Key result
Deep transfer learning for ECG classification reveals data split methods significantly alter model performance.
Why the study?
Available well-annotated public ECG datasets are often very small, and many existing deep learning models fail to avoid data leakage against AAMI recommendations.
Does a deep transfer learning framework using ResNet-18 improve ECG heartbeat classification on small datasets while adhering to AAMI EC57 standards?
Does a deep transfer learning framework using ResNet-18 improve ECG heartbeat classification on small datasets while adhering to AAMI EC57 standards?
Deep transfer learning with ResNet-18 can be used for ECG heartbeat classification, but adherence to AAMI EC57 standards is crucial to avoid data leakage in model evaluation.
Answers are grounded in this paper and cite their sources.
Proper data splitting is essential to validate ECG ML models and avoid leakage; leaves open clinical utility of ResNet-18 transfer learning.
Electrocardiogram (ECG) is a simple non-invasive measure to identify heart-related issues such as irregular heartbeats known as arrhythmias. While artificial intelligence and machine learning is being utilized in a wide range of healthcare related applications and datasets, many arrhythmia classifiers using deep learning methods have been proposed in recent years. However, sizes of the available datasets from which to build and assess machine learning models is often very small and the lack of well-annotated public ECG datasets is evident. In this paper, we propose a deep transfer learning framework that is aimed to perform classification on a small size training dataset. The proposed method is to fine-tune a general-purpose image classifier ResNet-18 with MIT-BIH arrhythmia dataset in accordance with the AAMI EC57 standard. This paper further investigates many existing deep learning models that have failed to avoid data leakage against AAMI recommendations. We compare how different data split methods impact the model performance. This comparison study implies that future work in arrhythmia classification should follow the AAMI EC57 standard when using any including MIT-BIH arrhythmia dataset.
No takes yet. Share an insight, caveat, or question.
Cao et al. (2023) studied Arrhythmia. Deep transfer learning framework (ResNet-18 with STFT) vs. Existing deep learning models was evaluated on Model performance and data leakage avoidance. A deep transfer learning framework using ResNet-18 was proposed for ECG heartbeat classification, demonstrating that different data split methods significantly impact model performance.
Synapse has enriched 10 related papers. Consider them for comparative context: