Key result
A novel algorithm using continuous wavelet transform and 2D convolutional neural networks achieved an overall accuracy of 99.23% for detecting atrial fibrillation using just five heartbeats.
Why the study?
Does an automatic detection algorithm based on continuous wavelet transform and 2D convolutional neural networks accurately detect atrial fibrillation episodes in ECG recordings?
Population
21 ECG recordings from the MIT-BIH Atrial Fibrillation Database, segmented into 100,000 five-beat samples.
Comparison
Automatic AF detection algorithm based on… vs Existing AF detection algorithms evaluated on…
Design
Other
Authors
Loading...
May enable rapid AF screening with limited ECG data; leaves open prospective validation before clinical adoption.
Does an automatic detection algorithm based on continuous wavelet transform and 2D convolutional neural networks accurately detect atrial fibrillation episodes in ECG recordings?
A novel algorithm combining continuous wavelet transform and 2D convolutional neural networks demonstrated >99% accuracy in detecting atrial fibrillation from short 5-beat ECG segments.
He et al. (2018) studied Atrial Fibrillation (n=21). Continuous Wavelet Transform and 2D Convolutional Neural Networks vs. Existing AF detection algorithms was evaluated on Overall accuracy (ACC) of AF detection. A novel algorithm using continuous wavelet transform and 2D convolutional neural networks achieved an overall accuracy of 99.23% for detecting atrial fibrillation using just five heartbeats.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: