Key result
MODWPT and Ada-IBL outperform CNNs for arrhythmia detection with ~99% classification accuracy.
Why the study?
Arrhythmia is highly imperceptible yet insidious and can result in severe complications, necessitating efficient ECG classification methods for detection.
The proposed Ada-IBL method combined with MODWPT provides highly accurate and computationally efficient ECG classification for arrhythmia detection.
Proposed ECG arrhythmia classifier using MODWPT and incremental broad learning requires clinical validation; leaves open diagnostic performance versus established algorithms.
Arrhythmia, which is deffned as a problem with the rhythm or rate at which the heart operates, is highly imperceptible yet insidious and can result in severe complications. In this study, in order to develop and evaluate an efffcient electrocardiogram (ECG) classiffcation method for arrhythmia, we propose a combination of maximum overlap discrete wavelet packet transform (MODWPT) and adaptive incremental broad learning for arrhythmia detection. In the proposed method, ECG signal is ffrst denoised and segmented. Then the energy spectrum features extracted by MODWPT is combined with morphological features and rhythmic features, are used as the input of Ada-IBL. Finally, the MIT-BIH arrhythmia database based on the AAMI ECAR1987 is used to compare the results. The accuracies for ventricular and supraventricular ectopic beats are 97.2% and 99.9%, respectively, indicating that the proposed method is superior to the existing methods. By compared with the convolutional neural network, Ada-IBL provides better computational efffciency and classiffcation accuracy. It requires only 0.31 s to achieve 99.15% classiffcation accuracy on the testing set while the training time is less than 2.4 s. Ada-IBL can well avoid the limitation of selecting parameters by experience, showing the ability of equilibrating time cost and accuracy.
No takes yet. Share an insight, caveat, or question.
Li et al. (2023) studied Arrhythmia (n=101,936). MODWPT and adaptive incremental broad learning (Ada-IBL) vs. Convolutional neural network (CNN) was evaluated on Overall classification accuracy on the testing set. The combination of MODWPT and adaptive incremental broad learning achieved 99.15% overall classification accuracy for arrhythmia detection, outperforming convolutional neural networks.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: