Blood vessel image photographing device and personal authentication system

Abstract

To provide a blood vessel image photographing device capable of preventing degradation of personal authentication accuracy due to local luminance saturation in finger blood vessel images, and a personal authentication system.SOLUTION: The blood vessel image photographing device is configured to acquire multiple finger blood vessel images to divide each finger blood vessel image into multiple division regions, and extract a blood vessel pattern from the finger blood vessel image. For the division regions where blood vessel patterns cannot be extracted due to local luminance saturation in a high-luminance finger blood vessel image, the blood vessel pattern of the corresponding division region of a low-luminance finger blood vessel image is selected, and for the division area with no local luminance saturation, the blood vessel pattern of the division region of the high-luminance finger blood vessel image is selected, thus, blood vessel patterns for personal authentication are generated.SELECTED DRAWING: Figure 4C

Description

Technical Field

[0001] The present invention relates to a vascular imaging device and a personal authentication system.

Background Art

[0002] One of the biometric authentication devices capable of highly accurate personal authentication uses a blood vessel pattern. This device utilizes the fact that blood vessel patterns vary from person to person. For example, it irradiates a finger with near-infrared light and compares the resulting blood vessel pattern that appears with the information of the pre-registered blood vessel pattern to perform personal authentication. Patent Document 1 describes an example of a personal authentication device that irradiates near-infrared light from above a finger held in front of the device, captures an image of the finger blood vessel pattern, and performs personal authentication.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The device described in Patent Document 1 assumes that the finger is held in the same posture as during registration at the time of authentication. If an attempt is made to perform collation with a finger posture different from that at the time of registration, correct personal authentication may not be possible. On the other hand, in the case of a user who is not accustomed to a personal authentication device using a blood vessel pattern, there may be a case where an attempt is made to perform personal authentication with a finger posture different from that at the time of registration unconsciously.

[0005] In a personal authentication device that irradiates near-infrared light from the back of a finger and performs personal authentication based on the finger blood vessel pattern, there is a difference in the near-infrared transmittance of the finger depending on the location, so luminance unevenness between bright and dark portions occurs in the captured finger blood vessel image. When the finger posture changes, the optical path through which the near-infrared light passes through the living body changes. Therefore, when the finger posture changes, the shape, position, and intensity of the luminance unevenness change.

[0006] At this time, depending on the dynamic range of the imaging device, local luminance saturation (white blooming) may occur in the finger vein image. If a blood vessel passes through a portion where luminance saturation has occurred in the finger vein image, part of the blood vessel pattern will not be captured in the image, resulting in a decrease in the authentication accuracy of the personal authentication and an inability to authenticate correctly.

[0007] The fact that the blood vessel pattern cannot be captured in the image due to local luminance saturation can be prevented by reducing the amount of light irradiated from the light source. That is, by reducing the luminance of the finger vein image, local luminance saturation will not occur, and the blood vessels passing through the portion where local luminance saturation has occurred will also be captured in the image. However, reducing the amount of light may cause the image to become dark overall, resulting in unclear blood vessel images in portions where luminance saturation did not occur or blackening. As a result, there is a problem that the accuracy of personal authentication may decrease and correct personal authentication may not be possible.

[0008] An object of the present invention is to provide a blood vessel image capturing device and a personal authentication system that do not reduce the accuracy of personal authentication due to local luminance saturation of a finger vein image.

Means for Solving the Problem

[0009] To solve the above problems, the blood vessel image capturing device of the present invention acquires a plurality of finger vein images, divides each finger vein image into a plurality of divided regions, extracts a blood vessel pattern from the finger vein image, and for a divided region where the blood vessel pattern cannot be extracted due to local luminance saturation of a high-luminance finger vein image, selects the blood vessel pattern of the corresponding divided region of a low-luminance finger vein image, and for a divided region that is not locally luminance-saturated, selects the blood vessel pattern of the above-mentioned divided region of a high-luminance finger vein image to generate a blood vessel pattern for personal authentication.

Effects of the Invention

[0010] According to the present invention, local luminance saturation does not occur in the finger vein image for personal authentication, so that the vein pattern is not missing and a decrease in the accuracy of personal authentication does not occur, and a vein image capturing device and a personal authentication system can be provided.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 4A

Figure 4B

Figure 4C

Figure 5A

Figure 5B

Figure 5C

Figure 5D

Figure 5E

Figure 5F

Figure 6A

Figure 6B

Figure 6C

Figure 7A

Figure 7B

Figure 7C

Figure 8

Figure 9A

Figure 9B

Mode for Carrying Out the Invention

[0012] Hereinafter, the blood vessel imaging device according to the embodiment of the present invention will be described in detail with reference to the drawings. According to the blood vessel imaging device of the embodiment, even when personal authentication fails due to local luminance saturation caused by finger posture variation, correct personal authentication can be performed.

[0013] FIG. 1 is a cross-sectional view of the imaging unit of the blood vessel imaging device 1 of the present embodiment as viewed from the side. The blood vessel imaging device 1 includes light sources 102, 103, 104, an upper cover 101, a support column 108, a fingertip-side finger placement table 109, a finger base-side finger placement table 110, an imaging means 112, and an optical filter 114.

[0014] The light sources 102, 103, and 104 are light sources that irradiate near-infrared light from above the finger to photograph the blood vessels of the finger, and are arranged in plurality along the finger. The upper cover 101 stores the light sources 102, 103, and 104. The support column 108 supports the upper cover 101. The fingertip-side finger rest 109 is a platform for placing the fingertip of the finger 113. The finger-base-side finger rest 110 is a platform for placing the base of the finger 113. The imaging means 112 photographs an image of the blood vessels of the finger. The optical filter 114 prevents the intrusion of dust and dirt and transmits near-infrared light.

[0015] When the finger 113 is placed on the fingertip-side finger rest 109 and the finger-base-side finger rest 110, the blood vessel imaging device 1 irradiates near-infrared light 105, 106, and 107 from the light sources 102, 103, and 104. The near-infrared light is blocked by the finger 113, but a part of it passes through the finger. Since hemoglobin contained in the blood blocks near-infrared light more strongly than other biological parts, the blood vessel pattern of the blood flow is photographed by the imaging means 112 as an image of a blackish line (finger blood vessel image).

[0016] Specifically, the near-infrared light that passes through the finger 113 enters the imaging means 112 through the optical filter 114 and is converted into an electrical signal to form a finger blood vessel image.

[0017] FIG. 2 is a configuration diagram of the personal authentication system 2 of the embodiment. The personal authentication system 2 in FIG. 2 has a configuration including a control unit of the blood vessel imaging device 1 that the authentication device 201 controls the imaging unit in FIG. 1.

[0018] The authentication device 201 is an information processing device composed of a CPU 206, a memory 207, and an interface 204. As an information processing device, the authentication device 201 connects an external storage device 205, a speaker 210, a keyboard 209, and a display device 208. By the CPU 206 executing the program stored in the memory 207, the authentication function of the authentication device 201 and the function of the control unit of the blood vessel imaging device 1 are realized.

[0019] Hereinafter, an outline of the operation of the authentication device 201 will be described. As one of the functions of the authentication device 201, the blood vessel imaging device 1 controls the light sources 102, 103, and 104 by the near-infrared light source control unit 203, captures the finger blood vessel image of the finger 113 by the imaging means 112 of the blood vessel imaging device 1 described in FIG. 1, acquires the finger blood vessel image by the near-infrared light image input unit 202, and extracts the blood vessel pattern. The authentication device 201 performs personal authentication using the blood vessel pattern as authentication information and outputs the authentication result to the speaker 210 and the display device 208.

[0020] Specifically, the blood vessel imaging device 1 controls the light sources 102, 103, and 104 by the near-infrared light source control unit 203 to irradiate weak near-infrared light, and acquires the output information of the imaging means 112 by the near-infrared light image input unit 202. When the output information is in a state where no near-infrared light is incident, the blood vessel imaging device 1 regards it as a state where the near-infrared light is blocked by the finger 113, and identifies that the user has held the finger 113 in front of the blood vessel imaging device 1.

[0021] When the blood vessel imaging device 1 identifies that the finger 113 has been held, it controls the light sources 102, 103, and 104 by the near-infrared light source control unit 203 to irradiate sufficient near-infrared light for capturing the finger blood vessel image. The blood vessel imaging device 1 acquires the finger blood vessel image captured by the imaging means 112 via the near-infrared light image input unit 202, extracts the blood vessel pattern, and records it in the memory 207.

[0022] The authentication device 201 collates the blood vessel pattern recorded in the memory 207 with one or more biometric information (blood vessel patterns) previously recorded in the external storage device 205 to perform personal authentication. The authentication device 201 notifies the user by displaying the authentication result on the display device 208 according to an instruction from the speaker 210 or the keyboard 209.

[0023] In the above description, as shown in FIG. 1, the imaging unit of the blood vessel imaging device 1 is configured separately from the authentication device 201 including the control unit of the blood vessel imaging device 1, but it may be configured as an integrated configuration including the imaging unit of the blood vessel imaging device 1 in the authentication device 201. Alternatively, a personal authentication system 2 in which the authentication device 201 and the blood vessel imaging device 1 are separate may be configured.

[0024] Here, a method for extracting a blood vessel pattern of the blood vessel imaging device 1 according to the embodiment will be described. First, with reference to FIGS. 3A and 3B, an example of a finger blood vessel image captured by the blood vessel imaging device 1 according to the embodiment will be described.

[0025] FIG. 3A is a diagram showing an example of a finger blood vessel image in which a local brightness saturation region has occurred. In the finger blood vessel image 401, a finger blood vessel 404 is shown between finger contours 402 and 403. The finger blood vessel image 401 is of high brightness and includes a local brightness saturation region 405 that is brightness-saturated. In this local brightness saturation region 405, a part of the finger blood vessel is not shown due to brightness saturation. The local dark region 407 is a dark region caused by brightness unevenness, is darker than the peripheral part, and the contrast of the finger blood vessel is lowered, but the finger blood vessel is still shown relatively clearly. Note that the region division line 406 is a division line when dividing the finger blood vessel image into a plurality of regions.

[0026] FIG. 3B is a diagram showing an example of a finger blood vessel image in which the occurrence of a local brightness saturation region is suppressed by suppressing the light amounts of light sources 102, 103, and 104 (see FIG. 1) in the same finger as in FIG. 3A.

[0027] The finger blood vessel image 408 is of low brightness, but a finger blood vessel 411 is shown between finger contours 409 and 410. In the portion of the region 412 of high brightness of the finger corresponding to the local brightness saturation region 405 in FIG. 3A, since brightness saturation does not occur in FIG. 3B, the finger blood vessel is shown. The local dark region 414 due to brightness unevenness is shown darker than the peripheral part, and the finger blood vessel overlapping the dark part is shown unclearly due to the decrease in the light amount. The region division line 413 is a division line when dividing the finger blood vessel image into a plurality of regions.

[0028] As shown in FIGS. 3A and 3B, when a local luminance saturation region occurs in the finger blood vessel image, simply suppressing the light amount of the light source may not be able to suppress the generation of the local luminance saturation region. In addition, a local dark region may occur, making the finger blood vessels unclear. Therefore, in the blood vessel image capturing apparatus 1 of the embodiment, not only the light amount is controlled, but also the blood vessel image is regionally divided to generate a blood vessel pattern. The generation method will be described below.

[0029] FIG. 4A is a diagram showing a blood vessel pattern generated from the blood vessel images of the divided regions by regionally dividing the finger blood vessel image 401 of FIG. 3A. In FIG. 4A, the finger blood vessel image 401 is divided into a plurality of divided regions by a region division line 406, and blood vessel patterns 507, 508, 509, 510, 511, and 512 are generated from the blood vessel images of the divided regions 501, 502, 503, 504, 505, and 506. Note that the blood vessel patterns 507, 508, 509, 510, 511, and 512 are shown as binary images obtained by inverting the blood vessel image, where the blood vessel portions are white and the non-blood vessel portions are black.

[0030] FIG. 4B is a diagram showing a blood vessel pattern generated from the blood vessel images of the divided regions by regionally dividing the low-luminance finger blood vessel image 408 that does not include the local luminance saturation region of FIG. 3B. In FIG. 4B, the finger blood vessel image 401 is regionally divided by a region division line 413, and blood vessel patterns 519, 520, 521, 522, 523, and 524 are generated from the blood vessel images of the divided regions 513, 514, 515, 516, 517, and 518. Note that FIG. 4B (FIG. 3B) corresponds to a blood vessel image in which the local luminance saturation region in FIG. 6B described later is not detected.

[0031] In the blood vessel image capturing apparatus 1, the blood vessel pattern 507 (FIG. 4A) is partially missing due to the portion where the finger blood vessels are not visible due to local luminance saturation, and is not suitable for personal authentication. In addition, in the blood vessel image capturing apparatus 1, the blood vessel patterns 523 and 524 (FIG. 4B) are partially missing or unclear due to the portion where the finger blood vessels are unclear due to the local dark region, and are not suitable for personal authentication.

[0032] Specifically, the blood vessel imaging device 1 of the embodiment determines the presence or absence of finger blood vessels that do not appear in the image due to local luminance saturation regions as follows.

[0033] First, for the divided regions 501, 502, 503, 504, 505, and 506 in FIG. 4A, the blood vessel imaging device 1 scans all the pixels in each region and obtains the number of pixels whose luminance is the upper limit value (saturation luminance). Then, it determines that local luminance saturation has occurred in the divided region 501 where pixels with saturation luminance exist at a certain ratio or more within the region. The other divided regions 502, 503, 504, 505, and 506 are determined not to have local luminance saturation.

[0034] Next, the blood vessel imaging device 1 generates a blood vessel pattern 507 from the divided region 501 where local luminance saturation has occurred (FIG. 4A), and generates a blood vessel pattern 519 from the divided region 513 of the finger blood vessel image 408 corresponding to the divided region 501 (FIG. 4B). Then, it performs template matching while slightly shifting the positions of the blood vessel pattern 507 and the blood vessel pattern 519, counts the number of non - matching pixels at the position where the number of matching pixels is the largest, and calculates the mismatch rate. When the mismatch rate is equal to or higher than a certain threshold determined by statistical processing, that is, when the difference in the blood vessel patterns is large, the blood vessel imaging device 1 determines that there are finger blood vessels that do not appear in the image due to local luminance saturation in the blood vessel pattern 507.

[0035] Similarly, the blood vessel imaging device 1 performs template matching on the blood vessel patterns 511 and 512 (FIG. 4A) and the blood vessel patterns 523 and 524 (FIG. 4B) respectively, and calculates the mismatch rate. Then, it determines that the blood vessel patterns 523 and 524 with a mismatch rate equal to or higher than a certain threshold are local dark regions where finger blood vessels that do not appear in the image exist.

[0036] When the vascular image capturing device 1 collates vascular patterns, it collates the vascular pattern 519 (Fig. 4B) with the vascular patterns 508, 509, 510, 511, 512 (Fig. 4A) with the registered vascular patterns for authentication registered in advance. That is, for the divided region 501 (Fig. 4A) where the vascular pattern cannot be extracted due to local luminance saturation of the high-luminance finger vascular image 401, it is collated with the vascular pattern 519 of the divided region 513 of the low-luminance finger vascular image 408 (Fig. 4B). And for the remaining divided regions that are not locally luminance-saturated, it is collated with the vascular patterns 508, 509, 510, 511, 512 of the high-luminance finger vascular image 401.

[0037] That is, for the vascular image capturing device 1, the vascular pattern used for collation preferably has no or few vascular defects or unclear parts. For the region where vascular defects occur due to local luminance saturation, the vascular pattern generated from the divided region of the low-luminance finger vascular image is used, and for the other regions, the vascular pattern generated from the divided region of the high-luminance finger vascular image is used.

[0038] Note that the vascular image capturing device 1 may not collate the vascular patterns of each divided region obtained by dividing the region, but may collate the finger vascular pattern as authentication information after synthesizing it into one vascular pattern in advance with the pre-registered finger vascular image for personal authentication. Specifically, as shown in Fig. 4C, the vascular pattern 519 and the vascular patterns 508, 509, 510, 511, 512 are synthesized into one vascular pattern.

[0039] Next, with reference to Figs. 5A to 5F, a method for setting the region dividing line of the finger vascular image will be described. One is a method of setting such that the center line in the finger length axis direction of the finger vascular image overlaps the center line of the region dividing line. The other is a method of setting the region dividing line of the finger vascular image to be collated at the position where the overlap of the vascular patterns is the largest.

[0040] FIG. 5A and FIG. 5D are diagrams showing pre-registered finger vein images for personal authentication. FIGS. 5B and 5E are diagrams showing finger vein images to be matched during personal authentication. FIG. 5C is a diagram showing a state where FIG. 5A and FIG. 5B are overlapped, and FIG. 5F is a diagram showing a state where FIG. 5D and FIG. 5E are overlapped.

[0041] In FIGS. 5A and 5B, the region dividing lines 601 and 604 are set so that the center lines 603 and 606 in the respective finger longitudinal axis directions of the corresponding fingers 602 and 605 overlap with the center lines of the region dividing lines in the finger longitudinal axis direction. In this case, since the finger is rotated (rolled) about the finger longitudinal axis direction due to posture variation and the positions of the finger veins shown in the captured image are shifted, as shown in FIG. 5C, the overlapping portions of the finger veins within each region divided by the region dividing lines become smaller.

[0042] On the other hand, in the case of FIGS. 5D and 5E, in FIG. 5D, similar to FIG. 5A, the center line 609 in the finger longitudinal axis direction of the finger 608 and the center line of the region dividing line 607 in the finger longitudinal axis direction are set to overlap. However, in FIG. 5E, the region dividing line 611 of FIG. 5E is set at the position where the overlap between the finger vein 610 of the finger 608 in FIG. 5D and the finger vein 613 of the finger 612 in FIG. 5E is the largest.

[0043] Thereby, the region dividing line 607 of FIG. 5D and the region dividing line 611 of FIG. 5E are set to overlap, and it is possible to prevent the overlapping portions of the finger veins within each region after region division from becoming smaller, and it is possible to suppress a decrease in matching.

[0044] When the blood vessel image photographing apparatus 1 of the embodiment divides the finger vein image and the blood vessel pattern into regions, the finger vein image and the blood vessel pattern are divided by the finger vein image and the blood vessel pattern pre-registered for personal authentication and the region dividing line set so that the overlap of the blood vessels becomes large.

[0045] Next, the operation of the blood vessel image photographing apparatus 1 (authentication apparatus 201) of the embodiment during personal authentication will be described with reference to the flowcharts of FIGS. 6A to 6C. The flowcharts of FIGS. 6A to 6C show the operation in a state where a finger is presented to the blood vessel image photographing apparatus 1.

