CAC-UNet-DigestPath2019

Implementation detail for our paper "Multi-level colonoscopy malignant tissue detection with adversarial CAC-UNet"

DigestPath 2019

The proposed scheme in this paper achieves the best results in MICCAI DigestPath2019 challenge (https://digestpath2019.grand-challenge.org/Home/) on colonoscopy tissue segmentation and classification task.

Dataset

Description of dataset can be found here: https://digestpath2019.grand-challenge.org/Dataset/

To download the the DigestPath2019 dataset, please sign the DATABASE USE AGREEMENT first at here: https://digestpath2019.grand-challenge.org/Download/

If you have problems about downing the dataset, please contact Prof. Hongsheng Li:[email protected]

Image sample:

sample-image

Envs

  • Pytorch 1.0
  • Python 3+
  • cuda 9.0+

install

$ pip install -r  requirements.txt

apex : Tools for easy mixed precision and distributed training in Pytorch

$ git clone https://github.com/NVIDIA/apex
$ cd apex
$ pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./

Dataset

├── data/
│   ├── tissue-train-neg/     
│   ├── tissue-train-pos-v1/

Preprocessing

$ cd code/
$ python preprocessing.py

Training

$ cd code/
$ python train.py --config_file='config/cac-unet-r50.yaml'

Citation

Please cite this paper in your publications if it helps your research:

@article{zhu2021multi,
  title={Multi-level colonoscopy malignant tissue detection with adversarial CAC-UNet},
  author={Zhu, Chuang and Mei, Ke and Peng, Ting and Luo, Yihao and Liu, Jun and Wang, Ying and Jin, Mulan},
  journal={Neurocomputing},
  volume={438},
  pages={165--183},
  year={2021},
  publisher={Elsevier}
}

About the multi-level adversarial segmentation part, you can read our ICASSP paper for more details:

@inproceedings{mei2020cross,
  title={Cross-stained segmentation from renal biopsy images using multi-level adversarial learning},
  author={Mei, Ke and Zhu, Chuang and Jiang, Lei and Liu, Jun and Qiao, Yuanyuan},
  booktitle={ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  pages={1424--1428},
  year={2020},
  organization={IEEE}
}

Author

Ke Mei, Ting Peng, Chuang Zhu

If you have any questions, you can contact me directly.

GitHub

https://github.com/bupt-ai-cz/CAC-UNet-DigestPath2019