WO2012029150A1

WO2012029150A1 - Biometric authentication system, biometric authentication method and program

Info

Publication number: WO2012029150A1
Application number: PCT/JP2010/064970
Authority: WO
Inventors: 福田充昭
Original assignee: 富士通株式会社
Priority date: 2010-09-01
Filing date: 2010-09-01
Publication date: 2012-03-08

Abstract

The disclosed biometric authentication system acquires registered biometric information in a position, orientation and form suitable for authentication without placing a large burden on the user. To this end, the biometric authentication system is provided with a biometric image acquisition unit, an evaluation unit, an evaluation value extraction unit, and a registered biometric information storage unit. The evaluation unit evaluates each of the multiple biometric images acquired by the biometric image acquisition unit for each of multiple evaluation criteria and outputs, as the evaluation result for each of said evaluation criterion, the evaluation values indicating the degree to which the biometric images match the evaluation criterion. From the evaluation values of the biometric images outputted by the evaluation unit, the evaluation value extraction unit extracts for each evaluation condition the evaluation values that indicate a degree of matching greater than or equal to a prescribed evaluation threshold value. Furthermore, if multiple evaluation values are extracted for the same evaluation criterion, the evaluation value with the highest degree of matching for that evaluation criterion is selected. The registered biometric information storage unit stores, as registered biometric information of the authenticated subject, biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit.

Description

Biometric authentication system, biometric authentication method, and program

The embodiment discussed in this specification relates to a biometric authentication technique for performing personal authentication using human body characteristics such as fingerprints of hands, palm veins, and faces.

In modern society, identity verification is required in various situations. In recent years, biometric authentication technology that performs personal authentication by using the characteristics of each part of the human body such as fingerprints, veins, faces, etc. has become widespread as one method for accurately performing identity verification. For example, biometric authentication is used to determine whether to enter or leave a restricted area where only authorized people can enter, whether to log in to a personal computer, whether to access various services such as online transactions, etc. It's being used.

In order to perform such biometric authentication, first, as a user's biometric information, for example, an image such as a fingerprint, a vein, or a face is acquired in advance and recorded as a registered biometric information of the user on a recording medium. . Thereafter, when the identity verification for using the service as described above becomes necessary, the biometric information of the user is acquired again in the same manner as at the time of registration. Then, the biometric information (authenticated biometric information) acquired at this time is compared with the registered biometric information recorded on the recording medium, and the level of similarity between the two is determined. Here, if the similarity is higher than a predetermined threshold, an authentication result indicating that the person is the person is obtained.

In order to perform such biometric authentication with high accuracy, an image of a biological part (biological image), which is biometric information used for authentication, can be taken at the same position and posture during registration and verification. preferable. However, there are actually many cases where such photographing cannot be performed, and this is one of the factors that reduce the authentication accuracy.

This is because, for example, it is usually extremely difficult to hold a living body part at the same position and the same posture with respect to a sensor (camera) that captures a biological image. Also, in many cases, it is difficult to take such an image because the user forgets at which position / posture the image of the living body part was imaged at the time of registration. Furthermore, since the human body has a large degree of freedom in posture and shape depending on the part, even if the same part is photographed, the obtained image is slightly different in posture and shape each time of photographing. End up. For example, since a part such as a palm is a living body part with a very high degree of freedom in posture and shape, the shape of the palm is slightly different every time an image is taken.

Several methods are known for avoiding a decrease in authentication accuracy due to the difference in position, orientation, and shape between registration and verification. One of them is to shoot with the body part firmly fixed at the same position, posture and shape. However, this method places a great burden on the user and is not generally preferred. In addition, it may take a considerable time to perform such fixing.

Another method is to capture and record multiple biometric images of the same user at various positions, postures, and shapes at the time of registration, and target multiple biometric images acquired at registration at the time of verification. Several methods have also been proposed for collating images on a single sheet to a plurality of sheets. This method can be classified into several types according to what kind of biometric images of the same user, which are registered biometric information, are to be used.

The first of these is to record a plurality of captured biometric images as registered biometric information about the user without any special consideration. Also, the second of these methods evaluates the similarity of the similarity between the captured biometric images, and determines the combination of bioimages that are not similar (that is, the similarity is low) as a registered biometric for the user. It is to record as information. In addition, the third method is to present a guidance message to the user at the time of registration to guide the living body part to an appropriate position and posture, and then shoot a plurality of images on the user. It is recorded as registered biometric information.

Special table 2007-521777 gazette JP 2007-156936 A

Since the first method described above is a method that relies entirely by chance, the registered biometric information is not always appropriate for biometric authentication.
In addition, even with the second method described above, it is not always possible to comprehensively acquire registered biometric information of positions and orientations suitable for collation with a user's biometric image captured during collation. In other words, this second method is also relied upon by chance, and appropriate registered biometric information is not always obtained.

On the other hand, the third method described above is an effective method in that there is a high possibility that appropriate registered biometric information can be obtained. However, it is possible to follow guidance to an appropriate position and posture of the living body part. This will put a burden on the user. In addition, a means for performing such guidance (for example, an assistant (operator), visual means such as a display device, an output device for emitting a guidance sound, etc.) is separately required.

As described above, it is desired to obtain registered biometric information having a position, posture, and shape suitable for collation without imposing a great burden on the user.

In the biometric authentication system described later in this specification, the person to be authenticated is based on the similarity between the registered biometric information of the person to be authenticated previously registered at the time of registration and the authentication biometric information of the person to be authenticated acquired at the time of authentication. There is something to verify the identity of. This biometric authentication system includes a biometric image acquisition unit, an evaluation unit, an evaluation value extraction unit, a registered biometric information storage unit, and an authentication unit. Here, the biometric image acquisition unit acquires a plurality of biometric images of the person to be authenticated. The evaluation unit evaluates each of the plurality of biological images acquired by the biological image acquisition unit for each of a plurality of predetermined evaluation conditions, and determines the degree of coincidence with the evaluation conditions for the biological image. The evaluation value to be shown is output as the evaluation result of the biological image for each of the plurality of evaluation conditions. The evaluation value extraction unit extracts, for each evaluation condition, an evaluation value indicating a high degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the evaluation values for each of the plurality of biological images output by the evaluation unit. When a plurality of evaluation values are extracted under the same evaluation condition by this extraction, the evaluation value extraction unit further has the highest degree of matching with the same evaluation condition from the plurality of extracted evaluation values. The evaluation value is extracted. The registered biometric information storage unit stores a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit as registered biometric information of the person to be authenticated. Then, the authentication unit determines whether the similarity between the registered biometric information stored in the registered biometric information storage unit and the authenticated biometric information is high or low, and verifies the identity of the person to be authenticated based on the determination result.

In addition, the biometric authentication method described later in this specification uses the biometric authentication method based on the degree of similarity between the registered biometric information of the authenticated person previously registered at the time of registration and the authenticated biometric information of the authenticated person acquired at the time of authentication. There is one that verifies the identity of the certifier. In this biometric authentication method, first, the evaluation unit evaluates each of the plurality of biometric images of the person to be authenticated for each of a plurality of predetermined evaluation conditions, and matches the evaluation conditions for the biometric image. An evaluation value indicating the height of the degree is output for each of the plurality of evaluation conditions as a biological image evaluation result. Next, the evaluation value extraction unit extracts, for each evaluation condition, an evaluation value indicating a degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the evaluation values output by the evaluation unit for each of a plurality of biological images. . When a plurality of evaluation values are extracted under the same evaluation condition by this extraction, the evaluation value extraction unit further has the highest degree of matching with the same evaluation condition from the plurality of extracted evaluation values. The evaluation value is extracted. Next, the registered biometric information storage unit stores a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit as registered biometric information of the person to be authenticated. Then, the authentication unit determines whether the similarity between the registered biometric information stored in the registered biometric information storage unit and the authenticated biometric information is high or low, and verifies the identity of the person to be authenticated based on the determination result.

In addition, in the program described later in this specification, the authenticated person based on the similarity between the registered biometric information of the authenticated person registered in advance at the time of registration and the authenticated biometric information of the authenticated person acquired at the time of authentication. There is a thing to make a computer perform identity verification. This program causes a computer to perform an evaluation process, an evaluation value extraction process, a registered biometric information storage process, and an authentication process. Here, the evaluation process evaluates each of the plurality of biometric images of the person to be authenticated for each of a plurality of predetermined evaluation conditions, and determines the degree of coincidence with the evaluation conditions for the biometric image. This is a process of outputting the indicated evaluation value as the evaluation result of the biological image for each of the plurality of evaluation conditions. The evaluation value extraction process is a process of extracting, for each evaluation condition, an evaluation value indicating a degree of coincidence that is equal to or higher than a predetermined evaluation threshold, from the evaluation values output by the evaluation process for each of a plurality of biological images. is there. If a plurality of evaluation values are extracted under the same evaluation condition by this extraction, the degree of matching with the same evaluation condition is the highest from the plurality of evaluation values extracted by the evaluation value extraction process. A process for extracting an evaluation value is performed. The registered biometric information storage process is a process of storing a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction process in the storage unit as registered biometric information of the person to be authenticated. In the authentication process, the degree of similarity between the registered biometric information stored in the storage unit and the authentication biometric information is determined, and the identity of the person to be authenticated is verified based on the determination result.

