JP2007115072A - Finger vein authentication system and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a biometrics authentication device, capable of keeping high authentication accuracy even in a device having the same level of performance as that of an existing authentication device. <P>SOLUTION: This system comprises a photographic device 25 for fluoroscopically taking a three-dimensional image of a finger vein pattern; a storage device 40 for storing three-dimensional image data 41 output by the photographic device 25 and a photographic characteristic parameter 41 of a specific finger vein authentication device; and an authentication image data generation means 32 for analyzing and arithmetically processing the three-dimensional image data 41, based on the photographic characteristic parameter 42, and generating a plurality of authentication two-dimensional image data 43 of the specific finger vein authentication device, which have different photographic conditions. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a finger vein authentication system and a finger vein authentication device used for personal authentication using finger veins.

The importance of personal authentication processing is increasing for financial institutions' ATM (automatic transaction equipment) and entrance / exit management. The biometric authentication process is an automatic personal authentication processing method using individual differences, has sufficiently high security as compared with existing authentication methods using a recitation number or the like, and has begun to spread rapidly. In particular, since authentication processing using a palm or finger vein pattern has higher security than authentication processing using a face or fingerprint, research and development has been actively promoted and is being put to practical use (Patent Document 1). (Patent Document 2).
JP 2005-168627 A JP 2005-071118 A

Here, the conventional technique has the following problems to be solved.
The known authentication device captures a vein image of the person's own index finger, for example. For example, the vein image data is converted into binary bitmap data and registered as authentication image data. The actual authentication uses the same device as the device that captured the authentication image data. This is because the closer the shooting condition is, the higher the recognition accuracy. When the vein image data requested for authentication is read, it is compared with the registered authentication image data. A corresponding portion of the image is cut out, normalization processing for correcting distortion is performed, and a difference value is calculated in units of pixels. If the difference value is less than or equal to the threshold value, the user is authenticated.

  The vein image data that actually requires authentication is photographed by the authentication device from the user, for example, every time the user enters or leaves the room. However, when the finger is put on the photographing apparatus, the photographing conditions slightly change from when the authentication image data is photographed. When photographing is performed under various conditions such as when a finger is not placed straight on the photographing apparatus, when blood circulation is good, or when a hand is cold, the difference from the authentication image data may exceed a threshold value. That is, even if it is the person himself, a situation occurs in which the gate does not open. In order to prevent such a situation, for example, authentication image data of four fingers is registered, and the gate is opened when the difference between the authentication image data and any of the fingers is equal to or less than a threshold value. Further, every time the user passes through the gate repeatedly, a difference value between the person's vein image data and the authentication image data is acquired, and when the difference value is always large, the threshold value is relaxed. This eliminates the problem of rejecting the person's passage, but increases the risk of passing another person.

  For a single finger, the design must reduce the accuracy to an error rate of 1 in 1000 even if the error rate is 1 in 100,000. Similarly, if the accuracy of the four fingers is lowered, the error rate is apparently reduced to 1/250. In this case, the error rate becomes too high compared to the number of residents in one apartment, and satisfactory security cannot be obtained. The rate of recognizing a person who is not the person is significantly increased. In other words, secure and reliable operation cannot be expected unless the error rate is further reduced by the authentication of each finger. Using a high-resolution device to shoot over time, acquiring a larger amount of data and performing high-speed computation can prevent errors, but the cost is too high and the authentication process takes time. It takes too much and is not practical.

  In addition, because it depends largely on the shooting conditions of the authentication device, even if the authentication image data was created for all residents of the apartment when the authentication device was newly installed, when the shooting device was repaired or the entire authentication device was replaced, All the authentication image data must be recreated. In addition, the authentication image data itself is personal information and requires strict management. As long as authentication image data is stored in the authentication apparatus main body, the data is also at risk of theft.

  The present invention has been made paying attention to the above points, and it is an object of the present invention to provide a biometric authentication apparatus that can maintain high authentication accuracy even with an apparatus having the same level of performance as an existing authentication apparatus.

In each embodiment of the present invention, the above-described problems are solved by the following configurations.
<Configuration 1>
An imaging device that performs fluoroscopic imaging of a 3D image of a finger vein pattern, a storage device that stores 3D image data output from the imaging device, an imaging characteristic parameter of a specific finger vein authentication device, and the 3D image An authentication image for analyzing and calculating data based on the imaging characteristic parameters and generating a plurality of authentication two-dimensional image data for the specific finger vein authentication device, each having different imaging conditions A finger vein authentication system comprising data generation means.

  Since very accurate three-dimensional image data of the subject is acquired in advance, and based on this, two-dimensional image data for authentication is generated and used. Therefore, the two-dimensional image for authentication using the finger vein authentication device itself is used. Data generation processing is unnecessary. Therefore, if a finger vein authentication device imaging characteristic parameter is prepared for an arbitrary finger vein authentication device, the two-dimensional image data for authentication of the subject can be automatically registered. In addition, since a plurality of two-dimensional image data for authentication with different shooting conditions are generated and registered for authentication, variations in shooting conditions can be absorbed without lowering the authentication accuracy.

<Configuration 2>
In the finger vein authentication system according to Configuration 1, the storage device stores data conversion arithmetic expressions unique to the specific finger vein authentication apparatus, and the plurality of authentication image data are converted by the data conversion arithmetic expressions, respectively. Then, the finger vein authentication system is output by an output device.

  If a data conversion arithmetic expression specific to a specific finger vein authentication device is prepared, the image data for authentication cannot be used for another finger vein authentication device. Even if the same model is used, a difference more than a mere design error occurs. Therefore, even if there is information leakage of authentication image data, it does not affect other finger vein authentication devices. It can also protect personal information. Accordingly, the unique data conversion formula is unique, and it is preferable to forcibly convert the contents of the authentication image data so that it can be distinguished from those of other finger vein authentication devices.

<Configuration 3>
In the finger vein authentication system according to Configuration 1, the plurality of two-dimensional image data for authentication are differently rotated about the reference line in a state where the finger is directed in parallel to the reference line facing the length direction. A finger vein authentication system comprising image data photographed from an angle and image data photographed at different inclination angles in a state where the finger is inclined with respect to the reference line.

  Even if a finger is photographed from various directions by the authentication device, each of the images is compared with accurate two-dimensional image data for authentication, so that authentication processing can be performed with high accuracy.

<Configuration 4>
In the finger vein authentication system according to Configuration 1, the imaging device captures a plurality of three-dimensional images of the finger vein pattern that have been captured under different physiological conditions, and the storage device includes a plurality of A finger vein authentication system, wherein a three-dimensional image of a vein pattern is stored, and the plurality of two-dimensional image data for authentication is generated for each of the three-dimensional images of the plurality of vein patterns.

  For example, the vein patterns when blood circulation is good or when the hands are cold may be significantly different. Taking this into consideration, three-dimensional images of vein patterns under different physiological conditions are taken, and two-dimensional image data for authentication is generated for each. As a result, authentication errors can be reduced without reducing accuracy.

<Configuration 5>
In the finger vein authentication system according to Configuration 3, the photographing device photographs a three-dimensional image of a three-dimensional finger model, and the storage device photographs the three-dimensional finger model with the specific finger vein authentication device. A plurality of two-dimensional image data having different shooting conditions are stored, and the authentication image data generating means generates the two-dimensional image data by calculation from the three-dimensional image data of the finger model. A finger vein authentication system, wherein an arithmetic expression is an arithmetic expression for generating the plurality of two-dimensional image data for authentication.

  Actually, the reference 3D image data is acquired from the 3D image of the 3D finger model, and the model 2D image data is acquired by photographing the 3D finger model with a specific finger vein authentication device. . By comparing these, an arithmetic expression for generating two-dimensional image data for authentication for authenticating the model two-dimensional image data by arithmetic processing can be acquired. This arithmetic expression is included in the photographing characteristic parameter.

<Configuration 6>
In the finger vein authentication system according to Configuration 1, the storage device updates a plurality of two-dimensional image data obtained by photographing the three-dimensional finger model with the specific finger vein authentication device already stored according to a correction request. When the two-dimensional image data of the three-dimensional finger model is updated, the authentication image data generation means updates the arithmetic expression for generating the plurality of authentication two-dimensional image data. Finger vein authentication system.

