Why the study?
Noise contamination introduces undesired intra-variability in ECG signals, creating the need for more robust biometric systems.
Deep learning applied to ECG spectrograms can achieve high within-session biometric identification accuracy, though across-session reliability requires further optimization.
Supports spectrogram-CNN exploration for robust ECG biometrics; leaves open validation in larger, diverse cohorts.
The Electrocardiogram (ECG) is considered as a physiological signature and has previously been used for biometric purposes. The contamination of the signal due to noise adds undesired intra-variability in the ECG signals, creating the need for more robust biometric systems (BSs). With the increase of interest in the application of Deep Neural Networks (DNN) to the medical field, new solutions are also being explored in the identification and authentication of individuals. The proposed architecture exploits the potential of Convolutional Neural Networks (CNN) to identify healthy subjects using temporal frequency analysis, i.e. spectrograms.
No takes yet. Share an insight, caveat, or question.
Bento et al. (2020) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: