CN1540568A - Identity recognition and authentication method - Google Patents

Abstract

The invention relates to a new identity identification method based on human body biological characteristic information, belonging to the field of biological characteristic identification and identification. The method provides that a photoelectric volume change signal (PPG) is applied to an identity authentication system, and an authentication conclusion is made according to a matching result by extracting a biological characteristic signal of an authentication object, generating a characteristic vector of the biological characteristic signal according to the biological characteristic signal and performing pattern matching with a characteristic template of the authenticator in a pre-stored database. The method is a non-invasive identity identification method and is easy to be accepted by people. The method can be effectively combined with other biological characteristic technologies, realizes multi-biological information identity identification and authentication, and improves the identification performance of an authentication system.

Description

Identification and Authentication Methods

technical field

The present invention relates to an identification and identification method, in particular to an identification and identification method based on human biological feature information, which extracts the biological feature signal of the identification object and generates a feature vector of the biological feature signal accordingly, and Pattern matching is performed on the feature template of the authenticator in a pre-stored database, and an identification conclusion is made according to the matching result.

Background technique

Human biological characteristics or behavioral characteristics are studied to be applied in the field of identification and authentication. What is widely used at present is to use face, fingerprint, voice, iris, lip movement, gait, EEG, ECG, etc. to carry out identification. However, these existing technologies have their own drawbacks. For example, fingerprints can be concealed in latex, faces can be faked through photos, and voices can be imitated, so the reliability of identification is poor; in addition, identification methods based on EEG or ECG require several electrodes to collect biometric signals , the operation is cumbersome and the cost is high.

Contents of the invention

Aiming at the deficiencies of the above-mentioned prior art, the present invention proposes an identification and identification method based on human biological feature information, which extracts the biological feature signal of the identification object and generates the feature vector of the biological feature signal, which is compared with the previously stored in the database. Pattern matching is performed on the feature templates, and identification conclusions are made based on the matching results.

In order to achieve the above object of the present invention, the present invention provides a method for identification and identification based on photoelectric volume signals, the method comprising the following steps:

a) Obtain the PPG signal waveform of the person being identified;

b) extracting the biological characteristics of the PPG signal according to the obtained PPG waveform;

c) calculating the characteristic parameter of the PPG signal according to the biological characteristics of the extracted PPG signal;

D) generating the PPG signal feature vector of the identified person according to the PPG signal feature parameter;

E) carrying out pattern matching between the PPG signal feature vector of the identified person and the PPG signal feature template of the identified person in a pre-generated database;

f) When the match exceeds a preset threshold value, it is judged that the identity of the authenticated person is qualified.

Compared with the prior art, the present invention has obvious advantages. First, the PPG signal is easy to extract from various parts of the human body such as fingers, earlobes, wrists, or arms, and the operator does not need to have special skills to master it, and the operation cost is low. ,efficient. Moreover, since the PPG signals of a single individual are basically the same, there are relatively large differences in the PPG signals of different individuals, and the PPG signals will not be easily faked or imitated like fingerprints, face, voice and other feature recognition. The reliability of identification and identification is improved.

Those skilled in the art will better understand the above-mentioned technical solutions and advantages of the present invention by describing specific embodiments of the present invention in conjunction with the accompanying drawings.

Brief description of the drawings

Figure 1 is a schematic diagram of a typical PPG waveform;

Fig. 2 is the PPG waveform detected by person A at two moments;

Fig. 3 is the PPG waveform detected by the appraiser B at the same two moments;

Fig. 4 is the block flow diagram of the identification and authentication method based on PPG signal according to the first embodiment of the present invention;

Fig. 5 is a characteristic schematic diagram of the PPG signal adopted in the method shown in Fig. 4;

Fig. 6 is used for extracting the module flowchart of PPG signal feature in the block diagram shown in Fig. 4;

Figure 7 and Figure 8 respectively provide the least squares curve fitting of the tenth order polynomial and the PPG waveform after the first derivative processing;

Fig. 9 is a flow chart of an identification and authentication method based on the combination of PPG signal and fingerprint according to the second embodiment of the present invention.

Detailed ways

PHOTOPLETHYSMOGRAPHIC SIGNAL (hereinafter referred to as PPG signal) has high diagnostic value in blood pressure measurement. Each heartbeat of the human body will generate a pulse, that is, the PPG waveform, as shown in Figure 1, which shows a schematic diagram of a typical PPG pulse, where the PPG pulse 100 reflects the blood volume (x-axis) versus time (t axis) changes. This pulse propagates within the body's blood circulatory system and can be detected non-invasively by a photoplethysmometer. Figure 2 and Figure 3 respectively show the PPG signals detected by test object A and test object B at two different times, the PPG waveforms of test object A are 200, 210, and the PPG waveforms of test object B are 300, 310. It can be seen from Figure 2 and Figure 3 that although the characteristics of the PPG signal of each individual will vary with the change of the detection site and detection time, the PPG signal of the same person remains basically stable, and the PPG signal of different individuals But there is a big difference. Based on the above knowledge, the inventors propose to use the PPG signal for identification and authentication.

Fig. 4 shows a first embodiment of the present invention, which illustrates a method for identification and authentication based on PPG signals. The method includes extracting the PPG signal of the identified person and preprocessing the extracted original PPG signal; detecting the biological characteristics of the preprocessed PPG signal; calculating the characteristic parameters of the PPG signal according to the monitored biological characteristics of the PPG signal, according to The calculated feature parameters are called feature vectors; the generated feature vectors are matched with the templates of the PPG database stored in advance; according to the matching results, it is judged whether the person to be identified belongs to the person predetermined by the database.

According to the first embodiment of the present invention, the PPG signal of the person to be authenticated can be extracted through the following exemplified methods, for example, the person to be authenticated first inputs his own ID or other identity-indicating information into the authentication system. A PPG detector is attached to the examinee's fingertip with a bandage, and the examinee's PPG signal is continuously recorded for at least one minute.

Because the PPG signal is more complicated, it has both high-frequency components and low-frequency components. In order to extract useful biological features conveniently, the extracted original PPG signal should be preprocessed. For example, the extracted signal is converted into a digital signal at a sampling frequency of 1 KHz, and then the high-frequency noise and low-frequency noise in the PPG signal are removed by smoothing technology.

The preprocessed PPG signal can be used to extract its biological characteristics. The extraction of PPG signal biometrics according to the method of the present invention will be analyzed in detail below.

Referring to Fig. 5, the main biological characteristics and characteristic parameters of the PPG signal are shown. The biological characteristics of PPG signal include PPG signal start point, PPG signal peak and valley point. The characteristic parameters corresponding to the biological characteristics of these PPG signals include: rising slope k ₁ , falling slope k ₂ , time interval t ₁ , time interval t ₂ and the number of vertices N, which are defined as follows:

a. Rising slope k ₁ : the slope between the valley point of each PPG waveform and the first apex thereafter;

b. Declining slope k ₂ : the slope between the last vertex of each PPG waveform and the next valley point;

c. Time interval t ₁ : the time interval between the valley point of each PPG waveform and the first apex thereafter;

d. Time interval t ₂ : the time interval between the first apex of each PPG waveform and the valley point of the subsequent PPG waveform;

e. The number of vertices N: the number of local maximum values on each PPG waveform.

Among them, the vertex of each PPG waveform is defined as the position of the local maximum contained in each PPG waveform, and the valley point of each PPG waveform is defined as the first local minimum before the first vertex of each PPG waveform The position of the value, as shown in Figure 5, the valley point of the PPG waveform is A, the first apex is B, and C is the valley point of the latter PPG waveform.

According to this embodiment, the maximum point among the first 2000 points of the received PPG signal is used as the starting point of the PPG signal. When the sampling frequency is 1KHz, choosing 2000 points can ensure that at least one vertex is included without wasting too much data. However, those skilled in the art can understand that the present invention is not limited thereto.

Referring now to FIG. 6 , after the collected PPG signal is input to the start point detection module 61 to identify the start point of the PPG signal, the apex and valley point monitoring module 62 will monitor the apex and valley point of the PPG signal. Specifically, the present invention applies the tenth-order polynomial least squares curve fitting to each PPG waveform, and utilizes this method to filter, smooth and denoise the PPG waveform. Figure 7 shows the waveforms before and after the least squares curve fitting of the tenth order polynomial. In Fig. 7, the upper waveform is the unprocessed signal waveform, and the lower waveform is the waveform after fitting the PPG waveform with the least square curve of the tenth-order polynomial. Then, the apex and valley points of the PPG signal can be calculated by calculating the first derivative of the fitted result. Figure 8 shows the PPG waveform after the first derivative processing. It can be seen from Fig. 8 that the intersection of the first-order derivative waveform and the horizontal axis (that is, where the first-order derivative is 0) represents the peak position of the PPG waveform. Figure 8 shows 4 peak locations, of which there are two apexes and two valleys.

Next, use the feature vector generation module 63 shown in Figure 6 to generate the feature parameters k ₁ and k ₂ of the PPG signal, the steps are:

a) Apply first-order curve fitting to the data between the valley point of each PPG waveform and the apex of the first PPG waveform to calculate the characteristic parameter rising slope k ₁ ;

b) Apply a first-order curve fitting to the data between the last vertex of each PPG waveform and the valley point of the subsequent PPG waveform to calculate the characteristic parameter descending slope k ₂ ;

c) average all _k1 and _k2 respectively;

d) subtracting the first valley point position of the PPG waveform from the first apex position of each PPG waveform, to calculate the characteristic parameter time interval t ₁ ;

e) subtracting the first apex position of the PPG waveform from the last waveform valley point position of each PPG waveform, to calculate the characteristic parameter time interval t ₂ ;

f) Carry out first-order derivation to the data between the first valley point of each PPG waveform and the valley point of the next PPG waveform, to calculate the number of feature parameters N (for example, N=2 among Fig. 8) .

Among the above methods, methods for performing least square curve fitting and first-order derivation of tenth-order polynomials are well known to those skilled in the art, so detailed descriptions are omitted here.

After the above characteristic parameters are calculated, it is only necessary to use the characteristic vector generation module 63 shown in FIG. 6 to arrange these characteristic parameters in a certain order to generate the characteristic vector of the PPG signal. For example, a form of feature vector may be: Feature Vector=[K ₁ , K ₂ , t ₁ , t ₂ , N].

Of course, in the PPG signal-based identification and identification method of the present invention, the characteristic parameters of the PPG signal are not limited to the above parameters, and other characteristic information can also be introduced according to the performance requirements of the identification system, such as the binary information of the PPG signal. Derivative features, etc.

Next, pattern matching is performed between the PPG signal feature vector of the person being identified and the feature template of the PPG signal of the person being identified in the pre-prepared database. In the matching comparison and authentication judgment, a threshold value is preset. Only when the match between the feature vector of the PPG signal of the person being authenticated and the template information of the system user stored in advance exceeds the threshold value, can it be confirmed that the person being authenticated belongs to the person predetermined by the database.

On the basis of the above-mentioned method for identification and identification using human PPG signals, and in consideration of practical application issues such as acquisition of biological feature signals, the present invention further proposes multi-biological feature information identification, that is, extracting two or more biometric information, and generate a joint feature vector corresponding to the above two or more than two kinds of feature information.

Fig. 9 is a flow chart of the identification and authentication method based on the combination of PPG signal and fingerprint according to the second embodiment of the present invention. As shown in Figure 9, in this embodiment, the PPG signal of the fingertip and the biometric information of the fingerprint image are combined for identification and identification of the identity, and the method includes:

a) Obtain PPG signal and fingerprint image and perform preprocessing;

b) extracting the biological characteristics of the PPG signal;

c) Extracting fingerprint features;

d) Matching and comparing the joint feature information used for identity verification with the template information stored in advance;

e) Identification judgment based on matching results.

The fingerprint image is acquired while acquiring the fingertip PPG signal, so that the identification system is convenient for the system user to operate. The feature extraction of the PPG signal is as mentioned above, and the extraction of fingerprint features is a relatively mature technology, so it will not be repeated here. It should be noted that although in the second embodiment of the present invention, the PPG signal of the fingertip is combined with the biological feature information of the fingerprint image, it can also be combined with other biological feature information (such as iris features) for identification and identification.

