This is an implementation of CycleGAN with a Blur Kernel for Deconvolution Microscopy: Optimal Transport Geometry.
- Python 3.7
- Pytorch, torch>=0.4.1, torchvision>=0.2.1
- To run the code, please install required packages by the following command
pip install -r requirements.txt
- Generate dataset
python generate_dataset.py --phase train --num_imgs 2000
python generate_dataset.py --phase test --num_imgs 500
- Rename the dataset as "dataset".
- To generate names of all the train and test data, run the file "readDatasetNames.py"
python readDatasetNames.py
python main.py --phase train --epoch 100 --gpu 0
python main.py --phase test --gpu 0
pad input image to 4x To do