Federated Cardiac Digital Twins: A Privacy-Preserving Continuous ECG Biometric Authentication Framework
Journal of Intelligent Decision Making and Information Science · 02.10.2026 · 3 (15s) · 71-90
DOI: 10.17762/jidmis.v3.4474 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 02.10.2026. Verfasst von Mohamed Hammad. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Electrocardiogram (ECG) signals offer a promising basis for continuous biometric authentication because they reflect individual cardiac physiology. However, practical deployment faces three main obstacles: template aging (slow drift of the enrolled signature), cross-device sensitivity (performance loss when sensors or lead placements change), and waveform level spoofing (replay or synthetic ECG attacks). This paper presents Federated Cardiac Digital Twins (FedCDT), a framework that redefines authentication by asking whether a heartbeat is a natural continuation of the individual’s cardiac dynamics rather than matching a fixed pattern. The method trains a shared convolutional encoder and a per client prototype on local devices using federated prototypical learning. Only compact 64 dimensional latent vectors and model weights are exchanged; raw ECG signals never leave the user’s device. The per client prototype acts as a cardiac digital twin that evolves with the user. During authentication, cosine similarity between a new ECG segment and the stored prototype produces a continuous, drift adaptive score. Experiments on the ECG ID and MIT BIH databases achieve an equal error rate of 0.1125 and an area under the ROC curve of 0.9613 while preserving data locality. The framework simultaneously suppresses template drift, generalizes across heterogeneous datasets, and raises resistance to spoofing because an attacker must reproduce the subject’s dynamical cardiac identity instead of a single waveform. These results demonstrate that FedCDT is the first complete integration of digital twins, federated learning, and ECG biometrics for real time, privacy compliant continuous authentication.
Zitieren
APA
Hammad, M. (2026). Federated Cardiac Digital Twins: A Privacy-Preserving Continuous ECG Biometric Authentication Framework. Journal of Intelligent Decision Making and Information Science, 3(15s), 71-90. https://doi.org/10.17762/jidmis.v3.4474
Vancouver
Hammad M. Federated Cardiac Digital Twins: A Privacy-Preserving Continuous ECG Biometric Authentication Framework. Journal of Intelligent Decision Making and Information Science. 2026;3(15s):71-90. doi:10.17762/jidmis.v3.4474
BibTeX
@article{Hammad2026,
author = {Hammad, Mohamed},
title = {Federated Cardiac Digital Twins: A Privacy-Preserving Continuous ECG Biometric Authentication Framework},
journal = {Journal of Intelligent Decision Making and Information Science},
year = {2026},
volume = {3},
number = {15s},
pages = {71--90},
doi = {10.17762/jidmis.v3.4474}
}
EUVIMED