  When the finger vein authentication device is distorted due to long-term use or the like, the model two-dimensional image data can be updated to update the arithmetic expression for generating the authentication two-dimensional image data. Thereby, even if the finger vein authentication device goes wrong, the two-dimensional image data for authentication can be automatically optimized.

<Configuration 7>
A storage device for storing a plurality of two-dimensional image data for authentication obtained by analyzing and calculating three-dimensional image data of a finger vein pattern based on its own imaging characteristic parameter, and different authentication conditions; The vein image data is obtained by individually comparing the imaging device that obtains vein image data by fluoroscopic imaging of the finger vein and the vein image data output from the imaging device and the plurality of two-dimensional image data for authentication. A finger vein authentication apparatus comprising: an authentication determination unit that performs authentication determination when approximating any one of a plurality of two-dimensional image data for authentication;

  The authentication device uses a plurality of two-dimensional image data for authentication for authentication of one finger. Even if there are variations in the shooting conditions, since there are a plurality of two-dimensional image data for authentication, it can be compared and determined individually with high accuracy. Since it is not necessary to lower the threshold value of each authentication two-dimensional image data, the accuracy is not lowered.

<Configuration 8>
In the finger vein authentication device according to Configuration 7, the storage device stores a plurality of authentication two-dimensional image data, and an authentication data group that is converted in advance by a data conversion arithmetic expression unique to the finger vein authentication device. Then, the authentication determining means compares the image data output from the imaging apparatus with the unique data conversion arithmetic expression and the authentication data group.

  Although the authentication two-dimensional image data corresponds to personal information, the authentication data group has been subjected to unique data conversion, and is, for example, data that is meaningless to a third party even if transmitted through a network or the like. Security is improved.

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

FIG. 1 is a block diagram of a finger vein authentication system according to the first embodiment.
This system is controlled by the personal computer 20, for example. The computer 20 includes a main body control unit 21, a display 22, a keyboard 23, and a mouse 24. An imaging device 25 is connected to the computer 20. As shown in the upper right circle, the photographing device 25 performs a fluoroscopic photographing of a three-dimensional image of the vein pattern of the finger 10. Specifically, the finger 10 is irradiated with the lamp 11, and the X-axis image line sensor 12 and the Y-axis image line sensor 13 are scanned in the direction of the arrow 14 to take an image. In order to photograph the blood flowing through the veins, it is only necessary to irradiate visible light having a wavelength suitable for the photographing with the lamp 11 and photograph the shadow.

  Various other methods for acquiring a three-dimensional image of a finger vein as described above can be considered. For example, optical image acquisition using a lens system or a reflecting mirror is also possible. Instead of visible light, ultraviolet light or X-rays can also be used. It is also possible to take a picture using reflected light or diffracted light, which is employed in a general nondestructive inspection apparatus. In addition, various devices used for photographing internal tissues of the human body in relation to medical equipment can be employed. For example, a micro cone beam type fluoroscopic imaging apparatus used by a dentist or the like can be employed. For example, it is possible to take a picture using ultrasonic waves with a finger placed in water. The purpose of the present invention is to acquire a three-dimensional image of a finger vein with as high accuracy as possible, so that authentication data can be acquired with any optimal device different from the finger vein authentication device used for authentication. There is an effect that can be done. In the present invention, imaging by any of the above methods will be referred to as “perspective imaging” and will be described.

  The computer 20 is provided with an arithmetic processing device 30 and a storage device 40 in order to process the three-dimensional image data output from the photographing device 25 in this way. The arithmetic processing unit 30 is provided with a scan control unit 31 and an authentication image data generation unit 32. These are all computer programs that give a predetermined function to the computer 20. The storage device 40 stores 3D image data 41 output from the imaging device 25. Further, a unique shooting characteristic parameter 42 of an authentication device (not shown) for performing authentication processing is stored here.

  The imaging characteristic parameter is data indicating imaging conditions unique to the imaging device of the finger vein authentication device. Even when exactly the same vein pattern of the finger is photographed, the two-dimensional image data for authentication to be acquired depends on the structure, position, photographing direction, illumination position, illumination light density distribution, and optical characteristics of the image sensor. There is a unique personality in the content. Accordingly, an arithmetic expression for obtaining unique two-dimensional image data for authentication by high-precision three-dimensional image data is different for each finger vein authentication device. The coefficients used in this calculation formula are called photographing characteristic parameters. Therefore, the imaging characteristic parameters are acquired for each finger vein authentication device by using a finger model described later.

  That is, the computer 20 collects and stores the photographing characteristic parameters 42 from a large number of authentication devices to be managed. As a result, authentication two-dimensional image data 43 for each authentication device can be generated from the same person's three-dimensional image data and distributed. The two-dimensional image data 43 for authentication is data obtained by analyzing the three-dimensional image data 41 based on the imaging characteristic parameter 42 of each finger vein authentication device, and calculating processing. Note that a plurality of authentication two-dimensional image data 43 is generated for each finger vein authentication device. These two-dimensional image data for authentication are assumed to have different shooting conditions. This photographing condition will be described later.

  The storage device 40 further stores a data conversion arithmetic expression 44. This is an arithmetic expression specific to each finger vein authentication device. The calculation formula setting method can be freely set for each finger vein authentication device as in the so-called encryption processing. The authentication two-dimensional image data 43 is converted by the data conversion arithmetic expression 44 and then output as authentication data 45 for each finger vein authentication device. This output is transferred by the output device 34 to the corresponding finger vein authentication device through the input / output interface 33. For example, it can be said that the service person uses the output device 34 to record the corresponding authentication data 45 on a CD ROM or the like and bring it to the user. The output device 34 may be externally attached to a USB interface provided in the main body control unit 21 of the computer 20.

FIG. 2 is an explanatory diagram of the contents of the three-dimensional image data.
As shown in the figure, when the vein 16 of the finger 10 is photographed, as shown on the right side, an image of a black dot 52 exists at a position where the vein exists as shown in a cross-sectional view 51 in which the finger is cut in a circle. In the present invention, this data is acquired as accurately as possible. Since this data is important personal information necessary for authenticating the person, the data is stored in a state of being strictly managed by the computer 20 of the service company. In order to acquire such 3D image data, for example, a person who is subject to entrance / exit management goes to a specific place and cooperates in acquiring accurate 3D image data.

  Conventionally, for example, when managing the entrance and exit of an apartment, authentication image data is registered using the finger vein authentication device of the apartment. However, in this case, for example, when the finger vein authentication device breaks down and a new finger vein authentication device is attached, it is necessary to retrieve the authentication data again. On the other hand, in the present invention, extremely accurate three-dimensional image data is acquired in advance, and two-dimensional image data for authentication that can be used for any finger vein authentication device is obtained each time using this three-dimensional image data. Therefore, it is not necessary to repeat the same shooting and registration process frequently. A finger vein authentication apparatus generally uses image data obtained by photographing a finger from one direction for authentication. That is, conversion to two-dimensional image data is necessary for the finger vein authentication device. For example, there are the following two types of conversion methods.

FIGS. 3A and 3B are explanatory diagrams of a method for generating two-dimensional image data for finger vein authentication devices having different methods.
(A) shows a two-dimensional image data generation method for a finger vein authentication device that acquires vein image data using the image line sensor 61. In this case, the image line sensor 61 is arranged with respect to the finger 10 in parallel with the longitudinal axis of the finger 10. The image line sensor 61 is rotated in the direction of the arrow 63 to acquire image data. At this time, as viewed from the rotation axis of the image line sensor 61, the image of the finger 10 is input radially as shown in the figure. Accordingly, when pixel values on the radiation as shown in the figure are cumulatively added, two-dimensional image data for one radiation is obtained. For the same sectional view 51, for example, N radiations 64 are set according to the resolution, and N accumulated pixel value groups are obtained. Also in the finger length direction, M number of cross-sectional views 51 corresponding to the resolution are set. Accordingly, in total, two-dimensional image data with N × M cumulative pixel value groups is obtained.

  (B) of the figure shows a two-dimensional image data generation method for a finger vein authentication device that acquires vein image data using a CCD image sensor. In this case, the pixel values on the parallel lines 65 are cumulatively added to obtain two-dimensional image data for one parallel line. For the same sectional view 51, for example, N parallel lines 65 are set according to the resolution, and N accumulated pixel value groups are obtained. Also in the finger length direction, M number of cross-sectional views 51 corresponding to the resolution are set. Accordingly, in total, two-dimensional image data with N × M cumulative pixel value groups is obtained. In addition, since the technique for acquiring 2D image data from such 3D image data should just use the existing arbitrary techniques, the further description is abbreviate | omitted.

FIG. 4 is an explanatory diagram of a three-dimensional finger model.
The finger vein exists substantially parallel to the longitudinal direction of the finger. Accordingly, as shown in the figure, as a very simple model, a plurality of opaque lines 73 (using metal lines or the like) arranged in the finger length direction and substrates 71 and 72 for fixing both ends thereof are provided. Use it. If the line 73 is assumed to be a vein and imaged by the imaging device 25 shown in FIG. 1, three-dimensional image data using a three-dimensional finger model can be acquired. Further, if this three-dimensional finger model is photographed with an arbitrary finger vein authentication device, two-dimensional image data V can be acquired. On the other hand, the two-dimensional image data W can be theoretically generated from the three-dimensional image data by paying attention to the photographing conditions of the finger vein authentication device, for example, the photographing method described in FIG. Actually, there is a difference between the two-dimensional image data V and W due to differences in various photographing conditions. Therefore, by comparing the two, for example, a correction coefficient for each pixel value is obtained in units of pixels. Thus, an arithmetic expression for generating highly accurate two-dimensional image data from the three-dimensional image data can be obtained. Since this arithmetic expression is different for each finger vein authentication device, it is stored in the storage device separately for each finger vein authentication device.

  For example, even if the same finger vein authentication device is used as it is, the recognition power may change due to the aging of the certification or the reading device. At this time, the three-dimensional finger model is photographed again, and the arithmetic expression is updated. Therefore, a plurality of two-dimensional image data necessary for authentication is newly generated by the arithmetic expression and used for the finger vein authentication device. ,Output. This completes the adjustment on the finger vein authentication device side. It is possible to manage the authentication device in a stable and highly accurate manner without changing the authentication accuracy or adjusting the photographing conditions of the finger vein authentication device with complicated operations. In addition, said finger model is an example. Depending on the characteristics of the finger vein authentication device, it is preferable to devise and optimize a more complicated cubic lattice structure or a plastic mold.

FIG. 5 is an explanatory diagram of a plurality of authentication two-dimensional image data acquisition methods.
A finger authentication device, that is, a photographing device of a finger vein authentication device, cannot always photograph a finger accurately from a certain direction. For example, as shown in (a) of this figure, there are cases where the finger 10 is photographed while being inclined by an angle α with respect to a reference line 50 oriented in the finger length direction. Considering this, two-dimensional image data for authentication in a state inclined by an angle α with respect to the reference line 50 is acquired. Also, as shown in (b) of the figure, when the finger 10 is photographed, the finger 10 may be photographed while being rotated by an angle β with respect to the reference line 50. Considering this, two-dimensional image data for authentication in a state rotated by an angle β with respect to the reference line 50 is acquired. In this way, when the shooting conditions are different, authentication image data to be shot under different shooting conditions is acquired from the three-dimensional image data 41 by arithmetic processing. Note that the rotation angle and the inclination angle are specific to each finger vein device. Conditions vary depending on the device, such as a finger that is bent and easily photographed, and a finger that is difficult to rotate. Therefore, this is stored as an imaging characteristic parameter. The angles α and β each have a unique value, including differences in the structure of the imaging device and differences in the illumination device. Thus, it is preferable to set the shooting characteristic parameter.

FIG. 6 is an explanatory diagram showing a specific example of the finger vein authentication device.
As shown in the figure, the finger vein authentication device includes an imaging device 80, a storage device 81, and an arithmetic processing device. The arithmetic processing unit shows only a main part, and includes an authentication unit 83 and an output device 84. The photographing apparatus 80 is provided with a U-shaped groove 85 for placing the finger 10. Here, the finger 10 is placed and photographed, and vein image data of the finger 10 is acquired. The storage device 81 stores authentication data 45 and vein image data 82. The vein image data 82 photographed by the photographing device 80 is temporarily stored in the storage device 81 and then compared with the authentication data 45 by the authentication means 83. As described with reference to FIG. 5, the authentication data 45 includes a plurality of two-dimensional image data generated under different conditions of the inclination angle α and the rotation angle β. The vein image data 82 is compared with all the two-dimensional image data. Then, the accumulated value of the difference between the pixel values is compared with the threshold value. If the threshold value is made sufficiently small, only the accumulated difference value with the two-dimensional image data generated under substantially the same conditions becomes smaller than the threshold value. The output device outputs this determination result and sends out a signal for releasing the lock of the auto-lock device, for example.

  The conventional finger vein authentication apparatus prepares one piece of authentication image data for one finger. The purpose of the device is to prevent other people from being authenticated as the principal, but it would be inconvenient for the user if the wrong person is mistakenly authenticated as the other person. Even if the shooting conditions vary, an error is allowed so that all the users are authenticated. That is, in terms of performance, the false authentication rate can be reduced to 1 / 100,000, but the threshold is relaxed and the authentication accuracy is set to 1 / 50,000. Make the same setting for each of the four fingers. As a result, the authentication accuracy is reduced to 1/12500. This is a theoretical number, and on a practical machine, the accuracy decreases as one person per thousand people, and the other person is authenticated.

  On the other hand, if a plurality of authentication image data are prepared in advance and can be compared with vein image data captured under different imaging conditions, the individual authentication accuracy does not have to be reduced to 1 / 100,000. That is, the threshold value may be small. Even if the number of authentication image data increases, the authentication accuracy does not decrease. In addition, in the past, four fingers were registered in order to reduce the rejection rate of the principal, but this is not necessary in the present invention. Thus, for example, registration of two fingers is acceptable. If one is used, there is a problem if the finger is injured, so two is the limit. Nevertheless, authentication can be performed with much higher accuracy than before. Of course, this can also reliably prevent an accident that authenticates another person.

FIG. 7 is an explanatory diagram of the finger vein authentication system according to the second embodiment.
In the finger vein authentication system of the present invention, there is a problem of how accurately a three-dimensional image is acquired. The problem here is the physiological condition of the subject. The state of the blood vessel is different between when the finger is warm and when the finger is cold. The photographed image may differ greatly when the finger is very cold in the cold season, or when the finger is warm or well-circulated after exercise. The vein pattern taken is different from the tolerance range. Therefore, for example, two types of three-dimensional image data are acquired under different physiological conditions. The apparatus shown in FIG. 7 is a specific example of such an image capturing apparatus.

First, the subject's hand 92 is cooled. For this, as shown in (a), a cooling spray 91 used for cooling the bruise in sports is used. Thus, the capillaries of the finger 10 are contracted. In this state, the hand 92 is put into the photographing apparatus 94 shown in FIG. For example, an infrared heater 95 is attached to the photographing device 94. Initially, the cooled finger 10 is photographed, and then the entire hand 92 is warmed by the infrared heater 95, and the finger 10 with improved blood circulation is photographed again. Thereby, two or more types of three-dimensional image data of different physiological states are acquired.

  Conventionally, in consideration of such differences, processing for increasing the threshold value of authentication image data has been performed. If this embodiment is used, a plurality of two-dimensional image data is generated from two or more types of three-dimensional image data in different physiological states and used as authentication data. High accuracy can be maintained.

  Further, for example, when a lamp provided in the photographing device of the finger vein authentication device provided at the entrance of the apartment breaks down, if the lamp is replaced, the photographing condition may be slightly different from before. In such a case, it is extremely difficult to re-photograph the fingers of all the residents of the apartment each time. When the system of the above-described embodiment is used, since the high-accuracy 3D image data of the person is stored in the computer 20, the imaging characteristic parameter 42 is newly acquired, and the authentication 2 suitable for the finger vein authentication device at any time. Dimensional image data can be generated. Therefore, the correction request may be made whenever necessary. The correction at this time may be performed by using a three-dimensional finger model and acquiring the imaging characteristic parameter of the corresponding finger vein authentication device in the manner already described. In this way, the arithmetic expression is updated. For example, when there are two entrances in an apartment, even if separate finger vein authentication devices are installed, the resident's finger only needs to be photographed once.

FIG. 8 is a flowchart showing the operation of the finger vein authentication system.
First, in step S <b> 11, fluoroscopic imaging of the subject's finger using the imaging device 25 is performed. In step S12 and step S13, signal correction processing and normalization processing are executed, and in step S14, the three-dimensional image data 41 is stored in the storage device 40 of the computer 20. Next, the imaging characteristic parameter 42 of each finger vein authentication device stored in the storage device 40 is read. Then, the two-dimensional image data 43 for authentication is generated using the photographing characteristic parameter 42 (step S16). In step S17, it is determined whether there is another shooting characteristic parameter. If there are other parameters, the process returns to step S15 to read the parameters again and generate two-dimensional image data for authentication. Even in the same finger vein authentication device, complicated parameters are used when various conditions differ depending on the angle of photographing.

  The shooting angle is also a kind of parameter. The shooting temperature is also a kind of parameter. In response to these, a plurality of two-dimensional image data for authentication is generated (step S16). When all necessary two-dimensional image data for authentication is generated, the process proceeds from step S17 to step S18. In step S18, the generated two-dimensional image data 43 for authentication is stored in the storage device 40. Next, in step S19, the data conversion arithmetic expression 44 of the corresponding finger vein authentication device is read from the storage device 40. Using this arithmetic expression, the authentication two-dimensional image data 43 is converted (step S20). Thus, the authentication data 45 is generated and stored in the storage device 40 (step S21). The authentication data 45 is taken out through the output device 34 and used.

FIG. 9 is a flowchart illustrating an operation example of the finger vein authentication device.
First, in step S31, a subject's authentication target finger is fluoroscopically photographed at an apartment gate or the like. In step S32, the vein image data is acquired and stored in the storage device of the finger vein authentication device. In step S33, the authentication image data 45 received from the computer 20 is read. In step S34, the difference value is calculated. When the difference value is calculated, a predetermined distortion correction is performed on the read vein image data, normalization for adjusting the aspect ratio and the like is performed, and a difference value calculation with each pixel of the authentication image data is performed. The comparison pixel is changed, or the difference value calculation for correcting the minute position and inclination of the image is repeated to obtain the minimum value of the difference values (step S35). In step S36, a comparison is made with a threshold value prepared in advance. This threshold value may be uniform or different for each authentication image data.

  In step S37, the comparison result is stored as a determination result. If the difference value is less than or equal to the threshold value, the person is determined. Otherwise, it is determined as someone else. In step S38, it is determined whether there is another authentication image data. If there is another authentication image data, the process returns to step S33, the image data is read, and the difference value calculation and the determination process after step S34 are executed. Thus, the determination process is repeated for all the authentication image data. As a result, when the difference value is obtained from the authentication image data having the closest photographing condition and the minimum value is obtained, the value may be equal to or less than the threshold value. At this time, the person can be authenticated.

  In step S39, the determination result group is searched in this way, and a determination is made that the determination result is the person himself / herself. If there is even one that is determined to be the person, the process proceeds to step S40, and the authentication result is determined to be the person. If all are determined not to be the person, the conclusion is that the person is not the person. When such processing is performed, determination is performed with high accuracy in steps S36 and S37 while the threshold value is kept high. Therefore, it is possible to prevent erroneous recognition of another person's vein image data as the original person. Conversely, since a large number of authentication image data are repeatedly extracted and determined, the probability that the shooting condition matches any of the authentication image data increases. This can reduce the rate of misrecognizing the person as another person. That is, as the number of authentication image data increases, the rate of misrecognizing the person as another person can be reduced. In either case, the difference value calculation is not performed and the threshold value is not lowered, but the authentication image data used for comparison determination is replaced to derive a condition that can always clear the threshold value, so that the person can be recognized with high accuracy. . In particular, for example, in the case of an anemic woman or a child whose finger is thin and the vein pattern is simple, erroneous recognition is likely to occur. It has been demonstrated that this can be maintained with high accuracy by the means described above. In addition, since only the threshold value is kept small, a biometric authentication device capable of maintaining high authentication accuracy is provided even with a device having the same level of performance as an existing authentication device.

1 is a block diagram of a finger vein authentication system according to a first embodiment. It is content explanatory drawing of three-dimensional image data. (A) And (b) is a two-dimensional image data generation method explanatory drawing for finger vein authentication apparatuses from which a system differs. It is explanatory drawing of a three-dimensional finger model. It is explanatory drawing of the several two-dimensional image data acquisition method for authentication. It is explanatory drawing which shows the specific example of a finger vein authentication apparatus. It is explanatory drawing of the finger vein authentication system of Example 2. It is a flowchart which shows operation | movement of a finger vein authentication system. It is a flowchart which shows the operation example of a finger vein authentication apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 20 Personal computer 21 Main body control part 22 Display 23 Keyboard 24 Mouse 25 Image pick-up device 10 Finger 11 Lamp 12 X-axis image line sensor 13 Y-axis image line sensor 14 Arrow 30 Arithmetic processing unit 40 Storage device 31 Scan control means 32 Image data for authentication Generation means 41 Three-dimensional image data 42 Shooting characteristic parameter 43 Authentication two-dimensional image data 44 Data conversion arithmetic expression 45 Authentication data 33 Input / output interface 34 Output device

Claims (8)

An imaging device for fluoroscopically capturing a three-dimensional image of a finger vein pattern;
A storage device for storing three-dimensional image data output by the imaging device and imaging characteristic parameters of a specific finger vein authentication device;
Analyzing and calculating the three-dimensional image data based on the photographing characteristic parameters, and generating a plurality of authentication two-dimensional image data for the specific finger vein authentication device, each having different photographing conditions A finger vein authentication system comprising: authentication image data generation means for performing authentication. The finger vein authentication system according to claim 1,
The storage device stores a data conversion arithmetic expression specific to the specific finger vein authentication device,
The finger vein authentication system, wherein each of the plurality of authentication image data is converted by the data conversion arithmetic expression and then output by an output device. The finger vein authentication system according to claim 1,
The plurality of two-dimensional image data for authentication is
In a state where the finger is directed in parallel to the reference line facing the length direction, image data taken from different rotation angles around the reference line, and the finger is inclined with respect to the reference line And a finger vein authentication system including image data photographed at different inclination angles. The finger vein authentication system according to claim 1,
The imaging device is a three-dimensional image of the finger vein pattern, and a plurality of images taken under different physiological conditions,
The storage device stores a three-dimensional image of a plurality of vein patterns,
The finger vein authentication system, wherein the plurality of two-dimensional image data for authentication is generated for each of the three-dimensional images of the plurality of vein patterns. The finger vein authentication system according to claim 3,
The photographing device photographs a three-dimensional image of a three-dimensional finger model,
The storage device stores a plurality of two-dimensional image data obtained by photographing the three-dimensional finger model with the specific finger vein authentication device, each having different photographing conditions,
The authentication image data generation means includes an arithmetic expression for generating the two-dimensional image data by calculation from the three-dimensional image data of the finger model, an arithmetic expression for generating the plurality of authentication two-dimensional image data, and A finger vein authentication system characterized by: The finger vein authentication system according to claim 1,
The storage device updates, in accordance with a correction request, a plurality of two-dimensional image data obtained by photographing the three-dimensional finger model that has already been stored with the specific finger vein authentication device,
The authentication image data generation means updates the arithmetic expression for generating the plurality of authentication two-dimensional image data when the two-dimensional image data of the three-dimensional finger model is updated. Finger vein authentication system. A storage device for storing a plurality of two-dimensional image data for authentication obtained by analyzing and calculating the three-dimensional image data of the finger vein pattern based on the own photographing characteristic parameter and having different photographing conditions;
An imaging device that obtains vein image data by fluoroscopic imaging of a finger vein to be authenticated;
The vein image data output from the imaging device is individually compared with the plurality of authentication two-dimensional image data, and if the vein image data approximates any one of the plurality of authentication two-dimensional image data, an authentication determination is made. And a finger vein authentication device, comprising: an authentication determination unit that performs a denial determination when none of them is similar. The finger vein authentication device according to claim 7,
The storage device is a plurality of authentication two-dimensional image data, and stores an authentication data group that is converted in advance by a data conversion arithmetic expression specific to the finger vein authentication device,
The finger vein authentication device, wherein the authentication determination unit compares the photographed image data output from the photographing device with the unique data conversion arithmetic expression and the authentication data group.

Images