PubLayNet is a large dataset of document images, of which the layout is annotated with both bounding boxes and polygonal segmentations. For more information, see PubLayNet original
 |
 |
|---|---|
| PMC4334925_00006.jpg | PMC538274_00004.jpg |
29/Feb/2020 - Add benchmarking for maskrcnn_resnet50_fpn.
22/Feb/2020 - Pre-trained Mask-RCNN model in (Pytorch) are released .
| Architecture | Iter_num (x16) | AP | AP50 | AP75 | AP Small | AP Medium | AP Large | MD5SUM |
|---|---|---|---|---|---|---|---|---|
| MaskRCNN-Resnet50-FPN | 196k | 0.91 | 0.98 | 0.96 | 0.41 | 0.76 | 0.95 | 393e6700095a673065fcecf5e8f264f7 |
Download trained weights in Benchmarking section above, locate it in maskrcnn directory
Run
python infer.py <path_to_image>
