Key result
The attention-based time-incremental convolutional neural network (ATI-CNN) achieved an overall classification accuracy of 81.2%, an average increase of 7.7% compared to a classical 16-layer CNN.
Why the study?
Convolutional neural networks lack consideration for temporal features of ECG signals, cannot accept varied-length signals, and have limited performance in detecting paroxysmal arrhythmias.
Does ATI-CNN improve classification accuracy for multi-class arrhythmia detection from varied-length ECGs compared to classical CNN?
Population
12-lead varied-length ECG signals
Comparison
Attention-based time-incremental CNN (ATI-CNN) vs classical 16-layer CNN (VGGNet)
Authors
Loading...
May enhance ECG arrhythmia classification accuracy; hypothesis-generating and should not yet change practice.
Does ATI-CNN improve classification accuracy for multi-class arrhythmia detection from varied-length ECGs compared to classical CNN?
ATI-CNN provides a computationally efficient and more accurate deep learning approach for detecting arrhythmias from varied-length ECGs compared to traditional CNN models.
Yao et al. (2019) studied Multi-class Arrhythmia. Attention-based time-incremental convolutional neural network (ATI-CNN) vs. Classical 16-layer CNN (VGGNet) was evaluated on Overall classification accuracy. The attention-based time-incremental convolutional neural network (ATI-CNN) achieved an overall classification accuracy of 81.2%, an average increase of 7.7% compared to a classical 16-layer CNN.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: