Key result
An end-to-end 12-lead ECG diagnosis system based on a deformable convolutional neural network achieved an overall diagnostic accuracy of 86.3% with good antinoise ability.
Why the study?
Existing CNN methods only consider local features and struggle with complex, noise-susceptible ECG signals.
Does a Deformable Convolutional Neural Network (Deform-CNN) improve diagnostic accuracy and antinoise ability for 12-lead ECGs compared to other deep learning algorithms?
Population
12-lead ECG data of CPSC-2018
Comparison
Deformable CNN vs other deep learning algorithms
Authors
Loading...
May support technical ECG algorithm development; leaves open prospective clinical validation before any practice change.
Does a Deformable Convolutional Neural Network (Deform-CNN) improve diagnostic accuracy and antinoise ability for 12-lead ECGs compared to other deep learning algorithms?
A novel Deformable CNN architecture demonstrates high diagnostic accuracy (86.3%) and good antinoise ability for automated 12-lead ECG interpretation.
Qin et al. (2021) studied ECG diagnosis. Deformable Convolutional Neural Network (Deform-CNN) vs. Other deep learning algorithms was evaluated on Overall diagnostic accuracy. An end-to-end 12-lead ECG diagnosis system based on a deformable convolutional neural network achieved an overall diagnostic accuracy of 86.3% with good antinoise ability.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: