WO2019180866A1

WO2019180866A1 - Authentication device, authentication method, and authentication program

Info

Publication number: WO2019180866A1
Application number: PCT/JP2018/011340
Authority: WO
Inventors: 利和金岡; 徹上和田
Original assignee: 富士通株式会社
Priority date: 2018-03-22
Filing date: 2018-03-22
Publication date: 2019-09-26

Abstract

[Problem] To improve the accuracy of authentication processing. [Solution] According to one aspect of the present invention, an authentication device acquires a plurality of images that were sequentially captured by an image capture device within a predetermined period, and uses an object image included in a first captured image of the acquired plurality of captured images to update a feature quantity used to authenticate a specific object, if the degree of similarity between a feature quantity of the object image included in the first captured image and the feature quantity used to authenticate the specific object does not meet a criterion and the degree of similarity between a feature quantity of an object image included in a second captured image of the plurality of captured images and the feature quantity used to authenticate the specific object meets the criterion.

Description

Authentication device, authentication method, authentication program

The present invention relates to authentication technology.

There is a face authentication technique for authenticating a user shown in a captured face image based on a face image picked up by an image pickup device and an authentication face image (template) registered in advance for each user. In the face authentication process, authentication is performed according to the degree of similarity (degree of correlation) between the feature amount of the captured face image (for example, the position of the feature point of the face) and the feature amount of the face image for authentication. Success or failure is determined. When the similarity between the feature amount of the captured face image and the feature amount of the authentication face image is high, the reliability of the captured face image is high, and the authentication is successful. On the other hand, when the similarity between the feature amount of the captured face image and the feature amount of the authentication face image is low, the reliability of the captured face image is low, and authentication fails.

Generally, as the number of authentication images (templates) for a user increases, the rejection rate for that user decreases (improved accuracy). This is because, for example, if a plurality of face images having different angles and facial expressions are registered, the amount of information used for authentication increases, and more cases can be covered. On the other hand, the smaller the face image for authentication, the smaller the amount of information used for authentication, and the identity rejection rate increases (accuracy deteriorates). Therefore, it is preferable that a certain number of face images for authentication with respect to the user is registered in an exhaustive manner.

However, it is not always easy to comprehensively register authentication images for the user in advance in view of versatility such as angles and facial expressions and user convenience. Therefore, there is a technique for adding an authentication image at the time of authentication. As an example of such a conventional technique, a technique for additionally registering highly reliable biometric information acquired from a user in authentication processing as biometric information for authentication when the authentication process determines that the user is the user is proposed. Has been.

International Publication No. 2012/144105

The prior art is capable of additionally registering biometric information with high reliability (similarity) with respect to biometric information already registered. However, in reality, if the influences such as aging of the face, variation of facial expressions, imaging angle, imaging environment (for example, illumination brightness and color) are taken into account, the reliability of the already registered biometric information is considered. A high degree of biological information alone is not always sufficient. According to this point, it is considered that the conventional technique in which a face image whose reliability is determined to be low is not additionally registered has few variations in biometric information (template) for authentication, and there is a shortage in the rejection rate. .

On the other hand, there is a problem in unconditionally registering face images that are judged to have low reliability. Registration of face images with low reliability may reduce the rejection rate of the person, but on the other hand, registration of other person's face image as biometric information for authentication may increase the acceptance rate of others There is. Thereby, it is assumed that authentication accuracy will fall.

In one aspect, an object of the present invention is to improve the accuracy of authentication processing.

According to an aspect of the present invention, the authentication device acquires a plurality of captured images sequentially captured within a predetermined period by the imaging device, and a subject included in the first captured image among the acquired plurality of captured images. The similarity between the feature amount of the image and the feature amount used for authentication of the specific subject does not satisfy the standard, and the feature amount of the subject image included in the second captured image among the plurality of captured images; When the similarity with the feature amount used for authentication of the specific subject satisfies the standard, the feature amount used for authentication of the specific subject is updated using the image of the subject included in the first captured image.

In one aspect, the present invention can improve the accuracy of authentication processing.

FIG. 1 is a diagram illustrating a configuration example of a network according to the first embodiment. FIG. 2 is a diagram illustrating a hardware configuration example of the authentication apparatus. FIG. 3 is a diagram illustrating an example of a functional configuration of the authentication apparatus. FIG. 4 is a diagram illustrating an example of the reliability table. FIG. 5 is a flowchart illustrating an example of the authentication processing procedure of the authentication apparatus. FIG. 6 is a diagram illustrating an example of the reliability table. FIG. 7 is a flowchart illustrating an example of the authentication processing procedure of the authentication apparatus. FIG. 8 is a flowchart illustrating an example of the authentication processing procedure of the authentication apparatus. FIG. 9 is a diagram illustrating a hardware configuration example of the authentication apparatus. FIG. 10 is a flowchart illustrating an example of the authentication processing procedure of the authentication apparatus.

Hereinafter, each embodiment will be described in detail with reference to the drawings. Each process in each embodiment can be appropriately changed. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiments, and the repetitive description thereof will be omitted.

In each embodiment, face authentication is described as an example of authentication. However, the present invention is not limited to this, and biometric authentication using other biometric information or object authentication using an image may be used. The same applies.

[First Embodiment]
The first embodiment will be described below. FIG. 1 is a diagram illustrating a configuration example of a network according to the first embodiment. The network according to the first embodiment includes a terminal 10 and an authentication device 11, and each is connected via a network 12. The number of terminals 10 and authentication devices 11 is not limited to the number illustrated, and a large number may be installed.

The terminal 10 is a general computer such as a personal computer, a smartphone, or a tablet terminal, and includes an imaging device 13. The terminal 10 is not limited to the above example, and may be a device used for a specific purpose such as a robot that communicates with a user or an ATM (Automatic Teller Machine) as long as it includes an imaging device. good.

For example, a CMOS (Complementary Metal Oxide Semiconductor) camera or a CCD (Charge Coupled Device) camera may be used as the imaging device 13.

The terminal 10, for example, captures a captured image of the user as a subject at a constant interval (for example, 30 images per second) by the imaging device 13, and transmits the captured image to the authentication device 11.

In the present embodiment, the terminal 10 and the authentication device 11 are described as separate devices. However, the present invention is not limited to this, and a single authentication device 11 having the imaging device 13 may be used.

FIG. 2 shows a hardware configuration example of the authentication device 11 according to the first embodiment. The authentication device 11 includes a control device 21, a storage device 22, and a network connection device 23, which are connected via a system bus 24 so that signals can be input and output.

The control device 21 is a device that controls the authentication device 11. For example, a CPU (Central Processing Unit) may be used as the control device 21. The control device 21 controls the operation of the authentication device 11 such as various operations and data input / output with each hardware component based on an OS (Operating System) and various programs stored in the storage device 22. Perform various processes. Various information required during the execution of the program can be acquired from the storage device 22, for example. Note that part of the processing performed by the control device 21 may be realized using dedicated hardware. The control device 21 executes the authentication process according to the first embodiment based on the authentication program according to the first embodiment.

The storage device 22 stores, for example, the authentication program according to the embodiment. The storage device 22 may include a main storage device and an auxiliary storage device. For example, the main storage device temporarily stores at least a part of the OS and application programs to be executed by the control device 21. The main storage device stores various data necessary for processing by the control device 21. As the main storage device, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), or the like may be used.

The auxiliary storage device can read various stored information and write information based on an instruction from the control device 21. As the auxiliary storage device, for example, a storage such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), or the like may be used. The auxiliary storage device may store information used in the authentication process according to the first embodiment and the result of the process. Further, the main storage device and the auxiliary storage device may share each other's functions.

The network connection device 23 is, for example, a NIC (Network Interface Card).

FIG. 3 is a diagram illustrating an example of a functional configuration of the authentication device 11 according to the first embodiment. The authentication device 11 has a control unit 31, a storage unit 32, and a network connection unit 33 as functions. Functions of the control unit 31, the storage unit 32, and the network connection unit 33 are obtained by the control device 21 reading a program stored in the storage device 22 and executing various processes using hardware included in the server device 11. Is realized.

具体 Illustrate the specific operation of the authentication device 11. The authentication device 11 accepts registration of a face image (template) for user authentication (user registration), and stores it in the storage unit 32 in association with user identification information. A face image is an example of a subject image. In user registration, a plurality of authentication face images may be registered for each user. The control unit 31 of the authentication device 11 calculates an authentication feature amount (for example, the position of a feature point on the face) for each user based on the user authentication face image stored in the storage unit 32. The feature value for authentication can be calculated by machine learning using a known technique such as principal component analysis or support vector machine.

The control unit 31 of the authentication device 11 detects the face image of the subject included in the captured image received from the terminal 10. The control unit 31 extracts a feature amount from the detected face image and compares it with the authentication feature amount stored in the storage unit 32. When the control unit 31 detects a specific user whose reliability (similarity or correlation) between the feature quantity of the face image and the authentication feature quantity is equal to or greater than a threshold as a result of the comparison, the subject included in the captured image is specified. Authenticate that you are a user. In addition, when detecting the specific user whose reliability is equal to or higher than the threshold, the control unit 31 detects the specific user whose average reliability of each of the plurality of face images sequentially captured is equal to or higher than the threshold. You may do it.

Note that although there are generally one-to-N authentication and one-to-one authentication as authentication methods, the first embodiment may be any authentication method. In the case of one-to-one authentication, the authentication device 11 receives user identification information, calculates the reliability between the authentication feature quantity corresponding to the received user identification information and the face image of the subject, and the reliability is a threshold value. You may determine whether it is above.

When the specific user is authenticated, the control unit 31 of the authentication device 11 according to the first embodiment is another captured image whose imaging time interval with the captured image used for authentication is within a predetermined range. Thus, the presence / absence of another captured image whose reliability with the feature amount of the face area included in the other captured image is less than a threshold is determined.

Here, the imaging time interval may be a relative time interval or an absolute time interval. As the relative time interval, for example, the number of images captured between two captured images may be used. As an absolute time interval, for example, a difference in real time between imaging timings of two captured images may be used.

When the control unit 31 of the authentication device 11 detects another captured image having a reliability less than the threshold value, the control unit 31 additionally registers the other captured image as a face image for authentication of a specific user. When a face image for authentication of a specific user is newly registered additionally, the control unit 31 authenticates the specific user based on the face image for authentication of the specific user including the additionally registered face image. Feature amounts may be calculated (updated).

In the above-described conventional technology, only biometric information with high reliability (similarity) with respect to already registered biometric information is additionally registered as biometric information for authentication, and variations in biometric information for authentication are scarce. It is likely that there is a lack of authentication accuracy. On the other hand, in the authentication method according to the first embodiment, even when biometric information with low reliability with respect to already registered biometric information is obtained, biometric information with high reliability is obtained before and after that. The biometric information for authentication is additionally registered. Thereby, since the variation of the biometric information for authentication is ensured, the person rejection rate can be reduced.

In the authentication method according to the first embodiment, among biometric information with low reliability alone, only information with an acquisition interval between biometric information with high reliability within a predetermined range is additionally registered as biometric information for authentication. Is done. As a result, only biometric information with a certain level of reliability is additionally registered based on the context of the biometric information with low reliability. Therefore, biometric information with low reliability is additionally registered unconditionally. The acceptance rate of others can be reduced compared to the case of doing so. Therefore, according to the first embodiment, biometric authentication having high authentication accuracy can be realized.

The following is a more specific explanation. FIG. 4 is a diagram illustrating an example of the reliability table. The reliability table 40 includes columns of face image ID and reliability. The face image ID column stores identification information of the face image. For example, the captured image including the face image with the face image ID t + 1 is a captured image captured immediately after the captured image including the face image with the face image ID t.

In the reliability column, the reliability (similarity) with the feature quantity for authentication of a specific user for each face image is stored. For example, the upper limit may be set to 100, and the reliability may be a value closer to 100 as the feature amount of the face image and the feature amount for authentication of a specific user are similar. Further, the threshold value for authentication may be set to 90, for example.

When the control unit 31 of the authentication device 11 authenticates that the subject included in the captured image is a specific user, the feature amount of the face image extracted from the captured image received from the terminal 10 and the feature amount for authentication of the specific user May be stored in the reliability table 40 in association with the face image ID. The control unit 31 may refer to the reliability table 40 and detect a face image with a face image ID t + 4 having a reliability less than a threshold value (for example, 90). The control unit 31 may additionally register the face image with the detected face image ID t + 4 in the storage unit 32 as a face image for authentication of a specific user. The control unit 31 may set a lower limit value of the reliability of the face image additionally registered as the authentication face image.

Also, the control unit 31 of the authentication device 11 may transmit candidate face images to be additionally registered to the terminal 10. The terminal 10 may display a face image, which is a candidate for additional registration, on the display device and accept whether or not additional registration is possible from the user. When the control unit 31 of the authentication device 11 receives a notification permitting additional registration from the terminal 10, the control unit 31 may perform additional registration.

Next, an example of an authentication processing procedure of the authentication device 11 according to the first embodiment will be described. FIG. 5 is a flowchart illustrating an example of the authentication processing procedure of the authentication device 11.

For example, the control unit 31 of the authentication device 11 receives a plurality of captured images sequentially captured by the imaging device 13 from the terminal 10 (S101). The control unit 31 extracts a subject image included in each of the plurality of captured images (S102). The control unit 31 calculates the feature amount of the subject image for each of the plurality of subject images (S103). The control unit 31 refers to the storage unit 32 that stores the feature value for authentication for each user, and whether there is a user whose reliability between the calculated feature value of the subject image and the feature value for authentication is equal to or greater than a threshold value. Is determined (S104).

When the specific user whose reliability is equal to or higher than the threshold is not detected (No in S104), the control unit 31 ends the authentication process. In this case, the control unit 31 may transmit information indicating that the authentication has failed to the terminal 10. The terminal 10 may display information indicating that the authentication has failed on the display device.

When a specific user whose reliability is equal to or higher than the threshold is detected (S104 Yes), the control unit 31 authenticates that the user shown in the subject image is the specific user (S105). In this case, the control unit 31 may transmit information indicating that the authentication has been successful or information identifying a specific authenticated user to the terminal 10. The terminal 10 may display information indicating that the authentication has been successful or information identifying a specific authenticated user on the display device. The control unit 31 determines whether or not a plurality of subject images include subject images whose reliability with a specific user authentication feature amount is less than a threshold (S106).

If the image of the subject whose reliability is less than the threshold is not included (No in S106), the control unit 31 ends the authentication process.

When an image of a subject whose reliability is less than the threshold is included (S106 Yes), the control unit 31 authenticates an image of the subject whose reliability is less than the threshold among the plurality of subjects. Is additionally registered in the subject image (S107). The control unit 31 calculates (updates) a feature amount for authentication of a specific user by using the subject image for authentication of the specific user after the additional registration (S108). The control unit 31 stores the calculated feature quantity for authentication of the specific user in the storage unit 32, and ends the authentication process.

As described above, according to the present embodiment, the authentication device 11 uses, for authentication, a face image whose reliability (similarity) is less than the threshold among the face images included in the captured image received when performing user authentication. Add to the face image. Thereby, although the subject is the user himself / herself, the face image whose reliability is calculated to be lower than the threshold value is used when calculating the feature value for authentication, so that the subsequent authentication accuracy is improved.

(Modification 1)
Next, Modification 1 of the first embodiment will be described. In the description of the first modification, the same portions as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted. In the first embodiment, the authentication apparatus 11 additionally registers a face image having a reliability less than the threshold as an authentication face image. However, in the first modification, a condition is further added.

For example, the control unit 31 of the authentication device 11 has at least the reliability of the face image of the captured image captured immediately before or the reliability of the face image of the captured image captured immediately after the face image with the reliability less than the threshold value. If the degree is greater than or equal to the threshold, a face image with a reliability less than the threshold may be additionally registered as a face image for authentication.

