Table Detection on Marmot Chinese
0.98RecallM-RCNN
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| M-RCNNTraining Dataset=TableBank-LaTeX (199K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.98 | 0.82 | 0.89 | — | |
| CDeC-NetTraining Dataset=Pascal VOC (16K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.966 | 0.988 | 0.977 | 0.959 | |
| CDeC-NetTraining Dataset=TableBank-LaTeX (199K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.966 | 0.994 | 0.98 | 0.962 | |
| YOLOTraining Dataset=Pascal VOC (16K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.96 | 0.95 | 0.96 | — | |
| CDeC-Net+Training Dataset=IIIT-AR-13K (9K images), Fine-tuning Dataset=Marmot (Chinese) (754 images), IoU=0.62020.08 | 0.944 | 0.988 | 0.966 | 0.935 | |
| YOLOTraining Dataset=TableBank-LaTeX (199K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.93 | 0.97 | 0.95 | — | |
| RetinaNetTraining Dataset=TableBank-LaTeX (199K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.87 | 0.87 | 0.87 | — | |
| RetinaNetTraining Dataset=Pascal VOC (16K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.85 | 0.78 | 0.81 | — | |
| M-RCNNTraining Dataset=Pascal VOC (16K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.83 | 0.52 | 0.64 | — | |
| CDeC-Net+Training Dataset=IIIT-AR-13K (9K images), IoU=0.62020.08 | 0.791 | 0.921 | 0.856 | 0.736 | |
| SSDTraining Dataset=Pascal VOC (16K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.7 | 0.57 | 0.63 | — | |
| SSDTraining Dataset=TableBank-LaTeX (199K), Fine-tuning Dataset=Marmot (Chinese) (754), IoU=0.62020.08 | 0.67 | 0.61 | 0.64 | — |