Key result
LOWESS preprocessing and self-adaptive learning improve deep learning model accuracy for ECG classification.
Why the study?
Deep learning methods have been proposed for automatic ECG diagnosis to provide a second opinion and detect pathologies early, but methods are needed to improve the quality of prediction.
Do non-architectural changes like signal preprocessing and self-adaptive learning improve the performance of deep neural networks for ECG classification?
Do non-architectural changes like signal preprocessing and self-adaptive learning improve the performance of deep neural networks for ECG classification?
Data preprocessing techniques and self-adaptive learning can significantly enhance the accuracy of automated ECG interpretation models without altering the underlying neural network architecture.
May enhance automated ECG interpretation; leaves open prospective clinical validation before practice change.
Recording and analyzing 12-lead electrocardiograms is the most common procedure for detecting heart disease. Recently, various deep learning methods have been proposed for the automatic diagnosis by an electrocardiogram. The proposed methods can provide a second opinion for the doctor and help detect pathologies at an early stage. Various methods are proposed in the paper to improve the quality of prediction of ECG recording pathologies. Techniques include adding patient metadata, ECG noise reduction, and self-adaptive learning. The significance of data parameters in training a classification model is also explored. Among the considered parameters, the influence of various ECG leads, the length of the electrocardiogram and the volume of the training sample is studied. The experiments carried out show the relevance of the described approaches and offer an optimal estimate of the input data parameters.
No takes yet. Share an insight, caveat, or question.
Ananev et al. (2021) studied ECG classification (n=74,931). Non-architectural model changes (signal preprocessing, self-adaptive learning) vs. Baseline model without preprocessing was evaluated on Quality of ECG classification (ROC-AUC, Average Precision, F1-score). Applying signal preprocessing (LOWESS) and self-adaptive learning improved the diagnostic performance of a deep learning model for ECG classification.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: