JP2007034380A - Identity determination method and identity determination system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To determine identity of determination objects in which different feature amounts are distributed at random. <P>SOLUTION: In order to raise determination accuracy of identity, there is a proper relation between an average cycle of random pattern originating in tangles of vegetable fibers, namely, the width of vegetable fibers, and observation areas per pixel. Necessary and sufficient information required, in order to determine the identity of determination objects is acquired from a reference pattern and a verification pattern by setting a reference pattern and verification pattern so that the relation 10w≥p≥w/5 is established, if the length of one side of a factor is set to p, and the average distribution gap of the different amount of features distributed on determination objects at random is set to w. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an identity determination method and an identity determination system for determining identity of a determination target.

For example, in order to prevent documents printed on paper such as securities, certificates, and confidential documents from being forged, there is a need for a technique for determining whether a printed document is an original document. Yes.
As an apparatus for performing this type of determination, a random fiber pattern created by the plant fibers forming the test paper is observed, and a feature amount (feature vector) of a pattern image obtained from the observation region and a prestored feature amount A device that determines whether or not the inspection paper is the original paper according to the distance is known (see Patent Document 1).

JP 2004-102562 A

  However, in the above conventional example, if the size of the mesh dividing the pattern image is not appropriately set with respect to the thickness of the plant fiber forming the test paper, is the test paper the original paper? There was a problem that the accuracy of determining whether or not to increase. In the above conventional example, there is no disclosure about a method for setting an appropriate size of a mesh that divides a pattern image.

  Therefore, an object of the present invention is to provide an identity determination method and an identity determination system that can accurately determine the identity of a determination target in which different feature quantities are randomly distributed.

  In order to achieve the above object, the first feature of the present invention is that one or more matching patterns indicating a distribution of matching feature values are generated for an unknown determination target, and different feature values are generated. In the identity determination method for determining the identity of a determination target based on a correlation between a reference pattern indicating a distribution of a reference feature quantity generated for a determination target that is randomly distributed and the matching pattern, at least the reference pattern One of the verification patterns is an identity determination method in which the sizes of elements constituting the pattern are set based on an average distribution interval of different feature amounts in the determination target. In other words, since the size of the elements that make up the pattern is set based on the period at which the randomly distributed feature quantity changes, the change in the feature quantity can be sufficiently expressed by the pattern, and different feature quantities can be randomly represented. It is possible to accurately determine the identity of distributed determination targets.

  In addition, the second feature of the present invention is that an unknown determination target is compared with a reference pattern indicating a distribution of reference feature values generated for a determination target in which different feature amounts are randomly distributed. A search region including a reference region to be matched with the reference pattern is specified, and one or more verification regions having substantially the same area as the reference region that can be selected in the specified search region are used for verification. In the identity determination method of generating one or more matching patterns indicating the distribution of the feature amount of the pattern and determining the identity of the determination target based on the maximum correlation value of the matching pattern with respect to the reference pattern, at least One of the reference pattern and the matching pattern has a size of an element constituting the pattern set based on an average distribution interval of different feature amounts in the determination target. That is in the identity determination method. That is, since it becomes possible to set a threshold value for determining the identity of the determination target based on the maximum correlation value of the matching pattern with respect to the reference pattern, the identity of the determination target in which different feature quantities are randomly distributed can be set. It can be determined with high accuracy.

  Further, the third feature of the present invention is that an unknown determination target is compared with a reference pattern indicating a distribution of reference feature values generated for a determination target in which different feature amounts are randomly distributed. A search region including a reference region to be matched with the reference pattern is specified, and one or more verification regions having substantially the same area as the reference region that can be selected in the specified search region are used for verification. Generating one or more matching patterns indicating the distribution of the feature quantity of the reference pattern, and determining the maximum correlation value of the matching pattern with respect to the reference pattern and a value obtained by normalizing the maximum correlation value In the identity determination method for determining identity, at least one of the reference pattern and the matching pattern is a pattern based on an average distribution interval of different feature amounts in the determination target. The size of the elements constituting at the same determination method is set. That is, since it is possible to set two threshold values for determining the identity of the determination target based on the maximum value of the correlation value of the matching pattern with respect to the reference pattern and the value obtained by normalizing the maximum value of the correlation value. It is possible to accurately determine the identity of a determination target in which different feature amounts are randomly distributed.

  Preferably, the element is a quadrangle whose length of at least one side is set to 10 times or less of the average distribution interval.

  Preferably, the element is a quadrangle whose length of at least one side is set to one fifth or more of the average distribution interval.

  Preferably, the reference pattern and the verification pattern are patterns having 16 or more elements in each orthogonal direction.

  Further, the fourth feature of the present invention is that a pattern generation unit that generates a reference pattern indicating a distribution of a feature amount serving as a reference for a determination target in which different feature amounts are randomly distributed, and the pattern generation unit Storing at least one of a reference pattern generated by the pattern generation unit and a matching pattern for an unknown determination target to be compared with the reference pattern is different in the determination target. In the identity determination system, the size of the elements constituting the pattern is set based on the average distribution interval of the feature amount.

  The fifth feature of the present invention is that pattern generating means for generating one or more matching patterns indicating a distribution of matching feature values for an unknown determination target, and different feature values are randomly selected. Determination means for determining the identity of the determination target based on the correlation between the reference pattern indicating the distribution of the feature amount serving as a reference for the determination target distributed in the pattern generation unit and the matching pattern generated by the pattern generation unit, At least one of the verification pattern and the reference pattern generated by the pattern generation unit is in the identity determination system in which the size of the elements constituting the pattern is set based on the average distribution interval of different feature quantities in the determination target. .

  According to the present invention, it is possible to accurately determine the identity of a determination target in which different feature amounts are randomly distributed.

Next, embodiments of the present invention will be described with reference to the drawings.
In FIG. 1, the outline | summary of the identity determination system 1 which concerns on embodiment of this invention is shown. The identity determination system 1 includes a printing device 2 and a determination device body 3. The printing apparatus 2 includes, for example, an electrophotographic image forming unit, and prints a document image on the paper 20 in accordance with the control of the determination apparatus main body 3.

  The determination apparatus main body 3 includes, for example, a communication unit 30, a storage unit 32, a user interface (UI) unit 34, a small reader 36, and a control unit 38. The communication unit 30 is connected to a network such as the Internet. The storage unit 32 stores a program executed by the control unit 38, image data, and the like. The storage unit 32 can supply a program from a storage medium 320 such as a CD-ROM. The UI unit 34 includes an input device and a display device.

  The small reader 36 has a reader main body 360 formed so as to be able to cover at least a part of the determination target surface such as the paper 20, and the image pickup device 362, the window 364, and the reading device main body 360. A reading switch 366 and the like are provided. For example, a state in which plant fibers forming the paper 20 are randomly entangled is captured (read) by the image sensor 362, and image data is output to the control unit 38.

  The image sensor 362 is a monochrome type CCD having, for example, 640 × 480 (approximately 300,000 pixels) effective pixels, and an area of 7.7 mm × 5.7 mm on the sheet 20 is formed with respect to the entire area occupied by the effective pixels. An image is formed by an optical system (lens) (not shown). That is, the observation area on the paper 20 per pixel is about 12 μm × 12 μm (7.7 mm / 640≈12 μm, 5.7 mm / 480≈12 μm). For example, when the image sensor 362 is a 1/3 inch CCD and the pixel size is a square lattice of 8.4 μm × 8.4 μm (the effective screen size of the CCD is 5.4 mm × 4.0 mm), the optical system (lens ) Is set to about 0.7 times (5.4 mm / 7.7 mm≈0.7, 4.0 mm / 5.7 mm≈0.7).

  The image sensor 362 is an 8-bit (256 gradation) black and white CMOS sensor having, for example, 367 × 291 (about 110,000 pixels) effective pixels, and the entire area occupied by the effective pixels is on the paper 20. An area of 7.3 mm × 5.8 mm may be imaged by an optical system (lens) (not shown). That is, the imaging device 362 may have an observation area on the paper 20 per pixel of about 20 μm × 20 μm (7.3 mm / 367≈20 μm, 5.8 mm / 291≈20 μm). For example, when the image sensor 362 is a 1/7 inch CMOS sensor and the pixel size is a square lattice of 5.6 μm × 5.6 μm (the effective screen size of the CMOS sensor is 2.1 mm × 1.6 mm), The magnification of the optical system (lens) is set to about 0.3 times (2.1 mm / 7.3 mm≈0.3, 1.6 mm / 5.8 mm≈0.3).

  The window 364 is configured so that the user can confirm the area on the paper 20 to be read by the small reader 36. The read switch 366 is a switch for the user to operate so that the small reader 36 reads a predetermined area on the paper 20.

  The control unit 38 includes a CPU 380, a memory 382, and the like, and controls each unit constituting the identity determination system 1. For example, the determination apparatus body 3 includes a PC (personal computer) and a small reader 36 connected to the PC via a USB cable. The control unit 38 and the storage unit 32 may be disposed in an external server (not shown) via the communication unit 30.

Next, a method by which the identity determination system 1 determines the identity of a determination target such as the paper 20 will be described.
FIG. 2 shows an example of a reading area on the paper 20 read by the small reader 36. As described above, the small reader 36 reads an area of 7.7 mm × 5.7 mm on the paper 20 by the image sensor 362 and outputs it to the control unit 38. For example, as indicated by a broken line in FIG. 2, the small reader 36 first picks up an image of a region located at a predetermined distance from the lower left of the paper 20 as a reading region 4 at the time of registration of the reference pattern by user operation. The image data is output to the control unit 38.

The control unit 38 collects a part of the image data received from the small reader 36 as the reference pattern 40 and stores it in the storage unit 32. For example, as illustrated in FIG. 3, the reference pattern 40 is an image pattern including an element 400 that is a square image indicating independent densities of 16 pixels in the x direction and the y direction. The reference pattern 40 indicates the density difference of an image derived from the fact that the plant fibers forming the paper 20 are randomly entangled by 256 independent density elements 400, and different feature amounts (feature vectors) are randomly assigned. It is used as a reference for the distributed pattern.
Hereinafter, an area on the paper 20 corresponding to the reference pattern 40 is referred to as a collation reference area A.

When the identity determination system 1 determines the identity of the paper 20, that is, whether or not it is an original, an area corresponding to the matching reference area A on the paper 20 whose identity is unknown according to a user operation. Whether or not it is possible to generate an image pattern having a high similarity to the reference pattern 40 from the specified search area 5. Based on the above, the identity of the unknown sheet 20 with respect to the sheet 20 from which the reference pattern 40 is collected is determined.
The search area 5 may be specified so as to include an area corresponding to the collation reference area A in the paper 20 whose identity is unknown. For example, the position of the collation reference area A with respect to the outer shape of the paper 20 is determined in advance by the user. Or a symbol or the like may be printed on the paper 20 so that the user can roughly grasp the position of the verification reference area A.

  When the search area 5 is specified, the control unit 38 sets a matching area B (observation area) having substantially the same area as the matching reference area A in the search area 5 as shown in FIG. A matching pattern 50 (FIG. 3) indicating the distribution of the feature amount for the region B is generated. In addition, the control unit 38 sets the matching area B by shifting by one pixel in the x direction or the y direction within the search area 5 and indicates the distribution of the feature amount for each of the set matching areas B. A pattern 50 is generated.

Then, the control unit 38 calculates a correlation value with respect to the reference pattern 40 of each of the matching patterns 50 generated from the search region 5, and normalizes the maximum correlation value and the maximum correlation value (normalized score). ), The identity of the sheet 20 is determined by determining whether an image pattern having a high similarity to the reference pattern 40 is generated. In other words, the control unit 38 generates a reference pattern 50 for each of the matching regions B having substantially the same area as the matching reference region A that can be selected in the specified search region 5, thereby generating a reference from the search region 5. Whether or not an image pattern having a high similarity to the pattern 40 is generated is searched. Here, the correlation value (i, j) when the matching region B is positioned so that the upper left pixel of the matching pattern 50 is the i-th pixel in the x direction and the j-th pixel in the y direction in the search region 5 is obtained. _Assuming that r _ij , the normalized score ns _ij is expressed as the following Expression 1.

FIG. 5 is a flowchart showing a process (S10) in which the identity determination system 1 determines the identity of a determination target such as the paper 20.
As shown in FIG. 5, in step 100 (S100), the identity determination system 1 determines an area including an area corresponding to the matching reference area A on the paper 20 whose identity is unknown in accordance with a user operation. The image is specified by picking up an image as the search area 5 (FIG. 2) of the pattern for verification by the small reader 36.

  In step 102 (S102), the control unit 38 sets a matching area B in the search area 5, and generates a matching pattern 50 indicating a distribution of feature amounts for the set matching area B.

  In step 104 (S104), the control unit 38 determines whether or not the search of all selectable matching areas B (observation areas) in the specified search area 5 has been completed, and the process has not ended. The process proceeds to S106, and if completed, the process proceeds to S108.

  In step 106 (S106), the control unit 38 sets the matching region B within the search region 5 by shifting (shifting) it by one pixel in the x direction or the y direction.

  In step 108 (S108), the control unit 38 calculates a correlation value for each reference pattern 40 of each of the matching patterns 50 generated from the search area 5, and extracts the maximum value of the correlation values.

  In step 110 (S110), the control unit 38 calculates a normalized score of the matching pattern 50.

  In step 112 (S112), the control unit 38 determines the identity of the paper 20 based on the maximum correlation value of the matching pattern 50 and the normalized score of the maximum correlation value.

  In step 114 (S114), the UI unit 34 displays (outputs) the determination result of the identity of the paper 20.

Next, the relationship between the elements 400 constituting the reference pattern 40 and the matching pattern 50 and the determination accuracy of the determination target identity will be described.
The identity determination system 1 pays attention to the fact that the density of the image data obtained by capturing the paper 20 by the small reader 36 varies depending on the entanglement of the plant fibers forming the paper 20 as described above. Determine. The feature amount of the paper 20 changes in units of the width of the plant fiber. The width of the plant fiber is generally 4 to 70 μm.

For example, when the average width of the plant fibers forming the paper 20 is 20 μm, when one side of the observation area per pixel of the image sensor 362 (one side of the element 400) is less than 4 μm, fibrils (internal fibrils due to fiber damage) In other words, an unstable random pattern (such as an exposed state) is observed, and the accuracy of matching between the reference pattern 40 and the matching pattern 50 is lowered.
On the other hand, when one side of the observation area per pixel of the image sensor 362 (one side of the element 400) exceeds 200 μm, the density on the image data in the observation area caused by the entanglement of the plant fibers is averaged, which is a feature. Differences in amount are less likely to occur.

FIG. 6 shows the relationship between the similarity of the matching pattern 50 with respect to the reference pattern 40 and the frequency with which the matching pattern 50 is normalized when different papers 20 are collated and when the same paper 20 is collated. ing.
When determining the identity of the sheets 20, there may be an error in determining the same sheet 20 as a different sheet 20 and an error in setting different sheets 20 as the same sheet 20. The identity determination system 1 has both an error rate (FRR: False Reject Ratio) for determining the same sheet 20 as a different sheet 20, and an error rate (FRR: False Accept Ratio) for setting different sheets 20 as the same sheet 20. The threshold value Th is preferably set so as to be substantially zero. That is, as shown in FIG. 6, when the identity determination system 1 determines the identity of the sheet 20 based on the threshold Th for the similarity, the distribution h _FRR corresponding to _FRR and the distribution h _FAR corresponding to _FAR are 0 respectively. It is preferable to set so as to be.

  In FIG. 7, an experiment showing the relationship between the maximum correlation value and the normalized score when the identity determination system 1 generates the reference pattern 40 and the matching pattern 50 at different resolutions and the accuracy of determining identity. Results are shown. In the experimental results shown in FIG. 7, the standard deviation is converted to correspond to 10.

  The paper used in the experiment is plain paper made from hardwood (plant fiber width 10-50 μm) and conifer (plant fiber width 20-70 μm). In FIGS. 7A to 7G, the resolution is 100, 200, 300, 400, 600, 800, 1200 dpi in order. That is, one side of the observation area per pixel (one side of the element 400) in each of FIGS. 7A to 7G is 254, 127, 85, 64, 42, 32, and 21 μm. In FIG. 7, a region a indicates a region where FRR is 0 even if FAR is 0, a region b indicates a region where both FAR and FRR are 0, and a region c indicates that FAR is 0 even if FRR is 0. It is shown that the area is not set, and the larger the area b, the higher the determination accuracy of the identity.

  In the comparative example (100 dpi: one side of the element 400 is 254 μm) shown in FIG. 7A, it is shown that it is difficult to increase the accuracy of identity determination. In addition, even if the resolution becomes too high (one side of the element 400 becomes too small), it means that a narrow area is observed in detail, and it becomes difficult to increase the accuracy of identity determination. Has been.

  As described above, the accuracy of determining the identity is increased between the average period of the random pattern derived from the entanglement of the plant fibers, that is, the width of the plant fibers and the observation area (the size of the element 400) per pixel. Therefore, when the length of one side of the element 400 is p, and the average distribution interval (such as the average width of plant fibers) of different feature quantities randomly distributed on the determination target is w, When the relationship is established, it can be seen that information necessary and sufficient for determining the identity of the determination target can be obtained from the reference pattern 40 and the matching pattern 50.

  10w ≧ p ≧ w / 5 (2)

  Further, in the reference pattern 40 and the matching pattern 50, the number of patterns that can be described is q to the d power, where d is the number of elements 400 (pixels) and q is the density level of each pixel. When the resolution is 2400 dpi, the length of one side of the element 400 is 11 μm and the collation reference area A and the collation area B are set narrowly because the width of the plant fiber forming the paper is 4 to 70 μm. The pattern 40 and the matching pattern 50 reflect the distribution of feature quantities derived from a very small number of plant fibers, and it is difficult to increase the accuracy of determining the identity. For example, when the reference pattern 40 and the matching pattern 50 are composed of 8 × 8 elements 400 (pixels), the matching reference area A and the matching area B have an area of about 85 μm × 85 μm, and only a few plant fibers are present. It is thought that it does not exist. Therefore, in order to obtain sufficient determination accuracy, the reference pattern 40 and the verification pattern 50 are preferably patterns having 16 or more elements 400 in each orthogonal direction.

  However, if the number of elements 400 constituting the reference pattern 40 and the matching pattern 50 increases, the amount of calculation necessary for determining the identity increases. Therefore, the reference pattern 40 and the matching pattern 50 are 16 × 16. It is preferably composed of ~ 256 x 256 elements 400. If the resolution is high, the identity determination accuracy cannot be increased unless the reference pattern 40 and the verification pattern 50 are composed of many elements 400. Therefore, when the determination target is paper, the length of one side of the element 400 is preferably 42 to 85 μm (600 to 300 dpi), and the number of the elements 400 is preferably 32 × 32 to 256 × 256.

  Further, when the resolution of the image sensor 362 is high and the length of one side of the observation area per pixel is less than 4 μm, the control unit 38 samples the image data (such as mosaic processing), The length of one side of the element 400 may be 4 μm or more.

  In the present embodiment, the case where the determination target is a sheet has been described as an example. However, the present invention is not limited to this, and the identity determination system 1 may be a determination target in which different feature amounts are randomly distributed. The identity of the determination target can be determined with high accuracy. For example, in the case where the determination target is a felt which is a kind of non-woven fabric, the thickness of the wool that is a felt fiber is 18 to 25 μm, and the average diameter of the fiber is 20 μm, as in the above embodiment. The length of one side of the element 400 is preferably set in the range of 4 μm to 200 μm.

  Further, when observing random scattering of toner around the toner pattern around the pattern of the toner fixed on the paper by the image forming apparatus, considering aggregation of toner particles (several μm to 10 μm), etc. It is preferable to set the length of one side of the element 400 within the range of 4 μm to 200 μm.

  In addition, since the average particle diameter of the metal particles dispersed in the paint such as the glossy painted surface is 15 to 30 μm, the length of one side of the element 400 is set to 4 μm even when determining the identity of the glossy painted surface. It is preferable to set within a range of ~ 200 μm.

It is a block diagram which shows the outline | summary of the identity determination system which concerns on embodiment of this invention. It is an example of the reading area on the paper which a small reader reads. It is an image pattern which shows the structure of a reference | standard pattern and the pattern for collation. It is a schematic diagram which shows the state in which a control part searches whether an image pattern with high similarity with a reference | standard pattern is produced | generated from a search area | region. It is a flowchart which shows the process (S10) which the identity determination system determines the identity of determination objects, such as a paper. It is a graph which shows the relationship between the similarity of the pattern for a collation with respect to the reference | standard pattern at the time of collating different paper, and the case where the same paper is collated, and the frequency which normalized the pattern for collation. It is a graph which shows the experimental result of the relationship between the maximum value of a correlation value, the normalized score, and the precision which determines an identity in case the identity determination system produces | generates a reference pattern and the pattern for collation with different resolution.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Identity determination system 2 ... Printing apparatus 20 ... Paper 3 ... Document processing apparatus main body 32 ... Memory | storage part 34 ... UI part 36 ... Small reader
360 ... Reading device body
362 ... Image sensor
364 ... Window
366 ... Reading switch 38 ... Control unit
380 ... CPU
382 ... Memory 4 ... Reading area 40 ... Reference pattern
400 ... element 5 ... search area 50 ... pattern for verification

