chu246 / inversehalftonecolorization

Inverse Halftone Colorization: Making Halftone Prints Color Photos, ICIP2021

Inverse Halftone Colorization: Making Halftone Prints Color Photos (ICIP 2021 accepted)

PyTorch implementaton of the following paper. In this paper, we propose a framework to recover colorful images from black and white halftone prints. You can visit our project website here.

Paper

Inverse Halftone Colorization: Making Halftone Prints Color Photos
Yu-Ting Yen, Chia-Chi Cheng, Wei-Chen Chiu
IEEE International Conference on Image Processing (ICIP), 2021.

Please cite our paper if you find it useful for your research.

@InProceedings{yen2021inverse,
    title={Inverse Halftone Colorization: Making Halftone Prints Color Photos},
    author={Yen, Yu-Ting and Cheng, Chia-Chi and Chiu, Wei-Chen},
    booktitle={2021 IEEE International Conference on Image Processing (ICIP)},
    pages={1734--1738},
    year={2021},
    organization={IEEE}
    }

Installation

• This code was developed with Python 3.8.5 & Pytorch 1.4.0 & CUDA 11.3.
• Other requirements: numpy, Pillow
• Clone this repo

git clone https://github.com/ccc870206/InverseHalftoneColorization.git
cd InverseHalftoneColorization

Testing

Download our pretrained models from here and put them under weights/.
Run the sample data provided in this repo:

python test.py

Run your own data:

python test.py --input_dir YOUR_INPUT_IMG_PATH
               --ref_dir YOUR_REFERENCE_IMG_PATH
               --target_img_path YOUR_TARGET_IMG_PATH

Training (TODO)

Acknowledgments

Our code is based on BicycleGAN and we re-implement Deep Inverse Halftoning via Progressively Residual Learning in Pytorch for inverse halftone network.