Method and system of adaptive biometric identification with determination of probability of false access taking into account user's location

Abstract

FIELD: information technologies.SUBSTANCE: invention relates to methods of adaptive biometric identification taking into account the user's localization. A computer-implemented method of adaptive biometric identification taking into account user localization, comprising the following steps: determining the number of registered biometric templates in the biometric system; obtaining the false positive probability value using retrospective data of the identification results; labelling the affiliation of biometric system users with regions based on the payment transactions they made over a specified period of time; determining the number of labelled users belonging to each region and the total number of labelled users; determining the number of registered biometric templates belonging to users in each region; determining the region in which the identification will be performed and obtaining a list of candidates and their scores; calculating the local FPIR-lk for each candidate from the list; filtering out FPIR-lk values exceeding the maximum permissible value, determining the minimum value, and confirming the identification of the candidate.EFFECT: increase in the accuracy of identification and a reduction in the number of erroneous access denials to genuine users.8 cl, 2 dwg

Description

AREA OF TECHNOLOGY

[0001] The present technical solution generally relates to the field of computational data processing, and in particular to methods of adaptive biometric identification taking into account the user's localization. LEVEL OF TECHNOLOGY

[0002] A computer-implemented method for the automated determination of an identification threshold for biometric samples in an access control system is known from the prior art, disclosed in patent RU 2779445 C1.

[0003] This patent describes a method for automatically determining the biometric identification threshold, which makes it possible to obtain accurate identification threshold values ​​based on the analysis of logs containing information on the results of real identifications.

[0004] The disadvantage of existing solutions in this field of technology is the inability to take into account the region of users’ location when carrying out identification, although this information can improve the accuracy of recognition. ESSENCE OF THE TECHNICAL SOLUTION

[0005] The declared technical solution offers a new approach in the field of adaptive biometric identification.

[0006] The technical problem or technical task to be solved is to create a new method and system of adaptive biometric identification taking into account the user's localization.

[0007] The technical result achieved by solving the above technical problem is to increase the accuracy of identification and reduce the number of erroneous access denials to genuine users.

[0008] The claimed result is achieved by a computer-implemented method of adaptive biometric identification taking into account the user's localization, performed using at least one processor and containing the following stages: a) obtaining retrospective data containing information on the results of comparing biometric samples undergoing identification with registered biometric templates, in which the result has the highest scoring score; b) selecting sample identification data for a selected period of time, during which the number of registered biometric templates changed by no more than 1% over the selected period of time; c) determining the number of identifications n and the number of registered biometric templates N over the selected period of time;d) determine for the identification data obtained in step b) information on at least two scoring points TOP-1 and TOP-2, where TOP-1 is the maximum value of the identification scoring point, TOP-2 is the next scoring point value after TOP-1, and the values ​​of TOP-1 and TOP-2 are above the authentication threshold; e) form a histogram consisting of m elements (X1; i ), distribution of the scoring score of TOP-1 identifications in the range from the authentication threshold to the maximum possible scoring score; f) form a histogram consisting of m elements (T2 i ), distribution of the scoring score of the TOP-2 identifications in the range from the authentication threshold to the maximum possible scoring score;g) determine for each point of the histograms the sum of the values ​​of all the elements of the histogram starting from the current one, obtaining new histograms of m elements;h) determine the FAR value i(false positive authentication probability) based on the data obtained in steps c) and g);i) assign users of the biometric system to regions based on the payment transactions they made over a given period of time;j) determine the number of assigned users belonging to each region (Nmark_J) and the total number of assigned users (Nmark);k) determine the number of registered biometric templates belonging to users in each region (Nrj) using the formula: Nrj=N * Nmark_j / Nmark;1) enter the parameter of reliability of user belonging to a region (d) based on the region of the last transaction;m) enter the parameter of maximum permissible error of the first kind (FPIR-max);n) determine the region in which the identification will be performed;o) perform identification and obtain a list of candidates and their scoreballs;p) calculate the local FPIR-lk for each candidate from the list obtained in step o) using the algorithm: o obtain the FAR-k value from the candidate's score using the data obtained in step h); o if the candidate's region is not determined, then FPIR_lk=1 - (1-FAR_k)^N; o if the candidate's region coincides with the region where the identification is performed, then FPIR_lk=(1 - (1-FAR_k)^(Nrj) / d; o if the candidate's region does not coincide with the region where the identification is performed, then FPIR_lk=(1 - (1-FAR_k)^(N-Nrj)) / (1-d); q) filter out from among all the FPIR-lk values ​​obtained in the previous steps, the values ​​that exceed the maximum permissible value FPIR_max; d) determine the minimum value of FPIR_lk among the values ​​selected as a result of filtering in step q); a) confirm the identification of the candidate based on the data obtained in the previous step.

[0009] In one of the particular implementation examples, if at stage q) no FPIR_lk value passes the specified filtering, then the identification result is that the user is not found.

[0010] In another particular example of implementation, at step i) the affiliation of users of the biometric system to regions is marked based on cellular network data or other available information about the users.

[0011] In another particular example of implementation, at step 1), a parameter of reliability of the user's belonging to a region (d) is introduced based on cellular network data or other available information about users.

[0012] In another specific implementation example, if a user has not had any payment transactions for a given period of time or other information by which the user's region is determined, then this user does not belong to any region.

[0013] Also, the claimed technical result is achieved through an adaptive biometric identification system taking into account the user's localization, containing: o at least one processor; o at least one memory connected to the processor, which contains machine-readable instructions that, when executed by at least one processor, ensure the implementation of the claimed method. DESCRIPTION OF DRAWINGS

[0014] The features and advantages of the present invention will become apparent from the following detailed description of the invention and the accompanying drawings, in which:

[0015] Fig. 1 illustrates the general scheme of the claimed method.

[0016] Fig. 2 illustrates an example of a general view of a computing system that ensures the implementation of the claimed solution. IMPLEMENTATION OF THE INVENTION

[0017] Today, biometric systems are commonplace and play an active role in our lives. Fingerprint scanners built into smartphones, facial recognition technologies, and other tools are gradually replacing traditional identification methods and are increasingly being implemented in large businesses such as banking and retail. Biometric systems offer several advantages over traditional methods, as they are designed for personal identification without the need to share a key and are in many ways more convenient for the user. However, the more these systems are implemented, the more pressing the issue of information security becomes.

[0018] Biometric identification is the process of comparing and determining the similarity between a person's data and their biometric "template." Biometrics allows for the identification and verification of a person based on a set of specific and unique traits inherent to them from birth. This recognition method is considered one of the most reliable, as, unlike standard logins and passwords, biometric data is much more difficult to misuse.

[0019] Biometric identification can be carried out: by fingerprint, by face, by iris, by hand geometry, by facial thermogram, by DNA, based on the acoustic characteristics of the ear, by vein pattern, etc., without limitation.

[0020] Fingerprint recognition is one of the first biometric methods. It is based on the structure of lines on the fingertips, also known as papillary patterns. After being scanned, the unique pattern is transformed into a digital biometric template, which the system uses to identify the person standing in front of it.

[0021] Finger / hand vein identification is an improved version of the previous identification method. Its algorithm is significantly more difficult to crack than other biometric scanning methods, as the veins are located deep under the skin. Infrared rays pass through the skin's surface, where they are absorbed by venous blood. A special camera captures the image, digitizes the data, and then either stores it or uses it for identity verification.

[0022] Hand geometry refers to the measurement of characteristics such as finger length and width, their curvature, and relative position. This method is now outdated and is rarely used, although it was once the dominant biometric identification option. Modern advances in fingerprint and facial recognition software have eclipsed its relevance. There is also a type of biometric palm recognition method known as "dactyloscopy."

[0023] Iris Identification. The iris, or colored part of the eye, is made up of thick, thread-like muscles. These muscles help shape the pupil to control the amount of light entering the eye. By measuring the unique folds and characteristics of these muscles, biometric verification tools can confirm identity with incredible accuracy. Dynamic scanning technologies (such as blink scanning) add an additional layer of accuracy and security.

[0024] Retinal scan scans the capillaries deep within the eye using near-infrared cameras. The resulting image is first pre-processed to improve its quality and then converted into a biometric template for new user registration and subsequent verification against the template during user identification attempts.

[0025] Facial recognition technology is undoubtedly one of the first forms of biometric identification systems. This type of software measures facial geometry, including the distance between the eyes and the distance from the chin to the forehead (to name just a few). After collecting the data, an advanced algorithm converts it into an encrypted code, or facial signature.

[0026] Ear shape identification. Unlike many other biometric methods that require special cameras, this type of identification measures the acoustics of the ear using special headphones and inaudible sound waves. A microphone inside each earphone measures how sound waves reflect off the ear and diverge in different directions depending on the curves of the ear canal. A digital copy of the ear shape is converted into a biometric template for further use.

[0027] Voice recognition technology encompasses both physiological and behavioral biometric data. Physiologically, these systems recognize the shape of a person's vocal tract, including the nose, mouth, and larynx, and determine the sound produced. Behaviorally, they capture the way a person speaks—variations in movement, tone, tempo, accent, etc.—which is also unique to each person.

[0028] A thermogram is a representation of infrared energy as an image of temperature distribution. Biometric facial thermography records the thermal patterns caused by the movement of blood under the skin. Since each person's blood vessels are unique, the corresponding thermograms are also unique, even among identical twins, making this biometric verification method even more accurate than traditional facial recognition.

[0029] DNA has long been used as an identification method. Furthermore, it is the only form of biometrics that can trace family ties. DNA matching is especially valuable in cases of missing persons, disaster victims, and potential human trafficking. Furthermore, besides fingerprints, DNA is the only biometric that cannot be inadvertently "forgotten." DNA collected from hair, saliva, etc., contains short tandem repeat sequences (STRs). These sequences can be used to definitively confirm an individual's identity by comparing them with other STRs in a database.

[0030] Biometric systems are widely used in various areas of society, such as banking, government services, and corporate security systems. One of the main obstacles to the effective operation of such systems is the high rate of false rejections for legitimate users (Type II errors). To address this issue, biometric identification providers can lower the decision threshold, thereby reducing Type II errors at the expense of significantly increasing Type I errors, which also negatively impacts the quality of services provided (increasing the likelihood of providing a product or service at the expense of someone else).

[0031] The proposed solution aims to address these shortcomings by integrating a mechanism that takes into account the local user context and the location of the identification process. The new approach is based on automated decision-making regarding the identification result, taking into account the region where the procedure is being conducted and information about the region of registered users.

[0032] As shown in Fig. 1, the claimed method of adaptive biometric identification taking into account the localization of the user (100) comprises a number of sequential steps performed using at least one processor.

[0033] At step (101), retrospective data is obtained containing information on the results of comparing the biometric samples undergoing identification with the registered biometric templates, in which the result has the highest scoring score.

[0034] At this stage, historical data on identification results is processed, for example, data obtained from identification logs for a specified time period (day, week, month, etc.). During each identification, each person's biometric sample, such as a facial or retinal image, is compared with all previously registered biometric templates. These comparisons yield scoring values, and the highest or highest score (in this case, the two highest scores are used) is the identification result.

[0035] At step (102), the identification data of biometric samples is selected for a selected period of time, during which the number of registered biometric templates changed by no more than 1% over the selected period of time.

[0036] The period of time for which the retrospective identification data is examined is determined, and the selection (filtering) is performed for the selected period.

[0037] At step (103), the number of identifications n and the number of registered biometric templates N for the selected period of time are determined.

[0038] This determines the number of identifications (n) performed during the selected period, as well as the number of biometric templates (N) registered in the system. Since the number of biometric templates during the selected period could change, the arithmetic mean of the number of templates at the beginning and end of the time period can be used as the number of templates (N). For example, n=25,000, N=10,000,000.

[0039] At step (104), for the identification data obtained at step (102), information is determined about at least two scoring points TOP-1 and TOP-2, where TOP-1 is the maximum value of the identification scoring point, TOP-2 is the next scoring point value after TOP-1, and the values ​​of TOP-1 and TOP-2 are above the authentication threshold.

[0040] At this stage, the result of each identification selected in stage (102) is analyzed. Subsequently, two values ​​are determined for each identification: the maximum scoring score for the sample compared to all samples registered in the system (TOP-1), and the second-highest comparison result (TOP-2). For example, if the following scoring scores were obtained during identification: 98, 95, 93, 92… (recorded in descending order). For this type of identification, only two values ​​are determined: 98 and 95.

[0041] It is important to note that when generating a list of scores for identification, a minimum score value condition may be imposed. This boundary (the minimum score value below which comparison results are not taken into account) will serve as the authentication threshold. For example, if the minimum score value of interest is 80, and the identification results in scores of 83, 78, 77, etc., then only the score 83 is determined as the identification result. The pair 83 and <80 (value less than 80) are determined as the TOP-1 and TOP-2 of this identification. If all comparisons as a result of identification are below the authentication threshold, for example, 75, 73, 70, etc., then it is determined that no scores above the required threshold were recorded, and the result of this identification is an empty set. The pair <80 and <80 (here 80 is the authentication threshold) are defined as TOP-1 and TOP-2 of such identification.

[0042] At step (105), a histogram consisting of m elements (T1 i ), distribution of the scoring score of TOP-1 identifications in the range from the authentication threshold to the maximum possible scoring score.

[0043] At this stage, the entire range of scores from the authentication threshold to the maximum possible value (an exact match of two samples) is divided into m parts (the larger m, the more points can be obtained on the graph). For example, the range from 80 to 100 can be divided into 20 equal parts: 80-81, 81-82, ... 99-100. The number of TOP-1 values ​​obtained in step (104) that fall within each range is determined. A histogram of the TOP-1 values ​​is formed. For example,

[0044] At step (106), a histogram consisting of m elements (T2) is formed i), distribution of the scoring score of the TOP-2 identifications in the range from the authentication threshold to the maximum possible scoring score.

[0045] At this stage, for the same ranges obtained in stage (105), similar actions are performed for the TOP-2 scores, and subsequently a histogram for the TOP-2 values ​​is formed. For example,

[0046] At step (107), the determination is made for each point of the histogram of the sums of the values ​​of all the elements of the histogram starting from the current one, obtaining new histograms from m elements (R1 i and R2 i )

[0047] At this stage, for each point of the histograms obtained in stages (105) and (106), the sum of the values ​​on the histogram is calculated, starting from the current score to the maximum score.

[0048] In step (108), the FAR value is calculated i(the probability of false positive during authentication) using the data obtained in steps (103), (107), according to the following formula:

[0049] At this stage, for each pair of values ​​of the two rows obtained in step (107) that are in the same positions (R1 i and R2 i ) and for a known value of the number of registered templates (N) and identifications over the period (n), the FAR value is determined i (the probability of a false positive during authentication), assuming q=1. For the above data, we obtain:

[0050] As can be seen from the table, the value of "y" for score 80 has the maximum value, that is, the condition for exiting the solution search cycle is met: If this condition is not met, then the value of q is reduced by 0.01 and the calculations of step (108) are repeated with the new value of q.

[0051] Next, at step (109), the biometric system users are assigned to regions based on the payment transactions they made over a specified period of time.

[0052] At this stage, user region assignments are determined using the region information of the user's last transaction within the selected time period. The longer the selected time period, the more users can be assigned, but the reliability of a user's actual region assignment will be described below.

[0053] Next, at step (110), the number of marked users belonging to each region (Nmark_j) and the total number of marked users (Nmark_j) are determined.

[0054] At this stage, the total number of users for whom the region of residence was determined based on the last transaction they made (Nmark) is calculated, as well as the number of users belonging to each region separately (Nmark_j).

[0055] For simplicity, let's assume we have only four regions. The first (Nmark_1) has 4 million users, the second (Nmark_2) has 2 million, the third (Nmark_3) has 1 million, and the fourth (Nmark_4) has 500,000 users. The total number of tagged users (Nmark) is 7.5 million.

[0056] Next, at step (111), the number of registered biometric templates belonging to users in each region (Nrj) is determined using the formula: Nrj=N * Nmark_j / Nmark.

[0057] At this stage, knowing the share of users in a particular region among all users with a known region, we find the number of users in each region among all users. We find Nr1 = 10,000,000 * 4,000,000 / 7,500,000 = 5,333,333. Similarly, we find Nr2 = 2,666,667, Nr3 = 1,333,333, Nr4 = 666,667.

[0058] Next, at step (112), the parameter of reliability of the user's belonging to the region (d) is entered based on the region of the last transaction.

[0059] At this stage, a parameter is set that evaluates the quality of user region detection. This parameter influences the algorithm's operation. If d = 1 (100%) is set, only users whose region matches the identification region will be identified. If d < 1 (< 100%), successful identification is permitted for users from other regions, but higher score requirements are imposed on their recognition than for users from the identification region. The lower d, the more lenient the score requirements for users from other regions. Parameter d can be determined both empirically and statistically by analyzing the quality of user region detection in stage (109). For example, let's set d = 0.9.

[0060] Next, at step (113), the parameter of the maximum permissible value of the first kind error (FPIR-max) is entered.

[0061] At this stage, the maximum permissible error rate (FPIR-max) is determined. This value is selected based on the riskiness of the transaction for which biometric identification is performed. GOST 19795-1-2007 "Biometric Identification" recommends a value of 0.01%. This parameter affects the algorithm's operation. The lower the FPIR-max value, the more stringent the requirements for the user's score during identification. For example, let's define FPIR-max=0.0005 (or 0.05%).

[0062] Next, at step (114), the region where the identification will be performed is determined.

[0063] At this stage, the region in which identification is performed is determined (where the terminal the user is interacting with is located). Below, we'll look at several examples with different calling regions.

[0064] Next, at step (115), identification is performed and a list of candidates and their scores is obtained.

[0065] At this stage, the user's sample (image) is compared with N biometric templates in the system database. The 5-10 biometric templates with the highest scores when compared to the user's sample, but not below the authentication threshold (in our example, 80), are selected as the identification result. Let's assume the identification result was as follows:

[0066] The examples below will use the same identification result for different calling regions, allowing comparison of the decision-making process when calling identification from different regions.

[0067] Next, at step (116), the local FPIR-lk is calculated for each candidate from the list obtained at step (115).

[0068] At this stage, the local FPIR-lk is calculated for each candidate from the list obtained in stage (115) using the algorithm:o the FAR-k value is obtained from the candidate's score using the data obtained in stage 108);o if the candidate's region is not determined, then FPIR_lk=1 - (1-FAR_k)^N;o if the candidate's region coincides with the region where the identification is performed, then FPIR_lk=(1 - (l1-FAR_k)^Nrj) / d;o if the candidate's region does not coincide with the region where the identification is performed, then FPIR_lk=(1 - (1-FAR_k)^(N-Nrj)) / (1-d).

[0069] For the identification call region 1 we get N=10,000,000, Nrj=5,333,333 and the following table: For the identification call region 2 we get N=10,000,000, Nrj=2,666,667 and the following table: For the identification call region 3 we get N=10,000,000, Nrj=1,333,333 and the following table: For identification call region 4 we get N=10,000,000, Nrj=666,667 and the following table:

[0070] Next, at step (117), among all the FPIR-lk values ​​obtained in the previous steps, the values ​​that exceed the maximum permissible value FPIR_max are filtered out.

[0071] At this stage, only those candidates whose FPIR-lk values ​​do not exceed the maximum FPIR value set at stage (113) are retained.

[0072] In the example above, when identifying in region 1, there are no suitable candidates, when identifying in region 2, there are also no suitable candidates, when identifying in region 3, only one candidate (B) satisfies the required condition, when identifying in region 4, only one candidate (D) satisfies the required condition.

[0073] Next, at step (118), the minimum value of FPIR_lk is determined among the values ​​selected as a result of filtering at step (117).

[0074] At this stage, the candidate with the minimum FPIR_lk value, but not exceeding the maximum allowable FPIR value, is selected.

[0075] And then at step (119) the identification of the candidate is confirmed based on the data obtained at the previous step.

[0076] At this stage, the candidate found in stage (118) is used as the identification result.

[0077] In the example above, when identifying in regions 1 and 2, the user will be denied access. When identifying in region 3, the identification result will be candidate (B), who is also located in region 3; when identifying in region 4, the identification result will be candidate (D), who is also located in region 4, even though their scores in the candidate list are not the highest. The difference in decision making is due to the availability of additional information about the user undergoing identification (their current region).

[0078] In another particular example of implementation, if at step (117) no FPIR_lk value passes the specified filtering, then the identification result is that the user is not found.

[0079] In another particular example of implementation, at step (109), the affiliation of users of the biometric system to regions is marked on the basis of cellular network data or other available information about users.

[0080] In another particular example of implementation, at step (112), a parameter of reliability of the user's belonging to a region (d) is entered based on cellular network data or other available information about users.

[0081] In another specific implementation example, if a user has not had any payment transactions for a given period of time or other information by which the user's region is determined, then this user does not belong to any region.

[0082] This method significantly reduces the Type II error (FNIR) while maintaining a fixed Type I error rate (FPIR). Research has provided compelling evidence of the proposed method's effectiveness. For example, testing showed that the FNIR rate was reduced by more than 60% for a region with one million registered biometric service users and a total of over 30 million users, while maintaining a FPIR of 0.01%.

[0083] Fig. 2 shows an example of a general view of a computing system (300) that ensures the implementation of the claimed method or is part of a computer system, for example, a server, a personal computer, part of a computing cluster, processing the necessary data for implementing the claimed technical solution.

[0084] In general, the system (300) comprises one or more processors (301), memory means such as RAM (302) and ROM (303), input / output interfaces (304), input / output devices (1105), and a device for network interaction (306), united by a common information exchange bus.

[0085] The processor (301) (or multiple processors, multi-core processor, etc.) can be selected from a range of devices that are widely used today, such as those from manufacturers such as: Intel TM , AMD TM , Apple TM , Samsung Exynos TM , MediaTEK TM , Qualcomm Snapdragon TMetc. The processor or one of the processors used in the system (300) must also include a graphics processor, for example, NVIDIA GPU or Graphcore, the type of which is also suitable for the full or partial implementation of the method, and can also be used for training and applying machine learning models in various information systems.

[0086] RAM (302) is random access memory (RAM) and is designed to store machine-readable instructions executed by the processor (301) to perform the necessary logical data processing operations. RAM (302) typically contains executable instructions of the operating system and corresponding software components (applications, software modules, etc.). The available memory of a graphics card or graphics processor may also serve as RAM (302).

[0087] ROM (303) is one or more permanent storage devices, such as hard disk drive (HDD), solid state drive (SSD), flash memory (EEPROM, NAND, etc.), optical storage media (CD-R / RW, DVD-R / RW, BlueRay Disc, MD), etc.

[0088] Various types of I / O interfaces (304) are used to organize the operation of system components (300) and organize the operation of external connected devices. The choice of the appropriate interfaces depends on the specific design of the computing device, which may include, but are not limited to: PCI, AGP, PS / 2, IrDa, Fire Wire, LPT, COM, SATA, IDE, Lightning, USB (2.0, 3.0, 3.1, micro, mini, type C), TRS / Audio jack (2.5, 3.5, 6.35), HDMI, DVI, VGA, Display Port, RJ45, RS232, etc.

[0089] To ensure interaction between the user and the computing system (300), various means (305) of I / O information are used, for example, a keyboard, a display (monitor), a touch display, a touchpad, a joystick, a mouse, a light pen, a stylus, a touch panel, a trackball, speakers, a microphone, augmented reality means, optical sensors, a tablet, light indicators, a projector, a camera, biometric identification means (a retinal scanner, a fingerprint scanner, a voice recognition module), etc.

[0090] The network communication means (306) provides data transmission via an internal or external computer network, such as an Intranet, the Internet, a LAN, etc. One or more means (306) may be, but are not limited to: an Ethernet card, a GSM modem, a GPRS modem, an LTE modem, a 5G modem, a satellite communication module, an NFC module, a Bluetooth and / or BLE module, a Wi-Fi module, etc.

[0091] The submitted application materials disclose preferred examples of the implementation of the technical solution and should not be interpreted as limiting other, particular examples of its implementation that do not go beyond the scope of the requested legal protection, which are obvious to specialists in the relevant field of technology.

Claims

1. A computer-implemented method for adaptive biometric identification taking into account the user's localization, performed using at least one processor and comprising the steps of: a) receive retrospective data containing information on the results of comparison of biometric samples undergoing identification with registered biometric templates, in which the result has the highest scoring score; b) sample identification data is collected for a selected period of time during which the number of registered biometric templates changed by no more than 1% over the selected period of time; c) determine the number of identifications n and the number of registered biometric templates N for a selected period of time; d) determine for the identification data obtained in step b), information on at least two scoring points TOP-1 and TOP-2, where TOP-1 is the maximum value of the identification scoring point, TOP-2 is the next scoring point value after TOP-1, and the values ​​of TOP-1 and TOP-2 are above the authentication threshold; e) form a histogram consisting of m elements (T1 i ), distribution of the scoring score of TOP-1 identifications in the range from the authentication threshold to the maximum possible scoring score; f) form a histogram consisting of m elements (T2 i ), distribution of the scoring score of the TOP-2 identifications in the range from the authentication threshold to the maximum possible scoring score; g) determine for each point of the histogram the sum of the values ​​of all the elements of the histogram starting from the current one, obtaining new histograms of m elements; h) determine the FAR value i(probability of false positive during authentication) based on the data obtained in steps c) and g); i) assign users of the biometric system to regions based on the payment transactions they have completed over a given period of time; j) determine the number of marked users belonging to each region (Nmark_j) and the total number of marked users (Nmark); k) determine the number of registered biometric templates belonging to users in each region (Nrj) using the formula: Nrj=N * Nmark_j / Nmark; l) introduce the parameter of reliability of the user’s belonging to a region (d) based on the region of the last transaction; m) introduce the parameter of the maximum permissible value of the first type error (FPIR-max); n) determine the region where the identification will be carried out; o) carry out identification and obtain a list of candidates and their scores; p) calculate the local FPIR-lk for each candidate from the list obtained in step o), using the algorithm: ○ obtain the FAR-k value for the candidate score using the data obtained in step h); ○ if the candidate region is not defined, then FPIR_lk=1 - (1-FAR_k)^N; ○ if the candidate’s region coincides with the region where the identification is performed, then FPIR_lk=(1 - (1-FAR_k)^Nrj) / d; ○ if the candidate’s region does not coincide with the region where the identification is performed, then FPIR_lk=(1 - (1-FAR_k)^(N-Nrj)) / (1-d); q) filter out from among all the FPIR-lk values ​​obtained in the previous steps the values ​​that exceed the maximum permissible value FPIR_max; r) determine the minimum value of FPIR_lk among the values ​​selected as a result of filtering at step q); and s) confirm the identification of the candidate based on the data obtained at the previous stage.

2. The method according to paragraph 1, characterized in that if at step q) not a single value of FPIR_lk passes the specified filtering, then the identification result is that the user is not found.

3. The method according to paragraph 1, characterized in that at step i) the affiliation of users of the biometric system to regions is marked on the basis of cellular network data or other available information about users.

4. The method according to paragraph 1, characterized in that at step 1) the parameter of reliability of the user’s belonging to a region (d) is entered based on cellular network data or other available information about users.

5. The method according to paragraph 1, characterized in that if the user has not had any payment transactions for a given period of time or other information by which the user’s region is determined, then this user does not belong to any region.

6. An adaptive biometric identification system taking into account the user’s location, comprising: ○ at least one processor; ○ at least one memory connected to the processor, which contains machine-readable instructions that, when executed by at least one processor, ensure the implementation of the method according to paragraphs 1-5.

Images