**5. ECG-based biometric identification structure**

Biometric identification involves the stage of user registration and the stage of user recognition. At the registration stage, the user takes biometric features and writes them to the database. At the recognition stage, biometric features are taken from an unknown person and consequently compared with the features stored in the database. If the features received from an unknown person by a certain criterion coincided with the features from the database, then a decision is made on the success of the identification. Biometric identification is a complex multi-stage process in which each stage can affect the final recognition accuracy.

While performing the project, we investigated the influence of various factors on the accuracy of biometric identification using electrocardiograms. To do this, a large-scale computational experiment was carried out using our programs written in Python. We used the popular libraries such sklearn, scipy, and matplotlib. Most digitalized electrocardiogram samples were taken from www.physionet.org website. When performing the digital signal processing, we used the biosppy and wfdb libraries. When classifying electrocardiograms, we used Multilayer Perceptron and Convolutional Neural Networks using TensorFlow technology.

The following main stages of biometric identification are follows: signal registration, signal preprocessing, biometric feature extraction, assessment of the informativeness of biometric features and selection of the most informative features (this is done to reduce dimensionality of input data), and classification of features. Consider each of the steps.

#### **6. Registration of an electrocardiogram**

Electrocardiogram (ECG or EKG [a]) is a graph of voltage versus time – of the electrical activity of the heart using electrodes placed on the skin. These electrodes detect the small electrical changes that are a consequence of cardiac muscle depolarization followed by repolarization during each cardiac cycle (heartbeat) (**Figure 3**).

From a technical point of view, Electrocardiographs are multichannel voltmeters that record electrical potentials in various areas of human surface. These devices differ in such characteristics as sampling frequency, bit depth, input voltage range, etc. A valuable resource for researchers in the field of analysis of biomedical signals is the website https://www.physionet.org/. PhysioNet is a repository of freely available medical research data, managed by the MIT Laboratory for Computational Physiology. The project is supported by the National Institute of General Medical Sciences (NIGMS) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB) under NIH grant number 2R01GM104987-09.

**25**

*Biometric Authentication Based on Electrocardiogram DOI: http://dx.doi.org/10.5772/intechopen.91172*

PTB Diagnostic ECG Database [10]

• Input resistance: 100 Ω (DC),

European ST-T Database [11]

20 millivolt input range.

ECG-ID Database [12, 13]

• ECG lead I, recorded for 20 s,

• 16 input channels,

**Figure 3.**

• Input voltage: ±16 mV,

tion is 3 minutes).

This site presents a large number of digitized electrocardiograms, for example,

• Bandwidth: 0–1 kHz (synchronous sampling of all channels, time of registra-

• each sampled at 250 samples per second with 12-bit resolution over a nominal

• digitized at 500 Hz with 12-bit resolution over a nominal ±10 mV range

As we can see, all the above databases used different electrocardiographs.

• Resolution: 16 bit with 0.5 μV/LSB (2000 A/D units per mV),

*ECG of a heart in normal sinus rhythm (https://en.wikipedia.org/wiki/Electrocardiography).*

• each record is two hours in duration and contains two signals,

*Biometric Authentication Based on Electrocardiogram DOI: http://dx.doi.org/10.5772/intechopen.91172*

*Biometric Systems*

but also speech.

signs [8, 9].

accuracy.

Consider each of the steps.

(heartbeat) (**Figure 3**).

number 2R01GM104987-09.

**6. Registration of an electrocardiogram**

Several patents have been published that propose the use of ultra-wideband radars for conducting banking operations without a credit card, as well as for controlling premises during confidential meetings (**Figure 2**). As it turned out, with the help of ultra-wideband radars, it is possible to restore not only the ECG,

Xiaolin et al. described using of ultra wide band radars for detecting of vital

Biometric identification involves the stage of user registration and the stage of user recognition. At the registration stage, the user takes biometric features and writes them to the database. At the recognition stage, biometric features are taken from an unknown person and consequently compared with the features stored in the database. If the features received from an unknown person by a certain criterion coincided with the features from the database, then a decision is made on the success of the identification. Biometric identification is a complex multi-stage process in which each stage can affect the final recognition

While performing the project, we investigated the influence of various factors on the accuracy of biometric identification using electrocardiograms. To do this, a large-scale computational experiment was carried out using our programs written in Python. We used the popular libraries such sklearn, scipy, and matplotlib. Most digitalized electrocardiogram samples were taken from www.physionet.org website. When performing the digital signal processing, we used the biosppy and wfdb libraries. When classifying electrocardiograms, we used Multilayer Perceptron and

The following main stages of biometric identification are follows: signal registration, signal preprocessing, biometric feature extraction, assessment of the informativeness of biometric features and selection of the most informative features (this is done to reduce dimensionality of input data), and classification of features.

Electrocardiogram (ECG or EKG [a]) is a graph of voltage versus time – of the electrical activity of the heart using electrodes placed on the skin. These electrodes detect the small electrical changes that are a consequence of cardiac muscle depolarization followed by repolarization during each cardiac cycle

From a technical point of view, Electrocardiographs are multichannel voltmeters that record electrical potentials in various areas of human surface. These devices differ in such characteristics as sampling frequency, bit depth, input voltage range, etc. A valuable resource for researchers in the field of analysis of biomedical signals is the website https://www.physionet.org/. PhysioNet is a repository of freely available medical research data, managed by the MIT Laboratory for Computational Physiology. The project is supported by the National Institute of General Medical Sciences (NIGMS) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB) under NIH grant

**5. ECG-based biometric identification structure**

Convolutional Neural Networks using TensorFlow technology.

**24**

This site presents a large number of digitized electrocardiograms, for example, PTB Diagnostic ECG Database [10]


European ST-T Database [11]


ECG-ID Database [12, 13]


As we can see, all the above databases used different electrocardiographs.

Previously, we investigated factors that influence the accuracy of biometric identification using an ECG. We have shown that the quality of an electrocardiograph affects the accuracy of biometric identification. Thus, the recognition accuracy during ECG classification using mixed Gaussian models of subjects from the ECG-ID database was 0.66, while for PTB this indicator was 0.8 [14].
