This repo is deprecated, please move to new repo
code for <Zhao, P., Wang, W., Zhang, G., & Lu, Y. (2021). Alleviating pseudo-touching in attention U-Net-based binarization approach for the historical Tibetan document images. Neural Computing and Applications, 1-12.>
python Train.py [imgs directory/] [masks directory/] (dont miss the last '/')
python Prediction.py [imgs dir] [out_dir] [model_pth] You are welcome to contact me anytime: oceanytech@gmail.com
The optimal magnifications of various data might be different. We suggest trying various magnifications on your own dataset.