This repository includes the official project of TFCNs, presented in our paper: TFCNs: A CNN-Transformer Hybrid Network for Medical Image Segmentation , which is accepted by ICANN 2022 (International Conference on Artificial Neural Networks).

paper link: https://arxiv.org/abs/2207.03450 or https://doi.org/10.1007/978-3-031-15937-4_65

Email: [email protected]

Please contact dihan or me if you need the further help.

model/ : save for the model you have train

networks/ : all the component that construct our TFCNs

preprocess.py : simple data augumentation

train_utils.py : some tools used for training

utils.py : some tools used for testing

you can run the train.py and test.py for training and testing.

Please prepare an environment with python=3.7, and then use the command "pip install -r requirements.txt" for the dependencies.

@inproceedings{li2022tfcns,
  title={TFCNs: A CNN-Transformer Hybrid Network for Medical Image Segmentation},
  author={Li, Zihan and Li, Dihan and Xu, Cangbai and Wang, Weice and Hong, Qingqi and Li, Qingde and Tian, Jie},
  booktitle={International Conference on Artificial Neural Networks},
  pages={781--792},
  year={2022},
  organization={Springer}
}