For example, the control unit 31 of the authentication device 11 has the reliability of the face image of the captured image captured before and the reliability of the face image of the captured image captured later of the face image with the reliability less than the threshold. If both are greater than or equal to the threshold value, a face image with a reliability less than the threshold value may be additionally registered as an authentication face image.

Here, the captured image captured before and the captured image captured later are not limited to the captured images captured immediately before or immediately after, but a plurality of images sequentially captured within the predetermined period by the terminal 10. It may be included in the captured image.

(Modification 2)
Next, a second modification of the first embodiment will be described. In the description of the second modification, the same portions as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted. In the first embodiment, the authentication apparatus 11 additionally registers a face image having a reliability less than the threshold value as a face image for authentication. However, in the second modification, a condition is further added.

For example, the control unit 31 of the authentication apparatus 11 includes, for example, the reliability of face images of a plurality of captured images captured before and after a face image with a reliability less than a threshold, and the reliability of a face image with a reliability less than a threshold. If the average reliability is greater than or equal to the threshold, a face image with a reliability less than the threshold may be additionally registered as a face image for authentication.

FIG. 6 is a diagram illustrating an example of the reliability table. The reliability table 60 is obtained by adding an average column to the reliability table 40.

The average column stores a value obtained by averaging the reliability of a plurality of face images taken before and after the corresponding face image. For example, the control unit 31 may calculate the average reliability of each of a plurality of face images including the five face images captured immediately before and the five face images captured immediately thereafter. For example, the control unit 31 calculates and stores the average reliability of the face images from the face image ID t + 1 to the face image ID t + 11 in the average column of the face image records with the face image ID t + 6. You may do it.

When authenticating a specific user based on the captured image received from the terminal 10, the control unit 31 of the authentication device 11 refers to the reliability table 60, is a face image having a reliability less than the threshold, and the average is the threshold The presence or absence of the face image as described above may be determined. For example, the face image with the face image ID t + 12 has a reliability of 80 and is less than a threshold value (for example, 90), and the average is 92 and is equal to or more than the threshold value. Are detected and registered in the face image for authentication of a specific user.

On the other hand, for example, the face image with the face image ID t + 19 has a reliability of 47 and is less than a threshold value (for example, 90), and the average is also 64.1 and is less than the threshold value. The t + 19 face image is not additionally registered as a face image for authentication of a specific user.

Next, an example of an authentication processing procedure of the authentication device 11 according to the second modification will be described. FIG. 7 is a flowchart illustrating an example of the authentication processing procedure of the authentication device 11.

For example, the control unit 31 of the authentication device 11 receives a plurality of captured images sequentially captured by the imaging device 13 from the terminal 10 (S201). The control unit 31 extracts a subject image included in each of the plurality of captured images (S202). The control unit 31 calculates the feature amount of the subject image for each of the plurality of subject images (S203). The control unit 31 refers to the storage unit 32 that stores the feature value for authentication for each user, and whether there is a user whose reliability between the calculated feature value of the subject image and the feature value for authentication is equal to or greater than a threshold value. Is determined (S204).

When the specific user whose reliability is equal to or higher than the threshold is not detected (No in S204), the control unit 31 ends the authentication process. In this case, the control unit 31 may transmit information indicating that the authentication has failed to the terminal 10. The terminal 10 may display information indicating that the authentication has failed on the display device.

When a specific user whose reliability is equal to or higher than the threshold is detected (Yes in S204), the control unit 31 authenticates that the user shown in the image of the subject is the specific user (S205). In this case, the control unit 31 may transmit information indicating that the authentication has been successful or information identifying a specific authenticated user to the terminal 10. The terminal 10 may display information indicating that the authentication has been successful or information identifying a specific authenticated user on the display device. The control unit 31 determines whether or not a plurality of subject images include images of subjects whose reliability with a specific user authentication feature amount is less than a threshold (S206).

When the image of the subject whose reliability is less than the threshold is not included (No in S206), the control unit 31 ends the authentication process.

When the image of the subject whose reliability is less than the threshold is included (S206 Yes), the control unit 31 includes each of a plurality of subject images including the image of the subject captured before and after the image of the subject whose reliability is less than the threshold. It is determined whether the average reliability is equal to or greater than a threshold value (S207).

When the average is less than the threshold value (No in S207), the control unit 31 returns to S206 and determines whether or not an image whose reliability is less than the threshold value is included.

When the average is equal to or greater than the threshold (Yes in S207), the control unit 31 determines the subject image for authentication of the authenticated specific user as the subject image whose reliability is less than the threshold and the average reliability is equal to or greater than the threshold. Are additionally registered in the image (S208). The control unit 31 calculates (updates) a feature amount for authentication of a specific user using the subject image for authentication of the specific user after the additional registration (S209). The control unit 31 stores the calculated feature quantity for authentication of the specific user in the storage unit 32, and ends the authentication process.

As described above, according to the second modification, in order to determine whether or not additional registration of a face image is possible based on the average reliability of a plurality of images captured before and after, the reliability is increased when the captured user is switched. It is possible to prevent the face image of another user calculated with a low degree from being additionally registered by mistake.

(Modification 3)
Next, Modification 3 of the first embodiment will be described. In the description of the modified example 3, the same parts as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted. In the first embodiment, the authentication apparatus 11 additionally registers a face image having a reliability less than the threshold as an authentication face image. However, in the third modification, a condition is further added.

For example, if the control unit 31 of the authentication apparatus 11 detects a face image whose reliability is less than the threshold after authenticating a specific user, the reliability with the face image whose reliability is less than the threshold is equal to or greater than the threshold. The presence or absence of a user may be determined. When there is no other user whose reliability is equal to or higher than the threshold value, the control unit 31 may additionally register a face image whose reliability is less than the threshold value to the authentication user's face image.

Next, an example of an authentication processing procedure of the authentication device 11 according to the modification 3 will be described. FIG. 8 is a flowchart illustrating an example of an authentication processing procedure of the authentication device 11.

For example, the control unit 31 of the authentication device 11 receives a plurality of captured images sequentially captured by the imaging device 13 from the terminal 10 (S301). The control unit 31 extracts a subject image included in each of the plurality of captured images (S302). The control unit 31 calculates the feature amount of the subject image for each of the plurality of subject images (S303). The control unit 31 refers to the storage unit 32 that stores the feature value for authentication for each user, and whether there is a user whose reliability between the calculated feature value of the subject image and the feature value for authentication is equal to or greater than a threshold value. Is determined (S304).

When the specific user whose reliability is equal to or higher than the threshold is not detected (No in S304), the control unit 31 ends the authentication process. In this case, the control unit 31 may transmit information indicating that the authentication has failed to the terminal 10. The terminal 10 may display information indicating that the authentication has failed on the display device.

When a specific user whose reliability is equal to or higher than the threshold is detected (Yes in S304), the control unit 31 authenticates that the user shown in the subject image is the specific user (S305). In this case, the control unit 31 may transmit information indicating that the authentication has been successful or information identifying a specific authenticated user to the terminal 10. The terminal 10 may display information indicating that the authentication has been successful or information identifying a specific authenticated user on the display device. The control unit 31 determines whether or not a plurality of subject images include subject images whose reliability with a specific user authentication feature amount is less than a threshold (S306).

If the image of the subject whose reliability is less than the threshold is not included (No in S306), the control unit 31 ends the authentication process.

When the image of the subject whose reliability is less than the threshold is included (S306 Yes), the control unit 31 determines the reliability between the feature amount of the subject image whose reliability is less than the threshold and the feature amount for authentication of other users. Then, it is determined whether there is another user whose reliability with the image of the subject whose reliability is less than the threshold is greater than or equal to the threshold (S307).

When there is another user whose reliability is greater than or equal to the threshold (S307 Yes), the control unit 31 does not include the subject included in the subject image whose reliability with the authenticated specific user is less than the threshold. The authentication process is terminated without performing additional registration.

