JP5229489B2 - Biometric authentication device - Google Patents

Description

  The present invention relates to a biometric authentication apparatus that performs personal authentication using a vein pattern of a biological part such as a finger.

  There is known a biometric authentication device that is thinned by using a lens array (see, for example, Patent Documents 1, 2, and 3).

  On the other hand, in a biometric authentication device mounted on a small device such as a mobile device, it is important that the device is thin and small. However, the devices disclosed in Patent Documents 1, 2, and 3 are difficult to reduce in size even though they can be reduced in thickness.

  That is, in the apparatus of Patent Document 1, since the lenses constituting the lens array and the pixels of the image sensor have a one-to-one correspondence, an image sensor having almost the same size as the finger region to be observed is required. In order to ensure the accuracy of authentication using sparsely existing vein patterns, it is necessary to observe a fairly large area of the finger, so an image sensor with a large imaging area is necessarily required. It will increase in size. In addition, since an image pickup element having a large image pickup area is expensive, there is a problem that the apparatus cost increases.

  Even in the devices of Patent Documents 2 and 3, since an image is taken with the finger in contact with or close to the lens array, an expensive image sensor with a large imaging area is required to image a wide range of the finger in order to ensure the accuracy of vein authentication. Therefore, an increase in the size and cost of the apparatus is inevitable.

  The present invention has been made in view of the above-described problems of the prior art, and enables highly accurate vein authentication using an inexpensive imaging device having a small imaging area. An object of the present invention is to provide a biometric authentication device that can be easily reduced in cost.

The biometric authentication device according to the invention of claim 1
A lens array in which a plurality of lenses are arrayed, illumination means for irradiating a biological part held in the field of view of the lens array with linearly polarized light in a specific polarization direction, and an object side of a part of lenses of the lens array Alternatively, a first analyzer that is provided on the image plane side and passes only a polarized component in a direction parallel to the specific polarization direction, and a lens in the field of view of the lens array that is substantially imaged by the lens of the lens array. An imaging device provided on the image plane side of the lens array for capturing a compound eye image that is a set of reduced images of the living body part (hereinafter referred to as a single eye image), and within the field of view of the lens array An image input device for inputting the compound eye image in a non-contact manner with a living body part held up;
In a compound eye image input by the image input device, for each pixel value of a single eye image by a lens not provided with the first analyzer of the lens array, the first analyzer of the lens array is A correction calculation means for performing a correction calculation for subtracting the corresponding pixel value of the single-eye image by the provided lens;
And authentication means for performing personal authentication using a single eye image after the correction calculation by the correction calculation means.

  According to a second aspect of the present invention, in the biometric authentication device according to the first aspect of the invention, the image input device is the specific polarization direction provided on the object side or the image plane side of a part of the lenses of the lens array. The first analyzer of the lens array in the compound eye image input by the image input device in the correction calculation means. Subtracting the corresponding pixel value of the single eye image by the lens provided with the first analyzer of the lens array from each pixel value of the single eye image by the lens without the second analyzer; and The correction operation of adding the corresponding pixel values of the single-eye image by the lens provided with the second analyzer of the lens array is performed.

  According to a third aspect of the present invention, in the biometric authentication device according to the first or second aspect of the invention, prior to the correction calculation by the correction calculation means, the parallax between the single-eye images with respect to the single-eye image related to the correction calculation. Parallax adjustment calculation means for performing parallax adjustment calculation for adjusting the image position so as to cancel the image position, and the single-eye image parallax used for the parallax adjustment calculation by the parallax adjustment calculation means to the compound eye image input by the image input device It further has parallax detection calculation means for calculation based on the above.

The biometric authentication device according to the invention of claim 4
A lens array in which a plurality of lenses are arrayed, illumination means for irradiating a biological part held in the field of view of the lens array with linearly polarized light in a specific polarization direction, and an object side of a part of lenses of the lens array Alternatively, a first analyzer that is provided on the image plane side and passes only a polarized component in a direction parallel to the specific polarization direction, and a lens in the field of view of the lens array that is substantially imaged by the lens of the lens array. An imaging device provided on the image plane side of the lens array for capturing a compound eye image that is a set of reduced images of the living body part (hereinafter referred to as a single eye image), and within the field of view of the lens array An image input device for inputting the compound eye image in a non-contact manner with a living body part held up;
A first single image using parallax from a plurality of single-eye images by a plurality of lenses not provided with the first analyzer of the lens array in a compound eye image input by the image input device. And reconstructing a second single image using the parallax from a plurality of single-eye images by a plurality of lenses provided with the first analyzer of the lens array. Computing means;
Correction calculation means for performing a correction calculation for subtracting the corresponding pixel value of the second single image for each pixel value of the first single image;
And authentication means for performing personal authentication using a single image after the correction calculation by the correction calculation means.

  According to a fifth aspect of the present invention, in the biometric authentication device according to the fourth aspect of the invention, the image input device is the specific polarization direction provided on the object side or the image plane side of a part of the lenses of the lens array. The second analyzer of the lens array in the compound eye image input by the image input device in the reconstruction calculation means includes a second analyzer that passes only a polarization component in a direction non-parallel to the lens. A third single image is reconstructed from a plurality of single-eye images by a plurality of provided lenses using the parallax thereof, and the correction calculation means performs each pixel value of the first single image with respect to each pixel value. , Performing a correction operation of subtracting the corresponding pixel value of the second single image and adding the corresponding pixel value of the third single image.

  According to a sixth aspect of the invention, in the biometric authentication device according to the fourth or fifth aspect of the invention, prior to the correction calculation by the correction calculation means, the parallax between the single images with respect to the single image related to the correction calculation. Used for parallax adjustment calculation means for performing parallax adjustment calculation for adjusting the image position so as to cancel out the image, parallax between single images used for parallax adjustment calculation by the parallax adjustment calculation means, and reconstruction calculation by the reconstruction calculation means It further has parallax detection calculation means for calculating parallax between single images based on a compound eye image input by the image input device.

  According to a seventh aspect of the present invention, in the biometric authentication device according to any one of the first to sixth aspects, the illuminating means is a linearly polarized light that is irradiated to a biological part held in the field of view of the lens array. It includes an illumination lens for making the amount of light uniform.

  According to an eighth aspect of the present invention, in the biometric authentication device according to the seventh aspect of the present invention, the illumination lens is formed on the lens array.

  In the biometric authentication device according to the present invention, the image input device uses a reduction optical system including a lens array, and picks up a compound eye image that is a set of reduced images of a biological part held in the field of view of the lens array. It is possible to input an image of a sufficiently wide range of a biological part using a small and inexpensive image pickup device, and this makes it possible to reduce the thickness and size of an image input device and a biometric authentication device including the image input device.

  In the image input device, linearly polarized light in a specific polarization direction is irradiated to the living body part as illumination light by the illuminating means, but this illumination light is reflected near the surface of the living body part or enters the inside of the living body part and is reflected / scattered. Therefore, the light returning from the living body part to the lens array includes reflected light near the surface of the living body part and light reflected / scattered inside the living body part. The reflected light from the vicinity of the surface of the living body part accounts for a large proportion of the polarization component in the direction parallel to the polarization direction of the illumination light, whereas the light reflected and scattered inside the living body part is the polarization direction of the illumination light. Polarized components in different directions occupy a large proportion. Therefore, in the compound eye image input by the image input device, the monocular image by the lens provided with the first analyzer is an image by reflected light near the surface of the living body part, and the first analyzer is provided. A single-eye image with no lens is an image in which an image of a vein reflected by light reflected and scattered inside a living body part is superimposed on an image of reflected light on the surface. When the body part is held away from the lens array, the reflected light near the surface of the body part increases compared to the case where the body part is in close contact with the lens array. The in-plane intensity distribution of the illumination light also changes, but the change in the in-plane intensity distribution is mainly due to the reflected light near the surface. Strong and easy to receive. According to the present invention, highly accurate personal authentication is possible by eliminating the influence of reflected light near the surface of a living body part.

  That is, according to the first aspect of the present invention, the component of the image reflected by the surface reflected light is removed from the monocular image by the lens not provided with the first analyzer by the correction calculation, and the light reflected and scattered inside the living body. A clear vein image can be obtained, and by using such a vein image, high-accuracy personal authentication is possible. Similarly, in the invention of claim 4, the component of the image by the surface reflected light is obtained from a single image of high resolution reconstructed from a plurality of single-eye images by a plurality of lenses not provided with the first analyzer. Since it is possible to obtain a clear high-resolution vein image by the light reflected and scattered inside the living body, it is possible to perform personal authentication with higher accuracy.

  In the image input device, the second analyzer provided on the object side or the image plane side of a part of the lenses of the lens array transmits a polarization component in a direction non-parallel to the polarization direction of the illumination light (linearly polarized light). However, the polarized component passing therethrough is a component in which the polarization direction is changed by the illumination light reflected by a vein or the like in the living body part. Therefore, according to the second aspect of the present invention, the component of the single eye image obtained by the lens provided with the first analyzer is subtracted from the single eye image obtained by the lens not provided with the analyzer by the correction calculation to obtain the second analyzer. By adding the components of a single-eye image by a lens provided with photons, the influence of reflected light near the surface of the living body part is removed, and a clear vein image with the vein edge emphasized is acquired for personal authentication. Can be used. Similarly, in the invention of claim 5, the first analyzer is provided from a high-resolution single image reconstructed from a plurality of single-eye images by a plurality of lenses not provided with an analyzer in the correction calculation. A single image reconstructed from a plurality of single-eye images by a plurality of lenses provided with a second analyzer by subtracting components of a single image reconstructed from a plurality of single-eye images by a plurality of lenses By adding these components, it is possible to remove the influence of the reflected light near the surface of the living body part, obtain a clear high-resolution vein image with emphasized vein edges, and use it for personal authentication.

  When the distance from the lens array of the living body member changes, the parallax between the single-eye images in the compound eye image changes, and the parallax between the single images reconstructed from the compound eye images also changes. On the other hand, according to the invention of claim 3 or 6, the single-eye image parallax or the single-image parallax used for the parallax adjustment calculation, and further the single-eye image parallax used for the single image reconstruction calculation However, since it is calculated based on the inputted compound eye image, it is possible to ensure high authentication accuracy even when the distance of the biological member from the lens array changes.

  If the illumination light with respect to the living body part has a characteristic in-plane intensity distribution, the intensity distribution of the input vein image changes depending on the position of the living body part held in the field of view of the lens array, which may reduce the authentication accuracy. . In the invention of claim 7, since the amount of illumination light is made uniform by the illumination lens, it is possible to suppress a change in the intensity distribution of the vein image due to the position of the living body part. In the invention of claim 8, since the illumination lens is formed in the lens array, it is possible to reduce the cost of the apparatus related to the illumination lens as compared with the case where the illumination lens independent from the lens array is provided. It is also advantageous for downsizing.

1 is a configuration explanatory diagram of a biometric authentication device according to a first embodiment of the present invention. FIG. It is a plane structure explanatory view of the image input device concerning a 1st embodiment of the present invention. It is a flowchart for operation | movement description of the biometrics apparatus which concerns on the 1st Embodiment of this invention. It is composition explanatory drawing of the biometrics apparatus which concerns on the 2nd Embodiment of this invention. It is a plane structure explanatory drawing of the image input device which concerns on the 2nd Embodiment of this invention. It is a flowchart for operation | movement description of the biometrics apparatus which concerns on the 2nd Embodiment of this invention.

[First Embodiment]
FIG. 1 is a configuration explanatory diagram of a biometric authentication apparatus according to the first embodiment of the present invention. This biometric authentication device uses a finger of a hand as a biological part and performs personal authentication by using vein pattern information inside the finger, and is in a non-contact state with a finger 1 at a distant position as shown in the figure. The image input device 2 for inputting the image of the finger 1 is provided. FIG. 1 schematically shows a cross-sectional structure of the image input apparatus 2. FIG. 2 schematically shows a planar structure of the image input device 2 as viewed from the side facing the finger. The longitudinal direction of the finger corresponds to the left and right direction in FIGS.

  The image input device 2 includes a lens array 3 in which a plurality of imaging lenses are arrayed in a plane substantially orthogonal to the lens optical axis. In the example shown here, as shown in FIG. 2, 6 lenses 4, 5, 6, 7, 8, and 9 are arranged in a 2 × 3 array.

  The image input device 2 includes an image pickup device 10 on the image plane side of the lens array 3 for picking up images that are substantially formed by the lenses 4 to 9. The image pickup device 10 has a large number of pixels two-dimensionally arranged on the image pickup surface 11, and for example, a general CMOS image sensor or CCD image sensor can be used. Since the image input device 2 captures a reduced image of the finger 1 away from the lens array 3, the image input device 2 uses a small and inexpensive image sensor 10 on the imaging surface 11, and the finger 1 is in close contact with the lens array 3. The image of a sufficiently wide area of the finger 1 can be input as compared with the form of imaging with. This makes it possible to reduce the size and cost of the image input apparatus 2.

  The image input device 2 also includes a light shielding wall member 12 for preventing crosstalk on the image plane of the light beams that have passed through the lenses 4 to 9. The lens array 3, the light shielding wall member 12, and the image sensor 10 are integrally held by a housing 13.

  The image input device 2 includes an illumination device 17 for irradiating a finger with linearly polarized light in the near infrared band. In the example shown here, the illuminating device 17 efficiently directs near-infrared light-emitting diodes (LEDs) 14 as light sources and near-infrared light emitted by the near-infrared LEDs 14 to the finger with a more uniform light amount. As an illuminating lens 15 provided integrally with the lens array 3 for irradiation, and a near-infrared light passing through the illuminating lens 15 as a polarizer for making linearly polarized light of a predetermined polarization direction. And a polarizing film 16 provided on the lens array. Near-infrared light has a transmissivity for a living body and is absorbed by reduced hemoglobin in blood, which is suitable for inputting a vein image inside a finger.

  As shown in FIG. 2, when the lens array 3 is observed from the side facing the finger, the light-shielding wall member 12 has rectangular areas 24, 25, 26, 27, 28 corresponding to the imaging lenses 4 to 9. , 29. The lens array 3 is provided with a polarizing film 30 as an analyzer that allows only a polarization component in a direction parallel to the polarization direction of the illumination light from the illumination device 17 to pass through the region 28 corresponding to the lens 8. In a region 29 corresponding to the lens 9, a polarizing film 31 is provided as an analyzer that allows only a polarization component in a direction orthogonal to the polarization direction of illumination light from the illumination device 17 to pass.

  Although not shown in FIGS. 1 and 2, regions 24, 26, 27, 28, and 29 corresponding to the lenses 4, 6, 7, 8, and 9 of the lens array 3 are provided on the upper or lower surface of the lens array. A visible light cut filter (or a near-infrared light transmission filter) made of a metal thin film or the like formed on is formed. Note that the region 25 corresponding to the lens 5 is not provided with a visible light cut filter (or near infrared light transmission filter), but this is to make the image by the lens 5 available for detecting the illuminance around the apparatus. It is. If such use is not considered, the same filter may be provided in the region 25 of the lens 5. Further, such a filter may be omitted if the biometric authentication apparatus is used in an environment where there is no possibility of incident visible light. Further, since each lens for imaging of the lens array 3 is circular, a light-shielding film (not shown) such as a metal foil film for preventing light from entering from an area outside the effective range of each lens is, for example, a lens. It is formed on the image side surface of the array 3.

  Note that it is possible to realize a similar polarizer or analyzer by forming a fine structure on the upper surface or the lower surface of the lens array 3 instead of the polarizing films 16, 30, and 31. Are encompassed by the present invention. Such an embodiment is generally advantageous in terms of production cost and processing accuracy, compared to a mode in which a polarizing film is attached to or applied to the lens array 3.

  With reference to FIG. 1, other components of the biometric authentication device will be described. In FIG. 1, an LED drive unit 41 is means for driving the near-infrared LED 14 of the illumination device 17. The image capture unit 42 captures the image data output from the image sensor 10 and extracts image (single-eye image) data S1, S2, I1, I2, and I3 from the lenses 4, 6, 7, 8, and 9. It is. The image data captured from the image sensor 10 is compound-eye image data in which single-eye images are collected by the lenses 4 to 9, and the periphery of each single-eye image is surrounded by the shadow of the light shielding wall member 12. Therefore, each eye image data can be easily extracted by comparing the pixel value of the compound eye image data with an appropriate threshold value and detecting the shadow portion of the light shielding wall member 12. If the accuracy of the lens pitch of the lens array 3 and the dimension of the light shielding wall member 12 is sufficiently high, a predetermined area may be simply cut out as a single eye image.

  The parallax detection calculation unit 43 calculates the parallax (image shift) between the two eye images from the single eye image data S1 and S2 corresponding to the lenses 4 and 6, and the individual corresponding to the lens 7 based on the parallax. This is a calculation means for calculating the parallax of the individual eye image data I2 and I3 of the lenses 8 and 9 with respect to the eye image data I1. For example, the luminance deviation between the single eye image data S2 of the single eye image data S1 and S2 shifted in the x direction and the y direction and the other single eye image data S1 is obtained. The calculation for obtaining the square sum is repeated while changing the shift amount, and the shift amount in the x direction and the y direction in which the square sum of the luminance deviation is minimized is obtained as the parallax in the x direction and y direction between the individual images. An operation such as If the parallax between the single-eye image data S1 and S2 is calculated in this way, the parallax of the single-eye image data I2 and I3 with respect to the single-eye image data I1 can be easily calculated based on the lens pitch or the like. it can.

  The parallax adjustment calculation unit 44 performs the parallax adjustment calculation for adjusting the image position so as to cancel the parallax based on the parallax with respect to the single-eye image data I1 of the single-eye image data I2 and I3 calculated by the parallax detection calculation unit 43. This is a means applied to the eye image data I2 and I3. Between the subject distance and the parallax, there is a relationship that the parallax decreases as the subject distance increases. However, in order to perform such parallax adjustment calculation, the distance in the optical axis direction from the lens array 3 to the finger 1 is constant. Normal authentication is possible even if it is not.

  The correction calculation unit 46 is a unit that performs a correction calculation using the single-eye image data I2 and I3 after the parallax adjustment calculation on the single-eye image data I1, and the first calculation mode is selected according to the calculation mode selection signal 47. Alternatively, the second calculation mode can be selected. The calculation mode selection signal 47 is set by, for example, operating a switch (not shown) provided in the biometric authentication device, operating a switch provided in a device such as an information terminal equipped with the biometric authentication device, or It is set by a program that runs on the device. When the first calculation mode is selected, the correction calculation unit 46 subtracts the corresponding pixel value of the single-eye image data I2 after the parallax adjustment calculation from each pixel value of the single-eye image data I1 (I1- I2) is performed. When the second calculation mode is selected, the corresponding pixel value of the single-eye image data I2 after the parallax adjustment calculation is subtracted from each pixel value of the single-eye image data I1, and the single eye after the parallax adjustment calculation An operation (I1−I2 + I3) for adding the corresponding pixel values of the image data I3 is performed.

  The single-eye image data after the correction calculation by the correction calculation unit 46 is input to the collation / registration switching unit 48. A verification / registration switching signal 50 is also input to the verification / registration switching unit 48. This verification / registration switching signal 50 is set by, for example, an operation of a switch (not shown) provided in the biometric authentication device, an operation of a switch provided in a device such as an information terminal equipped with the biometric authentication device, or Set by a program running on the device.

  When the collation / registration switching signal 50 is set to the registration side, the collation / registration switching unit 48 registers the single-eye image data ((I1-I2) or (I1-I2 + I3)) output from the correction calculation unit 46. The data is transferred to a database (DB) 51, and the individual eye image data is registered as registration data in the registration database 51. When the collation / registration switching signal 50 is set to the collation side, the collation / registration switching unit 48 outputs the single-eye image data ((I1-I2) or (I1-I2 + I3) output from the correction calculation unit 46. ) Is transferred to the collation calculation unit 52, and the collation calculation unit 52 performs collation calculation between the individual eye image data and the registration data of the registration database 51. The collation calculation unit 52 constitutes a means for performing vein authentication together with the registration database 51.

  Each computing unit described above may be realized by hardware, or may be realized by a program using hardware resources such as a microprocessor.

  In the registration database 51, vein feature information such as vein travel branch point coordinates extracted from individual eye image data is registered as registration data, and the collation operation unit 52 extracts vein feature information extracted from individual eye data. You may make it perform collation with the vein characteristic information registered in the registration database 51. FIG.

  The authentication / registration operation is started by a start signal 55. The start signal 55 is generated by, for example, an operation of a switch (not illustrated) provided in the biometric authentication device, an operation of a switch provided in a device such as an information terminal equipped with the biometric authentication device, or operates on the device. Occurs in the program that runs.

  FIG. 3 is a flowchart for explaining the overall operation flow of the biometric authentication apparatus. Hereinafter, the operation of the biometric authentication apparatus will be described with reference to FIG.

  As shown in FIG. 1, when a hand signal is held over the field of view of the lens array 3 and a start signal 55 is generated by a switch operation or the like, the operation starts.

  The LED drive unit 41 drives the near infrared LED 14 for a predetermined time to emit light. The finger 1 held over the lens array 3 is irradiated with linearly polarized light in the near-infrared band, and in this state, a compound eye image that is a set of single-eye images by the lenses 4 to 9 is picked up by the image pickup device 10, and the compound eye image data is obtained. The image is captured by the image capture unit 42 (step 1).

  The image capture unit 42 extracts single-eye image data S1, S2, I1, I2, and I3 corresponding to the lenses 4, 6, 7, 8, and 9 from the compound-eye image data (step 2).

  The parallax detection calculation unit 43 calculates the parallax between the single-eye image data S1 and S2, and based on this parallax, calculates the parallax of the single-eye image data I2 and I3 with respect to the single-eye image data I1 (step 3). .

  In the parallax adjustment calculation unit 44, parallax adjustment calculation for adjusting the image position so as to cancel the parallax of the single-eye image data I2 and I3 with respect to the single-eye image data I1 is performed on the single-eye image data I2 and I3 (step 4). ).

  In the correction calculation unit 46, if the calculation mode selection signal 47 selects the first mode (step 5, first mode), the corresponding pixel value of the individual image data I2 is calculated from each pixel value of the individual image data I1. Correction processing to be subtracted is performed, and the corrected single-eye image data is transferred to the collation / registration switching unit 48 (step 6). If the calculation mode selection signal 47 selects the second mode (step 5, second mode), the corresponding pixel value of the single-eye image data I2 is subtracted from each pixel value of the single-eye image data I1. A correction operation for adding the corresponding image values of the eye image data I3 is performed, and the corrected single-eye image data is transferred to the collation / registration switching unit 48 (step 7). Then, when the collation / registration switching signal 50 is set to collation (step 8, Yes), the collation / registration switching unit 48 executes a collation operation between the individual image data and the registered data in the collation operation unit 52. If the collation / registration switching signal 50 is set to registration (step 8, No), the single-eye image data is registered in the registration database 51 (step 10).

  The illumination light from the illumination device 17 is reflected near the surface of the finger, or enters the inside of the finger and is reflected / scattered. Therefore, the light returning from the finger to the lens array 3 is reflected near the surface of the finger. There are reflected light and light reflected and scattered inside the finger. The reflected light near the finger surface is dominated by the polarization component parallel to the polarization direction of the illumination light, while the light reflected and scattered inside the finger has a different polarization direction from the illumination light. The polarization component accounts for a large proportion.

  The polarizing film (analyzer) 30 provided in the region 28 of the lens 8 allows the polarization component in the direction parallel to the polarization direction of the illumination light to pass through, but blocks the passage of the polarization component in the other direction. An image formed by the reflected light near the surface is formed by the lens 8, and this is input as single-eye image data I2.

  On the other hand, since no analyzer is provided for the lens 7, an image is formed by the lens 7 by the light reflected and scattered inside the finger and the reflected light near the finger surface. Input as I1. This single-eye image is a vein image, but its sharpness is not good due to the influence of finger surface reflected light. In particular, when the finger is held away from the lens array 3, the reflected light near the surface of the finger increases compared to the case where the finger is brought into close contact with the lens array 3, and illumination occurs as the finger position changes. The in-plane intensity distribution of the light also changes, but the change in the in-plane intensity distribution is mainly caused by the reflected light near the finger surface, so in the case of vein authentication, the influence of the reflected light near the finger surface is strongly affected. Easy to receive. However, by the correction calculation in the first or second calculation mode by the correction calculation unit 46, the influence of the reflected light on the finger surface is effectively eliminated, and a clear finger vein image by the light reflected and scattered inside the finger is obtained. Since data can be acquired, highly accurate vein authentication can be performed using this data. The single-eye image data S1 and S2 corresponding to the lenses 4 and 6 are vein image data having slightly poor definition similar to the single-eye image data I1 corresponding to the lens 7.

  A polarizing film (analyzer) 31 provided corresponding to the lens 9 passes a polarized component in a direction orthogonal to the polarization direction of the illumination light, and the linearly polarized light applied to the finger is a vein as the passing polarized component. This is a component in which the polarization direction is largely changed by being reflected by. Therefore, when the correction calculation unit 46 performs the correction calculation in the second calculation mode in which the pixel values of the single-eye image data I3 corresponding to the lens 9 are added, the vein image data in which the edge of the vein is emphasized is obtained. Can be acquired. In some cases, it is more advantageous in terms of the accuracy of vein authentication to use a vein image in which such an edge is emphasized.

  In addition, the polarizing film 31 as an analyzer should just pass the polarization component of the direction non-parallel to the polarization direction of illumination light, and the polarization direction does not necessarily need to be a direction orthogonal to the polarization direction of illumination light.

  Further, when the distance from the lens array 3 to the finger is regulated to be constant, the parallax between the single-eye images is constant and can be calculated in advance based on the lens pitch or the like. Therefore, in this case, the parallax detection calculation unit 43 can be omitted. Further, it is not necessary to extract single-eye image data S1 and S2 for use in parallax detection, and the lenses 4 and 6 can be omitted. Such embodiments are also encompassed by the present invention.

[Second Embodiment]
With reference to FIG.4 and FIG.5, the structure of the biometrics apparatus which concerns on the 2nd Embodiment of this invention is demonstrated. 4 and 5, the same reference numerals are assigned to the same or corresponding elements as those of the biometric authentication apparatus according to the first embodiment.

  In the present embodiment, as shown in FIG. 5, the imaging lenses 4, 5, 6, 7, 8, and 9 of the lens array 3 are arranged so as to form a 2 × 2 array of four each. Although the number of imaging lenses is increasing in this way, a visible light cut filter (or a near-infrared light transmission filter, not shown) is provided in each region 24, 26, 27, 28, 29 of the imaging lens other than the lens 5. ), And a polarizing film 30 as an analyzer that passes only a polarized component in a direction parallel to the polarization direction of the illumination light by the illumination device 17 is provided in the region 28 corresponding to the lens 8. In the first embodiment, the polarizing film 31 serving as an analyzer that passes only the polarization component in the direction orthogonal to the polarization direction of the illumination light by the illumination device 17 is provided in the region 29 corresponding to the lens 9. It is the same as the case of the form. Further, a light-shielding film (not shown) such as a metal thin film for preventing light from entering from an area outside the effective range of each imaging lens is formed on the surface of the lens array 3 on the image plane side, for example. .

  Now, in a state where the illumination device 17 irradiates linearly polarized illumination light onto the finger 1 (biological part) separated from the lens array 3, a set (compound image) of images (single-eye images) by the imaging lenses 7 to 9 is formed. An image is picked up by the image sensor 10. The image capture unit 42 is means for capturing compound eye image data from the image sensor 10 and extracting and outputting single-eye image data from the compound eye image data. In the present embodiment, the image capture unit 42 corresponds to four lenses 7. Four single-eye image data I1, four four-eye image data I2 corresponding to four lenses 8, four four-eye image data I3 corresponding to four lenses 9, and one lens One single-eye image data S1 corresponding to 4 (for example, the upper left lens 4 in FIG. 5), and one single-eye image data corresponding to one lens 6 (for example, the upper right lens 6 in FIG. 5). S2 is extracted and output.

  The reconstruction calculation unit 61 is a means for performing a process of reconstructing a plurality of single-eye images with parallax into a single image with higher resolution using parallax. In the example shown here, the reconstruction calculation unit 61 performs a reconstruction process for each of four pieces of single-eye image data I1, four pieces of single-eye image data I2, and four pieces of single-eye image data I3. . Specifically, in the case of the single-eye image data I1 corresponding to the four lenses 7, for example, the remaining three lenses 7 with reference to the single-eye image corresponding to the lens 7 in the upper left position in FIG. Using the parallax of the single-eye image corresponding to, and placing the pixel brightness of the data of each single-eye image at a position determined according to the position of the single-eye image and the parallax in the reconstructed image space prepared on the memory, By repeating for all the pixels of all the single-eye images, a reconstructed image can be obtained on the reconstructed image space. Reconstructed images can be obtained by the same method for the data I2 and I3 of other single-eye images. The data of each reconstructed image is input to the parallax adjustment calculation unit 44.

  The parallax adjustment calculation unit 46 performs the same parallax adjustment calculation as in the first embodiment, except that the target is not a single-eye image but a reconstructed image generated by the reconstruction calculation unit 61. This is different from the case of the first embodiment. That is, the parallax adjustment calculation unit 46 cancels the parallax between the reconstructed image from the single-eye image data I2 and the reconstructed image from the single-eye image data I3. Thus, parallax adjustment calculation for adjusting the image position is performed. In this parallax adjustment calculation, in FIG. 5, for example, the parallax of the single-eye image by the upper left lens of the four lenses 8 on the basis of the single-eye image by the upper left lens of the four lenses 7. Is used as the parallax of the reconstructed image from the monocular image by the lens 8, and similarly, the parallax of the monocular image by the upper left lens of the four lenses 9 is reconstructed from the monocular image by the lens 9. Used as parallax.

  The parallax detection calculation unit 43 calculates the parallax (image shift) between the two eye images from the single-eye image data S1 and the single-eye image data S2 by the method described above, and re-based The above-described parallax used for the reconstruction calculation in the configuration calculation unit 61 and the above-described parallax used for the parallax adjustment calculation in the parallax adjustment calculation unit 44 are calculated, and the respective parallaxes are calculated by the reconstruction calculation unit 61 and the parallax adjustment calculation unit. 44.

  The correction calculation unit 46 reconstructs the reconstructed image data (for convenience, I2, I2) from the individual image data I2, I3 after the parallax adjustment calculation for the reconstructed image data (for convenience, expressed as I1) from the single-eye image data I1. The first calculation mode or the second calculation mode can be selected in accordance with the calculation mode selection signal 47. When the first calculation mode is selected, the correction calculation unit 46 performs a correction calculation (I1-I2) for subtracting the corresponding pixel value of the reconstructed image data I2 from each pixel value of the reconstructed image data I1. Is called. When the second computation mode is selected, the corresponding pixel value of the reconstructed image data I2 is subtracted from each pixel value of the reconstructed image data I1, and the corresponding pixel value of the reconstructed image data I3 is added. Correction calculation (I1-I2 + I3) is performed.

