JP6488853B2 - Authentication processing program, authentication processing apparatus, and authentication processing method - Google Patents

Description

  This case relates to an authentication processing program, an authentication processing device, and an authentication processing method.

  A 1: N authentication technology has been developed that matches registered biometric information of a plurality of users and biometric information acquired from a person to be authenticated without determining a person to be collated from a plurality of users registered with biometric information. Has been. In 1: N authentication, since the time required for authentication tends to be long, it is considered to suppress the number of candidates to be verified (see, for example, Patent Document 1).

JP 2010-049357 A

  However, in the technique of Patent Document 1, since candidate information is created when authentication fails, candidate information is not created unless authentication fails.

  In one aspect, an object of the present invention is to provide an authentication processing program, an authentication processing device, and an authentication processing method that can suppress the number of candidates to be verified in the authentication processing.

  In one aspect, the authentication processing program refers to a feature information storage unit that stores biometric information in association with a user and feature data, and stores the first of the users included in the feature information storage unit. The process of calculating the first similarity between the user's feature data and the biometric information, and the similarity with the first user's feature data when the calculated first similarity is equal to or greater than a first threshold value A process of extracting the second user by referring to a similar user storage unit that stores a second user associated with feature data having a threshold value equal to or greater than a second threshold in association with the first user; A process of calculating the second similarity between the feature data and the biological information, and a process of determining a user having a higher value among the first similarity and the second similarity as a user corresponding to the biological information When Cause the computer to execute.

  The number of candidates for verification can be reduced.

It is a table of the registration biometric information of a plurality of users registered beforehand. (A) is a block diagram for demonstrating the hardware constitutions of the biometrics apparatus which concerns on Example 1, (b) is a block diagram of each function implement | achieved by execution of an information processing program. It is an example of the flowchart showing a registration process. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. (A) is a figure which illustrates the table stored in the reference value storage part, (b) is a figure which illustrates the table stored in the takeover user storage part, (c) is a figure in the log storage part It is a figure which illustrates the table stored. It is an example of the flowchart showing an authentication process. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. It is a flowchart showing an example of a similar user collation process. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. It is a flowchart showing an example of a postponing process. (A) is a figure which illustrates a registration table, (b) is a figure which illustrates a score table, (c) is a figure which illustrates a similar user table. It is a flowchart showing an example of an adjustment process.

  Prior to the description of the embodiments, an outline of biometric authentication processing will be described. The biometric authentication process is performed when the verification result (similarity) between the biometric information for verification acquired from the authenticated person and the registered biometric information of the authenticated person registered in advance is equal to or greater than a predetermined value. This is the process of determining the authenticator as the principal. Biometric authentication processing includes 1: 1 authentication and 1: N authentication. In 1: 1 authentication, a person to be collated is determined in advance from a plurality of users whose biometric information is registered by inputting an ID or the like, and the biometric information of the person to be collated and the collation acquired from the person to be authenticated It is the process which collates with the biometric information. On the other hand, 1: N authentication refers to the registered biometric information of a plurality of users and the biometric information for verification acquired from the person to be authenticated without determining the person to be verified from among the plurality of users registered with the biometric information. Is a process of collating

  In 1: N authentication, it is possible to save time and labor for inputting IDs and the like, but it is necessary to collate with registered biometric information of a plurality of users, and the time required for authentication becomes longer. FIG. 1 is a table of registered biometric information of a plurality of users registered in advance. As illustrated in FIG. 1, registered biometric information (for example, feature data such as fingerprint information) is stored in association with the ID (user name or the like) of each user. In the example of FIG. 1, the number of registered users is 20. In 1: N authentication, for example, biometric information for verification acquired from the person to be authenticated and registered biometric information for these 20 people are sequentially verified. In FIG. 1, the collation score is also shown. A collation score is a similarity and shows the value of 0.0-10.0 in the example of FIG. A large value indicates a high degree of similarity.

  For example, the person to be authenticated is determined to be a user having a maximum matching score value that is equal to or greater than a predetermined value. In the example of FIG. 1, it is determined that the person to be authenticated is User001. However, although it is determined that the user to be authenticated is User001 that has been collated first, collation is performed with the registered biometric information of all remaining users. Therefore, authentication takes time. Therefore, in the following embodiments, an authentication processing program, an authentication processing device, and an authentication processing method that can suppress the number of candidates to be verified will be described.

  FIG. 2A is a block diagram for explaining a hardware configuration of the authentication processing apparatus 100 according to the first embodiment. As illustrated in FIG. 2A, the authentication processing device 100 includes a CPU 101, a RAM 102, a storage device 103, a display device 104, a biosensor 105, an input device 106, and the like. Each of these devices is connected by a bus or the like.

  A CPU (Central Processing Unit) 101 is a central processing unit. The CPU 101 includes one or more cores. A RAM (Random Access Memory) 102 is a volatile memory that temporarily stores programs executed by the CPU 101, data processed by the CPU 101, and the like.

  The storage device 103 is a nonvolatile storage device. As the storage device 103, for example, a ROM (Read Only Memory), a solid state drive (SSD) such as a flash memory, a hard disk driven by a hard disk drive, or the like can be used. The authentication processing program according to the present embodiment is stored in the storage device 103. The display device 104 is a liquid crystal display, an electroluminescence panel, or the like, and displays the processing result of the authentication processing device 100 and the like.

  The biometric sensor 105 is a sensor that acquires user biometric information. In the present embodiment, the biometric sensor 105 acquires a user's fingerprint image. The biological sensor 105 is an optical sensor, a capacitance sensor, or the like. The input device 106 is a device for inputting information such as a user ID, and is a mouse, a keyboard, or the like.

  The authentication processing program stored in the storage device 103 is expanded in the RAM 102. The CPU 101 executes an authentication processing program expanded in the RAM 102. Thereby, each process by the authentication processing apparatus 100 is performed.

  FIG. 2B is a block diagram of each function realized by executing the authentication processing program. As illustrated in FIG. 2B, the authentication processing apparatus 100 executes the authentication processing program so that the acquisition unit 10, the registration unit 20, the verification unit 30, the user extraction unit 40, the rearrangement unit 50, the storage unit 60, It functions as the output unit 70 and the update unit 80. The storage unit 60 functions as a registration table 61, a score table 62, a similar user table 63, a reference value table 64, a takeover user table 65, and a log table 66. Details of each part will be described below.

(registration process)
FIG. 3 is an example of a flowchart showing the registration process. The registration process is a process for registering in advance biometric information of a user who uses the biometric authentication process. An example of the registration process will be described with reference to FIGS. 2B and 3. First, the acquisition unit 10 acquires biometric information for registration and biometric information for verification from the image acquired by the biometric sensor 105 (step S1). The biometric information for registration and the biometric information for verification may be feature data extracted from the same image acquired by the biometric sensor 105, or may be feature data extracted from different images. In the present embodiment, the acquisition unit 10 acquires an average value of feature data extracted from a plurality of images acquired by the biometric sensor 105 as biometric information for registration. In this embodiment, in order for the biometric sensor 105 to acquire a fingerprint image, feature data such as a fingerprint image pattern and a fingerprint minutiae position is used. In this embodiment, the biometric information for verification is acquired from an image acquired by the biometric sensor 105, which is different from the image used for the biometric information for registration.

  Next, the registration unit 20 registers biometric information for registration in the registration table 61 as registered biometric information in association with the user name input from the input device 106 (step S2). Therefore, the registration table 61 of the storage unit 60 functions as an example of the feature information storage unit. FIG. 4A illustrates the registration table 61. In FIG. 4A, the fingerprint N of User014 is newly added as registered biometric information.

  Next, the collation unit 30 collates the biometric information for registration acquired in step S1 with the biometric information for collation, and registers the collation score as the highest score in the score table 62 (step S3). FIG. 4B illustrates the score table 62. In FIG. 4B, the matching score “9.0” of User014 is registered in the score table 62 as the highest score. The matching score can be “9.0” instead of “10.0” because the image used when acquiring biometric information for registration is different from the image used when acquiring biometric information for verification. It is.

  Next, the collation unit 30 collates the biometric information for registration with the registered biometric information of all other users already registered in the registration table 61. Next, the registration unit 20 reads the similar reference value from the reference value table 64, associates the user whose matching score is equal to or higher than the similar reference value, and User014, and stores it in the similar user table 63 (step S4). Therefore, the similar user table 63 of the storage unit 60 functions as an example of the similar user storage unit. FIG. 5A illustrates the reference value table 64. The similarity reference value is 7.0, for example. Note that the dissimilar reference values stored in the reference value table 64 will be described later. In the example of FIG. 4C, User001, User009, User012, and User014 are stored as similar users. Therefore, User 014, User 009, and User 012 are also associated with User 014 as a similar user.

  Next, the rearrangement unit 50 rearranges the registration table 61 in descending order of the number of similar users (step S5). In the example of FIG. 4C, since the number of similar users of User014 is the largest, the registration table 61 is not changed from FIG. Through the above processing, a new user is registered.

  By executing the registration process, users whose similarity between registered biometric information is equal to or greater than the similarity reference value are associated with each other as a similar user group. In addition, the registration table 61 is rearranged in descending order of the number of similar users. Furthermore, prior to the authentication processing described later, the similarity between the newly registered biometric information for the user and the biometric information for verification is registered as the highest score.

  Note that steps S3 to S5 do not have to be executed immediately after step S2. For example, for a plurality of registered users, steps S3 to S5 may be performed collectively by batch processing in a time zone where there are few users. Further, the rearrangement in step S5 may take into account not only the order in which the number of similar users is large but also the past number of authentications.

(Authentication process)
FIG. 6 is an example of a flowchart showing the authentication process. The authentication process is a process for the user registered by the registration process to perform identity verification. An example of the authentication process will be described with reference to FIG. 2B and FIG. First, the acquisition unit 10 acquires biometric information for verification from the image acquired by the biometric sensor 105 (step S11). Next, the user extraction unit 40 determines whether or not unmatched users remain in the registration table 61 (step S12). When it is determined as “No” in step S12, collation with the registered biometric information of all users is completed. Therefore, the output unit 70 outputs a signal related to the authentication failure (step S13). Thereby, the display device 104 performs display related to the authentication failure. In addition, the execution of the flowchart ends.

  When it is determined as “Yes” in step S <b> 12, the user extraction unit 40 extracts the first user among unmatched users. In the example of FIG. 7A, collation with User014 is performed first. Next, the collation unit 30 reads the registered biometric information of the user from the registration table 61 as target biometric information (step S14). Next, the collation part 30 collates target biometric information and biometric information for collation (step S15).

  Next, the collation unit 30 reads a similar reference value (for example, 7.0) from the reference value table 64, and determines whether or not the collation score is equal to or greater than the similar reference value (step S16). If it is determined as “Yes” in step S16, similar user collation processing is performed (step S17). If “No” is determined in step S16, the collation unit 30 reads the dissimilar reference value (for example, 3.0) from the reference value table 64 in FIG. 5A, and the collation score is equal to or less than the dissimilar reference value. It is determined whether or not (step S18). If “Yes” is determined in step S18, a post-processing is performed (step S19). If it is determined “No” in step S18, step S12 is executed again. In the example of FIG. 7A, verification with User001 is performed next to User014.

(Similar user verification process)
FIG. 8 is a flowchart illustrating an example of similar user verification processing. As illustrated in FIG. 8, the user extraction unit 40 refers to the similar user table 63 and extracts a similar user group of the target biological information (step S21). In the example of FIG. 7C, User001, User009, and User012 are extracted. Next, the collation unit 30 collates the biometric information (similar biometric information) of all users extracted in step S21 with the biometric information for collation (step S22). FIG. 7C illustrates a matching score with User001, User009, and User012. It is assumed that the collation score between the biometric information for collation and the registered biometric information of User014 is 7.3 as an example.

  Next, the user extraction unit 40 refers to the score table 62 in FIG. 7B, and in the collation score obtained in step S15 and step S22, whether there is a collation score that approximates each score maximum value. It is determined whether or not (step S23). Here, the comparison score obtained in step S15 is compared with the highest score of User014, and the comparison score obtained in step S22 is compared with the highest score of User001, User009, and User012. Approximating to the highest score means, for example, that the matching score is within a predetermined range (for example, highest score to highest score−0.5) with respect to the highest score. Within the predetermined range with respect to the highest score value corresponds to being above the first threshold value. The same applies hereinafter.

  In the example of FIG. 7C, since there is no matching score that approximates the highest score value, “No” is determined in step S23. When it is determined as “No” in Step S23, the user extraction unit 40 saves the user having the largest collation score other than User014 in the takeover user table 65 (Step S24). FIG. 5B illustrates the takeover user table 65. In the example of FIG. 7C, User012 is stored in the takeover user table 65 as the takeover user.

  Next, the user extraction unit 40 refers to the similar user table 63 and determines whether there is a user in the same group as the saved user (step S25). If it is determined “No” in step S25, step S12 is executed again. In this case, User014, User001, User009, and User012 are treated as collated. Therefore, collation with User007 is performed.

  When it is determined as “Yes” in Step S25, the user extraction unit 40 extracts a user in the same group as the saved user (Step S26). In the example of FIG. 9C, since User014 has been processed, User018 is extracted. Next, the rearrangement unit 50 moves the user extracted in step S26 to the top of the registration table 61 (step S27). In the example of FIG. 9A, User018 has moved to the top. Thereafter, the process is executed again from step S12.

  FIG. 10C shows another example of a matching score with User001, User009, and User012 in step S22. In the example of FIG. 10C, the matching score of User001 is 8.8, which is close to the highest score of User001 exemplified in FIG. 10B. Further, the matching score of User012 is 8.9, which is approximate to the highest score of User012 illustrated in FIG. Accordingly, “Yes” is determined in step S23.

  When it is determined as “Yes” in step S23, the user extracting unit 40 refers to the similar user table 63, and a user (second user) of the same group of the user (first user) having a matching score that approximates the highest score value. ) Is extracted (step S28). In the example of FIG. 11C, users in the same group as User001 and User012 are extracted as second users. That is, User 009, User 014, and User 018 are extracted as the second user. Since User 009 and User 014 have already been verified, User 018 is extracted as the second user. Next, the collation unit 30 collates the registered biometric information of all users (unmatched users) extracted in step S27 with the biometric information for collation (step S29). FIG. 11C illustrates a matching score with User001, User009, User012, User014, and User018.

  Next, the matching unit 30 selects the user with the highest matching score from the similar user group of User001. In the example of FIG. 11C, User001 is selected. Furthermore, the collation part 30 selects the user with the highest collation score from the similar user group of User012. In the example of FIG. 11C, User018 is selected. When there is one first user, the user with the highest matching score is selected from the similar user group of the first user. In the example of FIG. 11C, since there are a plurality of first users, the user with the largest matching score is selected. Next, the collation unit 30 refers to the score table 62, and determines whether or not the largest collation score approximates the highest score of the user related to the collation score (step S30).

  When it is determined as “No” in Step S30, the output unit 70 outputs a signal related to the authentication failure (Step S31). Thereby, the display device 104 performs display related to the authentication failure. When it is determined as “Yes” in step S30, the output unit 70 outputs a signal related to the authentication success, assuming that the person to be authenticated is the user (User018) having the highest matching score (step S32). ). Thereby, the display device 104 performs display related to the authentication success. If the collation score in this case is higher than the highest score value in the score table 62, the updating unit 80 may update the highest score value in the score table 62 with the collation score.

(Post-processing)
FIG. 12 is a flowchart illustrating an example of the post-processing. As illustrated in FIG. 12, the user extracting unit 40 refers to the similar user table 63 and extracts a user group of the target biological information (Step S41). In the example of FIG. 13C, User001, User009, and User012 are extracted. Next, the rearrangement unit 50 moves the user extracted in step S41 to the end of the registration table 61 (step S42). Thereafter, step S12 is executed. In this case, verification with User007 is performed next to User014.

(Adjustment process)
FIG. 14 is a flowchart illustrating an example of the adjustment process. The adjustment process is executed when authentication is successful. The log table 66 stores the postponed user and the collation score between the registered biometric information of the user and the biometric information for collation each time the postponing process described above is executed. FIG. 5C illustrates the log table 66.

  As illustrated in FIG. 14, the updating unit 80 refers to the log table 66 (step S51). Next, the update unit 80 determines whether or not the user related to the authentication success has been postponed (step S52). When it is determined as “Yes” in Step S52, the dissimilarity reference value is updated with the collation score of the postponed user (Step S53). After the execution of step S53 or when it is determined “No” in step S52, the execution of the flowchart ends.

  In the present embodiment, in the similar user collation process, the similarity (first similarity) between the registered biometric information of the user (first user) included in the similar user group and the biometric information for collation is calculated. Further, when the calculated first similarity approximates the highest score (when it is equal to or higher than the first threshold), the similar user table 63 indicates the similar user group associated with the user related to the first similarity. A user (second user) is extracted. Further, the similarity (second similarity) between the registered biometric information of the second user and the biometric information for verification is calculated. In addition, the user having a higher similarity among the first similarity and the second similarity is determined to be the person to be authenticated. In such a configuration, when the first similarity is equal to or greater than the first threshold, the collation target is narrowed down to similar user groups in the similar user table 63. That is, the number of candidates for verification can be suppressed. Therefore, the authentication time required per person can be shortened. The registered biometric information of similar users included in the similar user group has a similar relationship. Therefore, the selection accuracy of the candidate to be verified is improved, and the authentication accuracy can be maintained.

  Further, when the first similarity is equal to or less than the dissimilarity reference value, the similar user group of the user is postponed in the registration table 61, so that the time required for successful authentication can be shortened.

  Moreover, the reliability of a 1st threshold value becomes high by using the predetermined range of the highest score value for every user as a 1st threshold value. The first threshold value is not limited to the highest score value. A value within a predetermined range may be used for the similarity between the biometric information for verification acquired before the process of calculating the first similarity and the registered biometric information.

  The first threshold value is preferably a value indicating a state that is more similar to the second threshold value. This is because the number of similar users in the similar user group increases, and the authentication accuracy is maintained.

  Further, the registration table 61 is rearranged in the descending order of the number of similar users, so that the biometric information for collation and the registered biometric information are collated in the order of the number of similar users. As a result, the number of persons to be collated increases when the first similarity is equal to or greater than the first threshold. As a result, the authentication accuracy is maintained.

  In the above embodiment, the living body to be used is not limited to the fingerprint. A vein pattern, iris, face, palm shape, or the like may be used. The similarity reference value in the registration process may be different from the similarity reference value in the authentication process.

  In the above embodiment, the registration table 61 of the storage unit 60 functions as an example of the feature information storage unit. The collation unit 30 functions as an example of a first calculation unit, a second calculation unit, and a determination unit. The similar user table 63 of the storage unit 60 functions as an example of the similar user storage unit. The user extraction unit 40 functions as an example of an extraction unit.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and changes can be made within the scope of the gist of the present invention described in the claims. It can be changed.

DESCRIPTION OF SYMBOLS 10 Acquisition part 20 Registration part 30 Collation part 40 User extraction part 50 Rearrangement part 60 Storage part 61 Registration table 62 Score table 63 Similar user table 64 Reference value table 65 Takeover user table 66 Log table 70 Output part 80 Update part 100 Authentication process apparatus

Claims (9)

A process of obtaining biometric information;
The feature information storage unit that stores the user and the feature data in association with each other is referred to, and the first similarity between the feature data of the first user included in the feature information storage unit and the biological information is calculated. Processing,
When the calculated first similarity is equal to or greater than a first threshold, a second user associated with feature data whose similarity with the first user's feature data is equal to or greater than a second threshold is defined as the first user. A process of extracting the second user by referring to a similar user storage unit stored in association with
A process of calculating a second similarity between the feature data of the second user and the biological information;
An authentication processing program that causes a computer to execute a process of determining a user having a higher value of the first similarity and the second similarity as a user corresponding to the biological information.   When the calculated first similarity is equal to or less than a third threshold value, the second user's feature data and the biological information are collated with respect to another user included in the feature information storage unit. The authentication processing program according to claim 1, wherein the authentication processing program is executed after the verification processing.   The first threshold value is calculated between the biological information acquired from the first user and the feature data of the first user stored in the feature information storage unit before the process of acquiring the biological information. 3. The authentication processing program according to claim 1, wherein the similarity is a value within a predetermined range.   The first threshold value is calculated between the biological information acquired from the first user and the feature data of the first user stored in the feature information storage unit before the process of acquiring the biological information. 3. The authentication processing program according to claim 1, wherein the authentication processing program is a value within a predetermined range with respect to the highest value of the similarity.   The authentication processing program according to any one of claims 1 to 4, wherein the first threshold value is a value indicating a state that is more similar to the second threshold value. The similar user storage unit stores, in association with each other, a user whose similarity between the feature data stored in the feature information storage unit is equal to or greater than the second threshold as a similar user group,
The authentication processing program according to claim 1, wherein the first user is a user of any one of the similar user groups stored in the similar user storage unit.   The calculation process of the first similarity is performed by selecting the calculation target of the first similarity in descending order of the number of similar users from the similar user group stored in the similar user storage unit. 6. The authentication processing program according to 6. An acquisition unit for acquiring biological information;
A feature information storage unit that stores the user and feature data in association with each other;
A first calculation unit that refers to the feature information storage unit and calculates a first similarity between the feature data of the first user included in the feature information storage unit and the biological information;
A similar user storage unit that stores a second user associated with feature data whose similarity with the first user's feature data is equal to or greater than a second threshold value, in association with the first user;
An extraction unit that extracts the second user by referring to the similar user storage unit when the first similarity calculated by the first calculation unit is equal to or greater than a first threshold;
A second calculator that calculates a second similarity between the feature data of the second user and the biological information;
An authentication processing apparatus comprising: a determination unit that determines a user having a higher value among the first similarity and the second similarity as a user corresponding to the biological information. The biometric information is acquired by the acquisition unit,
The feature information storage unit that stores the user and the feature data in association with each other is referred to, and the first similarity between the feature data of the first user among the users included in the feature information storage unit and the biological information is first The calculation unit calculates,
When the calculated first similarity is equal to or greater than a first threshold, a second user associated with feature data whose similarity with the first user's feature data is equal to or greater than a second threshold is defined as the first user. The extraction unit extracts the second user by referring to the similar user storage unit stored in association with
A second calculator that calculates a second similarity between the feature data of the second user and the biological information;
The authentication processing method, wherein the determination unit determines a user having a higher value among the first similarity and the second similarity as a user corresponding to the biological information.

Images