JP5184279B2 - Finger vein authentication system - Google Patents

Description

  The present invention relates to a biometric authentication system that performs identity verification using a finger vein pattern as biometric information, and particularly relates to improvement of the authentication performance.

With the recent increase in security awareness and the expansion of network usage, there is a need for a personal authentication system that maintains strong security and is highly convenient in information security. As an authentication method, authentication using a password is the most common. However, in the authentication using a password, a simple character string is often set as a password so that the user can easily remember it, and it can be seen by others. Because it can be used by anyone other than the person by teaching or teaching, it is very vulnerable from a security perspective.
Various authentication methods have been conceived to solve these vulnerabilities, but biometric authentication that uses a part of the human body to verify the identity is stronger and more convenient. It is attracting attention as a method.
As biometric authentication technology, methods using fingerprints, voiceprints, irises, faces, and blood vessels of hands and fingers have been studied. Especially finger vein authentication using finger blood vessels is similar to fingerprints. It is difficult to counterfeit because it reads the internal features, not the surface of the living body, and there is less psychological resistance because it does not irradiate light like the iris, and the authentication performance is better than voiceprints and faces, Compared to the blood vessel pattern of the hand, there is an advantage that the device can be miniaturized.

Briefly describing finger vein authentication, near infrared light emitted from a light source is transmitted from a finger, and the transmitted light is imaged by a camera. The camera is equipped with an optical filter that passes only wavelengths in the near infrared region. In the captured image, the components in the blood absorb the near infrared light well, so the light of the blood vessel portion does not transmit and appears dark. A feature extraction process is performed on the image thus captured to create a template that emphasizes the blood vessel pattern of the finger.
The personal authentication is performed by pattern matching with a personal template created and registered in advance by the same method.

  What is important in personal authentication by pattern matching is the sharpness and reproducibility of the finger vein pattern. The clearer the finger vein pattern and the better the reproducibility of the finger vein pattern (the smaller the difference in pattern for each shooting), the easier it is to distinguish from other people's patterns, and the easier it is to identify them as their own patterns. . That is, by improving these, it is possible to improve the authentication accuracy (acceptance rate of others and rejection rate of identity) which is an index indicating the performance of biometric authentication.

As prior arts for improving the sharpness and reproducibility of the pattern in finger vein authentication, there are Patent Document 1 and Patent Document 2 below.
In Patent Document 1, the position of a finger is naturally guided to a specific position, and it is not necessary to perform alignment and rotation correction on a captured image, and stable matching can be performed, and a blood vessel pattern caused by finger compression or the like can be obtained. A finger vein authentication device that can improve the accuracy of authentication without omission and a finger placement guide are disclosed.
In Patent Document 2, feature point information of pattern information captured under an inappropriate imaging condition called an unauthorized template is stored in advance, and a template to be registered or authenticated is collated with an unauthorized template in advance. A biometric information authentication apparatus and an authentication method for disabling registration or authentication by considering it as an inappropriate template when the degree is high, and displaying a guidance message for solving a problem to a user are disclosed.

JP 2005-128936 A JP 2007-272501 A

As described above, in finger vein authentication, it is important that the clarity and reproducibility of the blood vessel pattern reflected in the template is good, but the following cases can be cited as factors that make them worse.
One is a factor related to how the user's finger is placed. For example, when placing a finger on the authentication device, if the finger is pressed strongly, the blood vessel is compressed and a template with a part of the blood vessel pattern missing is obtained. Further, in the state where the force is excessively stretched (a state where the finger is warped), the blood vessel is pressed by the tension of the epidermis, and the blood vessel pattern is partially lost in the same manner.
The other is a factor related to the usage environment. For example, when the external light that illuminates the entire authentication device changes, the brightness and contrast of the image to be captured change. Sunlight and illumination light also include near-infrared light, and the light wraps around and affects how the blood vessel pattern is imaged. Also, templates with partially missing blood vessel patterns due to the deterioration of blood flow due to blood vessel contraction and dryness due to dryness in women and the elderly in low temperature / humidity environments such as outdoors when the cold weather is severe in winter. Become.

The template created in such a case seems to be able to authenticate at first glance when registered. This is because the factors hardly change in a short time. For example, if the finger presses the same level of force, the template to be created is a part of the blood vessel pattern that is missing at the same level as when it was registered. However, the factors gradually change over time, and it is highly likely that it will be difficult to authenticate at a certain time in the future (the frequency of authentication failure increases). For example, as the finger strength changes as the finger gets used, the degree of missing blood vessel patterns reflected in the template differs from that at the time of registration.
However, in most cases, the user is unaware of the above factors, so it is very difficult to identify the cause of the difficulty of authentication and solve the problem. Another factor that makes it difficult to identify the cause is that the user cannot confirm the actually imaged blood vessel pattern. This is because finger vein authentication is authentication based on pattern matching, and in pattern matching, the pattern shape itself is a key for authentication. Therefore, if it is disclosed to the user, the confidentiality is significantly reduced. Therefore, the blood vessel pattern cannot be displayed on the screen. This also applies to system integrators as well as users.
The biometric authentication device is rarely used alone, and in most cases, is used in cooperation with other business systems by a system integrator. However, since the blood vessel pattern image is extremely important confidential information of the authentication device vendor, it is often not disclosed to the system integrator. Therefore, the authentication device vendor is ultimately responsible for pursuing the cause, and the time required to cope with it and the labor cost of maintenance personnel are enormous.
Therefore, in order to avoid such a situation, there is a means for preventing the template from being registered in a case where the sharpness and reproducibility of the blood vessel pattern is deteriorated, and the user himself / herself uses it after the registration. It is necessary to have a means for confirming how sharp and reproducible the template is, and a means for guiding the result so that it can be resolved by itself.

  The present invention has been made in view of the above problems, and provides a finger vein authentication system that realizes improvement in the clarity and reproducibility of a blood vessel pattern reflected in a template used for authentication in finger vein authentication. With the goal.

In order to achieve the above object, an authentication system according to the present invention is a finger vein authentication system that performs identity verification using biometric information related to a vein pattern of a human finger,
It is provided with a confirmation means for determining whether or not the finger vein pattern data registered in the storage means for performing identity verification is appropriate, displaying the determination result, and performing processing for prompting a re-registration operation according to the determination result. ,
The confirmation unit checks the degree of difference between the registered finger vein pattern data stored in the storage unit and the finger vein pattern data photographed by the photographing unit, and the sharpness of the finger vein pattern data. Based on the combination of sharpness data, determine whether the finger vein pattern data to be newly registered and the registered finger vein pattern data are appropriate,
The determination based on the difference between the finger vein patterns is determined by dividing the difference into four stages from the threshold determined from the difference frequency distribution when the finger vein patterns of the same fingers are collated.
The determination based on the sharpness of the vein pattern was obtained by binarizing the finger vein pattern image, extracting feature points from the thinned image, and calculating from the frequency distribution of the length of vein lines extracted based on the feature points. The sharpness of the vein pattern is determined in four stages .

  According to the authentication system of the present invention, it is possible to prevent a template from being registered in a case that deteriorates the sharpness and reproducibility of the finger vein pattern, and to what extent the template is still used by the user after registration. It is possible to improve the sharpness and reproducibility of the finger vein pattern reflected in the template used for authentication by confirming whether it is clear and reproducible, and guiding the result so that it can be resolved by itself. .

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a cross-sectional view illustrating the structure of a finger vein authentication device used in the finger vein authentication system of the present invention.
The finger vein authentication device is roughly composed of a light source unit 104 and an imaging unit 105.
The light source unit 104 is installed so as to irradiate the imaging unit 105 with near infrared light. In addition, a filter that transmits only near-infrared light is installed as an imaging window 106 between the light source unit 104 and the imaging unit 105. By inserting the finger 103 into the space between the light source unit 104 and the imaging window 106, the near-infrared light emitted from the light source unit 104 passes through the finger, and the imaging unit 105 captures the image through the imaging window 106. As a result, the blood vessel pattern image of the finger vein is obtained by the imaging unit 105.
Note that the light source unit 104 is installed in a lateral incidence method 101 in which near-infrared light is radiated from both sides of the finger 103 and an upper incidence method 102 in which the light source unit 104 is radiated from the upper surface of the finger 103. . In either method, the present invention can be similarly applied.

FIG. 2 is a system configuration block diagram showing an embodiment of a finger vein authentication system according to the present invention.
The finger vein authentication device 220 includes a camera module 221 that irradiates a finger with near-infrared light to image a vein, and an MPU 222 that controls the camera module and transmits a captured blood vessel pattern image of the finger vein to the PC 220. Is shown in FIG.
A PC (personal computer) 210 is a CPU 211 that executes software that provides a finger vein biometric authentication function, a memory 212 that temporarily reads software data and programs, and exchanges data with the CPU 211, and software data for a long time. An HDD (hard disk) 213 for storing the data and an I / O interface 214 for receiving a blood vessel pattern image of the finger vein from the finger vein authentication device 220.

FIG. 3 is a block diagram showing the configuration of the finger vein authentication system software and its functions according to the present invention.
There are many application examples of software using biometric authentication. As an embodiment of the present invention, software using finger vein authentication for logging in to a certain system will be described as an example.
The software includes registration software 310 and authentication software 320.
The registration software 310 includes a user information registration unit 311, a user information editing unit 312, and an authentication confirmation unit 313.
The user information registration unit 311 has a function of registering a finger vein pattern for logging in to the system by finger vein authentication. A finger vein pattern is acquired from the finger vein authentication device 220, a feature extraction process is performed, and a registration template is created. Further, login information (user ID, password, etc.) of the system to be logged in by finger vein authentication input by the user is associated with the registration template, and stored in the HDD 213 as user information.
The user information editing means 312 changes / deletes user information registered by the user information registration function 311. For example, there are re-acquisition of a registered template, change of login information of the target system, and deletion of specific user information.

The authentication confirmation means 313 is a function that forms the core of the present invention, and is a function that confirms whether or not the registered finger vein pattern is suitable for finger vein authentication. The difference from the authentication unit 321 described later is that the login to the target system is performed even if it is determined that the finger vein pattern is suitable for finger vein authentication only by checking whether the finger vein pattern is suitable for finger vein authentication. It is not. Instead, the appropriateness of the template is determined, the determination result is displayed on the screen, and the user is informed of how suitable the registered finger vein pattern is for finger vein authentication.
Further, the authentication confirmation unit 313 is executed immediately after the user information registration unit 311 ends, and if the determination result is not good, the user is notified to prompt re-registration, and the user information editing unit 312 is guided. To do. This is the first feature point of this embodiment.

The authentication software 320 includes an authentication unit 321, a login unit 322 to the target system, and an authentication confirmation unit 323.
The authentication unit 321 has a function of performing identity verification by finger vein authentication. A finger vein pattern of a finger to be authenticated is acquired by the finger vein authentication device 220, and a feature extraction process is performed to create a matching template. Then, a registered template is acquired from the user information stored in the HDD 213 by the user information registration function 311 and collated with the collation template to determine whether or not it matches the person's finger. If they match, the log-in means 323 for the target system is notified to that effect. Here, “match” means that the degree of difference between the finger vein pattern of the matching template and the registered template falls within a preset allowable range.
The log-in means 322 for the target system is a function for logging into the target system when the authentication is successful. Upon receiving a notification of authentication success from the authentication means 321, the login information of the target system associated with the registered template that has been successfully authenticated is acquired from the user information stored in the HDD 213 by the user information registration function 311. Log in to the target system using.

The authentication confirmation unit 323 has the same processing contents as the authentication confirmation unit 313 in the registered software 310. The registration software 310 differs from the authentication confirmation unit 313 in that the registration software 310 is used in combination with the user information registration unit 311, but here is used alone by the user. The purpose is to check how much the finger vein pattern already stored in the HDD 213 is suitable for finger vein authentication. If the determination result is not good, the user is notified to prompt re-registration and the user information is edited. Guide to means 312. This function is the second feature point of the present embodiment.
That is, the system according to the present embodiment activates the authentication confirmation unit 313 immediately after the registration procedure 311 in the registration software 310 is finished, confirms whether the newly registered finger vein pattern (template) is appropriate, and After the authentication confirmation means 323 in the authentication software 320 is activated and the finger vein pattern is registered in the HDD 213, how much the registered (registered template) is suitable for finger vein authentication of the same finger of the same user. It is characterized in that if the result of confirmation is not appropriate, the user is notified of this and the user is prompted to re-register.

Details of the authentication confirmation unit 313 will be described below with reference to the flowchart of FIG.
First, the finger vein of the finger to be authenticated by the finger vein authentication device 220 is photographed (step 401).
Next, a verification template is created from the photographed finger vein image (step 402).
Next, the registered template of the finger to be authenticated is read from the user information registered by the user information registration means 311 from the HDD 213 (step 403).
Next, the collation template created in step 402 and the registered template acquired in step 403 are collated, and the degree of difference is checked (step 404).
Here, the degree of difference is a value indicating how much difference there is between the collated matching template and the registered template, and indicates that the smaller the value is, the more matched.

If it is the authentication means 321, it is determined whether or not the authentication is successful (match) or authentication failure (mismatch) depending on whether or not it falls below a preset difference threshold based on the difference obtained so far. However, the authentication confirmation means 313 next determines the appropriateness of the template based on the matching template created in step 402, the registered template obtained in step 403, and the degree of difference obtained in step 404. Perform (step 405).
The determination result is acquired in four stages (good ← {◎, ○, △, ×} → bad). Details of template suitability determination will be described later.
Finally, as the authentication confirmation result, the content corresponding to the template appropriateness determination acquired in step 405 is displayed on the screen.

FIG. 5 is an example of a screen that displays the result of the authentication confirmation unit 313.
The screen displays a determination result symbol display portion (511, 521, 531, 541), a message display portion (512, 522, 532, 542) according to the determination result, and a user operation according to the determination result. Button (513, 523, 533, 534, 543).
If the determination result is ◎ or ○, the finger currently used by the user is suitable for finger vein authentication, so a message to that effect is displayed and only a button (513, 523) for ending authentication confirmation is displayed. To do.
Also, if the determination result is △, the finger currently used by the user may not be suitable for finger vein authentication, so the authentication confirmation function is executed again, and if the determination still does not change, A message is displayed stating that re-registration with your finger is recommended. Therefore, the button is a button 533 that leads to a screen for re-registering a finger (calling the user information registration means 311 in FIG. 3 or a separate registration screen), and a button for returning without re-registration for the time being Two of 534 are displayed.
If the judgment result is x, a message is displayed indicating that re-registration with another finger is necessary because the finger currently used by the user is not suitable for finger vein authentication, and re-registration is always performed. Only the button 543 for guiding to the screen for re-registering the finger is displayed.

Details of template appropriateness determination will be described with reference to the flowchart of FIG.
The appropriateness of the template is determined by integrating determinations from two different viewpoints.
First, a determination is made based on the difference (reproducibility) of the vein pattern (step 701).
The degree of difference obtained in step 404 in FIG. 4 is input to obtain a determination result in four stages (good ← {◎, ○, Δ, ×} → bad). Details of the determination based on the difference in vein pattern will be described later.
Next, a determination is made based on the sharpness of the vein pattern (step 702). The collation template and the registered template acquired in steps 402 and 403 in FIG. 4 are input, and a determination result in four stages (good ← {◎, ○, Δ, x} → bad) is obtained. Details of the determination based on the sharpness of the vein pattern will be described later.
Finally, the final determination result of the template suitability is determined from the above two types of determination results based on the table of FIG. 14 (step 703). The result of the final determination is four stages (good ← {◎, ○, △, ×} → bad).

Details of the determination based on the difference between the vein patterns will be described with reference to FIG.
FIG. 7 is a graph showing the frequency distribution of dissimilarities obtained by matching the same fingers. The horizontal axis represents the degree of difference obtained by matching the matching template with the registered template, and the vertical axis represents the frequency at which the degree of difference appears.
The authentication threshold is determined so that the authentication accuracy (person rejection rate) calculated from the frequency distribution obtained from a sufficient number of sample data to evaluate the target authentication accuracy satisfies the target.
In the determination of authentication, when the degree of difference is smaller than the threshold value for authentication, one determination is made for two cases of matching (authentication success), and when the degree of difference is large, the determination is one for two. Judgment is made in four stages (good ← {◎, ○, △, ×} → bad) by two thresholds A to C. The thresholds A and B are determined by experiments using sample data for each application system of the present invention. The threshold C is the same as the authentication threshold.

Details of the determination based on the sharpness of the vein pattern will be described with reference to the flowchart of FIG.
First, since the template is encoded, it is restored to the vein pattern image format (step 801).
Next, the vein pattern image is binarized (step 802). FIG. 9 shows an example 901 in which a vein pattern image of a template with clear veins is binarized and an example 902 in which a vein pattern image of a template with unclear veins is binarized. Here, white represents a region with veins and black represents a region without veins. A template with clear veins results in an image in which white areas are connected, whereas an unclear template results in an image in which some of the white areas are scraped and scattered.
Next, the binarized vein pattern image is thinned, and the white area is converted into a line having a width of 1 dot (step 803).
FIG. 10 shows an example 1001 in which a binarized image of a template with clear veins is thinned, and an example 1002 in which a binarized image of a template with unclear veins is thinned. A template with clear veins results in an image in which continuous lines with long white areas are combined, whereas a template with unclear veins results in an image in which short lines with finely divided white areas are scattered.

Next, feature points (end points, branch points, intersection points) are extracted from the thinned image (step 804).
FIG. 11 shows an example in which feature points are extracted from a thinned image of a template with clear veins.
Next, a vein line is extracted based on the extracted feature points (step 805). The vein line is extracted by tracking a point sequence between the feature points.
Next, the frequency distribution of length (number of dots) is totaled from all the extracted vein lines (step 806).

FIG. 12 shows an example 1201 of vein length frequency distribution of a template with clear veins and an example 1202 of vein line length frequency distribution of a template with unclear veins. The horizontal axis of the graph is the venous line length, and the vertical axis is the appearance frequency of the venous line length. Compared with the template distribution with clear veins, the template distribution with unclear veins has a steep distribution in which the frequency is concentrated on the shorter vein line length.
Finally, the value of the vein sharpness is calculated from the frequency distribution of the vein line length (step 807), and based on the value, there are four levels (good ← {◎, ○, Δ) with three thresholds A to C. , X} → bad) (step 808).
FIG. 13 shows a calculation formula 1301 for calculating vein sharpness and a table 1202 for determining a determination result from the calculated value. The greater the number of long vein lines, the greater the value of the vein definition.
The thresholds A to C are determined by experiments using sample data for each application system of the present invention.

It is sectional drawing which shows the structure of the finger vein authentication apparatus used for this invention. It is a system configuration block diagram showing an embodiment of a finger vein authentication system according to the present invention. It is a block diagram which shows the structure of the software of the finger vein authentication system which concerns on this invention, and its function. It is a flowchart which shows the process of the authentication confirmation means. It is a figure which shows an example of the screen which displays the determination result of authentication confirmation. It is a flowchart which shows the appropriateness determination process of a template. It is a figure which shows the determination method of the template appropriateness degree by frequency distribution of the difference degree by collation of the same fingers, and a difference degree. It is a flowchart which shows the appropriateness determination process of the template by the clarity of a vein pattern. It is a figure which shows the example of the image which binarized the vein pattern image decompress | restored from the template. It is a figure which shows the example of the image which thinned the binarized vein pattern image. It is a figure which shows the example which extracted the feature point from the thinned vein pattern image. It is a figure which shows the example which totaled the frequency distribution of the venous line length. It is a figure which shows the formula which calculates the clarity of a vein pattern, and the determination method of a template appropriateness from the result. It is a figure which shows the determination method of final template appropriateness.

Explanation of symbols

101 Lateral incidence type finger vein authentication device 102 Upper incidence type finger vein authentication device 103 Finger 104 Light source unit 105 Imaging unit 210 PC
211 CPU
212 Memory 213 HDD
214 I / O interface 220 Finger vein authentication device 221 Camera module 222 MPU
510 Authentication confirmation screen (judgment: ◎)
511 Judgment result display (judgment: ◎)
512 Message display (judgment: ◎)
513 End button 520 Authentication confirmation screen (judgment: ○)
521 Judgment result display (judgment: ○)
522 Message display (judgment: ○)
523 End button 530 Confirmation screen for authentication (judgment: △)
531 Judgment result display (judgment: △)
532 Message display (judgment: △)
533 Re-registration button 534 End button 540 Authentication confirmation screen (judgment: ×)
541 Judgment result display (judgment: x)
542 Message display (judgment: ×)
543 Re-registration button 701 Frequency distribution curve 901 of the degree of difference by collation of the same finger Binary vein pattern image (example where the vein is clear)
902 Binarized vein pattern image (example of unclear veins)
1001 Thinned vein pattern image (example of clear veins)
1002 Thinned vein pattern image (example of unclear veins)
1201 Venous line length frequency distribution (example of clear veins)
1202 Frequency distribution of vein line length (example of unclear veins)

Claims (1)

A finger vein authentication system that performs identity verification using biometric information related to a human finger vein pattern,
It is provided with a confirmation means for determining whether or not the finger vein pattern data registered in the storage means for performing identity verification is appropriate, displaying the determination result, and performing processing for prompting a re-registration operation according to the determination result. ,
The confirmation unit checks the degree of difference between the registered finger vein pattern data stored in the storage unit and the finger vein pattern data photographed by the photographing unit, and the sharpness of the finger vein pattern data. Based on the combination of sharpness data, determine whether the finger vein pattern data to be newly registered and the registered finger vein pattern data are appropriate,
The determination based on the difference between the finger vein patterns is determined by dividing the difference into four stages from the threshold determined from the difference frequency distribution when the finger vein patterns of the same fingers are collated.
The determination based on the sharpness of the vein pattern was obtained by binarizing the finger vein pattern image, extracting feature points from the thinned image, and calculating from the frequency distribution of the length of vein lines extracted based on the feature points. A finger vein authentication system characterized by determining the sharpness of a vein pattern in four stages .

Images