  In addition, each calculating part etc. which were demonstrated above may be implement | achieved by hardware, and may be implement | achieved by a program using hardware resources, such as a micro processor set.

  FIG. 6 is a flowchart for explaining the overall operation flow of the biometric authentication apparatus according to the present embodiment. Hereinafter, the operation of the biometric authentication apparatus will be described with reference to FIG.

  As shown in FIG. 4, when the finger 1 is held over the lens array 3 and a start signal 55 is generated by a switch operation or the like, the operation starts.

  The LED drive unit 41 drives the near infrared LED 14 for a predetermined time to emit light. A finger 1 held over the lens array 3 is irradiated with linearly polarized light in the near-infrared band, and in this state, a compound eye image that is a set of monocular images that are substantially formed by the lenses 4 to 9 is captured by the imaging device 10. The compound eye image data is captured by the image capture unit 42 (step 21).

  In the image capturing unit 42, single-eye image data S 1 corresponding to the lens 4 from the compound-eye image data S 1, single-eye image data S 2 corresponding to the lens 6, and four single-eye image data corresponding to the four lenses 7. The data I4 of four single-eye images corresponding to the four lenses I1 and the data I3 of four single-eye images corresponding to the four lenses 9 are extracted (step 22).

  In the parallax detection calculation unit 43, the parallax between the single-eye image data S1 and S2 is calculated, and based on this parallax, the parallax used in the reconstruction calculation in the reconstruction calculation unit 61 and the parallax adjustment in the parallax adjustment calculation unit 44 The parallax used for the calculation is calculated (step 23).

  In the reconstruction calculation unit 61, a calculation for reconstructing a single image from four single-eye images by the lens 7, a calculation for reconstructing a single image from four single-eye images by the lens 8, and four by the lens 9 An operation for reconstructing a single image from the single eye image is performed (step 24).

  In the parallax adjustment calculation unit 44, parallax adjustment calculation is performed on the reconstructed image data I2 and I3 (step 25).

  In the correction calculation unit 46, if the calculation mode selection signal 47 selects the first mode (step 26, first mode), the corresponding pixel value of the reconstructed image data I2 is calculated from each pixel value of the reconstructed image data I1. A correction calculation to be subtracted is performed, and the reconstructed image data after correction is transferred to the collation / registration switching unit 48 (step 27). If the calculation mode selection signal 47 selects the second mode (step 26, second mode), the corresponding pixel value of the reconstructed image data I2 is subtracted from each pixel value of the reconstructed image data I1, and further reconstruction is performed. A correction operation for adding the corresponding image value of the image data I3 is performed, and the reconstructed image data after correction is transferred to the collation / registration switching unit 48 (step 28). Then, when the collation / registration switching signal 50 is set to collation (step 29, Yes), the collation / registration switching unit 48 executes a collation operation between the reconstructed image data and the registered data in the collation operation unit 52. If the collation / registration switching signal 50 is set to registration (step 29, No), the reconstructed image data is registered in the registration database 51 (step 31). In the registration database 51, vein feature information such as the branch point coordinates of vein travel extracted from the reconstructed image data is registered as registration data, and the vein feature information extracted from the reconstructed image data in the matching calculation unit 52; You may make it perform collation with the vein characteristic information registered in the registration database 51. FIG.

  Now, when the finger is held away from the lens array 3, the reflected light near the surface of the finger increases as compared to the case where the finger is brought into close contact with the lens array 3. Further, the in-plane intensity distribution of the illumination light also changes with the change of the finger position, and the change in the in-plane intensity distribution is mainly caused by the reflected light near the finger surface. Therefore, when an image is taken with the finger away from the lens array 3, it is easily affected by the reflected light near the finger surface. In the present embodiment, the reconstructed image data of data I1 of a single eye image (a vein image with slightly low line brightness due to the influence of light reflected on the finger surface) from the lens 7 is an image based on reflected light near the finger surface. A clear and high-resolution finger in which the influence of reflected light in the vicinity of the finger surface is effectively removed by a correction calculation (first calculation mode) for subtracting the reconstructed image data of the single-eye image data I2 by the lens 8. Since a vein image can be acquired, highly accurate vein authentication is possible.

  Further, the edge of the vein is corrected by a correction operation (second operation mode) in which the reconstructed image data of the single-eye image data I3 by the lens 9 which is an image with a polarization component in a direction orthogonal to the polarization direction of the illumination light is added. The reconstructed image data as emphasized can be acquired. Utilizing such a high-resolution vein image that emphasizes the edge may be more advantageous in terms of the accuracy of vein authentication.

  In addition, the polarizing film 31 as an analyzer should just pass the polarization component of the direction non-parallel to the polarization direction of illumination light, and the polarization direction does not necessarily need to be a direction orthogonal to the polarization direction of illumination light.

  Further, when the distance from the lens array 3 to the finger is regulated to be constant, the parallax is also constant and can be calculated in advance based on the lens pitch or the like. Therefore, in this case, the parallax detection calculation unit 43 can be omitted. Further, it is not necessary to extract single-eye image data S1 and S2 for use in parallax detection, and the lenses 4 and 6 can be omitted. Such embodiments are also encompassed by the present invention.