[0046] In step S1, the blood vessel image capturing device 1 controls the light sources 102, 103, and 104 by the near-infrared light source control unit 203 to irradiate the finger with near-infrared light from above the finger. In step S2, the blood vessel image capturing device 1 captures a finger blood vessel image by the imaging means 112 and acquires the finger blood vessel image by the near-infrared light image input unit 202.

[0047] In step S3, the blood vessel image capturing device 1 detects the finger from the finger blood vessel image acquired in step S2. In step S4, the blood vessel image capturing device 1 determines whether or not a finger has been detected in step S3. If a finger has been detected (Yes in S4), the process proceeds to step S5. If no finger has been detected (No in S4), the process returns to step S1.

[0048] In step S5, the blood vessel image capturing device 1 calculates the average luminance of the detected finger based on the finger blood vessel image acquired in step S2. The average luminance of the finger is obtained, for example, by calculating the average luminance of the portion where the finger is always captured in the image taken when the finger is detected.

[0049] In step S6, the blood vessel image capturing device 1 checks whether or not the average luminance of the finger calculated in step S5 is equal to or higher than the target luminance lower limit and equal to or lower than the target luminance upper limit, that is, whether or not it is between the upper and lower limits of the target luminance. The target luminance is set, for example, as the intermediate value of the luminance range that the imaging device can represent.

[0050] Then, when the average luminance of the finger is between the upper and lower limits of the target value (Yes in S6), the blood vessel image capturing device 1 proceeds to step S8 in FIG. 6B. When the average luminance of the finger is not between the upper and lower limits of the target value (No in S6), the process proceeds to step S7.

[0051] In step S7, when the average luminance of the finger is below the target lower limit, the vascular image capturing device 1 performs light source control so as to increase the light quantity values of the light sources 102, 103, and 104 by the near-infrared light source control unit 203. When the average luminance of the finger exceeds the target upper limit, light source control is performed so as to decrease the light quantity value. Then, it returns to step S1.

[0052] In step S8 of FIG. 6B, the vascular image capturing device 1 extracts the finger contour of the acquired finger vascular image. In step S9, the vascular image capturing device 1 generates a blood vessel pattern from the finger vascular image.

[0053] In step S10, the vascular image capturing device 1 divides the finger vascular image into a plurality of regions by a predetermined region dividing line. At this time, the setting of the region dividing line of the finger vascular image is performed by the method of setting the region dividing line of the finger vascular image to be collated at the position where the overlap of the blood vessel patterns is the largest, which is described in FIGS. 5A to 5F.

[0054] Among the finger vascular images in step S10, the first finger vascular image subjected to the division process is the finger vascular image 401 in FIG. 4A. And the finger vascular image in step S10 after downwardly correcting the target luminance described later becomes the finger vascular image 408 in FIG. 4B.

[0055] In step S11, the vascular image capturing device 1 checks whether local luminance saturation has occurred in each region of the finger vascular image divided in step S10. Whether local luminance saturation has occurred is determined, for example, by scanning all the pixels in the region, and when the pixels whose luminance is the upper limit value (saturation luminance) exist at a certain ratio or more in the region, it is determined that local luminance saturation has occurred.

[0056] Then, the vascular image capturing device 1 determines whether local luminance saturation has been detected in any of the divided regions. If local luminance saturation has been detected (Yes in S11), it proceeds to step S12. If local luminance saturation has not been detected (No in S11), it proceeds to step S18 in FIG. 6C.

[0057] In step S12, the blood vessel image capturing device 1 performs downward correction of the target luminance. In step S13, the blood vessel image capturing device 1 checks whether the average luminance of the finger is equal to or higher than the lower limit of the target luminance downward-corrected in step S12 and equal to or lower than the upper limit of the target luminance, that is, whether it is between the upper and lower limits of the target luminance downward-corrected in step S12.

[0058] Then, when the average luminance of the finger is between the upper and lower limits of the downward-corrected target luminance (Yes in S13), the blood vessel image capturing device 1 returns to step S8. When the average luminance of the finger is not between the upper and lower limits of the downward-corrected target luminance (No in S13), it proceeds to step S14.

[0059] In step S14, when the average luminance of the finger is lower than the target lower limit, the blood vessel image capturing device 1 performs light source control so that the light quantity values of the light sources 102, 103, and 104 are increased by the near-infrared light source control unit 203. When the average luminance of the finger is higher than the target upper limit, light source control is performed so that the light quantity value is decreased. Then, it proceeds to step S15.

[0060] In step S15, the blood vessel image capturing device 1 controls the light sources 102, 103, and 104 by the near-infrared light source control unit 203 to irradiate the finger with near-infrared light from above the finger. In step S16, the blood vessel image capturing device 1 captures an image of the finger blood vessel of the finger by the imaging means 112 and acquires the finger blood vessel image by the near-infrared light image input unit 202.

[0061] In step S17, the blood vessel image capturing device 1 calculates the detected average luminance of the finger based on the finger blood vessel image acquired in step S16 and returns to step 13. That is, the light source control is repeated until the average luminance of the finger falls within the range of the target luminance downward-corrected in step S12.

[0062] In step S18 of FIG. 6C, the blood vessel image capturing device 1 divides the blood vessel pattern generated in step S9 (FIG. 6B) into each region by the region dividing line. That is, the divided blood vessel patterns 507 to 512 in FIG. 4A and the divided blood vessel patterns 519 to 524 in FIG. 4B are obtained.

[0063] In step S19, for each of the divided regions, the blood vessel image capturing device 1 selects which of the blood vessel pattern generated based on the finger blood vessel image in which local luminance saturation has occurred and the blood vessel pattern generated based on the finger blood vessel image in which the target luminance is corrected downward to suppress local luminance saturation is to be used as the blood vessel pattern for personal authentication. That is, it is selected whether to use the blood vessel pattern generated from the finger blood vessel image acquired in step S2 (FIG. 6A) or the blood vessel pattern generated from the finger blood vessel image acquired in step S16.

[0064] Specifically, as described with reference to FIGS. 4A and 4B, for the divided region 501 in which the blood vessel pattern cannot be extracted due to the local luminance saturation of the high-luminance finger blood vessel image 401 acquired in step S2, the blood vessel pattern 519 of the corresponding divided region 513 of the low-luminance finger blood vessel image 408 acquired in step S16 is selected. For the remaining divided regions that are not locally luminance-saturated, the blood vessel patterns 508, 509, 510, 511, and 512 of the high-luminance finger blood vessel image 401 are selected.

[0065] In step S20, the blood vessel image capturing device 1 collates the blood vessel pattern for personal authentication recorded in the external storage device 205 in advance and the blood vessel pattern selected in step S19 for each divided region, and calculates the score of the mismatch rate between the blood vessel patterns as the collation score. The score is calculated, for example, by performing template matching to check whether the pattern of a certain image exists in another image, checking the number of matching pixels while shifting the position of the image pattern to be checked little by little, and counting the number of non-matching pixels at the place where the most matching pixels are found.

[0066] In step S21, the blood vessel image capturing device 1 fuses the matching scores of each region to calculate the matching score of the entire finger. The fusion can be performed by simply adding them together, or by adding them together with weights varying according to the position.

[0067] In step S22, the blood vessel image capturing device 1 checks whether the matching score of the entire finger calculated in step S21 is below the threshold. That is, if the mismatch rate of the matching result is smaller than the threshold (Yes in S22), the blood vessel image capturing device 1 proceeds to step S23 as a successful authentication and, if the mismatch rate of the matching result is not smaller than the threshold (No in S22), the blood vessel image capturing device 1 proceeds to step S24 as a failed authentication.

[0068] In step S23, the blood vessel image capturing device 1 performs post - processing after successful authentication and ends the authentication process. In step S24, the blood vessel image capturing device 1 performs post - processing after failed authentication and ends the authentication process.

[0069] Note that in the process of generating the blood vessel pattern in step S9, a model can be generated by performing learning of the blood vessel pattern in advance, generating it as a bit string of a specific bit length by machine learning, and in the process of matching the blood vessel patterns in step S20, the Hamming distance between the bit strings may be obtained.

[0070] Also, instead of comparing the average luminance of the finger in step S13 and calculating the average luminance of the finger in step S17, the average luminance of the region where local luminance saturation has occurred may be referred to and calculated. Further, in this case, in step S12, instead of downward - correcting the target luminance, the target luminance of the region where local luminance saturation has occurred may be set to be the same as the target luminance of the finger.

[0071] Next, with reference to the flowcharts of FIGS. 7A to 7C, the operation at the time of registering the blood vessel pattern for personal authentication of the blood vessel image capturing device 1 (authentication device 201) of the embodiment will be described. The flowcharts of FIGS. 7A to 7C show the operation in a state where a finger is presented to the blood vessel image capturing device 1.

[0072] FIG. 7A is a flowchart of a process for setting the light quantity values of light sources 102, 103, and 104 so that the average luminance of a finger becomes within a predetermined target luminance range in the operation of the blood vessel imaging device 1 (authentication device 201) for registering a blood vessel pattern for personal authentication. Steps S1 to S7 in FIG. 7A are the same as those in FIG. 6A, and thus the description thereof is omitted.

[0073] FIG. 7B is a flowchart of a process for adjusting the light quantity values of light sources 102, 103, and 104 by downwardly correcting the target luminance so that local luminance saturation does not occur in a finger blood vessel image in the operation of the blood vessel imaging device 1 (authentication device 201) for registering a blood vessel pattern for personal authentication. The flowchart of FIG. 7B adds a process of step S25 for storing the light quantity value at the time of imaging before the processes of steps S8 to S17 in FIG. 6B. Steps S8 to S17 are the same as those in FIG. 6B, and thus the description thereof is omitted.

[0074] FIG. 7C is a flowchart of a process for generating a blood vessel pattern for personal authentication from a plurality of finger blood vessel images in the operation of the blood vessel imaging device 1 (authentication device 201) for registering a blood vessel pattern for personal authentication.

[0075] Step S26 in FIG. 7C proceeds from step S11 in FIG. 7B when local luminance saturation does not occur in any of the regions to be divided (No in S11). In step S26, the blood vessel imaging device 1 stores the light quantity value of the light source when the finger blood vessel image is captured.

[0076] In step S27, the blood vessel imaging device 1 divides the blood vessel pattern into each region by a region dividing line.

[0077] In step S28, the blood vessel imaging device 1 determines whether the number of captured finger blood vessel images exceeds a threshold value. If it does not exceed the threshold value (No in S28), the process proceeds to step S30, and if it exceeds the threshold value (Yes in S28), the process proceeds to step S29.

[0078] In step S30, the blood vessel image capturing device 1 displays an instruction to change the finger posture to the registered user on the display device 208 (Fig. 2), and returns to step S1 (Fig. 7A). In step S29, the blood vessel image capturing device 1 displays an instruction to change the finger posture of the registered user to the standard posture on the display device 208 (Fig. 2).

[0079] In step S31, the blood vessel image capturing device 1 sets the light source according to the plurality of light quantity values stored in steps S25 and S26, captures the finger, and acquires a plurality of finger blood vessel images.

[0080] In step S32, the blood vessel image capturing device 1 generates a blood vessel pattern from the plurality of captured finger blood vessel images and divides it into a plurality of regions. For the divided regions where the blood vessel pattern cannot be extracted due to local luminance saturation of the high-luminance finger blood vessel image, the blood vessel pattern of the divided region of the low-luminance finger blood vessel image is selected. For the remaining divided regions that are not locally luminance saturated, the blood vessel pattern of the high-luminance finger blood vessel image is selected. The blood vessel image capturing device 1 integrates the blood vessel patterns of the selected divided regions and generates them as a blood vessel pattern for personal authentication.

[0081] In step S33, the blood vessel image capturing device 1 stores the blood vessel pattern generated in step S32 in the external storage device 205 as a registered blood vessel pattern for personal authentication, and ends the registration process.

[0082] Through the above registration process of the blood vessel pattern for personal authentication, the blood vessel image capturing device 1 (authentication device 201) of the embodiment captures the finger blood vessel image by changing the finger posture, obtains the light quantity value that does not cause local luminance saturation in the finger blood vessel image, and registers the blood vessel pattern for personal authentication, so that the influence on the authentication rate due to the finger posture deviation at the time of authentication can be reduced.

[0083] Next, with reference to Fig. 8, other operations during the registration of the blood vessel pattern for personal authentication of the blood vessel image capturing device 1 (authentication device 201) of the embodiment will be described. The flowchart of Fig. 8 shows the operations in a state where a finger is presented to the blood vessel image capturing device 1

[0084] Steps S1 to S6 in FIG. 8 are the same as those in FIG. 6A, and thus the description thereof is omitted. In step S34, in step 6, when the average luminance of the finger is between the upper limit and the lower limit of the target value (Yes in S6), the process proceeds. In step S34, the blood vessel image capturing device 1 stores the light quantity value of the light source when capturing the finger blood vessel image.

[0085] In step S35, the blood vessel image capturing device 1 generates a blood vessel pattern from the finger blood vessel image, and divides the generated blood vessel pattern into each region by the region dividing line.

[0086] In step S36, the blood vessel image capturing device 1 acquires finger blood vessel images while changing the light quantity value from a preset low initial light quantity to the light quantity value stored in step S34.

[0087] In step S37, the blood vessel image capturing device 1 generates a blood vessel pattern from the plurality of captured finger blood vessel images and divides them into a plurality of regions. For the divided regions where the blood vessel pattern cannot be extracted due to local luminance saturation of the high-luminance finger blood vessel image, the blood vessel pattern of the divided region of the low-luminance finger blood vessel image is selected. For the remaining divided regions that are not locally luminance saturated, the blood vessel pattern of the high-luminance finger blood vessel image is selected. The blood vessel image capturing device 1 integrates the blood vessel patterns of the selected divided regions and generates them as a blood vessel pattern for personal authentication.

[0088] In step S38, the blood vessel image capturing device 1 stores the blood vessel pattern generated in step S37 in the external storage device 205 as a registered blood vessel pattern for personal authentication, and ends the registration process.

[0089] In step S36, instead of taking a plurality of finger vein images while changing the light amount value from a preset low initial light amount to the light amount value stored in step S34, from the image taken in step S2 and the light amount value stored in step S34, a light amount value at which local luminance saturation does not occur when the finger changes its posture is obtained in advance by machine learning. One or a plurality of finger vein images may be taken according to the light amount value at which local luminance saturation does not occur obtained from the learned model.

[0090] Next, with reference to FIGS. 9A and 9B, in the registration process of the blood vessel pattern for personal authentication described with FIGS. 7A to 7C, the treatment when the finger vein image in which local luminance saturation does not occur has a lower luminance than expected will be described.

[0091] FIG. 9A is an example of an image in which local luminance saturation does not occur and the captured image is darker than expected. Since the luminance of finger 901 is low, finger vein 902 is unclear. Also, when feature extraction is performed from an image with a lower luminance than the assumed luminance to generate a blood vessel pattern, a difference occurs in the local light and shade change compared to the finger vein image at the time of registration, and the mismatch rate at the time of collation becomes high, and there may be a case where correct personal authentication cannot be performed. For this reason, it is desirable that the average luminance of the finger is within a range where local luminance saturation does not occur and the finger vein can still be seen, and the average luminance of the finger is close to the average luminance of the finger in the finger vein image at the time of registration.

[0092] FIG. 9B is an example of an image in which the blood vessel image capturing device 1 of the embodiment adjusts the luminance of the finger vein image in FIG. 9A to a luminance close to the target luminance within a range where local luminance saturation does not occur and the finger vein can still be seen by image processing for level correction of luminance.

[0093] The image processing for luminance level correction divides the luminance range of the dark part, middle part, and bright part of pixels by two thresholds, namely the dark part threshold and the bright part threshold. For pixels with luminance (dark part) below the dark part threshold, the luminance value is set to the lower limit value; for pixels with luminance (bright part) above the bright part threshold, the luminance value is set to the upper limit value; and for pixels with luminance in the middle part, the luminance value is set to the intermediate value between the upper limit value and the lower limit value. Furthermore, each pixel in the middle part is adjusted so that the luminance changes smoothly. That is, for the blood vessel image, the luminance values of the dark part, bright part, and middle part are corrected to adjust the luminance of the finger blood vessel image.

[0094] As a result, the finger blood vessel 904 becomes clear, and by making the luminance of the finger 903 closer to the desired luminance, the difference in local shading changes can be reduced.

[0095] In addition, the present invention is not limited to the above-described embodiments and includes various modifications. The above embodiments have been described in detail for easy understanding of the present invention and are not necessarily limited to those having all the configurations described. Also, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can also be added to the configuration of one embodiment.

Explanation of Reference Numerals

[0096] 1 Blood vessel image capturing device 101 Upper cover 102, 103, 104 Light sources 108 Support column 109 Finger tip side finger placement table 110 Finger root side finger placement table 112 Imaging means 113 Finger 114 Optical filter 2 Personal authentication system 201 Authentication device 202 Near-infrared light image input unit 203 Near-infrared light source control unit 205 External storage device 206 CPU 207 Memory 208 Display device 209 Keyboard 210 Speaker

Claims

1. Obtain a plurality of finger blood vessel images and divide each finger blood vessel image into a plurality of divided regions, Extract a blood vessel pattern from the finger blood vessel image, For a divided region where the blood vessel pattern cannot be extracted due to local luminance saturation of the high-luminance finger blood vessel image, select the blood vessel pattern of the corresponding divided region of the low-luminance finger blood vessel image, For a divided region that is not locally luminance-saturated, select the blood vessel pattern of the divided region of the high-luminance finger blood vessel image, and Generate a blood vessel pattern for personal authentication, A blood vessel image capturing device.

2. In the blood vessel image capturing device according to Claim 1, For the finger blood vessel image, perform level correction on the luminance values of the dark part, bright part, and middle part to adjust the luminance of the finger blood vessel image, A blood vessel image capturing device.

3. Having the blood vessel image capturing device according to Claim 1, Perform personal authentication by comparing the blood vessel pattern generated from the high-luminance finger blood vessel image and the low-luminance finger blood vessel image with the registered blood vessel pattern, A personal authentication system.

4. In the personal authentication system according to Claim 3, The blood vessel image capturing device obtains the light quantity value of a light source where local luminance saturation does not occur in the finger blood vessel image in a finger posture different from the standard posture, Based on the finger blood vessel image of the finger in the standard posture captured by the light source with the light quantity value, the blood vessel image capturing device generates the registered blood vessel pattern, A personal authentication system.

Images