ECG-TFDS

Load common open source ECG databases as a Tensorflow dataset. The final format should be ECG signals of one second in length, representing a heartbeat. The R-peaks form the center of the signal, which is normalized in the range $[0, 1]$.

How to get started?

1. Install the requirements pip install -r requirements
2. Navigate into the desired dataset folder cd src/dataset/
3. Build the dataset tfds build

Available Datasets

1. Shaoxing (zheng): https://www.nature.com/articles/s41597-020-0386-x
2. PTB-XL: https://www.nature.com/articles/s41597-020-0495-6
3. Icentia11k: https://physionet.org/content/icentia11k-continuous-ecg/1.0/
4. MedalCare-XL: https://www.nature.com/articles/s41597-023-02416-4
5. ECGSYN: https://physionet.org/content/ecgsyn/1.0.0/
6. Custom: Add a custom dataset

How to Add a New Dataset?

To incorporate a new dataset into the collection, follow these steps:

1. Choose an open-source ECG database, e.g., PhysioNet. Ensure that you exclusively select datasets with an appropriate license.
2. Go to the src folder and execute the following command:

   tfds new DATASET_NAME

3. Open and edit the DATASET_NAME/DATASET_NAME_dataset_builder.py file following the provided instructions. Any generic preprocessing steps should be placed in the utils folder.
4. Include tests in the DATASET_NAME/DATASET_NAME_dataset_builder_test.py file.
5. Include metadata files: CITATIONS.bib, README.md, and TAGS.txt.
6. Confirm successful dataset building using the command:

   tfds build

7. During your initial complete build, register the checksum with:

   tfds build --register_checksums

8. Navigate to ./electrocardiogram and append the dataset to the collection in electrocardiogram.py.
9. Modify the requirements.txt and this README file accordingly
10. Create a pull request and provide a concise motivation, description, and dataset metadata, including details like count, size, dataset license, and source.

Ensure adherence to these steps to seamlessly integrate the new dataset into the collection.

How to cite?

If you are using this repository please reference the following article: (to be determined)

@article{kapsecker2025disentangled,
  title={Disentangled representational learning for anomaly detection in single-lead electrocardiogram signals using variational autoencoder},
  author={Kapsecker, Maximilian and Möller, Matthias C and Jonas, Stephan M},
  journal={Computers in Biology and Medicine},
  volume={184},
  pages = {109422},
  year = {2025},
  issn = {0010-4825},
  doi = {https://doi.org/10.1016/j.compbiomed.2024.109422},
  url = {https://www.sciencedirect.com/science/article/pii/S0010482524015075},
}