KR102276606B1 - Method for personal authentication using electrocardiogram - Google Patents

Abstract

There is provided a personal authentication method using an electrocardiogram for performing personal authentication through a comparison between a registered signal and a newly acquired signal after registering an electrocardiogram signal. A personal authentication method using an electrocardiogram according to an embodiment of the present invention includes collecting an ECG (electrocardiogram) signal; extracting the characteristics of the ECG signal using Auto Correlation Profile (ACP) and Linear Discriminant Analysis (LDA) algorithms; and comparing the characteristics of the extracted ECG signal with the pre-stored ECG template data to perform personal authentication. Accordingly, after registering the ECG (electrocardiogram) signal, personal authentication is performed through the comparison result between the registered signal and the newly acquired ECG signal, and during registration, the change width of the ECG signal is stored as well as whenever authentication is successful. Changing ECG information can be acquired by performing template updating. In addition, by performing personal authentication using the ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithms of the ECG signal, the overall characteristics of the electrocardiogram can be reflected, the amount of computation can be reduced, the measurement site in the real environment, Considering the impedance change between the skin and the sensor due to various factors such as measurement area and skin humidity level, it is possible to prevent a decrease in accuracy caused by changes in ECG signal due to changes in skin-to-electrode impedance during actual authentication. can

Description

{Method for personal authentication using electrocardiogram}

The present invention relates to a personal authentication method, and more particularly, to a personal authentication method using an electrocardiogram for performing personal authentication through a comparison between a registered signal and a newly acquired signal after registering an electrocardiogram signal.

Although the ECG signal has individual characteristics, it is constantly changed by the sympathetic/parasympathetic nerve, so it has the advantage of constantly changing passwords, unlike other biometric authentication methods (fingerprints, iris, etc.).

Conventionally, personal authentication was performed using a fiducial dependent authentication method and a fiducial independent authentication method. The fiducial dependent authentication method is based on the local characteristics of the electrocardiogram ( Amplitude of P, Q, R, T wave and Interval of PQ, QRS, QT, RT). , QT etc), it does not reflect changes in heart rate and compares only the duration and amplitude between local feature points, so it is greatly affected by noise, and there is a risk of missing the overall characteristics of the electrocardiogram, which is always changing. There is a big problem with this.

In addition, the fiducial independent authentication method is an authentication method that uses the characteristics of the overall model by analyzing the waveform of the ECG in the frequency domain. Since comparison with all registered persons is required, there is a disadvantage that a large amount of computation is required.

In addition, since the change in impedance between the skin and the sensor occurring in the actual use environment is not considered, there is a disadvantage in that the accuracy is greatly reduced when performing the actual authentication.

The present invention has been devised to solve the above problems, and an object of the present invention is to perform personal authentication through a comparison result between a registered signal and a newly acquired ECG signal after registering an ECG (electrocardiogram) signal, and registration It is intended to provide a personal authentication method using an electrocardiogram that can acquire changing ECG information by not only storing the change width of the ECG signal, but also performing template updating whenever authentication is successful.

Another object of the present invention is to perform personal authentication using the ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithms of the ECG signal to reflect the overall characteristics of the ECG, reduce the amount of computation, and Considering the impedance change between the skin and the sensor due to various factors such as the measurement area, measurement area, and skin humidity level in the environment, the accuracy caused by the change in the ECG signal due to the change in the skin-to-electrode impedance during actual authentication is determined. An object of the present invention is to provide a personal authentication method using an electrocardiogram capable of preventing a falling phenomenon.

According to an embodiment of the present invention for achieving the above object, a personal authentication method using an electrocardiogram includes: collecting an ECG (electrocardiogram) signal; extracting the characteristics of the ECG signal using Auto Correlation Profile (ACP) and Linear Discriminant Analysis (LDA) algorithms; and comparing the characteristics of the extracted ECG signal with the pre-stored ECG template data to perform personal authentication.

In addition, the method for personal authentication using an electrocardiogram according to an embodiment of the present invention may further include compensating for the collected ECG signal in consideration of the impedance between the skin and the ECG sensor.

In addition, the personal authentication method using an electrocardiogram according to an embodiment of the present invention further includes a pre-processing step of processing the ECG signal that has undergone a compensation step, wherein the pre-processing step is an amplification process for amplifying, passing only a specific frequency band It may include a band-pass filtering process that removes a specific frequency band and a notch filtering process that removes only a specific frequency band.

And in the extraction step, in order to minimize the effect of physiological factors, the ACP is divided into a sliding process of moving the ECG signal that has undergone the pre-processing step by the QRS interval and the ECG signal that has undergone the sliding process in a preset section unit, Using the LDA algorithm, it can be composed of an extraction process of extracting the amplitude (amplitude) for each section of each ACP and the projection angle for each section as features.

In addition, the personal authentication method using an electrocardiogram according to an embodiment of the present invention includes generating ECG template data based on the characteristics of an ECG signal at a user's request, and registering it as user authentication information used for personal authentication. ; may be further included.

In the registration step, a threshold value of each data is set using the average value and standard deviation value of each of the amplitude data and the projection angle data, and the standard deviation of each section of the amplitude data and the projection angle data is set. By comparing the values, weights are given to each data, and ECG template data including information on the threshold value of each data and the assigned weight can be registered as user authentication information.

In addition, in the authentication step, the matching score between the registered ECG signal and the ECG signal to be authenticated may be calculated by the following equation.

(Equation) Concordance = (amplitude weight)*(amplitude value)+(angle weight)*(angle value)

On the other hand, according to another embodiment of the present invention, a personal authentication system using an electrocardiogram, a signal measuring unit for collecting an ECG (electrocardiogram) signal; And ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithm to extract the characteristics of the ECG signal, and compare the extracted ECG signal characteristics with the pre-stored ECG template data, a processor for performing personal authentication; include

As described above, according to the embodiments of the present invention, after registering the ECG (electrocardiogram) signal, personal authentication is performed through the comparison result between the registered signal and the newly acquired ECG signal, and the change width of the ECG signal during registration In addition to saving, changing ECG information can be obtained by performing Template updating whenever authentication success is performed.

In addition, according to the embodiments of the present invention, personal authentication is performed using the Auto Correlation Profile (ACP) and Linear Discriminant Analysis (LDA) algorithms of the ECG signal to reflect the overall characteristics of the electrocardiogram and reduce the amount of computation. In the real environment, considering the impedance change between the skin and the sensor due to various factors such as the measurement site, measurement area, and skin humidity level, when performing the actual authentication, the change in the ECG signal due to the change in the skin to electrode impedance It is possible to prevent the occurrence of a decrease in accuracy.

1 is a diagram illustrating a basic model of an electrocardiogram;
2 is a view provided for explaining the configuration of a personal authentication system using an electrocardiogram according to an embodiment of the present invention;
3 is a view provided for explaining a process of personal authentication using a personal authentication system using an electrocardiogram according to an embodiment of the present invention;
4 to 5 are views illustrating a pre-processing process according to an embodiment of the present invention;
6 to 7 are diagrams provided for explanation of an ECG signal feature extraction process according to an embodiment of the present invention;
8 is a view provided for a description of a section unit according to an embodiment of the present invention;
9 is a view provided for explanation of biometric authentication efficiency according to an embodiment of the present invention, and
10 is a flowchart provided to explain a method for personal authentication using an electrocardiogram according to an embodiment of the present invention.

Hereinafter, the present invention will be described in more detail with reference to the drawings.

1 is a diagram illustrating a basic model of an electrocardiogram, FIG. 2 is a diagram provided to explain the configuration of a personal authentication system using an electrocardiogram according to an embodiment of the present invention, and FIG. 3 is an electrocardiogram according to an embodiment of the present invention It is a diagram provided to explain the process of personal authentication using a personal authentication system using

Referring to Fig. 1, the ECG signal is a recording of the electrical activity of the heart, and in the case of a single-lead ECG sensor, the basic ECG signal cycle is shown in Fig. 1. It may reflect the electrical activity of the heart whose signaling cycle consists of P waves and QRS complexes and T waves.

The personal authentication system using the electrocardiogram according to the present embodiment extracts the characteristics of the ECG signal from the QRS signal that is least affected by the physiological factors among the ECG (electrocardiogram) signals, and after registration, the registered signal and the newly acquired ECG signal Personal authentication can be performed through the comparison result.

In addition, the personal authentication system using the electrocardiogram according to this embodiment not only stores the change width of the ECG signal at the time of registration, but also performs template updating whenever authentication success is performed, thereby obtaining changing ECG information.

In addition, the personal authentication system using the electrocardiogram according to the present embodiment performs personal authentication using the ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithm of the ECG signal, thereby reflecting the overall characteristics of the electrocardiogram. , the amount of computation can be reduced, and the change in impedance between the skin and the sensor due to various factors such as the measurement area, measurement area, and skin humidity in the real environment is taken into account when performing the actual authentication, the ECG signal due to the change in skin to electrode impedance It is possible to prevent the deterioration of accuracy caused by the change of .

To this end, the personal authentication system using the electrocardiogram may include a signal measuring unit 110 , a signal processing unit 120 , a template generating unit 130 , a processor 140 , and a database unit 150 .

The signal measuring unit 110 may collect an ECG (electrocardiogram) signal, and compensate for the collected ECG signal by considering the impedance between the skin and the ECG sensor.

The signal processing unit 120 may pre-process the compensated ECG signal.

Specifically, as illustrated in FIGS. 4 to 5 , the signal processing unit 120 performs an amplification process of amplifying an ECG signal that has undergone a compensation step, and bandpass filtering that passes only a specific frequency band in order to remove or suppress noise. A notch filtering process of removing only the process and a specific frequency band may be performed.

The template generator 130 extracts the characteristics of the ECG signal by using an Auto Correlation Profile (ACP) and a Linear Discriminant Analysis (LDA) algorithm when a registration request is made, and based on the extracted characteristics of the ECG signal It is possible to extract unique feature sets from the ACP at intervals to create associations between users and their biometric characteristics, and to generate ECG template data, which is the extracted feature sets.

Also, the template generator 130 may update the template over time in order to mainly cope with intra-object changes in cardiac signals over time.

The processor 140 extracts the characteristics of the ECG signal using an Auto Correlation Profile (ACP) and a Linear Discriminant Analysis (LDA) algorithm when an authentication request is made, and compares the characteristics of the extracted ECG signal with the pre-stored ECG template data, Personal authentication can be performed.

The database unit 150 may register the ECG template data generated through the template generation unit 130 as user authentication information used for personal authentication.

6 to 7 are diagrams provided to explain an ECG signal feature extraction process according to an embodiment of the present invention.

The personal authentication system using the electrocardiogram according to the present embodiment performs a sliding process of moving the ECG signal that has undergone the pre-processing step by the QRS interval (m) as illustrated in FIG. 6 in order to extract the ECG signal characteristics. As illustrated in FIG. 7 , the ECG signal that has undergone the sliding process is divided into ACPs in units of preset intervals, and the amplitude (amplitude) of each ACP section and the projection angle of each section are extracted as features using the LDA algorithm. extraction process can be performed.

Specifically, in the sliding process, the ECG signal that has undergone the pre-processing step is divided into sections of a preset length (divided by 2-minute intervals in FIG. 6), and each training set X(n) is set, and each training set is It can be divided into a class (C), and these classes can be moved by the QRS interval (m).

In this case, assuming that the training set x(n) contains class C, we can define the ACP set for class C as

(m)을 아래 수식을 통해 획득할 수 있습니다. and the reference profile Φ as the average for all ACPs included in the training profile x(n).

(m) can be obtained through the formula below.

(Equation)

7 is an example of ACP for two different feature extraction targets (object A and feature extraction target B), assuming that each feature extraction target includes class C and that the class is displayed together within the same time as Q. shows

In Fig. 7, (a), (c) are intervals that are difficult to distinguish between the two objects by the amplitude ф of the ACP, but these two objects are more easily distinguishable by the angle θ of the ACP, and (b), (d) are the intervals of the ACP is the interval at which the two objects can be easily distinguished by the amplitude of . θ is considered to be the angle of the x-axis with respect to the ACP belonging to each object.

In this extraction process, since a section-based feature vector is generated, each class is preferably divided into short-time ACP segments called ACPlets in order to facilitate discrimination between different feature extraction targets.

8 is a view provided for explanation of a section unit according to an embodiment of the present invention.

