CLINet-ECG-Classification-2024

Source code of "CLINet: A Novel Deep Learning Network for ECG Signal Classification", accepted in Journal of Electrocardiology 2024

If you are using this code, please cite our paper: ```bash @article {ref199,

title = "CLINet: A Novel Deep Learning Network for ECG Signal Classification",

year = "2024",

author = "Ananya Mantravadi and Siddharth Saini and R Sai Chandra Teja and Sparsh Mittal and Shrimay Shah and R Sri Devi and Rekha Singhal",

journal = "Journal of Electrocardiology", } ```

Project Organization

├── LICENSE                         <- The LICENSE for developers using this project.
├── README.md                       <- The top-level README for developers using this project.
├── data                            <- Data used in the project.
│   ├── iccad                       <- Add ICCAD dataset with this path in the folder.
│   │   ├── tinyml_contest_data_training
│   |   │   ├──S01-AFb-1.txt
│   |   │   ├──S01-AFb-10.txt
│   |   │   ├──...
│   │   ├── data-indices
│   |   │   ├── train-indice    
│   |   │   ├── test-indice    
│   ├── mit-bih                     <- Add MIT-BIH dataset with this path in the folder.
│   │   ├── mitbih_database
│   |   │   ├──100.csv
│   |   │   ├──100annotations.txt
│   |   │   ├──...
├── src                             <- Source code for use in this project.
│   ├── iccad_dataloader.py         <- Source code for generating data loader for ICCAD dataset.
|   ├── mitbih_dataloader.py        <- Source code for generating data loader for MIT-BIH dataset.
│   ├── network.py                  <- Source code for the CLINet network.
│   ├── involution.py               <- Source code for definition of custom involution layer.
│   ├── tsne.py                     <- Source code for plotting t-SNE.
│   ├── main.py                     <- Source code for using CLINet on ICCAD and MIT-BIH
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────

Train model

To train CLINet, Run following command from /src directory.

bash python main.py Above command will train model for 50 epochs with given configuration.

License

MIT License Copyright (c) 2024 CandleLabAI