Tenserflow implementation for our conference paper "Unpaired Brain MR-to-CT Synthesis Using a Structure-Constrained CycleGAN".

Input: unpaired CT and MR volumes.
Output: the synthetic MR/CT volume corresponding to input CT/MR volume.

1. There are many settings in main.py that needs to be changed according to your applications. Besides, it contains three modes, respectively train, validation and test.
   
2. The experiments are performed on unpaired MR and CT data, including the training, validation and testing subjects.

If you use this code for your research, please cite our paper:

@inproceedings{yang2018,
title={Unpaired brain MR-to-CT synthesis using a structure-constrained CycleGAN},
author={Yang, Heran and Sun, Jian and Carass, Aaron and Zhao, Can and Lee, Junghoon and Xu, Zongben and Prince, Jerry},
booktitle={Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support},
pages={174--182},
year={2018}
}

The tensorflow implementation of CycleGAN: https://github.com/XHUJOY/CycleGAN-tensorflow