When there is no other user whose reliability is equal to or higher than the threshold (No in S307), the control unit 31 additionally registers an image of the subject whose reliability is less than the threshold as an image of the subject for authentication of the authenticated specific user. (S308). The control unit 31 calculates (updates) a feature amount for authentication of a specific user using the subject image for authentication of the specific user after the additional registration (S309). The control unit 31 stores the calculated feature quantity for authentication of the specific user in the storage unit 32, and ends the authentication process.

As described above, according to the third modification, after authentication of a specific user, an image of a subject whose reliability is less than the threshold is determined to determine whether the reliability with the feature amount for authentication of other users is greater than or equal to the threshold. For example, when the user is replaced, the face image of another user whose reliability is calculated to be low is prevented from being erroneously registered.

Modification 1 to Modification 3 can be combined within a range that does not contradict each other. Furthermore, Modification Examples 1 to 3 can also be applied to the second embodiment described below.

[Second Embodiment]
Hereinafter, a second embodiment will be described. In the description of the second embodiment, the same portions as those in the first embodiment are denoted by the same reference numerals, and description thereof is omitted. FIG. 9 illustrates a hardware configuration example of the authentication device 11 according to the second embodiment. In the second embodiment, the authentication device 11 includes the imaging device 13 in its own device.

Similarly to the first embodiment, the authentication device 11 according to the second embodiment also executes processing such as calculation of a feature amount for user authentication, calculation of the reliability of a face image, and additional registration of a face image. Further, after authentication of a specific user, the control unit 31 of the authentication device 11 may change the setting of the imaging device 13 and further acquire the captured image when a face image having a reliability less than the threshold value is not detected. The control unit 31 may restore the setting of the captured image 13 when a predetermined time (for example, 1 second) elapses after the setting is changed.

The change in the setting of the imaging device 13 includes, for example, changing the exposure, changing the imaging direction, defocusing, and the like. The brightness of the captured image can be changed by changing the exposure. By changing the imaging direction, the direction of the subject shown in the captured image can be changed. The degree of blur of the captured image can be changed by defocusing.

Note that the control unit 31 of the authentication device 11 may set an upper limit for the number of times the setting of the imaging device 13 is changed, and may end the authentication process when the number of times of the setting change reaches the upper limit.

Next, an example of an authentication processing procedure of the authentication device 11 according to the second embodiment will be described. FIG. 10 is a flowchart illustrating an example of the authentication processing procedure of the authentication device 11.

For example, the control unit 31 of the authentication device 11 acquires a plurality of captured images sequentially captured by the imaging device 13 (S401). The control unit 31 extracts a subject image included in each of the plurality of captured images (S402). The control unit 31 calculates the feature amount of the subject image for each of the plurality of subject images (S403). The control unit 31 refers to the storage unit 32 that stores the feature value for authentication for each user, and whether there is a user whose reliability between the calculated feature value of the subject image and the feature value for authentication is equal to or greater than a threshold value. Is determined (S404).

When the specific user whose reliability is equal to or higher than the threshold is not detected (No in S404), the control unit 31 ends the authentication process. In this case, the control unit 31 may transmit information indicating that the authentication has failed to the terminal 10. The terminal 10 may display information indicating that the authentication has failed on the display device.

When a specific user whose reliability is equal to or higher than the threshold is detected (S404 Yes), the control unit 31 authenticates that the user shown in the image of the subject is the specific user (S405). In this case, the control unit 31 may transmit information indicating that the authentication has been successful or information identifying a specific authenticated user to the terminal 10. The terminal 10 may display information indicating that the authentication has been successful or information identifying a specific authenticated user on the display device. The control unit 31 determines whether or not a plurality of subject images include subject images whose reliability with a specific user authentication feature amount is less than a threshold (S406).

When the image of the subject whose reliability is less than the threshold is not included (No in S406), the control unit 31 changes the setting of the imaging device 13 (S407). After changing the setting of the imaging device 13, the control unit 31 returns to the process of S 401, acquires the captured image again, extracts the subject image included in the captured image, and authenticates the subject image and the specific user. The reliability with the feature amount is calculated, and it is determined whether or not the image of the subject whose reliability is less than the threshold is included.

When an image of a subject whose reliability is less than the threshold is included (S406 Yes), the control unit 31 authenticates an image of the subject whose reliability is less than the threshold among a plurality of subjects. Is additionally registered in the subject image (S408). The control unit 31 calculates (updates) a feature quantity for authentication of a specific user using the subject image for authentication of the specific user after additional registration (S409). The control unit 31 stores the calculated feature quantity for authentication of the specific user in the storage unit 32, and ends the authentication process.

As described above, according to the second embodiment, the authentication device 11 changes the setting of the captured image when there is no face image whose reliability is less than the threshold after authentication of the specific user, and thus the reliability is less than the threshold. It is possible to increase the possibility of acquiring a face image.

Note that the same processing as in the first to third modifications of the first embodiment may also be applied to the face image captured after the setting of the imaging device 13 is changed. For example, the control unit 31 of the authentication device 11 selects a face image whose reliability with a feature amount for authentication of a specific user is less than a threshold among the face images captured after the setting of the imaging device 13 is changed. If the reliability of face images taken before and after the detection is greater than or equal to a threshold, a face image with the reliability less than the threshold may be additionally registered as a face image for authentication of a specific user. good.

The present invention is not limited to the configurations and procedures of the above-described embodiments and modifications, and the processing method can be appropriately changed and recombined without departing from the spirit of the present invention.

10: terminal 11: authentication device 12: network 13: imaging device 21: control device 22: storage device 23: network connection device 24: system bus 31: control unit 32: storage unit 33: network connection unit 40, 60: reliability table

Claims

Obtaining a plurality of captured images sequentially captured within a predetermined period by the imaging device;
The similarity between the feature amount of the image of the subject included in the first captured image and the feature amount used for authentication of the specific subject among the acquired captured images does not satisfy the standard, and the plurality of captured images The subject included in the first captured image when the similarity between the feature amount of the image of the subject included in the second captured image of the image and the feature amount used for authentication of the specific subject satisfies the criterion Update the feature value used for authentication of the specific subject using the image of
An authentication apparatus comprising a control unit that executes processing.
The second captured image is a captured image captured immediately before or immediately after the first captured image among the plurality of captured images.
The authentication apparatus according to claim 1.
The process of updating the feature amount used for authentication of the specific subject includes the feature amount of the subject image included in the third captured image among the plurality of captured images, and the feature amount used for authentication of the specific subject. Is executed if the similarity of
The second captured image is a captured image captured before the first captured image among the plurality of captured images, and the third captured image is the first captured image among the plurality of captured images. 1 is a captured image captured after the captured image of 1.
The authentication apparatus according to claim 1.
In the process of updating the feature amount used for authentication of the specific subject, the average of the similarity with the feature amount used for authentication of the specific subject of each of the plurality of subject images included in the plurality of captured images is Executed if the criteria are met,
The authentication apparatus according to claim 1.
When the image of the subject of the plurality of captured images does not include an image of a subject whose similarity with the feature amount used for authentication of the specific subject does not satisfy the criterion, the imaging condition of the imaging device is changed.
The authentication apparatus according to claim 1.
Obtaining a plurality of captured images sequentially captured within a predetermined period by the imaging device;
The similarity between the feature amount of the image of the subject included in the first captured image and the feature amount used for authentication of the specific subject among the acquired captured images does not satisfy the standard, and the plurality of captured images The subject included in the first captured image when the similarity between the feature amount of the image of the subject included in the second captured image of the image and the feature amount used for authentication of the specific subject satisfies the criterion Update the feature value used for authentication of the specific subject using the image of
An authentication method, wherein a computer executes processing.
Obtaining a plurality of captured images sequentially captured within a predetermined period by the imaging device;
The similarity between the feature amount of the image of the subject included in the first captured image and the feature amount used for authentication of the specific subject among the acquired captured images does not satisfy the standard, and the plurality of captured images The subject included in the first captured image when the similarity between the feature amount of the image of the subject included in the second captured image of the image and the feature amount used for authentication of the specific subject satisfies the criterion Update the feature value used for authentication of the specific subject using the image of
An authentication program for causing a computer to execute processing.