Skip to content

Latest commit

 

History

History
 
 

train_crnn_cinc2021

Will Two Do? Varying Dimensions in Electrocardiography: The PhysioNet/Computing in Cardiology Challenge 2021

docker-ci

Digest of Top Solutions (ranked by final challenge score)

  1. ISIBrno-AIMT: Custom ResNet + MultiHeadAttention + Custom Loss
  2. DSAIL_SNU: SE-ResNet + Custom Loss (from Asymmetric Loss)
  3. NIMA: Time-Freq Domain + Custom CNN
  4. cardiochallenger: Inception-ResNet + Channel Self-Attention + Custom Loss
  5. USST_Med: SE-ResNet + Focal Loss + Data Re-labeling Model
  6. CeZIS: ResNet50 + FlowMixup
  7. SMS+1: Custom CNN + Hand-crafted Features + Asymmetric Loss
  8. DataLA_NUS: EfficientNet + SE-ResNet + Custom Loss

Other teams that are not among official entries, but among unofficial entries:

  1. HeartBeats: SE-ResNet + Sign Loss + Model Ensemble

Aizip-ECG-team and Proton had high score on the hidden test set, but did not submitted papers, hence not described here.

Conference Website and Conference Programme

Website, Programme, IEEE Xplore, Poster

Data Preparation

One can download training data from GCP, and use python prepare_dataset -i {data_directory} -v to prepare the data for training

Deep Models

Deep learning models are constructed using torch_ecg, which has already been added as a submodule.

Final Results

Final results are on the leaderboard page of the challenge official website or one can find in the offical_results folder.

Citation

@inproceedings{wen_cinc2021,
      title = {{Hybrid Arrhythmia Detection on Varying-Dimensional Electrocardiography: Combining Deep Neural Networks and Clinical Rules}},
     author = {Hao Wen and Jingsu Kang},
  booktitle = {{2021 Computing in Cardiology (CinC)}},
        doi = {10.23919/cinc53138.2021.9662801},
       year = {2021},
      month = {9},
  publisher = {{IEEE}},
}
@article{Kang_2022_cinc2021_iop,
     author = {Jingsu Kang and Hao Wen},
      title = {{A Study on Several Critical Problems on Arrhythmia Detection using Varying-Dimensional Electrocardiography}},
    journal = {Physiological Measurement},
        doi = {10.1088/1361-6579/ac6aa3},
       year = {2022},
      month = {4},
  publisher = {{IOP} Publishing}
}