ECGTransForm: Empowering Adaptive ECG Arrhythmia Classification Framework with Bidirectional Transformer [Paper] [Cite]
Our proposed model, ECGTransForm, is a deep learning framework for ECG arrhythmia classification, featuring a novel Bidirectional Transformer mechanism and Multi-scale Convolutions for effective spatial and temporal feature extraction. The framework also includes a Context-Aware Loss to handle the class imbalance in ECG data, demonstrating superior performance in arrhythmia diagnosis.
We used two public datasets in this study (Download our preprocessed version of the datasets from Google Drive):
There are two configuration files:
- one for dataset configuration
configs/data_configs.py
- one for training configuration
configs/hparams.py
If you found this work useful for you, please consider citing it.
@ARTICLE{ecgTransForm,
title = {ECGTransForm: Empowering adaptive ECG arrhythmia classification framework with bidirectional transformer},
journal = {Biomedical Signal Processing and Control},
volume = {89},
pages = {105714},
year = {2024},
issn = {1746-8094},
doi = {https://doi.org/10.1016/j.bspc.2023.105714},
url = {https://www.sciencedirect.com/science/article/pii/S1746809423011473},
author = {Hany El-Ghaish and Emadeldeen Eldele},
}