RT Journal Article
T1 Human identification using compressed ECG signals
A1 Cámara Núñez, María Carmen
A1 Peris López, Pedro
A1 Estévez Tapiador, Juan Manuel
AB As a result of the increased demand for improved life styles and the increment of senior citizens over the age of 65, new home care services are demanded. Simultaneously, the medical sector is increasingly becoming the new target of cybercriminals due the potential value of users' medical information. The use of biometrics seems an effective tool as a deterrent for many of such attacks. In this paper, we propose the use of electrocardiograms (ECGs) for the identification of individuals. For instance, for a telecare service, a user could be authenticated using the information extracted from her ECG signal. The majority of ECG-based biometrics systems extract information (fiducial features) from the characteristics points of an ECG wave. In this article, we propose the use of non-fiducial features via the Hadamard Transform (HT). We show how the use of highly compressed signals (only 24 coefficients of HT) is enough to unequivocally identify individuals with a high performance (classification accuracy of 0.97 and with identification system errors in the order of 10(-2)).
PB Springer
SN 0148-5598
YR 2015
FD 2015-11-01
LK https://hdl.handle.net/10016/26174
UL https://hdl.handle.net/10016/26174
LA eng
NO This work was supported by the MINECO grant TIN2013-46469-R (SPINY: Security and Privacy in the Internet of You) and the CAM grant S2013/ICE-3095 (CIBERDINE: Cybersecurity, Data, and Risks).
DS e-Archivo
RD 9 jun. 2024