GCN-PN

GCN-PN

This code repository contains the our re-implementations of the method GCN-PN (ECCV2020).

Train From Scratch

If you want to train the model from scratch, please following these steps:

1.Firstly, prepare the dataset and datalist follows demo/reading_order_detection/datalist/DI/readme.md

2.Secondly, prepare the pretrained models from model_zoo of mmdetection:

• resnext101_64x4d-ee2c6f71.pth

3.Thirdly, direct run demo/reading_order_detection/GCN-PN/train.sh

Test

Given the trained model, direct run demo/reading_order_detection/GCN-PN/test.sh to test model.

Trained Model Download

For the released data is a subset, which smaller than paper reported. So the results might be slightly different from reported results. Moreover, paper takes sinkhorn method into training phase and get some improvements, but it works less in our implementation. Thus, we only release the base model.

Results on DI datasets and trained models are follows:

total_order_acc	DI_whole	DI_subset	Links
GCN-PN (report)	79	-	-
GCN-PN	-	72.23	config, pth (Access Code：QDHU)

Citation

@inproceedings{DBLP:conf/eccv/LiGBWYZ20,
  author    = {Liangcheng Li and
               Feiyu Gao and
               Jiajun Bu and
               Yongpan Wang and
               Zhi Yu and
               Qi Zheng},
  title     = {An End-to-End {OCR} Text Re-organization Sequence Learning for Rich-Text
               Detail Image Comprehension},
  booktitle = {ECCV},
  pages     = {85--100},
  year      = {2020},
}

License

This project is released under the Apache 2.0 license

Contact

If there is any suggestion and problem, please feel free to contact the author with qiaoliang6@hikvision.com.