The personal authentication system using the electrocardiogram according to the present embodiment may divide the ACP into sections having a specific length that is appropriately determined.

(m)의 ACPlet을 고려할 수 있다. At this time, a different feature vector can be applied to each ACPlet, and in order to determine the optimal size for each section, when the optimal size is determined as the maximum interval that satisfies the mean square error criterion between the ACPlet and the linear fit line, Φ

The ACPlet of (m) can be considered.

8 shows a typical example of MSE ζ (Mean Square Error) and identification accuracy η as a function of Number of Interval (NI). In Fig. 8, MSE ζ represents the ratio when it is assumed that MSE ζ is 100% when NI is set to 2.

(a) shows the minimum tolerance ε, which can be experimentally set to 20%, and (b) is the NI value satisfying (a), which corresponds to Q/L. (c) is the identification accuracy value when the NI value corresponds to (b).

Although MSE ζ decreases as NI increases, there is a trade-off between NI and computational cost, so NI is preferably set to Q/L.

9 is a diagram provided to explain biometric authentication efficiency according to an embodiment of the present invention.

Referring to FIG. 9 , FR (False Rejection) means a case of being rejected despite the self, and FA (False Acceptance) means a case of being accepted despite being a different person.

Threshold was selected as the intersection of FR (False Rejection) and FA (False Acceptance) depending on the subject. If the threshold is set to A section, the average of EER (Equal Error Rate) is 0%. You can see what comes out.

As such, the personal authentication system using the electrocardiogram according to the present embodiment performs personal authentication using the ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithm of the ECG signal, thereby reflecting the overall characteristics of the electrocardiogram. , the amount of computation can be reduced, and the change in impedance between the skin and the sensor due to various factors such as the measurement area, measurement area, and skin humidity in the real environment is taken into account when performing the actual authentication, the ECG signal caused by the change in the impedance of the skin to electrode. It is possible to prevent the deterioration of accuracy caused by the change of .

10 is a flowchart provided to explain a method for personal authentication using an electrocardiogram according to an embodiment of the present invention. Referring to FIG. 10 , the personal authentication method using an electrocardiogram according to the present embodiment includes a collection step of collecting an ECG (electrocardiogram) signal (S1010), an extraction step of extracting ECG signal features using ACP and LDA algorithms (S1020) and an authentication step (S1030) of performing personal authentication by comparing the characteristics of the extracted ECG signal with the pre-stored ECG template data.

In addition, the method may further include a registration step of generating ECG template data based on the characteristics of the ECG signal according to a user's request and registering the ECG template data as user authentication information used for personal authentication.

In the registration step, a threshold value of each data is set using the average value and standard deviation value of each of the amplitude data and the projection angle data, and the standard deviation value for each section of the amplitude data and the projection angle data is set. By comparing the , weights are given to each data, and ECG template data including information on the threshold value of each data and the assigned weight can be registered as user authentication information.

That is, in the registration step, the threshold value is set using the average value and STD value of the two factors of amplitude and angle (projection), registered as user information, and each section of the two factors Weights can be stored by comparing standard deviation (STD) values.

At this time, the higher the standard deviation value, the smaller the weight. Specifically, for example, if the weight of amplitude is 0.4 and the weight of angle is 0.6, weight of amplitude: 0.4, weight of angle: 0.6 is given.

In the authentication step, the matching score between the registered ECG signal and the ECG signal to be authenticated may be calculated by the following equation.

(Equation) Concordance = (amplitude weight)*(amplitude value)+(angle weight)*(angle value)

That is, in the authentication step, in order to generate ECG template data by collecting the ECG signal from the unknown ECG signal, the process of extracting the characteristics of the signal is performed in the same manner as in the registration step, and comparison is performed, and the registered ECG signal and the newly acquired ECG signal are compared. By calculating a match score between signals, if the calculation result is equal to or greater than the authentication reference value, it may be determined that personal authentication is successful.

In addition, in the authentication stage, changing ECG information can be obtained by performing template updating whenever authentication is successful.

In other words, since the ECG is constantly changing, the more the most recent ECG information is stored, the higher the coincidence. If the ECG signal matches and the newly acquired unknown ECG turns out to be the same as the registered ECG, the Enrolled information The newly acquired ECG information can be updated and saved.

Through this, after registering the ECG (electrocardiogram) signal, personal authentication is performed through the comparison result between the registered signal and the newly acquired ECG signal. Changing ECG information can be acquired by performing template updating.

On the other hand, it goes without saying that the technical idea of the present invention can also be applied to a computer-readable recording medium containing a computer program for performing the functions of the apparatus and method according to the present embodiment. In addition, the technical ideas according to various embodiments of the present invention may be implemented in the form of computer-readable codes recorded on a computer-readable recording medium. The computer-readable recording medium may be any data storage device readable by the computer and capable of storing data. For example, the computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, or the like. In addition, the computer-readable code or program stored in the computer-readable recording medium may be transmitted through a network connected between computers.

In addition, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention as claimed in the claims Various modifications are possible by those of ordinary skill in the art, and these modifications should not be individually understood from the technical spirit or prospect of the present invention.

110: signal measuring unit
120: signal processing unit
130: template generation unit
140: processor
150: database unit

Claims (8)

collecting, by a signal measuring unit, an ECG (electrocardiogram) signal;
extracting, by the processor, characteristics of the ECG signal using an Auto Correlation Profile (ACP) and a Linear Discriminant Analysis (LDA) algorithm; and
Comprising, by the processor, comparing the characteristics of the extracted ECG signal with the pre-stored ECG template data to perform personal authentication;
The personal authentication method is:
Compensating, by the processor, the collected ECG signal in consideration of the impedance between the skin and the ECG sensor; and
The processor further comprises a pre-processing step of processing the ECG signal that has undergone the compensation step,
The preprocessing step is
It includes an amplification process for amplifying, a band-pass filtering process for passing only a specific frequency band, and a notch filtering process for removing only a specific frequency band,
The extraction step is
In order to minimize the effect of physiological factors, the sliding process of moving the pre-processed ECG signal as much as the QRS interval and the ACP of the sliding process of the ECG signal are divided into preset sections, and the LDA algorithm is used to , an individual authentication method using an electrocardiogram, characterized in that it consists of an extraction process of extracting features of the amplitude (amplitude) for each section and the projection angle for each section of each ACP.
delete delete delete The method according to claim 1,
The processor generating ECG template data based on the characteristics of the ECG signal at the user's request, and registering it as user authentication information used for personal authentication; Personal authentication method using an electrocardiogram, characterized in that it further comprises: .
6. The method of claim 5,
The registration step is
Using the average value and standard deviation value of each of the amplitude data and the projection angle data, the threshold value of each data is set,
By comparing standard deviation values for each section of amplitude data and projection angle data, weights are given to each data,
A personal authentication method using an electrocardiogram, characterized in that the ECG template data including information on the threshold value of each data and the assigned weight is registered as user authentication information.
7. The method of claim 6,
The authentication step is
A personal authentication method using an electrocardiogram, characterized in that the matching score between the registered ECG signal and the ECG signal to be authenticated is calculated by the following equation.
(Equation) Concordance = (amplitude weight)*(amplitude value)+(angle weight)*(angle value)
a signal measuring unit collecting an ECG (electrocardiogram) signal; and
A processor that extracts ECG signal characteristics using ACP (Auto Correlation Profile) and LDA (Linear Discriminant Analysis) algorithms, compares the extracted ECG signal characteristics with pre-stored ECG template data, and performs personal authentication; includes; and,
The processor is
In consideration of the impedance between the skin and the ECG sensor, a compensation step for compensating the collected ECG signal and a pre-processing step for processing the ECG signal that has undergone the compensation step are performed,
The processor is
performing an amplification process for amplifying, a band-pass filtering process for passing only a specific frequency band, and a notch filtering process for removing only a specific frequency band,
The processor is
When extracting features of ECG signals, in order to minimize the effect of physiological factors, the sliding process of moving the pre-processed ECG signal as much as the QRS interval, and the sliding process of moving the ECG signal that has undergone the sliding process ACP in units of preset sections A personal authentication system using an electrocardiogram, characterized in that the extraction process is performed by dividing and extracting the amplitude (amplitude) for each section and the projection angle for each section by using the LDA algorithm.

Images