Claims (11)

  One or more matching patterns indicating a distribution of matching feature values for an unknown determination target are generated, and a reference feature amount distribution generated for a determination target in which different feature values are randomly distributed In the identity determination method for determining the identity of the determination target based on the correlation between the reference pattern indicating the reference pattern and the verification pattern, at least one of the reference pattern and the verification pattern is an average distribution of different feature quantities in the determination target An identity determination method, wherein the size of an element constituting a pattern is set based on an interval.   A search area including a matching reference area that should correspond to the reference pattern in an unknown determination target with respect to a reference pattern indicating a distribution of a reference feature quantity generated for a determination target in which different feature quantities are randomly distributed And one or more matching patterns indicating the distribution of feature quantities for matching with respect to one or more matching areas having substantially the same area as the matching reference area that can be selected in the specified search area In the identity determination method for determining identity of a determination target based on a maximum correlation value of the verification pattern with respect to the reference pattern, at least one of the reference pattern and the verification pattern is determined An identity determination method, wherein the size of an element constituting a pattern is set based on an average distribution interval of different feature values in a target.   A search area including a matching reference area that should correspond to the reference pattern in an unknown determination target with respect to a reference pattern indicating a distribution of a reference feature quantity generated for a determination target in which different feature quantities are randomly distributed And one or more matching patterns indicating the distribution of feature quantities for matching with respect to one or more matching areas having substantially the same area as the matching reference area that can be selected in the specified search area In the identity determination method for determining identity of a determination target based on a maximum value of a correlation value of the matching pattern with respect to the reference pattern and a value obtained by normalizing the maximum value of the correlation value, at least the reference Either of the pattern and the pattern for collation is configured such that the size of the elements constituting the pattern is set based on the average distribution interval of different feature quantities in the determination target. Identity determination method comprising.   4. The identity determination method according to claim 1, wherein the element is a quadrangle whose length of at least one side is set to 10 times or less of the average distribution interval. 5.   5. The identity determination method according to claim 1, wherein the element is a quadrangle whose length of at least one side is set to one fifth or more of the average distribution interval.   The identity determination method according to claim 1, wherein the reference pattern and the matching pattern are patterns having 16 or more elements in each orthogonal direction.   A pattern generation unit that generates a reference pattern indicating a distribution of a reference feature amount for a determination target in which different feature amounts are randomly distributed; and a storage unit that stores a reference pattern generated by the pattern generation unit. At least one of the reference pattern generated by the pattern generation unit and the matching pattern for an unknown determination target that is matched with the reference pattern forms a pattern based on the average distribution interval of different feature amounts in the determination target An identity determination system characterized by the element size being set.   Pattern generating means for generating one or more matching patterns indicating the distribution of matching feature quantities for an unknown judgment target, and a reference feature quantity for judgment targets in which different feature quantities are randomly distributed A determination unit that determines the identity of a determination target based on a correlation between a reference pattern indicating a distribution and a verification pattern generated by the pattern generation unit, and at least of the verification pattern and the reference pattern generated by the pattern generation unit One of the identity determination systems is characterized in that the size of an element constituting the pattern is set based on an average distribution interval of different feature amounts in the determination target.   9. The identity determination system according to claim 7 or 8, wherein the element constituting the pattern is a quadrangle whose length of at least one side is set to 10 times or less of an average distribution interval of different feature quantities in the determination target. .   The element constituting the pattern is a quadrangle whose length of at least one side is set to one fifth or more of an average distribution interval of different feature quantities in the determination target. Identity determination system.   11. The identity determination system according to claim 7, wherein the reference pattern and the verification pattern are patterns having 16 or more elements constituting the pattern in each orthogonal direction.

Images