Document-image-binarization-with-Cascaded-cGANs-generator

code for our paper "Document image binarization with cGANs based cascaded generator".

tools:

1.The python code is based on the python data science platform Anaconda3.

2.The python code is tested on Windows by PyCharm.

3.Install PyTorch and dependencies from http://pytorch.org

Testing:

1. few example test images are included in the testimg folder.

2. Run Test_All.py, test the binarization of an input image.

3. Please download the pre-trained model from here , and put it under ./checkpoints/

4. Executable file link ：exe

Training:

1. Please download the public binarization datasets

2. Run combine_A_and_B_sub.py to get the sub training data.

3. Run train.py to train a new model.

Acknowledgments:

This code borrows heavily from pytorch-CycleGAN-and-pix2pix. Cite:

Please cite this article as: J.Zhao, C.Shi and F.Jia et al.,Document image binarization with cascaded generators of conditional generative adversarial networks, Pattern Recognition, https://doi.org/10.1016/j.patcog.2019.106968