Key result
ResNet outperforms other deep networks for ECG biometric identification with up to ~98% accuracy.
Why the study?
Wavelet transformation of noisy ECG signals creates morphologically complex 2-D scalograms where simple classifiers may perform poorly, prompting evaluation of deep convolutional neural networks.
Population
ECG signals from the PTB-ECG and CU-ECG databases
Comparison
AlexNet vs GoogLeNet vs ResNet deep convolutional neural network models
Design
Comparative model evaluation study
Authors
Loading...
ResNet on ECG scalograms supports biometric feasibility; leaves open prospective validation before clinical adoption.
Deep learning models, particularly ResNet, can effectively identify individuals using ECG scalograms, offering a novel approach to biometrics.
Byeon et al. (2019) studied Biometric identification (n=390). Deep convolutional neural networks (ResNet, AlexNet, GoogLeNet) using ECG scalograms vs. Simple CNN and comparison among deep models was evaluated on Classification accuracy. ResNet achieved the highest classification accuracy of 98.10% on PTB-ECG and 93.20% on CU-ECG, outperforming AlexNet and GoogLeNet for ECG-based biometric identification.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: