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.
- 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.
- 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