Electrocardiogram signals-based user authentication systems using soft computing techniques
- PDF / 4,500,268 Bytes
- 43 Pages / 439.37 x 666.142 pts Page_size
- 38 Downloads / 292 Views
Electrocardiogram signals‑based user authentication systems using soft computing techniques Mehdi Hosseinzadeh1,2 · Bay Vo3 · Marwan Yassin Ghafour5 · Sajjad Naghipour4
© Springer Nature B.V. 2020
Abstract With the advent of various security attacks, biometric authentication methods are gaining momentum in the security literature. Electrocardiogram or ECG signals are one of the essential biometric features generated by the human heart’s electrical activities. Many authentication schemes apply these signals due to their uniqueness, resistance to fabrication attacks, and support for continuous authentication. This survey article focuses on the ECG-based authentication approaches and provides the required background knowledge about the ECG signals and authentication methods. Then, it presents a taxonomy of the ECG-based authentication approaches first based on the authentication factors and then according to the applied algorithms for conducting authentication. It then describes their key contributions, applied algorithms, and possible drawbacks. Furthermore, their employed evaluation factors, ECG datasets, and simulators are illuminated and compared. Finally, the concluding remarks and future studies directions in this context are provided. Keywords ECG · Authentication · Security · Feature selection · SVM · CNN · Deep learning
1 Introduction Over the past years, the increasing demand for more security against malicious behaviors and intrusions have inspired researchers to provide various security solutions (Karimian et al. 2017). To this end, many authentication schemes are introduced using different cryptographic (Li et al. 2018; Farash et al. 2017; He et al. 2016; Chaudhry et al. 2015; Sahoo and Mohanty 2018; Liao et al. 2017, 2019) and biometric methods (Kurogi et al. 2018) to mitigate vulnerabilities of the conventional authentication methods (Masdari et al. 2017). Biometric authentication is an interesting identification method, in which a person’s biological or behavioral traits should be compared with its templates recorded in the enrollment
* Bay Vo [email protected]
Sajjad Naghipour [email protected] Extended author information available on the last page of the article
13
Vol.:(0123456789)
M. Hosseinzadeh et al.
phase. Different biometric features such as fingerprint, palm, iris, electroencephalogram, and electrocardiogram (ECG) (Zeng et al. 2019; Nguyen et al. 2017) are employed for biometric authentication, which the latter one is focused on by this paper. In general, the electrical operations of the heart create the ECG signals, which can be recorded by some electrodes attached to particular locations of the body (Wu et al. 2018). Specialist cardiologists will monitor the patient’s ECG tape to diagnosing various heart diseases (Sahoo and Mohanty 2018). But, with the current progress in the VLSI context, it is possible to have ECG of subjects using their own smart mobile devices, earphones, smartwatches, implantable medical devices, and smart clothing. The availability of these s
Data Loading...