DESCRIPTION OF SYMBOLS 1 Finger 2 Image input device 3 Lens array 4-9 Lens 10 for imaging 10 Imaging element 11 Imaging surface 12 Light-shielding wall member 13 Case 14 Near-infrared light emitting diode (LED)
15 Illumination lens 16 Polarizing film (polarizer)
17 Illuminator 30, 31 Polarizing film (analyzer)
41 LED drive unit 42 image capture unit 43 parallax detection calculation unit 44 parallax adjustment calculation unit 46 correction calculation unit 48 collation / registration switching unit 51 registration database (DB)
52 collation calculation unit 61 reconstruction calculation unit

JP 2008-36058 A Japanese Unexamined Patent Publication No. 2009-17943 JP 2008-212311 A

Claims (8)

A lens array in which a plurality of lenses are arrayed, illumination means for irradiating a biological part held in the field of view of the lens array with linearly polarized light in a specific polarization direction, and an object side of a part of lenses of the lens array Alternatively, a first analyzer that is provided on the image plane side and passes only a polarized component in a direction parallel to the specific polarization direction, and a lens in the field of view of the lens array that is substantially imaged by the lens of the lens array. An imaging device provided on the image plane side of the lens array for capturing a compound eye image that is a set of reduced images of the living body part (hereinafter referred to as a single eye image), and within the field of view of the lens array An image input device for inputting the compound eye image in a non-contact manner with a living body part held up;
In a compound eye image input by the image input device, for each pixel value of a single eye image by a lens not provided with the first analyzer of the lens array, the first analyzer of the lens array is A correction calculation means for performing a correction calculation for subtracting the corresponding pixel value of the single-eye image by the provided lens;
A biometric authentication device comprising authentication means for performing personal authentication using a single eye image after correction calculation by the correction calculation means. The image input device includes a second analyzer that is provided on the object side or the image plane side of a part of the lenses of the lens array and passes only a polarization component in a direction non-parallel to the specific polarization direction,
In the correction calculation means, in the compound eye image input by the image input device, each pixel value of a single-eye image by a lens in which the first analyzer and the second analyzer of the lens array are not provided. On the other hand, the corresponding pixel value of the single eye image by the lens provided with the first analyzer of the lens array is subtracted, and the single eye image by the lens of the lens array provided with the second analyzer. The biometric authentication device according to claim 1, wherein a correction operation for adding the corresponding pixel values is performed. Prior to the correction calculation by the correction calculation means, parallax adjustment calculation means for performing a parallax adjustment calculation for adjusting the image position so as to cancel the parallax between the single-eye images with respect to the single-eye image related to the correction calculation;
The parallax detection calculation means which calculates the parallax between single eye images used for parallax adjustment calculation by the parallax adjustment calculation means based on a compound eye image input by the image input device. 2. The biometric authentication device according to 2. A lens array in which a plurality of lenses are arrayed, illumination means for irradiating a biological part held in the field of view of the lens array with linearly polarized light in a specific polarization direction, and an object side of a part of lenses of the lens array Alternatively, a first analyzer that is provided on the image plane side and passes only a polarized component in a direction parallel to the specific polarization direction, and a lens in the field of view of the lens array that is substantially imaged by the lens of the lens array. An imaging device provided on the image plane side of the lens array for capturing a compound eye image that is a set of reduced images of the living body part (hereinafter referred to as a single eye image), and within the field of view of the lens array An image input device for inputting the compound eye image in a non-contact manner with a living body part held up;
A first single image using parallax from a plurality of single-eye images by a plurality of lenses not provided with the first analyzer of the lens array in a compound eye image input by the image input device. And reconstructing a second single image using the parallax from a plurality of single-eye images by a plurality of lenses provided with the first analyzer of the lens array. A reconstruction calculation means for performing calculations;
Correction calculation means for performing a correction calculation for subtracting the corresponding pixel value of the second single image for each pixel value of the first single image;
A biometric authentication apparatus comprising: authentication means for performing personal authentication using a single image after correction calculation by the correction calculation means. The image input device includes a second analyzer that is provided on the object side or the image plane side of a part of the lenses of the lens array and passes only a polarization component in a direction non-parallel to the specific polarization direction,
In the reconstruction calculation means, using a parallax from a plurality of single-eye images by a plurality of lenses provided with the second analyzer of the lens array in a compound eye image input by the image input device. Reconstructing a third single image,
In the correction calculation means, the corresponding pixel value of the second single image is subtracted from each pixel value of the first single image, and the corresponding pixel value of the third single image is added. The biometric authentication device according to claim 4, wherein a correction calculation is performed. Prior to the correction calculation by the correction calculation means, parallax adjustment calculation means for performing parallax adjustment calculation for adjusting the image position so as to cancel the parallax between the single images with respect to the single image related to the correction calculation;
Based on the compound eye image input by the image input device, the parallax between single images used for the parallax adjustment calculation by the parallax adjustment calculation means and the single-eye image parallax used for the reconstruction calculation by the reconstruction calculation means. 6. The biometric authentication device according to claim 4, further comprising a parallax detection calculation means for calculating.   7. The illumination device according to claim 1, wherein the illumination unit includes an illumination lens for uniformizing the amount of linearly polarized light irradiated to a living body part held in the field of view of the lens array. The biometric authentication device according to item.   The biometric authentication device according to claim 7, wherein the illumination lens is formed on the lens array.

Images