In summary, the identification and identification method based on human biometric information proposed by the present invention can easily and simply improve the reliability of identification, and can also combine with other biometric technologies to realize identification and identification with multiple biometric information. The above descriptions of the specific implementation modes and examples of the present invention are only exemplary, and those skilled in the art can understand that the various feature extraction methods of the present invention can make various changes and modifications, for example, according to the requirements of the recognition system performance And introduce other feature information, such as the second order derivative feature of the PPG signal. These changes and modifications that do not deviate from the concept of the present invention all fall within the scope of the present invention defined by the claims of the present invention.

Claims (12)

1. identification and authentication method based on a PPG signal may further comprise the steps:

A) obtain by assessor's PPG signal waveform;

B) according to the biological characteristic of the PPG waveform extracting PPG signal that is obtained;

C) calculate the characteristic parameter of PPG signal according to the biological characteristic of the PPG signal that is extracted;

D) generate by assessor's PPG signal characteristic vector according to described PPG signal characteristic parameter;

E) should be carried out pattern match between assessor's the PPG signal characteristic template in the database that will be generated in advance by assessor PPG signal characteristic vector and;

F) when coupling surpassed a default threshold value, it was qualified by assessor's identity to judge.

2. method according to claim 1 comprises further in step (a) that wherein the PPG signal to being obtained carries out pretreated step.

3. method according to claim 2, to comprise the PPG conversion of signals be digital signal and remove high frequency noise and low-frequency noise wherein the PPG signal that is obtained to be carried out pretreated step.

4. method according to claim 1, the biological characteristic of wherein said PPG signal waveform comprise PPG waveform starting point, the summit of PPG signal waveform and valley point.

5. method according to claim 4, the position of the local maximum that the summit of wherein said PPG signal waveform is comprised for each PPG waveform; The valley point of described PPG signal waveform is the position of first local minimum before first summit of each PPG waveform.

6. method according to claim 5, the extraction of described PPG signal biological characteristic comprises:

A) detect PPG signal starting point;

B) detect PPG signal summit and valley point.

7. method according to claim 6, the step that wherein detects described PPG signal summit and valley point comprises:

A) to each PPG waveform application ten rank curve fitting;

B) result after the curve fitting is asked first order derivative.

8. method according to claim 1, the characteristic parameter of wherein said PPG signal comprises rate of rise k ₁, descending slope k ₂, time interval t ₁, time interval t ₂With number of vertices N.

9. method according to claim 8, wherein said rate of rise k ₁The valley point that is defined as each PPG waveform is to the slope between its first summit; Descending slope k ₂Last summit that is defined as each PPG waveform is to the slope between PPG waveform valley point thereafter; Time interval t ₁The valley point that is defined as each PPG waveform is to the time interval between its first summit; Time interval t ₂First summit that is defined as each PPG waveform is to the time interval between the valley point of a PPG waveform thereafter; And number of vertices N is defined as the number of local maximum on each PPG waveform.

10. method according to claim 9, the step that wherein obtains described each characteristic parameter of PPG signal is:

A) the first order curve match is applied to the valley point of each PPG waveform to the data between its first summit, calculated characteristics parameter k ₁

B) first order curve match last summit of being applied to each PPG waveform is to data between PPG waveform valley point thereafter, calculated characteristics parameter k ₂

C) respectively to all k ₁And k ₂Average;

D) calculated characteristics parameter time interval t is come in first vertex position of each PPG waveform position, valley point of deducting this PPG waveform ₁

E) position, a back waveform valley point of each PPG waveform first vertex position of deducting this PPG waveform comes calculated characteristics parameter time interval t ₂

F) data that are applied between the valley point of the valley point of each PPG waveform and a back PPG waveform of first order derivative are come calculated characteristics parameter number of vertices N.

11. method according to claim 1 wherein in described step c), is extracted two or more biological information, and produces corresponding to above-mentioned two or more the associating proper vector of characteristic information.

12. method according to claim 11 is wherein extracted PPG signal and fingerprint image biological information.

Images