This is the official PyTorch implementation for the DC-MT method to handle automatic knee cartilage defect assessment based on T2-weighted MRI.

Paper

Automatic Grading Assessments for Knee MRI Cartilage Defects via Self-ensembling Semi-supervised Learning with Dual-Consistency
by Jiayu Huo, Xi Ouyang, Liping Si, Kai Xuan, Sheng Wang, Weiwu Yao, Ying Liu, Jia Xu, Dahong Qian, Zhong Xue, Qian Wang, Dinggang Shen, Lichi Zhang

We propose a semi-supervised framework to effectively use unlabeled data for better evaluation of knee cartilage defect assessment. The framework contains two separate models (DC-MT model and Aggregation model). Experimental results show that our proposed method can significantly improve both classification and localization performances of knee cartilage defects.

Clone this repo.

git clone https://github.com/King-HAW/DC-MT.git
cd DC-MT/

This code requires PyTorch 1.1+ and python 3+. Please install Pytorch 1.1+ environment, and install dependencies (e.g., visdom and h5py) by

pip install -r requirements.txt

For the usage of DC-MT network and aggregation network, please go to DC-MT and Aggregation for the detailed information.

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

@article{huo2022automatic,
  title={Automatic Grading Assessments for Knee MRI Cartilage Defects via Self-ensembling Semi-supervised Learning with Dual-Consistency},
  author={Huo, Jiayu and Ouyang, Xi and Si, Liping and Xuan, Kai and Wang, Sheng and Yao, Weiwu and Liu, Ying and Xu, Jia and Qian, Dahong and Xue, Zhong and Wang, Qian and Shen, Dinggang and Zhang, Lichi},
  journal={Medical Image Analysis},
  volume={80},
  pages={102508},
  year={2022},
  issn={1361-8415},
  publisher={Elsevier}
}

• This code repository refers to HAM, MT, and RSNA Intracranial Hemorrhage Detection

• We thank Xi Ouyang, Tao Shen, and Antti Tarvainen for their awesome and efficient code bases.