JP6960738B2 - Face recognition system, face recognition method and face recognition program - Google Patents

Description

The present invention relates to a face recognition system, a face recognition method and a face recognition program.

There is known a face recognition system that collates face data consisting of a face image or its feature amount and authenticates the person to be authenticated as a registered authentication target person. In such a system, for example, at the entrance of a security-controlled area (such as a server room in a data center), the face of the person to be authenticated is photographed to acquire face data, and the face of the person to be authenticated registered in the system is acquired. Authentication is performed by collating with the data.

In such a face recognition system, a method has been proposed in which learning contents are added to the face data of the person to be authenticated at any time to improve the authentication accuracy.

In the biometric authentication device of Patent Document 1, dictionary information (characteristic information unique to a person) is updated in order to reduce the influence of factors that deteriorate authentication performance (changes in input environment, changes over time, etc.).

In the face image matching device of Patent Document 2, in addition to the initially registered face image registered as the reference face image in advance, an input face image having a certain degree of similarity with the initially registered face image is additionally registered as the reference face image. By doing so, the definition range (person definition range) to be authenticated with the face image of the registrant is expanded. As a result, the deterioration of the authentication accuracy is suppressed in response to changes in facial features and the environment.

Japanese Unexamined Patent Publication No. 2006-31103 Japanese Patent No. 5390228 Japanese Unexamined Patent Publication No. 2003-109010

The entry / exit authentication device of Patent Document 3 controls re-entry by registering a user's face pattern when temporarily exiting the security area and collating it with the face pattern at the time of re-entry. When the user re-enters, the registered face pattern is no longer needed and is deleted.

When the registration as the authentication target person is deleted in this way, the face data of the authentication target person is usually deleted. For example, if a person is allowed to enter the server room for a one-week maintenance period, at the end of the maintenance period, the person's registration as a person to be authenticated will be deleted and for authentication. Face data is no longer needed and will be deleted.

Here, in the above example, when the next maintenance period arrives and the person in charge is registered as the person to be authenticated, the same person as the person in charge of the previous maintenance period may be the person in charge. In this case, the same person will be registered as the person to be authenticated, but since the learning content related to the face data is lost due to the deletion of the face data at the time of (previous) deregistration, the previous maintenance period Authentication accuracy is lower than at the end.

The biometric authentication device of Patent Document 1 described above does not consider the handling of dictionary information when the registrant is deregistered. Then, even if the dictionary information is updated many times for a certain registrant and the authentication performance is improved, the dictionary information is deleted by the deregistration of the registrant. When the person is registered in the biometric authentication device again, the updated dictionary information is lost, so that the authentication performance is lower than that at the time of deregistration.

In the face image matching device of Patent Document 2 described above, the handling of the reference face image (added learning content) when the registrant is deregistered is not considered. Then, a large number of reference face images are added for a certain registrant, and even if the authentication accuracy is improved, the reference face image is deleted by deregistering the registrant. Then, when the person is registered in the face image matching device again, the added reference face image is lost, so that the authentication performance is lower than that at the time of deregistration.

The present invention has been made in view of the above problems, and provides a method capable of maintaining the past authentication performance when a person is re-registered as an authentication target person, in other words, recovering the past authentication performance. With the goal.

The face recognition system of the present invention
It is a face recognition system that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
A storage unit that stores the face data of the authentication target person and the face data of the non-authentication target person.
An authentication unit that collates the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
A deregistration unit that stores the face data of the authentication target person to be stored in the storage unit as the face data of the non-authentication target person.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific It is provided with a registration unit that stores the face data as the face data of the processing target person in the storage unit.

The face recognition method of the present invention
It is a face recognition method that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
An authentication process of collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
The deregistration process of storing the face data of the authentication target person in the storage unit as the face data of the non-authentication target person.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific It has a registration process of storing the face data as the face data of the person to be processed in the storage unit.

The face recognition program of the present invention
It is a face recognition program for collating face data consisting of face images or their feature amounts to authenticate the person to be authenticated as a registered authentication target person.
An authentication processing procedure for collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing procedure for storing the face data of the authentication target person as the face data of the non-authentication target person in the storage unit, and a deregistration processing procedure.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific The computer is made to execute the registration processing procedure of storing the face data as the face data of the processing target person in the storage unit.

According to the above configuration, when the registration of the authentication target person is deleted, the face data of the authentication target person is stored in the storage unit as the face data of the non-authentication target person. Then, the face data of the person who was the authentication target person remains in the storage unit as the face data of the non-authentication target person. And from the face data of the non-authenticated person, data on the processing subject (persons registered as authentication target person) are identified and registered as the processing subject of the face data to be authenticated person. As a result, since the face data of the person to be authenticated in the past is used for the authentication of the person to be processed, the authentication performance is improved as compared with the case where the past face data is not reused. That is, according to the above configuration, when a person is re-registered as an authentication target person, the past authentication performance can be maintained, in other words, the past authentication performance can be restored. Note that "maintaining past authentication performance" or "recovering" is not limited to the case of achieving exactly the same performance as past authentication performance. If months or years have passed since the person was previously certified, the hairstyle and physiognomy will change to some extent. Even in such a case, according to the above-mentioned feature configuration, the effect of improving the authentication performance can be obtained as compared with the case of registering the face data as a completely new authentication target person without reusing it.

In or face authentication system of the present invention, the registration unit collates with the face data of the processing subject, and the face data of the non-object's stored in the storage unit, the processing target It may be configured to identify the face data for a person.

According to the above configuration, since the face data relating to the processing target person is specified by collating the face data, it is preferable that the face data relating to the processing target person is surely specified.

In or face authentication system of the present invention,
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
It is preferable that the registration unit is configured to store the specified face data in the storage unit as the face data of the processing target person in a state of being associated with the personal identification data of the processing target person. The personal identification data is information that can identify an individual, such as a name and affiliation, and is information other than face data.

Further, in the face recognition system of the present invention,
The storage unit stores the face data of the non-authentication target person in association with the personal identification data of the non-authentication target person.
The registration unit collates the personal identification data of the processing target person with the personal identification data of the non-authentication target person stored in the storage unit to identify the face data of the processing target person. It may be configured as follows.

According to the above configuration, since the face data relating to the processing target person is specified by collating the personal identification data, it is preferable that the face data relating to the processing target person is surely specified.

Further, in the face recognition system of the present invention, the storage unit includes a first storage unit that stores the face data of the authentication target person and a second storage unit that stores the face data of the non-authentication target person. It may be configured as follows.

The face recognition system of the present invention
It is a face recognition system that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
A storage unit that stores the face data of the authentication target person and the face data of the non-authentication target person.
An authentication unit that collates the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
A deregistration unit that stores the face data of the authentication target person to be stored in the storage unit as the face data of the non-authentication target person.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. A registration unit that identifies the face data related to the above and stores the specified face data in the storage unit as the face data of the processing target person.
With
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
When the face data of the authentication target person whose registration is deleted is stored in the storage unit as the face data of the non-authentication target person, the registration deletion unit deletes the registration of the individual of the authentication target person. The identification data is deleted from the storage unit.
The face recognition method of the present invention
It is a face recognition method that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
An authentication process of collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing for storing the face data of the authentication target person in the storage unit as the face data of the non-authentication target person, and deregistration processing.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. It has a registration process of specifying the face data relating to the above and storing the specified face data in the storage unit as the face data of the processing target person.
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
In the deregistration process, when the face data of the authentication target person whose registration is to be deleted is stored in the storage unit as the face data of the non-authentication target person, the individual of the authentication target person whose registration is to be deleted is stored. The identification data is deleted from the storage unit.
The face recognition program of the present invention
It is a face recognition program for collating face data consisting of face images or their feature amounts to authenticate the person to be authenticated as a registered authentication target person.
An authentication processing procedure for collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing procedure for storing the face data of the authentication target person as the face data of the non-authentication target person in the storage unit, and a deregistration processing procedure.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. A computer is made to execute a registration processing procedure of specifying the face data relating to the above and storing the identified face data as the face data of the processing target person in the storage unit.
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
The authentication target person whose registration is deleted when the face data of the authentication target person whose registration is deleted is stored in the storage unit as the face data of the non-authentication target person in the registration deletion processing procedure. The personal identification data is deleted from the storage unit.

According to the above configuration, when the registration of the authentication target person is deleted, the face data of the authentication target person is stored in the storage unit as the face data of the non-authentication target person. Then, the face data of the person who was the authentication target person remains in the storage unit as the face data of the non-authentication target person. Then, from the face data of the non-authentication target person, the data related to the processing target person (the person registered as the authentication target person or the person registered as the authentication target person) is specified, and the processing target person who is the authentication target person is identified. It is registered as face data. As a result, since the face data of the person to be authenticated in the past is used for the authentication of the person to be processed, the authentication performance is improved as compared with the case where the past face data is not reused. That is, according to the above configuration, when a person is re-registered as an authentication target person, the past authentication performance can be maintained, in other words, the past authentication performance can be restored. Note that "maintaining past authentication performance" or "recovering" is not limited to the case of achieving exactly the same performance as past authentication performance. If months or years have passed since the person was previously certified, the hairstyle and physiognomy will change to some extent. Even in such a case, according to the above-mentioned feature configuration, the effect of improving the authentication performance can be obtained as compared with the case of registering the face data as a completely new authentication target person without reusing it.
Further, according to the above configuration, the personal identification data is deleted from the storage unit when the registration of the authentication target person is deleted, and the face data of the non-authentication target person is stored in a state not associated with the personal identification data. It will be. This is preferable because it is possible to reduce the risk related to the storage of face data (for example, the risk when information leakage occurs).

According to the present invention, it is possible to provide a method capable of maintaining (recovering) past authentication performance when a person is re-registered as an authentication target person.

Block diagram showing the configuration of the face recognition system Diagram explaining the outline of the face recognition system and the authentication process The figure explaining the deregistration process in a face recognition system The figure explaining the registration process in a face recognition system Flowchart showing the processing procedure of the authentication process Flowchart showing the processing procedure of deregistration processing Flowchart showing the processing procedure of the registration process Flowchart showing the processing procedure of the update process Diagram explaining the outline of the face recognition system and the authentication process The figure explaining the deregistration process in a face recognition system The figure explaining the registration process in a face recognition system The figure explaining the registration process in a face recognition system

(First Embodiment)
The face recognition system, face recognition method, and face recognition program according to the present embodiment will be described below with reference to FIGS. 1 to 8.

The face recognition system 1 is a system that collates face data composed of a face image or a feature amount thereof and authenticates the person to be authenticated A as a registered authentication target person B. The face recognition system 1 is installed, for example, in a server room of a data center to manage admission to a management area. Specifically, the face recognition system 1 allows only the authenticated person who has been authenticated as the person to be authenticated to enter the management area.

As shown in FIG. 1, the face recognition system 1 includes a terminal device 10, an electric control door 20, and a control device 30.

In the face recognition system 1 of the present embodiment, as shown in FIG. 2, when the person P01 (certified person A) stands in front of the electric control door 20, the terminal device 10 photographs the face of the person P01 and faces. Generate data F01. The control device 30 collates the face data F01 of the person P01 with the face data F02 to F05 of the authentication target person B stored in the first storage unit 31. Then, when the face data matches, the person P01 (certified person A) is authenticated as the authentication target person B. Then, the electric control door 20 is unlocked, and the person P01 can enter the management area. In the face recognition system 1 of the present embodiment, the face data F06 to F08 of the non-authentication target person C are stored in the second storage unit 32.

The person to be authenticated A is a person who is authenticated by the face recognition system 1. The authentication target person B is a person registered in the face recognition system 1 as a person who is permitted to enter the management area. The non-certified person C is a person who does not correspond to the certified person B and is not permitted to enter the management area.

The face data F is a face image or a feature amount thereof. The face image is an image of a person's face taken with a camera or the like. The feature amount is an amount obtained by quantifying the facial features of a person in a face image. The face data F may include a plurality of face images and feature amounts. For example, the face data F may include a plurality of face images having different shooting angles, light amounts, and the like, and feature amounts based on the face images. The personal identification data I is information that can identify an individual, such as a name and affiliation, and is information other than face data. For example, the personal identification data I02 is information including the name, affiliation company name, department, job title, employee number, management number, contact telephone number, etc. of the person P02.

Hereinafter, each configuration of the face recognition system 1 will be described with reference to FIG. 2 as appropriate.

The terminal device 10 is a device that is arranged near the entrance of the management area and acquires face data from the person to be authenticated A. Specifically, the terminal device 10 is configured to have a camera, photographs the face of a person P01 standing in front of the terminal device 10, and transmits face data F01 to the control device 30.

In the present embodiment, in detail, the face image taken by the terminal device 10 is transmitted to the control device 30, and the control device 30 calculates the feature amount from the face image. Hereinafter, the face image and the feature amount will be collectively referred to as “face data” without distinction.

The terminal device 10 may be configured to have a card reader. Then, the face recognition system 1 shall be used in combination with face recognition or authentication using an ID card or the like using a card reader, and if one of the authentications is successful, the person to be authenticated A shall be authenticated as the person to be authenticated B. It may be configured.

The electric control door 20 is a door whose unlocking / locking (or opening / closing) is controlled by the control device 30. The electric control door 20 is installed at the entrance of the management area. Then, when the person to be authenticated A is authenticated as the person to be authenticated B, the electric control door 20 is unlocked (or opened) by the control device 30, and the person to be authenticated A can enter the management area. Become.

The control device 30 controls the overall operation of the face recognition system 1 and performs various processes. The control device 30 includes a first storage unit 31, a second storage unit 32, a control unit 33, and a communication unit 34.

The first storage unit 31 and the second storage unit 32 are composed of storage devices such as an HDD (Hard Disk Drive) and a non-volatile RAM (Random Access Memory). The first storage unit 31 and the second storage unit 32 may be physically separate storage devices, or may be two storage areas provided in one storage device. Further, the first storage unit 31 and the second storage unit 32 may be provided in a place where the face recognition system 1 is provided (for example, a server room), or may be a so-called storage area on the cloud. Hereinafter, the first storage unit 31 and the second storage unit 32 may be collectively referred to as a storage unit M.

In the present embodiment, as conceptually shown in FIG. 1, the face data F and the personal identification data I of the authentication target person B are stored in the first storage unit 31. In the example of FIG. 2, for the four persons P02 to P05 who are the authentication target persons B, the face data F02 to F05 and the personal identification data I02 to I05 are associated with each person in the first storage unit. It is stored in 31. That is, the "registration" in the face recognition system 1 as the person B to be authenticated means that the face data F and the personal identification data I of the person are stored in the storage unit M in a state of being associated with each other. means.

The face data F of the non-authentication target person C is stored in the second storage unit 32. In the example of FIG. 2, face data F06 to F08 of each of the three persons P06 to P08, which are the non-authentication target persons C, are stored in the second storage unit 32. In the present embodiment, the personal identification data I of the non-authentication target person C is not stored in the storage unit. That is, the face data F06 to F08 are stored in the second storage unit 32 in a state in which they cannot be associated with personally identifiable information such as a name, in other words, in a state where the face data F is alone. In FIG. 2, each face data F and the person P are represented in a state of being connected by a line, but this simply conceptually shows that the person P, which is the source of the face data F, exists. I'm just there.

The control unit 33 includes an authentication unit 33a, an update unit 33b, a deregistration unit 33c, and a registration unit 33d. The control unit 33 is composed of, for example, a software program for realizing various processes, a CPU for executing the software program, various hardware controlled by the CPU, and the like. Software programs and data necessary for the operation of the control unit 33 are stored in the storage unit M, but the storage destination is not particularly limited, and the software programs and data are stored in a storage device such as a disk or a flash memory provided separately. It may be present, or it may be stored in a server, a storage device, or the like connected so as to be communicable.

The authentication unit 33a collates the face data F of the authenticated person A with the face data F stored in the storage unit M, and authenticates the authenticated person A as the authentication target person B. In the present embodiment, the face data F01 of the person P01 who is the authenticated person A is collated with the face data F02 to F05 of the authentication target person B stored in the first storage unit 31, and the face data F matches. If so, the person P01 is authenticated as the authentication target person B. For example, as shown in FIG. 2, when the face data F01 of the person to be authenticated A matches the face data F04, the person to be authenticated A authenticates to be the person P04 of the person to be authenticated B.

In this embodiment and subsequent embodiments, "matching face data" is not only when the face image and feature amount are completely the same, but also when it is determined by face recognition that the face data is the same person. including.

The update unit 33b updates the face data F of the authentication target person B by using the face data F of the authenticated person A and the processing target person D. For example, the update unit 33b additionally registers the face data F of the person to be authenticated A and the person to be processed D to the face data F of the person to be authenticated B. In the example of FIG. 2, the person P01 who is the person to be authenticated A is authenticated as the same person as the person P04 who is the person to be authenticated B. Therefore, the face data F01 of the person P01 is added to the face data of the person P04. Since the face data F01 is face data related to an image newly taken by the terminal device 10, the facial expression and the amount of light rays may be different from those taken in the past. Therefore, by adding the face data F01 to the face data F04 related to the past captured image, the authentication performance of the person P04 (person P01) can be improved. The update unit 33b updates the face data F when the authenticated person A is authenticated as the authentication target person B (the above example) or when the authentication target person B is registered by the registration unit 33d. .. The face data F of the authentication target person B can be updated by the update unit 33b in various modes such as the above-mentioned additional registration of the face data F and the change or adjustment of the feature amount related to the face data F.

The registration deletion unit 33c stores the face data F of the authentication target person B whose registration is deleted in the storage unit M as the face data F of the non-authentication target person C. In the present embodiment, the registration deletion unit 33c deletes the personal identification data I of the authentication target person B whose registration is deleted from the storage unit M.

The operation of the deregistration unit 33c will be described in detail with reference to FIG. A case of canceling the registration of the person P05 who is the authentication target person B will be described. The face data F05 and the personal identification data I05 of the person P05 are stored in the first storage unit 31 as data related to the authentication target person B.

The deregistration unit 33c stores (copies) the face data F05 of the person P05 in the second storage unit 32. Then, the deregistration unit 33c deletes the face data F05 and the personal identification data I05 stored in the first storage unit 31. That is, the face data F05 is moved from the first storage unit 31 to the second storage unit 32. As a result, the face data F05 is subsequently treated as the face data of the non-authentication target person C. That is, the person P05 is deregistered as the authentication target person B and becomes the non-authentication target person C. In other words, the handling (or attribute) of the person P05 is changed from the authentication target person B to the non-authentication target person C. Then, the personal identification data I05 is deleted from the storage unit M (first storage unit 31 and second storage unit 32).

In the present embodiment, the registration unit 33d sets the person P registered as the authentication target person B as the processing target person D, and from the face data F of the non-authentication target person C stored in the storage unit M, the processing target person D The face data related to the data is specified, and the specified face data F is stored in the storage unit M as the face data of the processing target person D. Specifically, the registration unit 33d collates the face data F of the processing target person D with the face data F of the non-authentication target person C stored in the storage unit M, and identifies the face data F related to the processing target person D. .. Then, the registration unit 33d stores the specified face data F in the storage unit M as the face data F of the processing target person D in a state of being associated with the personal identification data I of the processing target person D.

The operation of the registration unit 33d will be described in detail with reference to FIG. A case where the person P09 is registered as the authentication target person B will be described. That is, the person P09 is the processing target person D. First, the face data F09 and the personal identification data I09 of the person P09 are prepared for registration and input to the control device 30.

The registration unit 33d collates the face data F09 of the processing target person D with the face data F stored in the second storage unit 32 to specify the face data F related to the processing target person D. As shown in FIG. 4, when the face data F09 of the processing target person D matches the face data F08, the registration unit 33d identifies that the face data F08 is the face data related to the processing target person D. That is, the person P08 related to the face data F08 was actually the person P09 related to the processing target person D.

Then, the registration unit 33d stores the specified face data F08 in the first storage unit 31 as the face data F09 of the processing target person D in a state of being associated with the personal identification data I09 of the processing target person D. That is, as shown in FIG. 4, the face data F08 and the personal identification data I09 are stored in the first storage unit 31 as the data of the person P09. As a result, the face data F08 is subsequently treated as the face data of the authentication target person B. That is, the person P09 of the processing target person D becomes the authentication target person B. The face data F08 stored in the second storage unit 32 is deleted from the second storage unit 32 by the registration unit 33d.

If the face data F09 of the processing target person D does not match the face data F stored in the second storage unit 32, the registration unit 33d sets the face data F09 of the processing target person D and the personal identification data I09. , Are stored in the first storage unit 31 in a state of being associated with each other. That is, in this case, the processing target person D has never been the authentication target person B of the face recognition system 1 in the past, and is the person who is registered in the face recognition system 1 for the first time.

At this time, the update unit 33b may update the face data F08 by using the face data F09 input for registration of the person P09 as the authentication target person B. It is specified by the registration unit 33d that the face data F09 and the face data F08 are the face data F of the same person P09. Therefore, by updating, the authentication performance of the person P09 can be improved.

The face authentication method performed by the face authentication system 1 described above will be described with reference to the flowcharts of FIGS. 5 to 8. The face recognition method of the present embodiment includes an authentication process, a deregistration process, a registration process, and a renewal process.

<Authentication processing>
In the authentication process, the authentication unit 33a collates the face data F of the person to be authenticated A with the face data F stored in the storage unit M, and authenticates the person to be authenticated A as the person to be authenticated B. Specifically, the authentication process is performed as follows according to the flowchart shown in FIG.

In the input step of step S101, the face data F of the person to be authenticated A is input to the face recognition system 1. Specifically, the terminal device 10 photographs the face of the person to be authenticated A (person P01) and transmits the face data F01 to the control device 30. Then, the process proceeds to step S102.

In the authentication step of step S102, the authentication unit 33a collates the input face data F of the authenticated person A with the face data F stored in the storage unit M, and sets the authenticated person A as the authentication target person B. Authenticate as. In the present embodiment, the authentication unit 33a collates the face data F01 of the person P01 who is the person to be authenticated A with the face data F02 to F05 of the authentication target person B stored in the first storage unit 31. When the face data F matches, the person P01 is authenticated as the authentication target person B. Then, the process proceeds to step S103.

In the unlocking step of step S103, the control unit 33 unlocks the electric control door 20. Then, the authentication process is completed.

<Deregistration process>
In the deregistration process, the deregistration unit 33c stores the face data F of the authentication target person B whose registration is to be deleted in the storage unit M as the face data F of the non-authentication target person C. Specifically, the deregistration process is performed as follows according to the flowchart shown in FIG.

In the input step of step S201, the face recognition system 1 is input to specify the authentication target person B whose registration is to be deleted. Specifically, the control device 30 accepts the input of the personal identification data I (name, management number, etc.) of the authentication target person B whose registration is to be deleted. Then, the process proceeds to step S202.

In the specific step of step S202, the registration deletion unit 33c identifies a person (for example, person P05) whose registration is deleted from the authentication target person B based on the input content received in step S201. Specifically, the deregistration unit 33c collates the personal identification data I of the authentication target person B whose registration is deregistered with the personal identification data I stored in the first storage unit 31, and the person whose data matches. Identify P. Then, the process proceeds to step S203.

In the copying step of step S203, the deregistration unit 33c stores (copies) the face data F of the person whose registration is to be deleted in the second storage unit 32. For example, the face data F05 of the person P05 is stored (copied) in the second storage unit 32. Then, the process proceeds to step S204.

In the deletion step of step S204, the deregistration unit 33c deletes the face data F05 and the personal identification data I05 of the person P05 from the first storage unit 31. By the above copying step and deletion step, the personal identification data I of the authentication target person B whose registration is deleted is deleted from the storage unit M, and the face data F is moved from the first storage unit 31 to the second storage unit 32. Then, the deregistration process is completed.

<Registration process>
In the registration process of the present embodiment, the registration unit 33d processes the face data F of the non-authentication target person C stored in the storage unit M with the person P registered as the authentication target person B as the processing target person D. The face data related to the target person D is specified, and the specified face data F is stored in the storage unit M as the face data of the processing target person D. In the present embodiment, the registration unit 33d collates the face data F of the processing target person D with the face data F of the non-authentication target person C stored in the storage unit M, and obtains the face data F related to the processing target person D. Identify. Then, the registration unit 33d stores the specified face data F in the storage unit as the face data F of the processing target person D in a state of being associated with the personal identification data I of the processing target person D. Specifically, the registration process is performed as follows according to the flowchart shown in FIG.

In the input step of step S301, information about the person P (processing target person D) to be registered as the authentication target person B is input to the face recognition system 1. Specifically, the control device 30 receives the input of the face data F09 and the personal identification data I09 regarding the person P09 who is the processing target person D. Then, the process proceeds to step S302.

In the specific step of step S302, the registration unit 33d collates the face data F09 of the processing target person D with the face data F stored in the second storage unit 32, and obtains the face data F related to the processing target person D. Identify. For example, when the face data F09 of the processing target person D matches the face data F08, the registration unit 33d identifies that the face data F08 is the face data related to the processing target person D. Then, the process proceeds to step S303.

In the registration step of step S303, the registration unit 33d associates the specified face data F09 with the personal identification data I09 of the processing target person D, and sets the face data F09 of the processing target person D in the first storage unit 31. Remember. That is, the face data F08 and the personal identification data I09 are stored in the first storage unit 31 as the data of the person P09. As a result, the face data F08 is subsequently treated as the face data of the authentication target person B. That is, the person P09 of the processing target person D becomes the authentication target person B. Then, the registration process is completed.

<Update process>
In the update process, the update unit 33b updates the face data F of the authentication target person B using the face data F of the authenticated person A and the processing target person D. Specifically, the update unit 33b additionally registers the face data F of the person to be authenticated A and the person to be processed D to the face data F of the person to be authenticated B. In this update process, when the authentication process is performed and the face data F of the person to be authenticated A is input to the face recognition system 1, or when the registration process is performed and the face data F of the person D to be processed is the face recognition system 1. It is done when it is entered in. Specifically, the update process is performed as follows according to the flowchart shown in FIG.

In the input process of step S401, the face data F for performing the update process is input to the face authentication system 1. When the update process is performed when the authentication process is performed, the face data F of the person to be authenticated A becomes the input data. When the update process is performed when the registration process is performed, the face data F of the process target person D becomes the input data. Then, the process proceeds to step S402.

In the update step of step S402, the update unit 33b updates the face data F of the authentication target person B using the input data. Specifically, the update unit 33b additionally registers the face data F of the person to be authenticated A and the person to be processed D to the face data F of the person to be authenticated B. When the update process is performed when the authentication process is performed, the face data F is updated for the authentication target person B who has been authenticated as the same person as the authenticated person A. When the update process is performed when the registration process is performed, the face data F is updated for the person P registered as the authentication target person B. Then, the update process is finished.

(Second Embodiment)
In the above-described first embodiment, the face recognition system 1 is configured in such a form that the personal identification data I of the non-authentication target person C is not stored in the storage unit M. In the second embodiment, the storage unit M stores the face data F of the non-authentication target person C in association with the personal identification data I of the non-authentication target person C. Along with this, the operation of the deregistration unit 33c and the registration unit 33d becomes a mode different from that of the first embodiment. In the following description of the second embodiment, the same reference numerals may be given to the same configurations as those of the first embodiment, and the description may be omitted.

The face data F and personal identification data I of the non-authentication target person C are stored in the second storage unit 32. In the example of FIG. 9, face data F06 to F08 and personal identification data I06 to I08 are stored in the second storage unit 32 for each of the three persons P06 to P08 who are the non-authentication target persons C. That is, the face data F06 to F08 are stored in the second storage unit 32 in a state of being associated with personally identifiable information such as a name.

The registration deletion unit 33c stores the face data F of the authentication target person B whose registration is deleted in the storage unit M as the face data F of the non-authentication target person C. In the present embodiment, the registration deletion unit 33c does not delete the personal identification data I of the authentication target person B whose registration is deleted from the storage unit M.

The operation of the deregistration unit 33c will be described in detail with reference to FIG. A case of canceling the registration of the person P05 who is the authentication target person B will be described. The face data F05 and the personal identification data I05 of the person P05 are stored in the first storage unit 31 as data related to the authentication target person B.

The deregistration unit 33c stores (copies) the face data F05 and the personal identification data I05 of the person P05 in the second storage unit 32. Then, the deregistration unit 33c deletes the face data F05 and the personal identification data I05 stored in the first storage unit 31. That is, the face data F05 and the personal identification data I05 are moved from the first storage unit 31 to the second storage unit 32. As a result, the face data F05 is subsequently treated as the face data of the non-authentication target person C. That is, the person P05 is deregistered as the authentication target person B and becomes the non-authentication target person C. In other words, the handling (or attribute) of the person P05 is changed from the authentication target person B to the non-authentication target person C. Note that, unlike the first embodiment, the personal identification data I05 is not deleted from the storage unit M, but is simply moved from the first storage unit 31 to the second storage unit 32.

Similar to the first embodiment, the registration unit 33d processes the person P registered as the authentication target person B as the processing target person D from the face data F of the non-authentication target person C stored in the storage unit M. The face data related to the target person D is specified, and the specified face data F is stored in the storage unit M as the face data of the processing target person D. Unlike the first embodiment, the registration unit 33d collates the personal identification data I of the processing target person D with the personal identification data I of the non-authentication target person C stored in the storage unit M, and the processing target person The face data F regarding D is specified. Then, the registration unit 33d stores the specified face data F in the storage unit as the face data F of the processing target person D in a state of being associated with the personal identification data I of the processing target person D.

The operation of the registration unit 33d will be described in detail with reference to FIG. A case where the person P09 is registered as the authentication target person B will be described. That is, the person P09 is the processing target person D. First, the face data F09 and the personal identification data I09 of the person P09 are prepared for registration and input to the control device 30.

The registration unit 33d collates the personal identification data I of the processing target person D with the personal identification data I of the non-authentication target person C stored in the second storage unit 32, and obtains the face data F regarding the processing target person D. Identify. In other words, the registration unit 33d selects the processing target person D from the face data F of the non-authentication target person C stored in the second storage unit 32 based on the personal identification data I such as the name and employee number, for example. The face data F is specified. As shown in FIG. 11, when the personal identification data I09 of the processing target person D matches the personal identification data I08, the registration unit 33d identifies the face data F08 as the face data related to the processing target person D. That is, the person P08 related to the face data F08 was actually the person P09 related to the processing target person D.

Then, the registration unit 33d stores the specified face data F09 in the first storage unit 31 as the face data F09 of the processing target person D in a state of being associated with the personal identification data I09 of the processing target person D. That is, as shown in FIG. 11, the face data F08 and the personal identification data I09 are stored in the first storage unit 31 as the data of the person P09. As a result, the face data F08 is subsequently treated as the face data of the authentication target person B. That is, the person P09 of the processing target person D becomes the authentication target person B. Then, the face data F08 becomes the face data F09 of the person P09. The face data F08 and the personal identification data I08 stored in the second storage unit 32 are deleted from the second storage unit 32 by the registration unit 33d.

The face recognition method performed by the face recognition system 1 according to the second embodiment is slightly different from the first embodiment due to the difference in the aspect of the face recognition system 1. This will be described below.

<Deregistration process>
In the deregistration process of the present embodiment, the personal identification data I of the person whose registration is to be deregistered is not deleted from the storage unit M, but is moved to the second storage unit 32.

In the copying step of step S203, the deregistration unit 33c stores (copies) the face data F and the personal identification data I of the person whose registration is to be deleted in the second storage unit 32. For example, the face data F05 and the personal identification data I05 of the person P05 are stored (copied) in the second storage unit 32. Then, the process proceeds to step S204.

In the deletion step of step S204, the deregistration unit 33c deletes the face data F05 and the personal identification data I05 of the person P05 from the first storage unit 31. By the above copying step and deleting step, the face data F and the personal identification data I are moved from the first storage unit 31 to the second storage unit 32. Then, the deregistration process is completed.

<Registration process>
In the registration process of the present embodiment, unlike the first embodiment, the registration unit 33d displays the personal identification data I of the processing target person D and the personal identification data I of the non-authentication target person C stored in the storage unit M. By collating, the face data F relating to the processing target person D is specified.

In the specific step of step S302, the registration unit 33d collates the personal identification data I09 of the processing target person D with the personal identification data I stored in the second storage unit 32, and the face data regarding the processing target person D. Identify F. For example, when the personal identification data I09 of the processing target person D matches the personal identification data I08, the registration unit 33d identifies that the face data F08 is the face data related to the processing target person D. Then, the process proceeds to step S303.

(Third Embodiment)
In the first embodiment and the second embodiment described above, the registration process is performed with the person P registered in the face recognition system 1 as the authentication target person B as the processing target person D. It is also possible to modify this and perform the registration process with the person P registered in the face recognition system 1 as the authentication target person B as the processing target person D. FIG. 12 shows an example in which the registration process is performed with the person P09, who is the authentication target person B, as the processing target person D.

The registration unit 33d collates the face data F09 of the person P09 who is the processing target person D with the face data F stored in the second storage unit 32, and identifies the face data F related to the processing target person D. As shown in FIG. 12, when the face data F09 of the processing target person D matches the face data F08, the registration unit 33d identifies that the face data F08 is the face data related to the processing target person D. That is, the person P08 related to the face data F08 was actually the person P09 related to the processing target person D.

Then, the registration unit 33d stores the specified face data F09 in the first storage unit 31 as the face data F09 of the processing target person D in a state of being associated with the personal identification data I09 of the processing target person D. That is, as shown in FIG. 12, the face data F08 and the personal identification data I09 are stored in the first storage unit 31 as the data of the person P09. As a result, the face data F08 is subsequently treated as the face data of the authentication target person B.

As described above, in the face recognition system 1 according to the third embodiment, when the registration unit 33d performs the registration process, the person P who has already been registered in the face recognition system 1 as the authentication target person B is processed. Let it be D. That is, the face data F of the processing target person D is in a state of being stored in the first storage unit 31. For example, when registering a certain person P in the face recognition system 1, the face image taken at that time is used for registration, and the registration unit 33d registers the person P at night when the processing load of the face recognition system 1 is small. It is conceivable to execute the processing in batch processing. In such a case, the registration process according to the third embodiment is useful.

(Other embodiments)
(1) In the above-described embodiment, the authentication unit 33a has described an embodiment in which the face data F of the person to be authenticated A is collated with the face data F of the authentication target person B stored in the first storage unit 31. .. By modifying this, the authentication unit 33a modifies the face data F of the person to be authenticated A and the face data F of the authentication target person B stored in the first storage unit 31 and the second storage unit 32, that is, the entire storage unit M. May be matched with. In this case, even if the face data F of the authenticated person A and the face data F of the non-authenticated person C match, the authenticated person A is not allowed to enter the management area, and the electric control door 20 is opened. Can't be opened.

(2) In the above-described embodiment, the control device 30 is configured to have two storage units, a first storage unit 31 and a second storage unit 32. That is, the storage unit M was divided into two. Then, the face data F of the authentication target person B is stored in the first storage unit 31, and the face data F09 of the non-authentication target person C is stored in the second storage unit 32. A configuration in which the control device 30 has one storage unit or a configuration in which the control device 30 has three or more storage units is also possible. That is, the face data F09 of the authentication target person B and the non-authentication target person C may be stored in one storage unit, or may be distributed and stored in three or more storage units.

(3) There are no particular restrictions on the installation locations of the first storage unit 31 and the second storage unit 32. For example, the first storage unit 31 is installed in the vicinity of a plurality of management areas, and the face data F of the authentication target person B related to the installed management area is stored in each of the first storage units 31, and each management area is stored. It is also possible to set up a second storage unit 32 in the headquarters that controls the above and centrally manage the face data F of the non-authentication target person C.

(4) In the above-described embodiment, the case where the face recognition system 1 manages admission to a management area such as a server room of a data center has been described as an example. The face recognition system 1 is applicable not only to such admission management. For example, it can be applied to a case where a person passing by is photographed by a camera and a specific person is detected / notified. In this case, the "passing person" corresponds to the authenticated person A, and the detected "specific person" corresponds to the authenticated person B. In this case, the face data of the "specific person" taken by the camera or the face data of the "specific person" prepared separately is registered. At this time, the "specific person" corresponds to the processing target person D, and the processing target that is the "specific person" from the face data F of the non-authentication target person C registered as the "specific person" in the past. The face data F of the person D is specified and stored in the first storage unit 31 as the face data F of the person D to be processed.

(5) In the above-described embodiment, in the "registration" of the person as the authentication target person B in the face recognition system 1, the storage unit M is in a state where the face data F and the personal identification data I of the person are associated with each other. Personal identification data I is not essential, although it means that it is stored in. For example, if at least the face data F of the person who is permitted to enter is stored in the storage unit M, the entrance can be permitted by collating with the face data F of the person to be authenticated A. Further, if at least the face data F of the "specific person" to be detected is stored in the storage unit M, the "specific person" can be detected / notified by collating with the face data F of the "passing person". .. That is, "registration" means that at least the face data F is stored in the storage unit M.

Further, each configuration shown in the above-described embodiment is a schematic function, and does not necessarily have to be physically illustrated. That is, the form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically distributed / integrated in an arbitrary unit according to various loads and usage conditions. Can be configured.

As described above, the face recognition system or the like according to the present invention can provide a method capable of improving the authentication performance when a person is re-registered as an authentication target person.

1: Face recognition system 10: Terminal device 20: Electric control door 30: Control device 31: First storage unit 32: Second storage unit 33: Control unit 33a: Authentication unit 33b: Update unit 33c: Deregistration unit 33d: Registration Part 34: Communication unit A: Authenticated person B: Authentication target person C: Non-authentication target person D: Processing target person F: Face data I: Personal identification data M: Storage unit P: Person

Claims (10)

It is a face recognition system that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
A storage unit that stores the face data of the authentication target person and the face data of the non-authentication target person.
An authentication unit that collates the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
A deregistration unit that stores the face data of the authentication target person to be stored in the storage unit as the face data of the non-authentication target person.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific A registration unit that stores the face data as the face data of the processing target person in the storage unit, and a registration unit.
Face recognition system with. The registration unit collates the face data of the processing target person with the face data of the non-authentication target person stored in the storage unit to identify the face data of the processing target person. The face recognition system according to 1. The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
The first or second aspect of claim 1 or 2, wherein the registration unit stores the specified face data in the storage unit as the face data of the processing target person in a state of being associated with the personal identification data of the processing target person. Face recognition system. The storage unit stores the face data of the non-authentication target person in association with the personal identification data of the non-authentication target person.
The registration unit collates the personal identification data of the processing target person with the personal identification data of the non-authentication target person stored in the storage unit to identify the face data of the processing target person. The face recognition system according to claim 1 or 2. Any one of claims 1 to 4, wherein the storage unit has a first storage unit that stores the face data of the authentication target person and a second storage unit that stores the face data of the non-authentication target person. The face recognition system described in the section. It is a face recognition method that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
An authentication process of collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing for storing the face data of the authentication target person in the storage unit as the face data of the non-authentication target person, and deregistration processing.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific The registration process of storing the face data as the face data of the person to be processed in the storage unit, and
Face recognition method with. It is a face recognition program for collating face data consisting of face images or their feature amounts to authenticate the person to be authenticated as a registered authentication target person.
An authentication processing procedure for collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing procedure for storing the face data of the authentication target person as the face data of the non-authentication target person in the storage unit, and a deregistration processing procedure.
The persons registered as before Symbol object's as the processing subject, from among the face data of the non-object's stored in the storage unit to identify the face data related to the processing subject, specific A registration processing procedure for storing the face data as the face data of the processing target person in the storage unit, and
A face recognition program that lets your computer run. It is a face recognition system that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
A storage unit that stores the face data of the authentication target person and the face data of the non-authentication target person.
An authentication unit that collates the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
A deregistration unit that stores the face data of the authentication target person to be stored in the storage unit as the face data of the non-authentication target person.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. A registration unit that identifies the face data related to the above and stores the specified face data in the storage unit as the face data of the processing target person.
With
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
When the face data of the authentication target person whose registration is deleted is stored in the storage unit as the face data of the non-authentication target person, the registration deletion unit deletes the registration of the individual of the authentication target person. A face authentication system that deletes identification data from the storage unit. It is a face recognition method that authenticates the person to be authenticated as a registered authentication target person by collating face data consisting of a face image or its feature amount.
An authentication process of collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing for storing the face data of the authentication target person in the storage unit as the face data of the non-authentication target person, and deregistration processing.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. It has a registration process of specifying the face data relating to the above and storing the specified face data in the storage unit as the face data of the processing target person.
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
In the deregistration process, when the face data of the authentication target person whose registration is to be deleted is stored in the storage unit as the face data of the non-authentication target person, the individual of the authentication target person whose registration is to be deleted is stored. A face authentication method for deleting identification data from the storage unit. It is a face recognition program for collating face data consisting of face images or their feature amounts to authenticate the person to be authenticated as a registered authentication target person.
An authentication processing procedure for collating the face data of the person to be authenticated with the face data stored in the storage unit to authenticate the person to be authenticated as the person to be authenticated.
Deregistration processing procedure for storing the face data of the authentication target person as the face data of the non-authentication target person in the storage unit, and a deregistration processing procedure.
The processing target person is selected from the face data of the non-authentication target person stored in the storage unit, with the person registered as the authentication target person or the person registered as the authentication target person as the processing target person. A computer is made to execute a registration processing procedure of specifying the face data relating to the above and storing the identified face data as the face data of the processing target person in the storage unit.
The storage unit stores the face data of the authentication target person in association with the personal identification data of the authentication target person.
In the deregistration processing procedure, when the face data of the authentication target person whose registration is to be deleted is stored in the storage unit as the face data of the non-authentication target person, the authentication target person whose registration is to be deleted is said to be the same. A face authentication program that deletes personal identification data from the storage unit.

Images