 | 91.15 | 94.3 | |
| DiT-BBackbone=DiT-B2022.03 | 96.29 | 95.61 | 95.39 | 92.46 | 94.74 | |
| BEIT-BBackbone=BEIT-B2022.03 | 96.06 | 95.39 | 95.16 | 91.34 | 94.25 | |
| DeiT-BBackbone=DeiT-B2022.03 | 95.51 | 94.61 | 93.48 | 89.89 | 93.07 |