The biometric authentication device described later in this specification can acquire registered biometric information having a position, posture, and shape suitable for collation without imposing a great burden on the user.

It is a functional block diagram of one Example of a biometrics authentication system. It is a figure explaining operation | movement of one Example of a biometrics authentication system. It is the figure which illustrated the 1st example of the correspondence of a biometric image and an evaluation value. It is the figure which illustrated the 2nd example of the correspondence of a biometric image and an evaluation value. It is a figure explaining an example of the estimation method of the inclination angle of the image of the hand reflected in the biological image. It is the figure which illustrated the 3rd example of the correspondence of a biometric image and evaluation value. It is a figure explaining an example of the estimation method of the index value showing the opening degree of the finger in the image of the hand reflected in the biological image. It is a figure explaining the utility of the biometric authentication system of FIG. It is the figure which illustrated an example of the specific structure of the biometric information registration part in the biometric authentication system of FIG. It is the figure which illustrated the example of a structure of the computer. 10 is a flowchart illustrating an example of processing content of a biometric information registration process performed by a biometric information registration unit in FIG. 9. It is a 1st modification of the flowchart of FIG. It is a 2nd modification of the flowchart of FIG. It is the figure which illustrated an example of the specific structure of the authentication part in the biometric authentication system of FIG. It is the flowchart which illustrated an example of the processing content of the collation process performed in the authentication part of FIG. It is a 3rd modification of the flowchart of FIG. It is a 1st modification of the flowchart of FIG. It is a modification of the concrete structure of the biometric information registration part of FIG.

First, FIG. 1 will be described. FIG. 1 is a functional configuration diagram of an embodiment of a biometric authentication system.
The biometric authentication system 1 confirms the identity of the authenticated person based on the similarity between the registered biometric information of the authenticated person registered in advance at the time of registration and the authenticated biometric information of the authenticated person acquired at the time of authentication. I do.

The biometric authentication system 1 includes a biometric information registration unit 10 that registers biometric information of a person to be authenticated, and authenticates the person to be authenticated using the biometric information registered by the biometric information registration unit 10. And an authenticating unit 20 for performing identity verification.

The biometric information registration unit 10 includes a biometric image acquisition unit 11, an evaluation unit 12, an evaluation value extraction unit 13, and a registered biometric information storage unit 14.
The biometric image acquisition unit 11 acquires a plurality of biometric images of the person to be authenticated.

The evaluation unit 12 evaluates each of the plurality of biological images acquired by the biological image acquisition unit 11 for each of a plurality of predetermined evaluation conditions, and has a high degree of coincidence with the evaluation conditions for the biological image. An evaluation value indicating the accuracy is output as the evaluation result of the biological image for each of the plurality of evaluation conditions.

The evaluation value extraction unit 13 calculates an evaluation value indicating a degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the evaluation value for each of the plurality of biological images output by the evaluation unit 12 for each evaluation condition described above. To extract. The extraction of the evaluation value is performed in order to shake off a biometric image that cannot be used for the later-described authentication of the person to be authenticated by the authenticating unit 20 from the registered biometric information candidates. In addition, when a plurality of evaluation values are extracted under the same evaluation condition by this extraction, the evaluation value extraction unit 13 further has the highest degree of match with the same evaluation condition from the plurality of extracted evaluation values. Extract high evaluation values.

The registered biometric information storage unit 14 stores a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit 13 as registered biometric information of the person to be authenticated.
The authentication unit 20 determines the degree of similarity between the registered biometric information stored in the registered biometric information storage unit 14 and the authentication biometric information described above, and authenticates the person to be authenticated based on the determination result. Do it.

The biometric information registration unit 10 of the biometric authentication system 1 may further include an evaluation result check unit 15 and a guidance output unit 16 as illustrated in FIG.
The evaluation result check unit 15 determines whether or not there is an evaluation condition in which the evaluation value is not extracted in the extraction of the evaluation value for each evaluation condition by the evaluation value extraction unit 13. If the evaluation result check unit 15 determines that there is an evaluation condition for which no evaluation value has been extracted, the evaluation result check unit 15 acquires the biological image by the biological image acquisition unit 11 and evaluates the biological image by the evaluation unit 12. And the control for causing the evaluation value extraction unit 13 to extract the evaluation value again.

When the evaluation result check unit 15 determines that there is an evaluation condition in which the evaluation value is not extracted in the extraction of the evaluation value for each evaluation condition by the evaluation value extraction unit 13, the guidance output unit 16 A guide to the person to be authenticated for guiding the living body part is output. This guidance is for the biological image acquisition unit 11 to obtain a biological image that satisfies the evaluation condition.

Next, the operation of the biometric authentication system 1 will be further described with reference to FIG.
In FIG. 2, the biometric image 101 is a plurality of images about the person to be authenticated acquired by the biometric image acquisition unit 11. In the example of FIG. 2, each of the plurality of biometric images 101 includes images of hands having various positions, postures, and shapes that are biometric parts of the person to be authenticated.

The evaluation unit 12 evaluates each of the biological images 101 with respect to a plurality of predetermined evaluation conditions 102, and calculates an evaluation value indicating the degree of matching with each of the evaluation conditions 102 for each of the biological images 101. Is output as an evaluation result for each evaluation condition 102.

The evaluation condition 102 is preferably such that various image patterns that are typical in the usage scene of the biometric authentication system 1 are arranged as registered biometric information. In the example of FIG. 2, the first evaluation condition 102a, the second evaluation condition 102b, the third evaluation condition 102c, the fourth evaluation condition 102d, the fifth evaluation condition 102e, the sixth evaluation condition 102f, and the seventh evaluation condition 102g are calculated. Seven evaluation conditions 102 are defined. Here, the first evaluation condition 102a is a condition that the size of the hand image shown in the image is a predetermined first size, and the second evaluation condition 102b is shown in the image. The condition is that the size of the hand image is a predetermined second size. The third evaluation condition 102c is a condition that the inclination of the hand image shown in the image is a predetermined first inclination, and the fourth evaluation condition 102d is the image of the hand shown in the image. Is a predetermined second inclination. The fifth evaluation condition 102e is a condition that the image of the hand shown in the image has a shape with all fingers closed, and the sixth evaluation condition 102f is the image of the hand shown in the image. It is a condition that all fingers are in a shape with the maximum opening. Furthermore, the seventh evaluation condition 102g is a condition that each finger of the image of the hand shown in the image is bent to a predetermined angle.

The evaluation value extraction unit 13 determines, for each evaluation condition 102, an evaluation value that indicates the degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the evaluation values for each of the plurality of biological images 101 output by the evaluation unit 12. To extract. Further, when a plurality of evaluation values are extracted under the same evaluation condition 102 by this extraction, the evaluation value extraction unit 13 further determines the degree of match with the same evaluation condition 102 from the plurality of extracted evaluation values. The highest evaluation value is extracted. The extracted biological image 103 in FIG. 2 is a biological image 101 for each evaluation condition 102 corresponding to the evaluation value extracted by the evaluation value extraction unit 13 in this way.

When the biometric authentication system 1 includes the evaluation result check unit 15, the evaluation result check 104 in FIG. 2 is performed. In this evaluation result check 104, it is determined whether or not there is an evaluation condition 102 for which the evaluation value is not extracted in the extraction of the evaluation value for each evaluation condition 102 by the evaluation value extraction unit 13. Here, when it is determined that the evaluation condition 102 from which the evaluation value has not been extracted exists, the biological image 101 is acquired again by the biological image acquisition unit 11, and further, the biological image 101 by the evaluation unit 12 is acquired. And evaluation value extraction by the evaluation value extraction unit 13 are performed again. With this evaluation result check 104, the evaluation values for all the evaluation conditions 102 are extracted, and the extracted biological images 103 for all the evaluation conditions 102 are obtained. When the biometric authentication system 1 further includes the guidance output unit 16, in the above-described case, the biometric part of the person to be authenticated is guided so that the biometric image acquisition unit 11 can obtain the biometric image 101 that satisfies the evaluation condition 102. The guidance output unit 16 outputs guidance to the person to be authenticated for this purpose.

The extracted biometric image 103 obtained for each evaluation condition 102 as described above is stored in the registered biometric information storage unit 14 as the registered biometric information 105 of the person to be authenticated. The authentication unit 20 determines whether the degree of similarity between the registered biometric information of the person to be authenticated stored in the registered biometric information storage unit 14 and the authentication biometric information of the person to be authenticated acquired at the time of authentication is high or low. The identity verification of the person to be authenticated is performed based on the determination result.

Next, the operation of the evaluation unit 12 for each evaluation condition 102 illustrated in FIG. 2 will be described.
First, FIG. 3 will be described. FIG. 3 is a first example of the correspondence between the biological image 101 and the evaluation value. This first example represents the correspondence between the size of the hand image shown in the biological image 101 and the evaluation value. In the first example, in the case of the first evaluation condition 102a and the second evaluation condition 102b described above, that is, the evaluation condition 102 includes information on the size of the image of the person to be authenticated on the biometric image 101. It can be used when it has a condition corresponding to the evaluation value.

In FIG. 3, the horizontal axis represents information on the size of the hand image shown in the biological image 101, more specifically, the area of the hand image, and the vertical axis represents the evaluation value for the area. Represents. In the graph of FIG. 3, as the area of the hand image is closer to a predetermined value set in advance as the evaluation condition 102, a larger evaluation value is associated with the value of the area of the hand image. It can be seen from this evaluation value that the degree of coincidence with the size information included in the condition 102 is high. The evaluation unit 12 includes a table corresponding to the graph of FIG. 3 in which the correspondence between the area value of the hand image and the evaluation value is represented.

Note that the area of the hand image shown in the biological image 101 is, for example, counting the number of pixels whose luminance value is estimated to represent the hand image among the pixels constituting the biological image 101. Can be estimated. The evaluation unit 12 refers to the above-described table, acquires an evaluation value associated with the area value estimated as described above from the table, and outputs the acquired evaluation value.

Next, FIG. 4 will be described. FIG. 4 is a second example of the correspondence between the biological image 101 and the evaluation value. This second example represents the correspondence between the inclination of the hand image shown in the biological image 101 and the evaluation value. In the second example, in the case of the third evaluation condition 102c and the fourth evaluation condition 102d described above, that is, the evaluation condition 102 evaluates information on the inclination of the image of the user's hand on the biometric image 101. It can be used when it has a condition corresponding to a value.

In FIG. 4, the horizontal axis represents information on the inclination of the hand image shown in the biological image 101, more specifically, the inclination angle of the hand image, and the vertical axis represents the evaluation value for the angle. Represents. In this graph, the closer the inclination angle of the hand image is to a predetermined value set in advance as the evaluation condition 102, the larger the evaluation value is associated with the inclination angle of the hand image. It can be seen from this evaluation value that the degree of coincidence with the inclination information included in 102 is high. The evaluation unit 12 is provided with a table corresponding to the graph of FIG. 4 in which the correspondence between the inclination angle of the hand image and the evaluation value is represented.

Note that the inclination angle of the hand image shown in the biological image 101 can be estimated, for example, as shown in FIG. In this method, first, feature points in the image of the hand shown in the biological image 101 are extracted (step 1), and then a plane imitating a palm is formed based on the positional relationship of the feature points ( Step 2). Then, about this plane, around the X axis (the horizontal axis passing through the central position of the biological image 101) and the Y axis (the vertical axis passing through the central position of the biological image 101) with respect to the image plane of the biological image 101. A rotation angle is calculated (step 3). The evaluation unit 12 refers to the table described above, acquires two evaluation values associated with the two rotation angles estimated as described above from the table, and, for example, obtains an average value of the evaluation values. Output.

Next, FIG. 6 will be described. FIG. 6 is a third example of the correspondence relationship between the biological image 101 and the evaluation value. This third example represents the correspondence between the shape of the hand image shown in the biological image 101 and the evaluation value. In the third example, in the case of the fifth evaluation condition 102e and the sixth evaluation condition 102f described above, that is, the evaluation condition 102 evaluates information on the shape of the image of the person's hand on the biometric image 101. It can be used when it has a condition corresponding to a value.

In FIG. 6, the horizontal axis represents information on the shape of the hand image shown in the biological image 101, more specifically, an index value indicating the degree of finger opening in the hand image, and the vertical axis represents the index value. It represents an evaluation value for the index value. In the graph of FIG. 4, the closer the index value in the hand image is to a predetermined value set in advance as the evaluation condition 102, the higher the evaluation value is associated with the index value. It can be seen from this evaluation value that the degree of coincidence with the included shape information is high. The evaluation unit 12 includes a table corresponding to the graph of FIG. 6 in which a correspondence relationship between the index value and the evaluation value in the hand image is represented.

Note that the index value representing the degree of finger opening in the hand image shown in the biological image 101 can be estimated as shown in FIG. 7, for example. In this method, first, from the biological image 101, the direction from the base of each finger in the image of the hand shown there is extracted as the direction of each finger, and the angle difference between the directions of adjacent fingers is calculated. The opening angles θ ₁ , θ ₂ , θ ₃ , and θ ₄ are calculated (step 1). Next, an index value for evaluating the finger opening degree is calculated based on the finger opening angle (step 2). The evaluation of the finger opening degree based on the finger opening angle is performed by the evaluation unit 12 as follows, for example.

First, the evaluation unit 12 calculates the sum of squares θ ₁ ² + θ ₂ ² + θ ₃ ² + θ ₄ ² as a first index value I1 as an index value for representing the magnitude of the finger opening angle. Next, the evaluation unit 12 uses (θ ₂ −θ ₃ ) ² + (θ ₃ −θ ₄ ) ² + (θ ₄ − as an index value for expressing the degree of opening of the four fingers excluding the thumb. θ ₂ ) ² is calculated as the second index value I2. Note that the evaluation unit 12 is provided with two tables in advance, in which the correspondence between the two index values I1 and I2 and the evaluation value in the hand image is represented. Then, the evaluation unit 12 refers to this table, obtains two evaluation values E1 and E2 associated with the index values I1 and I2 from the table, performs subtraction of E1-E2, and performs the subtraction. The result is output as an evaluation value of the degree of finger opening.

When the evaluation unit 12 performs the evaluation based on the seventh evaluation condition 102g, for example, the length from the base of each finger in the image of the hand shown in the biological image 101 and the evaluation condition 102 are set in advance. The evaluation unit 12 outputs an evaluation value corresponding to the closeness of the value to the predetermined value.

In addition, you may make it the evaluation part 12 perform the evaluation about the biometric image 101 based on another viewpoint. For example, an evaluation is made as to whether or not the image of the hand, which is the body part of the person to be authenticated, is deviated from the center position of the living body image 101, and the center position and the pixel whose luminance value is estimated to represent the image of the hand You may make it perform based on the positional relationship with the gravity center position of a group.

Next, the utility of storing the registered biometric information of the person to be authenticated in the registered biometric information storage unit 14 in the biometric authentication system 1 of FIG. 1 will be described with reference to FIG.
In various image examples illustrated in FIG. 8 in which an image of a hand as biometric information of the person to be authenticated is shown, the registered biometric information 105 arranged in the lower part of FIG. It is extracted from the biological image 101 by the operation and stored in the registered biological information storage unit 14. On the other hand, the registered biometric information 100 according to the conventional method arranged in the upper part of FIG. 8 is selected from the biometric images 101 by the second method described in the background section, that is, the similarity between the biometric images 101 is evaluated. This is a combination of images that are not similar to each other. The authentication biometric information 106 is acquired at the time of authentication. Here, it is assumed that the same authentication biometric information 106 is obtained by the conventional technique and the biometric authentication system 1 of FIG.

There are various positions, postures, and shapes of living body parts shown in the living body image 101. However, what kind of position / posture / shape of the biometric image 101 can be obtained by taking the biometric image 101 without particularly guiding the subject is subject to many subjects. It can be generally grasped by the experiment. The probability values attached to the biological image 101 in FIG. 8 are examples of probability values obtained by such an experiment. If there are images of position / posture / shape obtained with a relatively high probability, the probability values are relatively low. There are also images of position, posture, and shape that can only be obtained with probability.

Under such circumstances, when the above-described second method, that is, a method of selecting a combination of images that are not similar to each other from the biological image 101, a living body having a position, posture, and shape that can be obtained only with low probability. It is possible that the part showing the part is selected as the registered biometric information 100. In the registered biometric information 100 according to such a conventional method, it is naturally likely that the position, posture, and shape of the image of the biometric part are different from the authenticated biometric information 106. There is concern about the deterioration of authentication accuracy. In the upper example of FIG. 8, the registered biometric information 100 and the authenticating biometric information 106 according to the conventional method are 15% when the image of the body part of the person to be authenticated is similar in position / posture / shape. , 80% of cases are not similar.

On the other hand, in the biometric authentication system 1 of FIG. 1, the evaluation condition 102 is set so that an image of a biological part image having a position / posture / shape obtained with a relatively high probability by the above-described experiment is extracted. Can do. If the biometric image 101 extracted in this way is registered biometric information 105, there is a high possibility that the authentication biometric information 106 and the position / posture / shape of the image of the biometric part are similar. It can be expected that the authentication with the information 106 can be performed with high accuracy. In the lower example of FIG. 8, the registered biometric information 105 and the authenticated biometric information 106 are 70% when the image of the body part of the person to be authenticated is similar in position, posture, and shape. This means that 25% of cases do not exist.

As described above, in the biometric authentication system 1 of FIG. 1, by setting the evaluation condition 102 that exhibits an evaluation value having a higher degree of match as the position / posture / shape of the biological part that the person to be authenticated will perform at the time of verification, The possibility of performing biometric authentication with high accuracy can be increased. Therefore, by using the biometric authentication system 1 shown in FIG. 1, it is possible to perform authentication with high accuracy in accordance with an actual usage scene.

In addition, instead of setting the evaluation condition 102 as described above, an image of a living body part having a position / posture / shape that the algorithm for matching the registered biometric information and the authenticating biometric information used in the authentication unit 20 is good at. The evaluation condition 102 may be set so as to be extracted. Furthermore, an evaluation condition 102 is set so that an image of a biological part having a position, posture, and shape that is not suitable for an algorithm for collating registered biometric information and authenticated biometric information used in the authentication unit 20 is extracted. Also good.

Next, FIG. 9 will be described. FIG. 9 illustrates an example of a specific configuration of the biometric information registration unit 10 in the biometric authentication system 1 of FIG. In this example, the biometric information registration unit 10 includes an imaging unit 31, an imaging control unit 32, a biological image temporary storage unit 33, an image evaluation unit 34, a registration candidate storage unit 35, an evaluation result check unit 36, a feature extraction unit 37, and A registered biometric information storage unit 38 is provided.

The imaging unit 31 is a camera that captures a biological part of the person to be authenticated and acquires a biological image 101. The imaging control unit 32 controls the imaging unit 31 to acquire a plurality of biometric images 101 of the person to be authenticated. The imaging unit 31 and the imaging control unit 32 provide the function of the biological image acquisition unit 11 in FIG.
The biological image temporary storage unit 33 is a storage unit that temporarily stores the biological image 101 acquired by the imaging unit 31.

First, the image evaluation unit 34 reads the biological image 101 stored in the biological image temporary storage unit 33, evaluates each of a plurality of predetermined evaluation conditions 102, and evaluates each of the evaluation conditions 102 for the biological image 101. Obtain an evaluation value. Next, the image evaluation unit 34 extracts, for each evaluation condition 102, an evaluation value 107 indicating a high degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the acquired evaluation value 107. When a plurality of evaluation values are extracted under the same evaluation condition 102 by this extraction, the image evaluation unit 34 further determines the degree of match with the same evaluation condition 102 from the plurality of extracted evaluation values. The highest evaluation value 107 is extracted. Further, the image evaluation unit 34 reads the biological image 101 whose evaluation value 107 extracted as described above is an evaluation result from the biological image temporary storage unit 33 and associates it with the evaluation value 107 in the registration candidate storage unit 35. Each evaluation condition 102 is temporarily stored. The image evaluation unit 34 provides the functions of the evaluation unit 12 and the evaluation value extraction unit 13 in FIG.

The registration candidate storage unit 35 stores the evaluation value 107 extracted by the image evaluation unit 34 and the biological image 101 whose evaluation value 107 is an evaluation result in association with each other, and temporarily stores each evaluation condition 102. Part.

The evaluation result check unit 36 determines whether there is an evaluation condition 102 in which the evaluation value 107 is not stored in the registration candidate storage unit 35. Here, when it is determined that such an evaluation condition 102 exists, the evaluation result check unit 36 gives an instruction to the imaging control unit 32 to control the imaging unit 31 to acquire a further biological image 101. Let it be done. The biometric image 101 acquired again in this way is stored in the biometric image temporary storage unit 33, and then the above-described evaluation is performed by the image evaluation unit 34 to obtain the evaluation value 107 for each evaluation condition 102. Based on the registration candidate storage unit 35. The evaluation result check unit 36 repeats the above-described operation until there is no evaluation condition 102 in which the evaluation value 107 is not stored in the registration candidate storage unit 35. The evaluation result check unit 36 provides the function of the evaluation result check unit 15 in FIG.

The feature extraction unit 37 first reads all the combinations of the biological image 101 and the evaluation value 107 for each evaluation condition 102 from the registration candidate storage unit 35 after the above-described operation by the evaluation result check unit 36 is completed. Then, this combination is stored in the registered biometric information storage unit 38 as the biometric feature 108.

The registered biometric information storage unit 38 stores, as registered biometric information, a biometric feature 108 that is a combination of the evaluation value 107 extracted by the image evaluation unit 34 and the biometric image 101 whose evaluation value 107 is an evaluation result. It is a storage unit. The registered biometric information storage unit 38 corresponds to the registered biometric information storage unit 14 in FIG.

The biometric information registration unit 10 in FIG. 9 is configured as described above. Note that some components of the biometric information registration unit 10 illustrated in FIG. 9 can be configured using a computer having a standard configuration.

Here, FIG. 10 will be described. FIG. 10 illustrates an example of the configuration of a computer.
The computer 40 includes an MPU 41, ROM 42, RAM 43, hard disk device 44, input device 45, display device 46, interface device 47, and recording medium drive device 48. These components are connected via a bus line 49, and various data can be exchanged under the management of the MPU 41.

An MPU (Micro Processing Unit) 41 is an arithmetic processing unit that controls the operation of the entire computer 40.
A ROM (Read Only Memory) 42 is a read-only semiconductor memory in which a predetermined basic control program is recorded in advance. The MPU 41 reads out and executes this basic control program when the computer 40 is activated, thereby enabling operation control of each component of the computer 40.

A RAM (Random Access Memory) 43 is a semiconductor memory that can be written and read at any time and used as a working storage area as necessary when the MPU 41 executes various control programs.

The hard disk device 44 is a storage device that stores various control programs executed by the MPU 41 and various data. The MPU 41 can perform various control processes by reading and executing a predetermined control program stored in the hard disk device 44.

The input device 45 is, for example, a keyboard device or a mouse device. For example, when operated by the administrator of the biometric authentication system 1 in FIG. 1, the input device 45 acquires various information input from the administrator associated with the operation content. Then, the acquired input information is sent to the MPU 41.

The display device 46 is a liquid crystal display, for example, and displays various texts and images according to display data sent from the MPU 41.
The interface device 47 manages the exchange of various information with various devices connected to the computer 40.

The recording medium driving device 48 is a device that reads various control programs and data recorded in the portable recording medium 50. The MPU 41 can read out and execute a predetermined control program recorded on the portable recording medium 50 via the recording medium driving device 48 to perform various control processes described later. As the portable recording medium 50, for example, a flash memory equipped with a CD-ROM (Compact Disc Read Only Memory), DVD-ROM (Digital Versatile Disc Disc Read Only Memory), or USB (Universal Serial Bus) standard connector. and so on.

In order to configure the biometric information registration unit 10 of FIG. 9 using such a computer 40, for example, a control program for causing the MPU 41 to perform biometric information registration processing described below is created. The created control program is stored in advance in the hard disk device 44 or the portable recording medium 50. In this program, for example, the hard disk device 44 is caused to function as the biological image temporary storage unit 33, the registration candidate storage unit 35, and the registered biological information storage unit 38. Further, a camera which is an example of the imaging unit 31 is connected to the interface device 47 of the computer 40 so that the computer 40 can control the imaging unit 31 to acquire the biological image 101 and take it into the computer 40. Keep it. Then, a predetermined instruction is given to the MPU 41 to read and execute this control program. By doing so, the imaging control unit 32, the biological image temporary storage unit 33, the image evaluation unit 34, the registration candidate storage unit 35, the evaluation result check unit 36, the feature extraction unit 37, and the registered biological information storage unit 38 respectively have. It is possible to provide the function with the computer 40.

Next, FIG. 11 will be described. FIG. 11 is a flowchart illustrating an example of the contents of the biometric information registration process performed by the biometric information registration unit 10 in FIG.
When this biometric information registration process is started, in FIG. 11, first, in S101, the imaging unit 31 is controlled to capture the biometric part of the person to be authenticated, thereby acquiring the biometric image 101, and the biometric image temporary storage unit 33. The imaging control unit 32 performs processing to be stored in. In FIG. 11, the biological image 101 stored in the biological image temporary storage unit 33 at this time is expressed as “CaptImage”. This notation is used in the following description.

Next, in S102, the image evaluation unit 34 performs a process of substituting the initial value “0” for the variable i.
Next, in S103, "CaptImage" is evaluated with a value corresponding to the value of the variable i at this point in time among a plurality of predetermined evaluation conditions 102, and an evaluation value as an evaluation result is obtained. The image evaluation part 34 performs the process to perform.

In FIG. 11, among the plurality of evaluation conditions 102, the one corresponding to the variable i is expressed as “Func [i]”, and the above-mentioned “CaptImage” is obtained by evaluating with this “Func [i]”. The obtained evaluation value is written as “Result [i]”. This notation is used in the following description.

Next, in S104, “Result [i]” acquired by the processing of S103 executed most recently is more than a predetermined evaluation threshold “Thresh [i]” predetermined for “Func [i]”. The image evaluation unit 34 performs a process of determining whether the evaluation value is large and higher than a predetermined value. When the image evaluation unit 34 determines that “Result [i]” acquired in the most recently executed S103 is “Thresh [i]” (when the determination result is Yes), the process proceeds to S105. . On the other hand, when the image evaluation unit 34 determines that “Result [i]” acquired by the most recently executed S103 is equal to or less than “Thresh [i]” (when the determination result is No), the process proceeds to S108. Advances.

Next, in S105, it is determined whether or not the biological image 101 is stored in a predetermined storage position in the registration candidate storage unit 35 where the biological image 101 having the highest evaluation value for “Func [i] i” is stored. The image evaluation part 34 performs the process to perform.

In FIG. 11, the biological image 101 having the highest evaluation value for “Func [i]” is denoted as “TempImage [i]”, and this notation is used in the following description.
In the determination process of S105, when the image evaluation unit 34 determines that “TempImage [i]” is not stored in the predetermined storage position in the registration candidate storage unit 35 (is empty) (when the determination result is Yes) ) Proceeds to S107. On the other hand, when the image evaluation unit 34 determines that “TempImage [i]” is stored at a predetermined storage position in the registration candidate storage unit 35 (when the determination result is No), the process proceeds to S106.

In S106, “Result [i]” obtained by the most recently executed processing of S103 is larger than the evaluation value for “Func [i]” stored in the registration candidate storage unit 35 at the time of this processing. The image evaluation unit 34 performs a process for determining whether or not.

In FIG. 11, the evaluation value for “Func [i]” stored in the registration candidate storage unit 35 at the time of this processing is indicated as “TempResult [i] 、”, and this notation is also used in the following description. We will use it.

In the determination process of S106, when the image evaluation unit 34 determines that “Result [i]” acquired by S103 executed most recently is larger than “TempResult [i]” (when the determination result is Yes). In step S107, the process proceeds. On the other hand, when the image evaluation unit 34 determines that “Result [i]” acquired by the most recently executed S103 is equal to or lower than “TempResult [i]” (when the determination result is No), the process proceeds to S108. Advances.

In S107, “CaptImage” is stored in the registration candidate storage unit 35 as a new “TempImage [i]”, and “Result [i]” obtained by the most recent execution of S103 is changed to a new “TempResult [i]”. Is stored in the registration candidate storage unit 35 by the image evaluation unit 34.

In S108, the image evaluation unit 34 performs a process of adding “1” to the current value of the variable i and substituting the addition result into the variable i.
In S109, the image evaluation unit 34 performs a process of determining whether or not the value of the variable i at the time of this process is equal to or greater than the number FuncNum of the evaluation conditions 102. Here, when the image evaluation unit 34 determines that the value of the variable i is equal to or greater than FuncNum (when the determination result is Yes), the process proceeds to S110. On the other hand, when the image evaluation unit 34 determines that the value of the variable i is less than FuncNum (when the determination result is No), the process returns to S103, and the above-described value of the variable i is below the value at the time of this process. The processing after S103 is performed again.

In S110, it is determined whether or not the biometric image 101 is stored in all the storage positions of “TempImage [i]” (where i = 0,..., (FuncNum−1)) in the registration candidate storage unit 35. The evaluation result check unit 36 performs the processing. Here, when the evaluation result check unit 36 determines that the biological image 101 is stored in all of the storage positions of “TempImage [i]” in the registration candidate storage unit 35 (when the determination result is Yes), S111. The process proceeds. On the other hand, when all the storage positions of “TempImage [i]” in the registration candidate storage unit 35 remain where the biological image 101 is not stored (when the determination result is No), the process proceeds to S101. Will return. In this case, the above-described processing after S101 is repeated until the biological image 101 is stored in all the storage positions of “TempImage [i]” in the registration candidate storage unit 35.

In S111, all combinations of “TempImage [i]” and “TempResult [i]” stored in the registration candidate storage unit 35 by the process of S107 (where i = 0,..., (FuncNum-1)) The feature extraction unit 37 performs a process of reading and storing it in the registered biometric information storage unit 38. In this process, the feature extraction unit 37 also stores information for specifying “Func [i]” in the registered biometric information storage unit 38 in association with the above combination. Thereafter, the biometric information registration process ends.

The process up to and above is the biometric information registration process. By performing this processing, the evaluation value extracted by the image evaluation unit 34, the evaluation condition 102 when the image evaluation unit 34 outputs the evaluation value, and the biological image 101 that is the evaluation target are associated with each other. The biometric feature 108 is stored in the registered biometric information storage unit 38.

In addition, the function of the biological image acquisition unit 11 in FIG. 1 is provided by the imaging control unit 32 executing the process of S101 in FIG. Moreover, the functions of the evaluation unit 12 and the evaluation value extraction unit 13 in FIG. 1 are provided by the image evaluation unit 34 executing the processing from S102 to S109 in FIG. Furthermore, the function of the evaluation result check unit 15 in FIG. 1 is provided by the evaluation result check unit 36 executing the process of S110 in FIG. 11, and the registration in FIG. 1 is performed by the feature extraction unit 37 executing the process of S111. The function of the biological information storage unit 14 is provided.

In the biometric information registration unit 10 in FIG. 9, the biometric image 101 acquired by the imaging unit 31 is temporarily stored in the biometric image temporary storage unit 33. Here, every time the imaging unit 31 acquires one biological image 101, the biological image 101 may be evaluated by the image evaluation unit 34. Alternatively, the imaging unit 31 may acquire a plurality of biological images 101, and then the image evaluation unit 34 may collectively evaluate the plurality of biological images 101.

Further, in the flowchart illustrated in FIG. 10, the value of the variable i is changed by “1” by the process of S108, so that the evaluation of “CaptImage による” by “Func [i]” is sequentially performed. Instead of this, the evaluation processing of “CaptImage” by “Func [i]” may be executed in a batch as parallel processing. For example, when a processing environment such as a multi-core multi-processor or a cloud system is used, processing time can be shortened by performing such parallel processing.

Moreover, the evaluation result check unit 36 in FIG. 9 further extracts the biological image until the evaluation values 107 for all the evaluation conditions 102 are extracted in the extraction of the evaluation values 107 for each evaluation condition 102 by the image evaluation unit 34. 101 is acquired by the imaging unit 31. Instead, even if the evaluation condition 102 for which the evaluation value 107 has not been extracted exists, if the number of the evaluation values 107 reaches a predetermined number RegistNum, the evaluation result check unit 36 uses the imaging unit 31 The control for acquiring the image 101 may be terminated.

Here, FIG. 12 will be described. FIG. 12 is a first modified example of the flowchart of FIG. 11, in which the above-described operation is performed by the evaluation result check unit 36. This determination process of S120 replaces the determination process of S110 in the flowchart of FIG.

The determination process in S120 is a process executed by the evaluation result check unit 36 when the result of the determination process in S109 in FIG. 11 is Yes. In this process, the evaluation result check unit 36 stores the biological image 101 stored in each storage position of “TempImage [i]” (where i = 0,..., (FuncNum−1)) in the registration candidate storage unit 35. To determine whether there are more than RegistNum copies. Here, when the evaluation result check unit 36 determines that at least RegistNum biometric images 101 are stored (when the determination result is Yes), the process proceeds to S111 in FIG. On the other hand, when the evaluation result check unit 36 determines that the number of stored biological images 101 is less than RegistNum (when the determination result is No), the process returns to S101 in FIG. In this case, the processing from S101 described above is repeated until at least RegistNum biometric images 101 are stored in the registration candidate storage unit 35.

When the process of S120 is executed, in subsequent S111, the feature extraction unit 37 stores all “TempImage [i]” and “TempResult [i]” stored in the registration candidate storage unit 35 at that time. The combination is read out and stored in the registered biometric information storage unit 38. However, i = 0,..., (RegistNum-1).

9 performs an operation of reading all combinations of the biological image 101 and the evaluation value 107 for each evaluation condition 102 from the registration candidate storage unit 35 and storing them in the registered biological information storage unit 38. Instead, the feature extraction unit 37 selects only a predetermined number of RegistNum sets in the descending order of the evaluation value 107 (in descending order of coincidence with the evaluation condition 102) from among the combinations of the biological image 101 and the evaluation value 107. You may make it memorize | store in the registration biometric information memory | storage part 38. FIG.

Here, FIG. 13 will be described. FIG. 13 is a second modification of the flowchart of FIG. 11 and is a case where the feature extraction unit 37 performs the above-described operation. The process of S121 replaces the process of S111 in the flowchart of FIG.

The process of S121 is a process executed by the feature extraction unit 37 when the result of the determination process of S110 of FIG. 11 is Yes. In this process, the feature extraction unit 37 firstly combines all “TempImage [i]” and “TempResult [i]” stored in the registration candidate storage unit 35 by the process of S107 (provided that i = 0, ..., (FuncNum-1)) is read. Subsequently, the feature extraction unit 37 causes the registered biometric information storage unit 38 to store only the RegistNum sets in descending order of the value of “TempResult [i]” in this combination. In this process, the feature extraction unit 37 also stores information for specifying “Func [i]” in the registered biometric information storage unit 38 in association with the above combination. After that, the biometric information registration process in FIG. 11 ends.

Next, FIG. 14 will be described. FIG. 14 illustrates an example of a specific configuration of the authentication unit 20 in the biometric authentication system 1 of FIG. In this example, the authentication unit 20 includes an imaging unit 51, an imaging control unit 52, a verification biometric image temporary storage unit 53, an image evaluation unit 54, a feature extraction unit 55, a verification biometric information storage unit 56, an evaluation order determination unit 57, And a feature matching unit 58. 14 is provided in the biometric information registration unit 10 in FIG. 9, and the biometric feature 108 described above is stored as registered biometric information.

The imaging unit 51 is a camera that captures a biometric part of a person to be authenticated and acquires a biometric image 111 for verification during authentication. The imaging control unit 52 controls the imaging unit 51 to acquire a plurality of verification subject biometric images 111.

The verification biometric image temporary storage unit 53 is a storage unit that temporarily stores the verification biometric image 111 acquired by the imaging unit 51.
The image evaluation unit 54 reads the verification biometric image 111 stored in the verification biometric image temporary storage unit 53 and uses the same evaluation conditions as those used by the image evaluation unit 34 in the biometric information registration unit 10 of FIG. 102 is evaluated, and an evaluation value 112 for each evaluation condition 102 is obtained. Next, the image evaluation unit 34 extracts, for each evaluation condition 102, an evaluation value 107 indicating a high degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the acquired evaluation value 107.

The feature extraction unit 55 reads the biometric image 111 for collation stored in the biometric image temporary storage unit 53 for collation, and the biometric feature for collation, which is information on the characteristics of the biological part shown in the biometric image 111 for collation. 113 is extracted and stored in the collation biometric information storage unit 56.

The collation biometric information storage unit 56 is a storage unit that stores the biometric feature 113 for collation extracted from the biometric image 111 for collation by the feature extraction unit 55.
The evaluation order determination unit 57 obtains the order in which the feature matching unit 58 reads the biometric feature 108 from the registered biometric information storage unit 38 and performs the later-described matching operation, and the image evaluation unit 54 evaluates the matching biometric image 111. This is determined based on the evaluation value 112. More specifically, the feature matching unit 58 first arranges the evaluation values 112 for each evaluation condition 102 in descending order of matching between the matching biological image 111 and the evaluation condition 102. Next, the feature matching unit 58 obtains the order of the evaluation conditions 102 by replacing the evaluation values 112 arranged in this way with the evaluation conditions 102 corresponding to the evaluation values 112. Then, the feature matching unit 58 determines the order of the evaluation conditions 102 as the reading order of the biometric image 101 associated with the evaluation condition 102 as the biometric feature 108 in the registered biometric information storage unit 38.

The feature matching unit 58 first reads the matching biometric feature 113 from the matching biometric information storage unit 56. Next, the feature matching unit 58 reads the biometric image 101 in the biometric feature 108 stored in the registered biometric information storage unit 38 in the order determined by the evaluation order determining unit 57 and matches it with the biometric feature 113 for matching. The feature matching unit 58 determines the level of similarity between the matching biometric image 111 and the biometric image 101 read out from the registered biometric information storage unit 38 in this way, and based on the determination result, the identity of the person to be authenticated Confirm and output the confirmation result.

As described above, in the authentication unit 20 in FIG. 10, the image evaluation unit 54 evaluates the biometric image 111 for verification by applying the same evaluation condition 102 as the image evaluation unit 34 in the biometric information registration unit 10 in FIG. 9. . Then, according to the determination by the evaluation order determination unit 57, the feature matching unit 58 is registered in association with the evaluation condition 102 when the evaluation value 112 indicates that the matching degree with the matching biological image 111 is high. The collation with the biological image 101 stored in the storage unit 38 is performed with priority.

In this way, the time required for authentication can be shortened. This is because the images of living body parts having the same position, posture, and shape can be expected to have high authentication accuracy, so the probability of correctly identifying the person as the person to be authenticated increases when the person authenticates. This is because it is often the case that a collation result is obtained immediately after performing one collation. On the other hand, in the verification of images of biological parts with greatly different positions, postures, and shapes, the authentication accuracy is low, so if the person authenticates as the person to be authenticated, there are many cases where the person cannot be correctly determined. As a result, the verification needs to be repeated many times.

The authentication unit 20 in FIG. 14 is configured as described above. Note that some components of the authentication unit 20 illustrated in FIG. 14 can be configured using a computer having a standard configuration as illustrated in FIG. For this purpose, for example, a control program for causing the MPU 41 to perform a collation process described below is created. The created control program is stored in advance in the hard disk device 44 or the portable recording medium 50. In this program, for example, the hard disk device 44 is caused to function as the verification biometric image temporary storage unit 53, the verification biometric information storage unit 56, and the registered biometric information storage unit 38. In addition, a camera which is an example of the imaging unit 51 is connected to the interface device 47 of the computer 40, and the computer 40 controls the imaging unit 51 to acquire the biometric image 111 for verification and causes the computer 40 to Make sure you can capture it. Then, a predetermined instruction is given to the MPU 41 to read and execute this control program. By doing so, the imaging control unit 52, the biometric image temporary storage unit 53, the image evaluation unit 54, the feature extraction unit 55, the verification biometric information storage unit 56, the evaluation order determination unit 57, the feature verification unit 58, and the registered biometrics The functions that the information storage unit 38 has can be provided by the computer 40.

Next, FIG. 15 will be described. FIG. 15 is a flowchart illustrating an example of the processing content of the collation processing performed by the authentication unit 20 of FIG.
When this verification process is started at the time of authentication, in FIG. 15, first, in S201, the imaging unit 51 is controlled to capture the biometric part of the person to be authenticated to acquire the verification biometric image 111. The imaging control unit 32 performs processing to be stored in the image temporary storage unit 53. In FIG. 15, the biometric image 111 for collation stored in the biometric image temporary storage unit 53 for collation at this time is denoted as “SampleImage”. This notation is used in the following description.

In step S <b> 202, the image evaluation unit 54 performs a process of substituting the initial value “0” for the variable i.
Next, in S203, "SampleImage" is evaluated with a value corresponding to the value at the time of processing of the variable i among the same plurality of evaluation conditions 102 used in the image evaluation unit 34 in FIG. The image evaluation unit 54 performs processing for obtaining an evaluation value that is the evaluation result.

In FIG. 15, among the plurality of evaluation conditions 102, the one corresponding to the variable i is expressed as “Func [i]” as in FIG. 11, and the above-described “SampleImage 、” is expressed as “Func [i]”. "EvalResult [i] .value" is expressed as the evaluation value obtained by the evaluation. This notation is used in the following description.

Next, in S204, the image evaluation unit 54 performs a process of substituting the value of the variable i at the time of execution of this process into “EvalResult [i] .index”. Note that “EvalResult [i] .index” is information for specifying “Func [i]” when “EvalResult [i] .value” is obtained.

In S205, the image evaluation unit 54 performs a process of adding “1” to the current value of the variable i and substituting the addition result into the variable i.
In S <b> 206, the image evaluation unit 54 performs a process of determining whether or not the value of the variable i at the time of this process is equal to or greater than the number FuncNum of the evaluation conditions 102. Here, when the image evaluation unit 54 determines that the value of the variable i is greater than or equal to FuncNum (when the determination result is Yes), the process proceeds to S207. On the other hand, when the image evaluation unit 54 determines that the value of the variable i is less than FuncNum (when the determination result is No), the process returns to S203, and the above-described value of the variable i is below the value at the time of this process. The processing after S203 is performed again.

In S207, the combination of “EvalResult [i] .index” and “EvalResult [i] .value” (where i = 0,..., (FuncNum-1)) is used as the value of “EvalResult [i] .value”. The evaluation order determination unit 57 performs a process of rearranging in descending order. In FIG. 15, a combination of “EvalResult [i] .index” and “EvalResult [i] .value” is expressed as “EvalResult [i]”. With this sort,
EvalResult [0] .value ≧ EvalResult [1] .value ≧
EvalResult [2] .value ≧ …… ≧ EvalResult [FuncNum−1] .value
It becomes.

In S208, “SampleImage” stored in the verification biometric image temporary storage unit 53 is read out, and information on the characteristics of the biological part shown in the image is extracted as the verification biometric feature 113, and the verification biometric information storage unit The feature extraction unit 55 performs processing to be stored in 56. In FIG. 15, the biometric feature 113 for matching is represented as “SampleFeature”, and this notation is used in the following description.

In S209, the feature matching unit 58 performs a process of substituting the initial value “0” for the variable j.
In S210, the feature matching unit 58 performs a process of substituting the value of “EvalResult [j] .index” into the variable k.

In S211, the feature matching unit 58 performs a process of matching the matching biometric feature 113 extracted from the matching biometric image 111 with the biometric image 101 read from the registered biometric information storage unit 38. In this process, first, “SampleFeature” is read from the collation biometric information storage unit 56 and the biometric image 101 associated with “Func [k]” in the biometric feature 108 stored in the registered biometric information storage unit 38. Is read out. Then, the read “SampleFeature” and the feature extracted from the biometric image 101 are collated, and a process of obtaining the similarity between the biometric image 111 for collation and the biometric image 101 is performed.

In FIG. 15, the feature extracted from the biometric image 101 associated with “Func [k] において” in the biometric feature 108 stored in the registered biometric information storage unit 38 is denoted as “Template [k]”. is doing.

Next, in S212, the feature matching unit 58 performs a process of determining whether or not the person to be authenticated has been confirmed to be the person. This determination process is performed by determining whether the similarity between the biometric image 111 for collation obtained by the collation process in S211 and the biometric image 101 is equal to or greater than a predetermined threshold value.

In this determination process, when the feature collation unit 58 determines that the person to be authenticated has been confirmed as the principal (when the determination result is Yes), the process proceeds to S213. On the other hand, in this determination process, when the feature matching unit 58 determines that the person to be authenticated cannot be confirmed as the person (when the determination result is No), the process proceeds to S214.

In S213, the feature matching unit 58 performs a process of outputting information indicating “person” as a matching result. Thereafter, the collation process ends.
In S214, the feature matching unit 58 performs a process of adding “1” to the current value of the variable j and substituting the addition result into the variable j.

In S215, the feature matching unit 58 performs a process of determining whether or not the value of the variable j at the time of this process is equal to or greater than the number FuncNum of the evaluation conditions 102. Here, when the feature matching unit 58 determines that the value of the variable j is equal to or greater than FuncNum (when the determination result is Yes), the process proceeds to S216. On the other hand, when the feature matching unit 58 determines that the value of the variable j is less than FuncNum (when the determination result is No), the process returns to S210, and the above-described value of the variable j is below the value at the time of this process. The processes after S210 are performed again.

In S <b> 216, the feature matching unit 58 performs a process of outputting information indicating “other” as a matching result. Thereafter, the collation process ends.
The process so far is the collation process. By performing this process, the degree of similarity between the registered biometric information stored in the registered biometric information storage unit 38 and the authenticated biometric information is determined, and the authentication subject's person based on the determination result is determined. Identity verification is performed. In addition, according to this process, as described above, since matching is performed preferentially between images showing images of living body parts having the same position, posture, and shape at the time of registration and at the time of authentication, The time required for verification can be shortened.

In the authentication unit 20 in FIG. 14, the feature matching unit 58 registers the biometric features 108 corresponding to the evaluation conditions 102 in the order of the evaluation conditions 102 having the highest evaluation value 112 when the biometric image 111 for matching is evaluated. The collation operation is performed by reading from the information storage unit 38. Instead, the biometric features 108 are arranged in the descending order of the evaluation value 107 and stored in the registered biometric information storage unit 38, and the feature matching unit 58 reads the biometric features 108 from the registered biometric information storage unit 38 in this storage order. You may make it perform collation operation | movement with the biometric image 111 for collation. In this way, it is possible to preferentially collate with the biometric image 111 for collation by giving priority to the biometric image 101 that is more appropriate as the registered biometric information, so that it can be expected to shorten the time required for collation of identity verification.

Here, FIG. 16A and FIG. 16B will be described. FIG. 16A is a third modification of the flowchart of FIG. 11, in which the feature extraction unit 37 performs an operation for storing the biometric features 108 in the registered biometric information storage unit 38 in the order described above. FIG. 16B is a first modification of the flowchart of FIG. 15, and is a case where the above-described operation is performed by the feature matching unit 58.

The process of S131 in FIG. 16A replaces the process of S111 in the flowchart of FIG. The process of S131 is a process executed by the feature extraction unit 37 when the result of the determination process of S110 of FIG. 11 is Yes. In this process, the feature extraction unit 37 firstly combines all “TempImage [i]” and “TempResult [i]” stored in the registration candidate storage unit 35 by the process of S107 (provided that i = 0, ..., (FuncNum-1)) is read. Subsequently, the feature extraction unit 37 stores this combination in the registered biometric information storage unit 38 in the descending order of the value of “TempResult [i]”. In this process, the feature extraction unit 37 also stores information for specifying “Func [i]” in the registered biometric information storage unit 38 in association with the above combination. After that, the biometric information registration process in FIG. 11 ends.

Also, the process from S221 to S223 in FIG. 16B replaces the process from S207 to S212 in the flowchart of FIG. However, if the result of the determination process in S215 in FIG. 15 is No, the process returns to S223 in FIG. 16B.

In FIG. 16B, in S221, “SampleImage” stored in the verification biometric image temporary storage unit 53 is read out, and the verification biometric feature 113 shown in the image is extracted and stored in the verification biometric information storage unit 56. The feature extraction unit 55 performs the processing. Note that the processing content of S221 is the same as the processing of S208 in FIG.

Next, in S222, the feature matching unit 58 performs a process of substituting the initial value “0” for the variable j. Note that the processing content of S222 is the same as the processing of S209 in FIG.
In step S223, the feature matching unit 58 performs a process of matching the matching biometric feature 113 extracted from the matching biometric image 111 with the biometric image 101 read from the registered biometric information storage unit 38. In this process, first, “SampleFeature” is read from the collation biometric information storage unit 56 and the biometric image 101 associated with “Func [j]” in the biometric feature 108 stored in the registered biometric information storage unit 38. Is read out. Then, the read “SampleFeature” and the feature extracted from the biological image 101 (that is, “Template [j]”) are collated to obtain the similarity between the collating biological image 111 and the biological image 101. Processing is performed. Thereafter, the process proceeds to S212 in FIG.

Next, FIG. 17 will be described. FIG. 17 is a modification of the specific configuration of the biometric information registration unit 10 illustrated in FIG.
In FIG. 17, the same components as those in FIG. 9 are denoted by the same reference numerals, and detailed description thereof will be omitted.

The configuration example illustrated in FIG. 17 is different from the configuration example illustrated in FIG. 9 only in that a guidance message display unit 39 is added.
In FIG. 17, the evaluation result check unit 36 determines whether there is an evaluation condition 102 in which the evaluation value 107 is not stored in the registration candidate storage unit 35, as in the configuration example of FIG. 9. Here, when it is determined that such an evaluation condition 102 exists, the evaluation result check unit 36 gives an instruction to the imaging control unit 32 to control the imaging unit 31 to acquire a further biological image 101. Let it be done. However, in the configuration example of FIG. 17, the evaluation result check unit 36 gives a predetermined instruction to the guidance message display unit 39 in parallel with the instruction to the imaging control unit 32.

The guidance message display unit 39 is a display device that displays a predetermined message when receiving a predetermined instruction from the evaluation result check unit 36. This message has a content for guiding the body part of the person to be authenticated so that the imaging unit 31 can acquire the biological image 101 that satisfies the evaluation condition 102 in which the evaluation value 107 is not stored in the registration candidate storage unit 35. The guidance message display unit 39 provides the function of the guidance output unit 16 of FIG.

When the guidance message display unit 39 displays such a message, the biometric image 101 that satisfies the evaluation condition 102 that is insufficient in the registration candidate storage unit 35 is quickly obtained because the authenticated person follows the content of the message. Acquisition can be expected.

DESCRIPTION OF SYMBOLS 1 Biometric authentication system 10 Biometric information registration part 11 Biometric image acquisition part 12 Evaluation part 13 Evaluation

value extraction part

14, 38 Registration biometric

information storage part

15, 36 Evaluation result check part 16 Guidance output part 20 Authentication part 31 Imaging part 32 Imaging control Unit 33 biological image

temporary storage unit

34, 54 image evaluation unit 35 registration candidate storage unit 37, 55 feature extraction unit 39 guidance message display unit 40 computer 41 MPU
42 ROM
43 RAM
44 Hard Disk Device 45 Input Device 46 Display Device 47 Interface Device 48 Recording Medium Drive Device 49 Bus Line 50 Portable Recording Medium 51 Imaging Unit 52 Imaging Control Unit 53 Biological Image Temporary Storage Unit for Verification 56 Verification Biometric Information Storage Unit 57 Evaluation Order Determination Unit 58 Feature collation unit 100 Registered biometric information by conventional method 101 Biometric image 102 Evaluation condition 103 Extracted biometric image 104 Evaluation result check 105 Registered biometric information 106 Authentication

biometric information

107, 112 Evaluation value 108 Biometric feature 111 Biometric image for collation 113 For collation Biological features

Claims

A biometric authentication system that performs identity verification of a person to be authenticated based on a similarity between the registered biometric information of the person to be authenticated registered in advance at the time of registration and the authentication biometric information of the person to be authenticated acquired at the time of authentication ,
A biometric image acquisition unit that acquires a plurality of biometric images of the person to be authenticated;
Each of the plurality of biological images acquired by the biological image acquisition unit is evaluated for each of a plurality of predetermined evaluation conditions, and the degree of coincidence with the evaluation conditions for the biological images is indicated. An evaluation unit that outputs an evaluation value for each of the plurality of evaluation conditions as an evaluation result of the biological image;
From the evaluation values for each of the plurality of biological images output by the evaluation unit, an evaluation value indicating the degree of coincidence that is equal to or higher than a predetermined evaluation threshold is extracted for each evaluation condition, and evaluation is performed by the extraction. When a plurality of values are extracted under the same evaluation condition, an evaluation value extraction unit that extracts an evaluation value having the highest degree of matching with the same evaluation condition from the plurality of extracted evaluation values;
A registered biometric information storage unit that stores a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit as registered biometric information of the person to be authenticated;
An authentication unit that determines the level of similarity between the registered biometric information stored in the registered biometric information storage unit and the authentication biometric information, and performs identity verification of the person to be authenticated based on the determination result; A biometric authentication system.
The evaluation condition includes information on the size of the image of the body part of the person to be authenticated on the biometric image as a condition that corresponds to an evaluation value,
The evaluation unit matches the size of the image of the body part of the person to be authenticated in the biometric image acquired by the biometric image acquisition unit on the biometric image and the size information included in the evaluation condition Output an evaluation value indicating the height of the degree,
The biometric authentication system according to claim 1.
The evaluation condition includes information on the inclination of the image of the biometric part of the person to be authenticated on the biometric image as a condition that corresponds to the evaluation value.
The evaluation unit is a degree of coincidence between an inclination on the biological image of an image of the biological part of the person to be authenticated in the biological image acquired by the biological image acquisition unit, and inclination information included in the evaluation condition. Output an evaluation value indicating the height of
The biometric authentication system according to claim 1.
The evaluation condition includes information on the shape of the image of the body part of the person to be authenticated on the biometric image as a condition that corresponds to the evaluation value.
The evaluation unit is a degree of coincidence between the shape on the biological image of the image of the biological part of the person to be authenticated in the biological image acquired by the biological image acquisition unit and information on the shape included in the evaluation condition Output an evaluation value indicating the height of
The biometric authentication system according to claim 1.
In the extraction of the evaluation value for each of the evaluation conditions by the evaluation value extraction unit, it is determined whether there is an evaluation condition for which the evaluation value has not been extracted, and there is an evaluation condition for which the evaluation value has not been extracted. If it is determined that the biometric image is acquired by the biometric image acquisition unit, the biometric image is evaluated by the evaluation unit, and the evaluation value is extracted by the evaluation value extraction unit. The biometric authentication system according to claim 1, further comprising an evaluation result check unit.
The evaluation result check unit is extracted by the evaluation value extraction unit even when there is an evaluation condition in which the evaluation value is not extracted in the evaluation value extraction for each of the evaluation conditions by the evaluation value extraction unit. The biometric authentication system according to claim 5, wherein when the evaluation value reaches a predetermined number, the control is terminated.
The registered biometric information storage unit has a degree of coincidence with the evaluation condition indicated by the evaluation value among the biometric images in which the evaluation value extracted by the evaluation value extraction unit is an evaluation result by the evaluation unit. The biometric authentication system according to claim 1, wherein a predetermined number of biometric images are stored in descending order.
The registered biometric information storage unit associates a biometric image corresponding to the evaluation value extracted by the evaluation value extraction unit with an evaluation condition when the evaluation unit outputs the evaluation value as the registered biometric information. Remember,
The authentication unit performs the same evaluation as the evaluation unit on the biometric image at the time of authentication of the person to be authenticated acquired as the authentication biometric information, and obtains an evaluation value for the biometric image at the time of the authentication. The degree of coincidence indicated by the evaluation value is high for a biometric image acquired for each evaluation condition and associated with the evaluation condition when the evaluation value is obtained and stored in the registered biometric information storage unit Read in order to determine the level of similarity with the biometric image at the time of authentication,
The biometric authentication system according to claim 1.
The registered biometric information storage unit has a high degree of coincidence with a biometric image whose evaluation value extracted by the evaluation value extraction unit is an evaluation result by the evaluation unit and the evaluation condition indicated by the evaluation value. Store them in the order of the highest
The authentication unit reads out biometric images stored in the registered biometric information storage unit in the order arranged in the registered biometric information storage unit, and determines whether the similarity with the authentication biometric information is high or low.
The biometric authentication system according to claim 1.
When the evaluation result check unit determines that there is an evaluation condition in which the evaluation value is not extracted in the extraction of the evaluation value for each evaluation condition by the evaluation value extraction unit, a biological image satisfying the evaluation condition is obtained. The biometric authentication system according to claim 1, further comprising a guidance output unit that outputs a guidance to the person to be authenticated for guiding the biological part of the person to be authenticated so that the biometric image acquisition unit is obtained.
A biometric authentication method for performing identity verification of a person to be authenticated based on a similarity between registered biometric information of a person to be authenticated registered in advance at the time of registration and the authentication biometric information of the person to be authenticated acquired at the time of authentication. ,
The evaluation unit evaluates each of the plurality of biological images of the person to be authenticated for each of a plurality of predetermined evaluation conditions, and determines the degree of coincidence with the evaluation conditions for the biological image. An evaluation value to be output as an evaluation result of the biological image for each of the plurality of evaluation conditions,
An evaluation value extraction unit extracts, for each of the evaluation conditions, an evaluation value indicating a high degree of coincidence that is equal to or higher than a predetermined evaluation threshold value from the evaluation values output by the evaluation unit for each of the plurality of biological images. In the case where a plurality of evaluation values are extracted by the extraction under the same evaluation condition, an evaluation value having the highest degree of matching with the same evaluation condition is further extracted from the plurality of extracted evaluation values,
A registered biometric information storage unit stores a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction unit as registered biometric information of the person to be authenticated,
The authentication unit determines the level of similarity between the registered biometric information stored in the registered biometric information storage unit and the authenticated biometric information, and performs identity verification of the person to be authenticated based on the determination result ,
Biometric authentication method.
A program for causing a computer to perform identity verification of a person to be authenticated based on the similarity between the registered biometric information of the authenticated person registered in advance at the time of registration and the authenticated biometric information of the authenticated person acquired at the time of authentication And the computer
Each of the plurality of biological images of the person to be authenticated is evaluated for each of a plurality of predetermined evaluation conditions, and an evaluation value indicating a high degree of coincidence with the evaluation conditions for the biological image is obtained. An evaluation process for outputting the evaluation result of the biological image for each of the plurality of evaluation conditions;
For each of the plurality of biological images, an evaluation value indicating a degree of coincidence that is equal to or higher than a predetermined evaluation threshold is extracted for each evaluation condition from the evaluation values output by the evaluation process, and evaluation is performed by the extraction. When a plurality of values are extracted under the same evaluation condition, an evaluation value extraction process for extracting an evaluation value having the highest degree of coincidence with the same evaluation condition from the plurality of extracted evaluation values;
A registered biometric information storage process for storing a plurality of biometric images corresponding to the evaluation values extracted by the evaluation value extraction process in the storage unit as registered biometric information of the person to be authenticated;
An authentication process for determining the degree of similarity between the registered biometric information stored in the storage unit and the authentication biometric information, and performing identity verification of the person to be authenticated based on the determination result;
A program to let you do.