JP4207937B2 - Personal authentication device - Google Patents

Description

The present invention relates to an apparatus and a method for identifying an individual using human biometric information, particularly a finger vein pattern.

As security technology for personal property and information is regarded as important, biometric authentication using human biometric information is attracting attention as a personal authentication technology that is highly convenient and confidential. Conventional biometrics authentication technologies have been devised using fingerprints, irises, voices, faces, veins on the back of the hand, and finger veins. In particular, authentication technology using finger veins has the feature that psychological resistance is low because authentication can be performed simply by shining light on the finger, and that anti-counterfeiting is excellent because internal information of the living body is used.

Authentication using the finger vein is realized as follows. When infrared light is irradiated from the back side or side surface of the finger toward the inside of the finger, the light is scattered inside the finger and then emitted to the outside. At this time, since hemoglobin in the blood absorbs infrared light more than the surrounding tissue, when the transmitted light emitted from the ventral side of the finger is imaged, blood vessels distributed subcutaneously on the ventral side of the finger, that is, the finger The veins are visualized as dark shadow patterns. This image is registered in advance, and it is determined whether or not the user is a registrant by obtaining a correlation with the presented finger image, and personal authentication is performed.

Regarding an authentication device using a finger vein, Patent Document 1 discloses a personal authentication device using a finger vein, and in particular, in order to reduce a loss of light amount when photographing, a finger image with an optical fiber in close contact with the finger. Is disclosed. Patent Document 2 discloses an apparatus and method for performing personal authentication using a finger vein pattern image in an environment where non-contact property is required. Patent Document 3 discloses an apparatus using a finger guide part and a fingertip part button switch for making the photographing conditions for each authentication in finger vein pattern authentication uniform.

Japanese Unexamined Patent Publication No. 7-21373

Japanese Patent Laid-Open No. 2002-83298 JP 2003-30632 A

In the conventional technique, in order to capture a finger vein image with high reproducibility, a guide unit for fixing or guiding the finger position is installed in the authentication apparatus. By inserting the finger along this guide part, the finger shape fluctuation factors such as the finger joint angle, the three-dimensional rotation angle fluctuation with respect to the finger central axis, the camera distance fluctuation, etc. are suppressed. The shape of the vein pattern can be imaged with good reproducibility.

However, a user who is not proficient in using the authentication device may mistakenly insert a finger, such as bringing the finger into contact with the finger guide, bending the joint by applying a strong force to the finger, or the finger insertion angle being unstable. Or deformation may occur. In addition, a user who has been using the device for a long time can present his / her finger with good reproducibility, but may change in a manner different from the manner in which the finger is placed at the time of registration, and the correlation with the registration pattern may become low. Further, there are cases where a malicious user inserts a forgery finger that is not a living body or a living finger affixed with a forgery pattern into an authentication device, and registration / authentication is performed.

An object of the present invention is to provide an authentication device that is easy to use and has high accuracy by preventing erroneous insertion and deformation of a finger and insertion of a forged finger by a malicious user.

In order to achieve the above object, the outline of representative ones of the inventions disclosed in the present application will be briefly described as follows.

An imaging unit that captures a finger vein image, a light source that emits light transmitted through the finger, an image calculation unit that performs image matching, a guide unit that indicates an imaging position of the finger, and a contact between the finger and the guide unit are detected. A personal authentication device provided with a detection unit.

In addition, a light source that emits reflected light that illuminates the ventral side of the finger, a light source that transmits through the tip of the finger, a light receiving element that receives the light source that transmits through the tip of the finger, and information that guides the device to regular use. Display means. Furthermore, a configuration is disclosed such as means for switching between collation for all registered images and collation for only registered images corresponding to the user in accordance with the stability of authentication, means for updating registered data, and the like.

According to the present invention, it is possible to prevent a reduction in the recognition rate of a finger vein authentication device at a low cost from an erroneous insertion of a finger by a user who is not familiar with the device or an insertion of a forged finger by a malicious user. Can do.

Hereinafter, embodiments of the present invention will be described in detail.

FIG. 1 is a schematic diagram of a system configuration for realizing the present invention. User is finger 1 during authentication
Is inserted into the authentication device 6. At this time, the tip of the finger 1 is placed on the fingertip guide portion 4, and the base side of the finger 1 is placed on the finger placement guide portion 9. When the finger 1 is inserted into the apparatus, the authentication process is started. The authentication device 6 is a cross-sectional view from the side.

Authentication is performed by the following procedure. Infrared light is irradiated from the back side of the finger 1 from the light source 3. The light passes through the finger 1 and the finger insertion guide portion 5 and passes through the optical filter 7 that transmits only the infrared wavelength light, and reaches the imaging device 2. The light is converted into an electrical signal by the imaging device 2 and is taken into the computer 10 as an image via the image input means 18. The captured image is once stored in the memory 12. Then, the finger vein image registered in advance is stored in the memory 12 from the storage device 14, and the CPU 11 collates the registered image with the input image by the program stored in the memory 12. In the collation processing, a correlation value between the two images to be compared is calculated, and it is determined whether the image matches the registered image according to the value. An individual is authenticated in accordance with the result, and when the authentication is correctly performed, a process at the time of authentication is performed on the control target of the authentication system.

The light source 3 is composed of components that emit infrared light, such as LEDs and halogen lamps. In this embodiment, it is assumed that a plurality of LEDs are arranged in the longitudinal direction of the finger, but a light-emitting element having one elongated shape may be used. By arranging the light sources in the longitudinal direction of the finger, the entire finger can be illuminated evenly,
The finger vein pattern can be acquired clearly overall. Each LED may be controlled to have the same brightness, or each LED may be controlled individually. When each LED is controlled individually, it is possible to increase the amount of light in the thick portion of the finger while suppressing the amount of light in the portion with high light transmittance of the finger 1 such as the joint portion, and a clear finger vein pattern can be obtained.

An example of a method for controlling the amount of light emitted from the light source 3 will be described below. First, at a certain initial value, L
When the ED is irradiated, the luminance value at the position of each area where the LED appears on the image is examined. However, it is assumed that the position of each LED on the image is known. If the brightness value is within the range where it is recognized that the brightness is appropriate, the amount of light continues to be irradiated. If it is brighter than that, the light amount of the LED corresponding to the image position is lowered, and if it is dark, the amount of light is increased. . Such a determination process is performed in the area of each LED. Here, the amount of change in the amount of light to be controlled may be a fixed value or a variable value.

An example of the light amount control when the change amount of the light amount control becomes a variable value is shown below. First, each LED
The initial value of the maximum value and the minimum value of the light intensity control range is determined. Here, the light amount control range is a range in which the control value of the light amount output signal given to the LED can be taken, and the light amount is controlled so as not to exceed this range. The initial values of the maximum value and the minimum value are determined according to the characteristics of each LED. Next, the control value of the light amount of all the LEDs is set at the center of the control range, light is irradiated to acquire an image, and the luminance value around each LED position is examined.

If the luminance value is dark, it is necessary to increase the light quantity of the LED located at that location. Therefore, the light amount control value when the brightness value around the LED position is first examined is set as a new minimum value of the light amount control range of the LED located at that location. After narrowing the control range, the control value of the amount of light emitted in the next stage is set at the center of the control range. Thereby, in the next stage, a stronger light amount is irradiated than in the previous stage.

Further, when the luminance value is bright, it is necessary to weaken the light amount of the LED located at that location.
Therefore, the control value of the light amount when the luminance value around the LED position is first examined is set as a new maximum value of the light amount control range of the LED located at that location. After narrowing the control range, the control value of the amount of light emitted in the next stage is set at the center of the control range. As a result, a weaker amount of light is irradiated in the next stage than in the previous stage.

If the luminance value is appropriate, the irradiation is continued without changing the light quantity of the LED. Such brightness determination and control are repeated for all the LEDs until the light quantity of all the LEDs does not change.

According to the above method, the amount of change in the amount of light is large at the beginning of the control, and becomes a minute amount of change as the control proceeds. Therefore, compared with the case where the amount of change is fixed, fine control is possible and the processing speed can be improved. However, as the control proceeds, the light intensity control range becomes narrower.
If there is no optimal solution in the range, control may not converge. In order to solve this, a process for resetting the maximum value and the minimum value of the light amount control range to the initial state is provided for an area that has not converged after several times of control. Thereby, the control range can be expanded again.

As a method for starting the authentication, the authentication switch 8 is pressed, the image of the finger 1 is always captured by the image input means 18, and the CPU 11 determines that the finger 1 is completely inserted. There is. In this embodiment, the former method is used. When the user presses down the fingertip guide unit 4 with the fingertip, the authentication switch 8 is pressed at the same time. At this time, a signal that the authentication switch 8 has been conducted is transmitted to the inside of the computer 10 via the interface 13, and the CPU 11 detects the pressed state of the authentication switch 8 to start authentication.

The fingertip guide unit 4 and the finger placement guide unit 9 support the finger 1 at two points to suppress the horizontal displacement of the finger 1, keep the distance between the finger 1 and the imaging device 2 constant, and further the finger 1 To keep the ventral side of the device from touching the device. If the ventral side of the finger 1 comes into contact with the device, the veins distributed on the ventral side of the finger are compressed so that blood cannot pass through and the pattern cannot be seen.

Image processing for extracting a finger vein pattern is performed as a previous step for obtaining a correlation between finger vein patterns. An example of the technique is shown below. The finger vein pattern has a darker luminance value than the surrounding area without blood vessels. In other words, in the luminance value profile of the cross section perpendicular to the running direction of the blood vessel, a valley or a depression having the peak at the center of the blood vessel is seen. Therefore, for all profiles in the vertical and horizontal directions of the image, the center of the valley of the luminance value or the depression is found. To find it, the opening angle or curvature of the profile curve may be calculated, or the depression may be detected based on the average value theorem. Then, the luminance value at that position is increased so that only that point is emphasized. Then, the center of the dark line of the entire image is emphasized, and a finger vein pattern is extracted. By this process, even when the dark line is thin, when there is a close dark line, or when the light amount varies, the dark line can be emphasized stably and clearly. Furthermore, since the center of the blood vessel is extracted, it is possible to perform authentication without being affected by the fluctuation of the blood vessel width with respect to the expansion and contraction of the blood vessel due to a change in temperature, a change in physical condition, and the like.

If the authentication is started in a state where the finger 1 is not in contact with the finger placement guide unit 9, the magnification rate of the finger on the image is different from the state in which the finger 1 is in contact. In order to absorb such fluctuations in the enlargement ratio, the following image processing is performed. First, the outline of the finger shown on the image is extracted. As the contour extraction method, conventional general image processing, such as a method of applying an edge enhancement filter to the entire image to check the connection of lines, a method of sequentially tracking edge portions and obtaining a locus thereof, can be used. Next, the width of the acquired finger outline is examined. As a method for determining the contour width, there are a method for detecting a joint portion of a finger by image processing and obtaining the contour width at that position, a method for obtaining the contour width at the center position of the image, and the like. Thereafter, the enlargement ratio of the finger is normalized based on the contour width. As a normalization method, when the contour width is larger than a specified value, the entire image is reduced, and when the contour width is smaller, the entire image is enlarged to match the contour width to the specified value, or the contour width is uniformly specified. There is a method of elastically expanding and reducing the width so as to be a width of 2 and correcting the two contour lines to be parallel straight lines. As a result, even if enlargement / reduction is included in the imaging state of the finger, it is possible to correct to a constant enlargement ratio, and even when authentication is started in a state where the finger 1 is away from the finger placement guide unit 9, it is recognized correctly can do.
FIG. 2 is an example of an authentication device that detects a state in which the finger 1 is in contact with the finger insertion guide unit 5.

At the time of imaging the finger vein pattern, the finger 1 is supported at two points by the finger placement guide portion 9 and the fingertip guide portion 4 so that the ventral side of the finger 1 is not in contact. At this time, when stress is applied to the joint of the finger 1 with the fingertip as a fulcrum, the finger is bent so that the ventral side of the finger is warped. Then, since the joint position of the finger 1 fluctuates up and down, the reproducibility for each authentication is lowered. In order to prevent this, the finger insertion guide portion 5 is provided, and the space is limited so that fluctuation does not occur.

However, when the finger joint is moved up and down as described above, the ventral side of the finger 1 is moved to the finger insertion guide portion 5.
To touch. Then, since the contact part and joint pressure are added, blood will escape from the peripheral part. In this state, when authentication is started by pressing down the fingertip guide unit 4 or confirming finger insertion by the image processing unit, finger veins are not reflected around the first joint in the captured finger image. . Accordingly, information characterizing the individual is lost, and the recognition rate is significantly reduced.

In order to detect such a contact of the finger 1 with the finger insertion guide portion 5, a push switch 20 is installed in a portion that supports the finger insertion guide portion 5. When the finger insertion guide portion 5 is entirely pushed down by the contact of the finger 1, the push switch 20 is also pushed down. The CPU 11 detects this state,
If the push switch 20 is depressed, the authentication is not started even if the authentication switch 8 is depressed. Moreover, the finger insertion guide part 5 can also be made into the shape and arrangement | positioning which are touched when stress is applied to a joint by using a fingertip as a fulcrum, and it is bent and bent so that the belly side of a finger may warp. This prevents authentication from being performed when the vein pattern is missing due to finger surface contact,
Decrease in recognition rate can be prevented.

As a means for detecting a finger contact, instead of the push switch 20, an electric type, a capacitance type, a pressure type, etc. made of a material that does not block infrared light or a shape that does not hinder the transmission of infrared light. The contact sensors may be distributed on the surface of the finger insertion guide portion 5. As a result, the push switch 2
It is not necessary to move the stroke only by depressing 0, and the sensitivity to contact can be increased. Further, the push switch 20 may be provided only on the support portion on the fingertip side of the finger insertion guide portion 5 or only on the support portion on the finger base side. At this time, the finger insertion guide portion 5 and the authentication device 6 are joined, and the joint portion on the side where the push switch 20 is not provided is located at the lower side with the finger insertion guide portion 5 centered on the joint portion when the finger 1 is contacted. It is assumed that the shape is such that it is slightly rotated and the push switch 20 is depressed. As a result, it is possible to reduce the processing for sensing the pressing of the push switch 20 and the components necessary for it, and to reduce the device manufacturing cost.

Further, the contact between the finger 1 and the finger insertion guide portion 5 may be detected by image processing. As an example of a method based on image processing, an image with contact has a tendency that the finger area of the finger joint portion becomes brighter than an image without contact, and the contact is detected by evaluating this brightness. can do. In the image processing method, the push switch 20 is completely unnecessary, so that the device structure is simplified and the device manufacturing cost can be reduced.

When it is detected that the finger 1 and the push switch 5 are in contact, or when authentication is started in that state, a warning is issued from the display means 15 or the speaker 17 to notify the user of the contact of the finger 1. Can do. This prompts the user to re-authenticate, improves the proficiency level of the device, and prevents the recognition rate from decreasing.

FIG. 3 is a configuration example of an authentication device that detects the state of the surface on the ventral side of the finger. The reflected light source 22 is installed at a position that irradiates the ventral surface of the finger 1. The reflected light source 22 emits light having a wavelength that can pass through the optical filter 7. The amount of light from the reflected light source 22 is adjusted to the amount of light that can be captured by the finger surface.

When the light source 3 is irradiated with the reflected light source 22 turned off, an image showing a finger vein is captured. When the reflected light source 22 is irradiated with the light source 3 turned off, an image showing the finger surface is captured. Is done. That is, when performing authentication, the light source 3 is turned on and the reflected light source 22 is turned off. However, the state of turning on and off is switched immediately before or after that, and the finger image by the reflected light on the finger surface is acquired and analyzed. It is possible to evaluate whether or not the finger image obtained for performing is a valid finger.

FIG. 4 is an example of an image of the finger surface imaged by the authentication device of FIG. When the reflected light source 22 is irradiated without irradiating the light source 3, the finger vein pattern is almost invisible, but information seen on the finger surface, for example, wrinkles present on the joints of the finger, dirt on the finger surface, and the like are imaged.

FIG. 4A shows an image of the surface of the finger when there is no rotation with respect to the central axis in the longitudinal direction of the finger. The joint wrinkles 30 are distributed up and down with respect to the central axis of the finger 1. FIG. 4B is an image of the surface of the finger when there is rotation of the central axis of the finger. The joint wrinkle 30 is largely biased downward, and an end point 31 of the joint wrinkle is seen. Further, the fingertip 32 has a nail 32 in the direction opposite to the direction of the bias of the joint wrinkle 30.
It can be seen that the image is also captured. That is, by examining the imaging state of the finger joint wrinkle 30 or the nail 32, it can be determined whether the finger 1 is rotating about the central axis. When this rotation is detected, a warning is issued to the user and authentication processing is not performed. This makes it possible to reduce misrecognition. Below, an example of the method of detecting the position bias of the finger joint wrinkle 30 using image processing is shown. First, an edge enhancement filter is applied to the obtained image to enhance the vertical component line of the image. Next, the emphasized image is binarized and thinned. afterwards,
It is evaluated by the line tracking process whether or not the lines are continuously arranged in the vertical direction. If the line is a wrinkled part of the joint, tracking over a relatively long distance will be performed. Based on this tracking length,
Determine if it is a joint wrinkle. Thereafter, the position of the end point of the traced line segment is obtained, and the distance relationship with the contour position of the finger is calculated to evaluate whether the position of the end point 31 of the joint wrinkle is biased. Further, as a method of detecting rotation without obtaining the end point 31 of the joint wrinkle, there is a method of calculating the position of the center of gravity of the joint wrinkle distribution and evaluating whether the position in the vertical direction is biased. In addition, the bias of the joint wrinkle 30 may be detected using a conventional technique based on image recognition. By the above image processing, it is possible to detect the rotation of the finger without using a special sensor, and it is possible to prevent the recognition rate from being lowered without increasing the device manufacturing cost.
Below, an example of the image processing which detects the nail | claw 32 is shown. First of all, as a result of image texture analysis using general techniques such as density histogram analysis and frequency component analysis of small regions and executing region division, divided regions of a certain area or more are connected to the upper or lower part near the fingertip. If
Alternatively, if there is a similar region in comparison with the statistically calculated texture feature of the nail 32, that portion is determined to be the nail 32. However, you may detect the nail | claw 32 with the other method based on the conventional image recognition.

FIG. 5 is an image of a finger with a real biometric finger and a copied finger vein pattern attached, taken by the authentication apparatus of FIG.

FIG. 5A is an example of a finger image when the light source 3 is irradiated and the reflected light source 22 is not irradiated.
Here, a clear finger vein pattern 41 is visible. However, it cannot be determined from this image whether the finger vein is genuine or a copy. From this state, the light source 3 is turned off and the reflected light source 22 is irradiated. At this time, when the imaged finger vein pattern is a real living body, the finger vein pattern 40 is almost invisible or unclear as shown in FIG. 5B. However, a counterfeit finger vein pattern is drawn on an object made of a material that transmits the light from the light source 3 or a material that blocks the light from the light source 3 on a sheet that is translucent to the light. When the pattern is drawn with an object that blocks light, the finger vein pattern is clearly confirmed even when only the reflection light source 22 is irradiated as shown in FIG. Therefore, by determining whether a clear finger vein pattern is imaged with only reflected light, it is possible to determine whether or not a forgery pattern is applied.
If a high correlation is found between an image obtained when only the light source 3 is irradiated and an image obtained when only the reflected light source 22 is irradiated, it is determined that a forged pattern is pasted or directly drawn and used. A warning may be issued to the person through the display means 15 or the speaker 17 so that the authentication process is not performed. Thereby, it is possible to prevent erroneous recognition of the forged finger vein pattern.

The reflected light source 22 is not only applied with a forged finger vein pattern but also when the surface of the finger or any of the finger insertion guide unit 5, the optical filter 7, and the imaging device 2 is contaminated or dusty.
A clear pattern may be reflected on the finger image by only irradiation. This can be determined by detecting whether the finger image irradiated with only the reflected light source 22 has a darker portion that is conspicuous than the surroundings. As an example of the detection method, when a value obtained by differentiating the finger region of the image exceeds a certain threshold value, it can be determined that a clear pattern exists at that position. However, a pattern that is clearly displayed may be detected by other conventional image processing methods. If the position of the clearly reflected pattern is the same every time, it can be determined that dirt or dust has accumulated in the apparatus, and if the position fluctuates, it can be determined that dirt has adhered to the finger. The result is displayed on display means 15 or speaker 17.
Notify the user or administrator via Based on this notification, measures such as washing and cleaning can be taken, and a reduction in recognition rate can be prevented.

FIG. 6 is a configuration example of a finger vein authentication device that detects a biological reaction of a finger. A living body detection light source 51 is installed on the fingertip guide unit 4, and a light receiving element 50 is installed on the light source 51. The living body detection light source 51 emits infrared light, passes through the fingertip, and reaches the light receiving element 50. At this time, a light shielding member 52 is provided between the light source 3 and the light receiving element 50 in order to prevent the light emitted from the light source 3 but not the light emitted from the living body detection light source 51 from reaching the light receiving element 50. May be installed. Whether or not the finger is a living finger is detected from the change in the amount of light received by the light receiving element 50.

FIG. 7 is a configuration example of a finger vein authentication device that detects a biological reaction of a finger. The difference from FIG. 6 is that the light source 3 used for imaging the finger vein pattern is used as a living body detection light source, and the light receiving element 50 is installed in the fingertip guide unit 4. Thereby, the number of parts can be reduced.

FIG. 8 is an example of a timing chart of the amount of light received by the authentication switch 8 and the light receiving element 50 in FIG. 6 or FIG. Whether or not the finger is a living finger is determined by checking the pressing timing of the authentication switch 8 and the fluctuation timing of the amount of light received by the light receiving element 50.
When the finger 1 is a living finger, the amount of light received by the light receiving element 50 is not greatly changed and is synchronized with the pulse until the finger 1 is placed on the fingertip guide unit 4 and immediately before the authentication switch 8 is pressed. Variations due to minute pulsations or finger shake are confirmed, and no variation occurs. When the authentication switch 8 is pressed with the fingertip, the authentication switch 8 is turned on, and the amount of current flowing instantaneously increases. At the same time, the fingertip is pressed by the fingertip guide unit 4, the blood volume at the fingertip is reduced, and the transmittance of light emitted from the living body detection light source 51 is increased. Accordingly, the amount of light received by the light receiving element 50 increases in synchronization with the pressing of the authentication switch 8. When the pressure for depressing the authentication switch 8 is weakened, the authentication switch 8 becomes non-conductive, and accordingly, the blood volume at the fingertip increases and the amount of light received by the light receiving element 50 gradually decreases.
When the finger 1 is a fake finger that is not a living body, there is no pulsation before the authentication switch is pressed because the blood flow does not flow in the finger, and the blood volume at the fingertip at the time of pressing does not change. Therefore, there is no change in the amount of light received by the light receiving element 50 as shown in FIG. In this way, by evaluating the amount of change in the amount of light received by the light receiving element 50 when the authentication switch 8 is pressed, it is possible to determine whether the finger is a living finger or a forged finger.

FIG. 9 illustrates the detection of contact with the finger insertion guide unit 5 described above, detection of the state of the ventral side of the finger, detection of a living finger, in the shape of an authentication device that images the finger vein pattern by irradiating transmitted light from the side of the finger. It is an example of 1 structure of the authentication apparatus which has a function. The authentication device 6 is shown as a side sectional view and a view from above. The light source 3 is located on the side of the finger, and the upper part of the authentication device 6 is open. Light incident from the side surface is scattered inside the finger and reaches the imaging device 2 through the finger insertion guide 5 and the optical filter 7.

Even in such an authentication apparatus using side-surface transmitted light, only the position of the light source 3 is different from that of the above-described authentication apparatus. As in the above-described embodiment, detection of contact with the finger insertion guide portion 5, State detection and biometric finger detection can be performed.

FIG. 10 shows an example of a registration screen that displays the user's finger insertion state during registration and guides correct insertion. At the time of registration, most users are not familiar with the authentication device, so it is necessary to guide the finger insertion method. The imaging state of the finger 1 is displayed on the captured image monitor 60 by the display means 15, and an outline guide 61 serving as a guide for matching the outline of the finger 1 is displayed in an overlapping manner. The user can adjust the position of his / her finger while viewing the captured image monitor 60. At this time, finger 1
The finger 1 is rotating about the central axis, the finger 1 is in contact with the finger insertion guide 5, the finger 1 is a forged finger, and a forged finger vein pattern is affixed to the finger 1. The finger state such as “Yes” is expressed as a sentence on the guidance sentence display unit 65. Further, a guidance diagram 64 showing the finger state is displayed inside the guidance display unit 63. Based on these pieces of information, the user can know how to correct the finger state. As described above, when the user correctly performs the registration work according to the guidance on the screen, the work performed by the administrator at the time of registration can be reduced, or the administrator need not be present, and the burden on the administrator can be reduced.

It is possible to improve the reliability of the registered data by presenting the finger several times and selecting data suitable for the registration from among them, rather than acquiring the registered data in a single trial. In addition, it is possible to efficiently increase the proficiency level by showing the user how much the current insertion method is different from the previous insertion method. An example of registration for presenting such information is shown below.
Each time the user presents a finger for registration, pattern matching with the finger vein pattern in the first or past trial is performed, the degree of pattern variation is displayed on the matching result display unit 62, and the finger It is displayed in time series on the variation graph display section 66 for insertion. Further, information such as the state of the finger and the reason for the change may be displayed on the guidance display unit 63. The user can try many times while viewing such history information, and can visually understand the variation of finger insertion according to how the finger 1 is inserted. Also, using this history, do not register data when there are large fluctuations in how you put your fingers,
A finger vein pattern can be registered when a state with small fluctuations continues. This registration may be performed manually by the user, or the system may automatically determine and perform registration. The registration data may be one or more.

Instead of the display unit 15 of FIG. 1 in which the registered image of FIG. 10 is an information transmission unit, information that guides the user how to enter the information correctly may be transmitted by a transmission unit using sound from the speaker 17 or the like. With such a registration method, the user's proficiency level can be improved and highly reliable registration data can be acquired.

As an authentication mode, 1-N authentication for collation for all registered images, and 1-1 authentication for collation only for registered images corresponding to the ID number for identifying the user in advance are input. An embodiment for providing a mode will be described. In the 1-N authentication mode, the authentication is started immediately after the finger is inserted into the apparatus. In the 1-1 authentication, the finger is inserted after the ID number is input using the input unit 16, and the authentication is performed.
The authentication result and information necessary for authentication can be presented to the user using the display means 15 or the speaker 17. As the display means, a display, a liquid crystal, an LED lamp, or the like can be used. The sound emitted from the speaker 17 includes voice and beep sound.
FIG. 11, FIG. 12, and FIG. 13 are examples of flowcharts of an authentication mode for transitioning between 1-N authentication and 1-1 authentication. 1-N authentication is an authentication mode in which authentication is performed simply by presenting a registered finger without inputting information such as an ID number for identifying an individual, and is convenient for the user. Is expensive. However, since the threshold value of the correlation value of the registrant determination in the collation with the registration data is common to all the registrants, the user who tends to have a low correlation with the personal data is likely to be rejected. On the other hand, 1-1 authentication is an authentication mode in which authentication is executed by presenting a finger after inputting information such as an ID number for identifying an individual to be authenticated in advance, and convenience is reduced. However, since a threshold can be set for each individual, the recognition rate can be increased for users who are not familiar with the apparatus. In this embodiment, the registration data is provided with a 1-N authentication valid / invalid attribute, and the 1-N authentication mode operation is automatically validated / invalidated so that 2 in one system.
One authentication mode is switched for each individual, and when it is determined that the correlation with the registered data is lowered, the registered data is automatically updated to improve convenience and realize a high recognition rate.

FIG. 11 is an example of a flowchart of switching between two authentication modes, 1-N authentication and 1-1 authentication. First, initialization is performed in step S100. The initialization includes, for example, initialization of the imaging unit 2, the light source 3, the memory 12, and the like. In the next step S110, it is determined whether the authentication switch 8 has been pressed. When pressed, the 1-N authentication mode of S120 is executed. 1 of S120
Details of the -N mode will be described later with reference to FIG. If not, it is determined in step S130 whether an ID number has been input. If it is determined that the ID number has been input and the authentication switch has been pressed in S140, the 1-1 authentication mode in S150 is executed. S12
Details of the 0 1-N mode will be described later with reference to FIG. When each authentication mode ends, S16
If the authentication mode of the authentication system itself is exited with 0, the process ends. If not, the authentication mode is repeated again. According to this procedure, the highly convenient 1-N authentication and the highly reliable 1-1 authentication are switched, and a system including both convenience and reliability can be realized.

FIG. 12 is an example of a flowchart of a 1-N authentication mode having a function of invalidating 1-N authentication for each user. In step S200, all registered data for which 1-N authentication is valid is collated with the inserted finger, and a pattern correlation is calculated. In S210, it is determined whether the registrant is registered using the correlation value. If it is determined that it is not registered, this mode is terminated. If it is determined that the user is registered, the person authenticated in S220 is identified and an authentication process is performed. Here, the authentication process refers to a process according to an authentication system, such as opening a door, logging in to a PC, or leaving an authentication log. Next, in S230, S
It is determined whether the correlation value calculated in 220 is sufficiently high. If it is sufficiently high, this mode is terminated. If the correlation is not high enough, authentication may be performed but stable authentication may not be possible. Therefore, in S240, the number of unstable authentications corresponding to the authenticator is read, and S
At 250, add 1 to the number of times and save again. Further, in S260, it is determined whether or not the number of unstable authentications exceeds a certain threshold value T1. If not, exit this mode. If exceeded, it is determined that unstable authentication has been repeated continuously or frequently, and S2
At 70, the 1-N authentication attribute of the registration data of the certifier is invalidated. As a result, the authenticator can no longer authenticate in the 1-N authentication mode, and can automatically authenticate only in the 1-1 authentication mode. Thereafter, the number of unstable authentications is initialized to 0 in S280, and this mode is terminated. At this time, the user cannot be authenticated in the 1-N authentication mode due to voice, text, lamp lighting, etc., and only the 1-1 authentication mode in which the ID number is input each time can be supported. You can also let them know. As described above, according to the procedure shown in FIG. 12, a user who cannot perform stable authentication with 1-N authentication can be automatically changed to the 1-1 authentication mode, so that reliability can be improved.

FIG. 13 is an example of a flowchart of a 1-1 authentication mode having a function of enabling 1-N authentication for each user and replacing registered data. In S300, the registered data corresponding to the input ID number is collated with the inserted finger data to calculate a correlation value. Thereafter, S310 to S330 correspond to S210 to S230. If it is determined that the correlation value calculated in S330 is sufficiently high, the number of stable authentications of the authenticator is increased by one in S340. Furthermore, in S350, it is determined whether the number of stable authentications exceeds a certain threshold value T2. If the threshold is not exceeded, this mode is terminated. If the threshold value is exceeded, stable authentication can be performed continuously or frequently, the 1-N authentication attribute of the registration data corresponding to the ID number is enabled in S360, and stable / unstable in S370. Both authentication times are initialized to 0. And this authentication mode is complete | finished.

If it is determined that the correlation calculated in S330 is not sufficiently high, 1 is added to the number of unstable authentications of the authenticator. Further, in S390, it is determined whether this number exceeds a certain threshold value T3. If not exceeded, this authentication mode is terminated. If exceeded, both stable / unstable authentication times are initialized to 0 in S400. In S410, the registration data corresponding to the ID number is
Replace with the currently entered data. Since this data has been correctly confirmed as the data of the certifier, the data of another person cannot be intentionally registered, and the reliability of the data is maintained. In step S420, the authentication system administrator is notified of replacement of registration data.
And this authentication mode is complete | finished.

In S410 of FIG. 13, the currently input data is replaced with registered data. This corresponds to automatically updating the registration data. As a result, 1-N authentication mode and 1
The correlation value is increased in the -1 authentication mode, and the user who cannot perform stable authentication can have the highest correlation in the current finger placement. This effect increases the recognition rate of users who have changed their finger placement during registration and finger placement during operation, as well as the finger vein pattern over time, changes in physical condition, disease, and growth process. It is effective even when it fluctuates,
The automatic update of data makes it possible to keep the recognition rate high without incurring system operating costs. In addition, depending on the user, the blood flow may vary due to a change in temperature due to seasonal variations, and the sharpness of the finger vein pattern image may change. Especially during the low temperature season, the blood flow decreases.
Correlation with the registered pattern registered in the high temperature period decreases. This is considered to fluctuate gradually in units of several months. By replacing the registration data when the correlation decreases, it is possible to use the registration pattern according to the current season and prevent the recognition rate from decreasing. .
Here, in the 1-N authentication mode of FIG. 12, when the number of unstable authentications exceeds the threshold T1, instead of invalidating the 1-N authentication as in S270, the registration data is not passed through the 1-1 authentication. It may be replaced. According to this, it becomes unnecessary to input the ID number in the 1-1 authentication, and the operation without impairing the convenience becomes possible. However, by replacing the registration data through 1-N authentication and 1-1 authentication as in the present embodiment, the registrant can be reliably identified, and the registration data can be updated more safely.

Also, in this embodiment, the registration database used for 1-N authentication and 1-1 authentication is shared, and a 1-N authentication attribute is provided for each registration data to record that 1-N authentication is valid or invalid. However, two databases for 1-N authentication and 1-1 authentication are prepared, and switching between valid / invalid of 1-N authentication is realized by writing / deleting data to / from the database for 1-N authentication. Good. As a result, it is possible to switch between 1-N authentication and 1-1 authentication without providing special attributes for the data to be stored, and the amount of information to be stored can be reduced.

In the authentication mode, a plurality of registration data is held for one finger. In the 1-N authentication mode, one representative registration data is set as a verification target. In the 1-1 authentication mode, all of the registration data is stored. Registration data may be used as a verification target. At this time, in the process S410 of replacing the above-described registration data with input data, one of a plurality of registration data of one finger is replaced.
As a method for selecting the registration data to be replaced, a method of selecting the oldest registration data or selecting data having the lowest sum of correlation values of all other registration data can be used. Thus, by collating with a plurality of registered data, the tolerance of collation with respect to fluctuations in finger presentation can be increased, and the recognition rate can be improved. Further, by providing a plurality of registered data classified for each registered season, it is possible to perform authentication in accordance with the seasonal variation of the finger vein pattern, thereby preventing a reduction in recognition rate.

In S300 of FIG. 13, the ID number corresponds to one registration pattern. However, a specific registration pattern may be regarded as one group and a group ID number may be input. In this case, all the patterns in a plurality of registered patterns specified by the group ID number are collated, but the user is authenticated by inputting a short ID number, and the number of data to be collated can be reduced and high speed is achieved. Processing becomes possible.

Although the invention made by the present inventor has been specifically described based on the embodiments, it is needless to say that the present invention is not limited to the above embodiments and can be variously modified without departing from the gist thereof. . For example, the transition between 1-N authentication and 1-1 authentication shown in FIG. 11-13 is not limited to finger authentication, but for fingerprint authentication and iris authentication in which the matching pattern varies with physical condition, disease, growth process, and season. It can be applied to representative biometric authentication.

As described above, according to the invention made by the present inventor, erroneous insertion and deformation of a finger can be automatically detected. Therefore, the wrong finger vein pattern has been registered.To prevent this, the system administrator does not check whether the fingers have been inserted correctly. It is possible to completely prevent erroneous insertion.

In addition, it is possible to prevent a reduction in the recognition rate of the finger vein authentication device at a low cost from a wrong finger insertion by a user who is not familiar with the device or a forged finger insertion by a malicious user. .

The system configuration example of the authentication apparatus which implements this invention. The structural example of the authentication apparatus which detects the contact of the finger | toe to a finger guide part. The structural example of the authentication apparatus which detects the state of a finger | toe surface. An example of the finger image by reflected light when a finger axis rotation occurs. An example of the finger image by reflected light at the time of sticking a forgery pattern on the finger surface. An example of composition of an authentication device which detects a living body finger. An example of composition of an authentication device which detects a living body finger. An example of the fluctuation | variation of the light reception amount of a light receiving element when an authentication switch is pushed down. The structural example of the authentication apparatus which performs the finger insertion guide part contact which has a light source using the side surface transmitted light of a finger, finger surface state detection, and biometric finger detection. An example of a screen that shows the variation of the inserted finger and guides correct insertion. An example of the flowchart which switches 1-N authentication mode and 1-1 authentication mode. An example of the flowchart of 1-N authentication mode which invalidates 1-N authentication. An example of the flowchart of 1-1 authentication mode which validates 1-N authentication and updates registration data.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Finger, 2 ... Imaging part, 3 ... Light source, 4 ... Finger tip guide part, 5 ... Finger insertion guide part, 6 ... Authentication apparatus, 7 ... Optical filter, 8 ... authentication switch, 9 ... finger placement guide, 11 ... CPU, 12 ... memory, 13 ... interface, 1
DESCRIPTION OF SYMBOLS 4 ... External storage device, 15 ... Display means, 16 ... Input means, 17 ... Speaker, 18 ... Image input means, 20 ... Push switch, 22 ... Reflection light source, 30 ...
Joint wrinkle, 31 ... joint wrinkle end point, 32 ... nail, 40 ... unclear finger vein pattern, 41 ... fake finger vein pattern, 50 ... light receiving element, 51 ... for living body detection Light source, 5
2 ... light shielding member, 60 ... photographed image monitor, 61 ... contour guide, 62 ... collation result display unit, 63 ... guidance diagram display unit, 64 ... finger insertion state Guidance diagram, 65... Guidance text display section, 66... Finger insertion variation graph display section.

Claims (10)

An imaging unit;
A guide unit for setting a finger to be imaged;
A light source that irradiates light that passes through the installed finger and enters the imaging unit;
An image calculation unit that generates a blood vessel pattern for use in personal authentication from the image captured by the imaging unit,
The personal authentication device, wherein the image calculation unit generates the blood vessel pattern by detecting a center of a blood vessel from the captured image. The personal authentication device according to claim 1, wherein the center of the blood vessel is detected by detecting a valley or a depression of a luminance value profile of the captured image. An imaging unit;
A guide unit for setting a finger to be imaged;
A light source that irradiates light that passes through the installed finger and enters the imaging unit;
An image calculation unit that generates a blood vessel pattern from the image captured by the imaging unit and performs personal authentication using the blood vessel pattern;
The personal authentication device, wherein the image calculation unit generates the blood vessel pattern by detecting a valley or a depression of a luminance value profile of the captured image and detecting a center of the blood vessel from the image. The personal authentication apparatus according to claim 2 or 3, wherein the valley or depression of the luminance value is calculated from an opening angle of the luminance value profile curve. The personal authentication device according to claim 2 or 3, wherein the valley or depression of the luminance value is calculated from a curvature of the luminance value profile curve. The personal authentication apparatus according to claim 1, wherein the image calculation unit generates the blood vessel pattern by changing a luminance value of the detected blood vessel center. An imaging unit;
A guide unit for setting a finger to be imaged;
A light source that irradiates light that passes through the installed finger and enters the imaging unit;
An image calculation unit that generates a blood vessel pattern for use in personal authentication from the image captured by the imaging unit,
The image calculation unit detects a contour of the installed finger from the captured image , calculates a contour width at a predetermined position of the contour, and uses an enlargement ratio determined using the contour width A personal authentication device for normalizing the captured image. The personal authentication apparatus according to claim 7, wherein the image calculation unit calculates a width of the contour in a joint portion of the installed finger. The personal authentication device according to claim 7, wherein the image calculation unit calculates a width of the contour at a center of the installed finger. The personal authentication device according to claim 7, wherein the image calculation unit detects the contour by tracking an edge portion of the captured image.

Images