Pytorch implementation of the paper SwinDocSegmenter: An End-to-End Unified Domain Adaptive Transformer for Document Instance Segmentation. This model is implemented on top of the detectron2 framework. The proposed model can be used to analysis the complex layouts including magazines, Scientific Reports, historical documents, patents and so on as shown in the following examples.
Magazines | Scientific Reports |
![]() |
![]() |
Tables | Others |
![]() |
![]() |
git clone https://github.com/ayanban011/SwinDocSegmenter.git
cd SwinDocSegmenter
follow the installation instructions
python ./train_net.py \
--config-file maskdino_R50_bs16_50ep_4s_dowsample1_2048.yaml \
--eval-only \
--num-gpus 1 \
MODEL.WEIGHTS ./model_final.pth
python train_net.py --num-gpus 1 --config-file config_path SOLVER.IMS_PER_BATCH SET_TO_SOME_REASONABLE_VALUE SOLVER.BASE_LR SET_TO_SOME_REASONABLE_VALUE
In this section we release the pre-trained weights for all the best DocEnTr model variants trained on benchmark datasets.
Dataset | Config-file | Weights | AP |
---|---|---|---|
PublayNet | config-publay | model | 93.72 |
Prima | config-prima | model | 54.39 |
HJ Dataset | config-hj | model | 84.65 |
TableBank | config-table | model | 98.04 |
DoclayNet | config-doclay | model | 76.85 |
If you find this useful for your research, please cite it as follows:
@article{banerjee2023swindocsegmenter,
title={SwinDocSegmenter: An End-to-End Unified Domain Adaptive Transformer for Document Instance Segmentation},
author={Banerjee, Ayan and Biswas, Sanket and Llad{\'o}s, Josep and Pal, Umapada},
journal={arXiv preprint arXiv:2305.04609},
year={2023}
}
Many thanks to these excellent opensource projects
Thank you for interesting in our work, and sorry if there is any bugs.