Authenticity determination device and program

Abstract

To provide an authenticity determination apparatus, or the like, configured to properly determine authenticity of an identification certificate.SOLUTION: An authenticity determination apparatus 3 includes: receiving means 301 which receives a captured image of an ID card from a user terminal 5; and authenticity determination means 302 which determines authenticity of the ID card by comparing a contour part of a character to be inspected that is formed in a predetermined range of the ID card with a contour part of an authentic character, using the captured image of the ID card. The character to be inspected is a character in a pre-printed part of the ID card.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to an authenticity determination device and a program therefor. [Background technology]

[0002] When opening an account at a financial institution, identification is required using an ID card (personal identification) such as a driver's license, but there has been an increase in cases where ID cards are forged or altered in order to open an account under the guise of another person for criminal purposes such as special fraud.

[0003] At typical financial institutions, users go to the institution to submit their ID card, and the financial institution then uses the submitted ID card to verify the user's identity in person. In this case, the ID card can be carefully observed, and any anomalies can be detected by touch alone.

[0004] Furthermore, Patent Documents 1 and 2 describe an authenticity determination device that determines the authenticity of an ID card from the image obtained by reading the face of the ID card with a reading device such as a scanner, and it is also possible to determine the authenticity of an ID card by reading a submitted ID card with a reading device such as a scanner. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-34535 [Patent Document 2] Japanese Patent Application Publication No. 2019-117549 Summary of the Invention [Problem to be solved by the invention]

[0006] Recently, online financial institutions such as online banks have become common. Some of these financial institutions verify identity by taking a photo of the face of an ID card with a camera on a smartphone or other device and sending the image to the financial institution.

[0007] In this case, identity verification is performed only by image rather than face-to-face, so a method for detecting counterfeits is desired. However, it is difficult to detect counterfeits from photographs of ID cards, which has been an issue when applying online.

[0008] The present invention has been made in view of the above problems, and has as its object to provide an authenticity determination device and the like that can suitably determine the authenticity of an identification card. [Means for solving the problem]

[0009] The first invention for solving the above-mentioned problems is an apparatus for determining the authenticity of an identification card, which has an authenticity determination means for determining the authenticity of the identification card by using a photographed image of the identification card and comparing the outline of a character to be inspected formed in a predetermined area of the identification card with the outline of a genuine character. The authenticity determination means calculates a vector from the origin to a contour pixel in an image of a character region of a positive character, detects a corresponding pixel at the end of the vector starting from the origin in an image of the character region of the character to be inspected, and if the corresponding pixel is not a contour pixel, calculates a difference between the position of the detected contour pixel and the corresponding pixel by searching for pixels in a range outside the corresponding pixel as a difference vector, and determines the authenticity of the identification card based on a determination value based on the difference vector. The authenticity determination device is characterized by the above.

[0010] The font of the characters formed on the identification card is predetermined, and therefore, by comparing the outline of the character to be inspected formed in a predetermined area on the identification card with the outline of the correct character to determine authenticity, it is possible to suitably distinguish between a genuine identification card in which the character to be inspected is formed in the predetermined font and a false identification card in which the character to be inspected is formed in a font other than the predetermined font. Furthermore, the difference between the outline of the character to be inspected and the outline of the correct character can be appropriately evaluated using the difference vector, and the authenticity can be suitably determined.

[0011] The characters to be inspected are preferably characters in the pre-printed portion of the personal identification card. For characters in the pre-printed portion, it is predetermined which characters will be formed in which ranges, so the positive characters to be compared are known in advance, and only a small number of positive data are required.

[0012] The predetermined range may be a range in which numbers or letters are formed as on-demand information on the personal identification card. Regarding on-demand information, the characters formed within that range differ depending on the individual ID. However, there are also on-demand information ranges that consist only of numbers 0 to 9, such as date of birth, and ranges that consist only of letters A to Z. Therefore, when using characters within those ranges to determine authenticity, a relatively small amount of positive data is required.

[0014] The contour pixels of the character to be inspected are obtained as pixels of the second gradation value by processing the target pixel to the second gradation value if the target pixel has the first gradation value and all eight pixels surrounding the target pixel are not the first gradation value in a binary image in which the character is set to a first gradation value and the background is set to a second gradation value, and otherwise the target pixel is set to the first gradation value.It is also desirable that the contour pixels of a positive character are obtained as pixels of the second gradation value by processing the target pixel to the first gradation value if the target pixel has the first gradation value and all four pixels above, below, left, and right of the target pixel are not the first gradation value in a binary image in which the character is set to a first gradation value and the background is set to a second gradation value, and otherwise the target pixel is set to the first gradation value. Since the above-mentioned method of calculating the difference vector searches for pixels in the character area to be inspected that correspond to the contour pixels of the positive character, the number of pixel searches corresponds to the number of contour pixels of the positive character, and the larger the number of contour pixels of the character to be inspected, the earlier one search will be completed. Therefore, as described above, by reducing the contour pixels of the positive character using strict criteria and increasing the contour pixels of the character to be inspected using looser criteria, it is possible to quickly determine whether the identity card is genuine and to prevent erroneous determination of a genuine identity card as fake.

[0015] It is desirable that the authenticity determination device has a receiving means for receiving the captured image from a user terminal. This makes it possible to construct an authenticity determination system using a network in which an authenticity determination device determines the authenticity of a photographed image of an identification card sent from a user terminal.

[0016] The second invention is a program for causing a computer to function as an authenticity determination device for an identification card, the authenticity determination device having an authenticity determination means for determining the authenticity of the identification card by using a photographed image of the identification card and comparing the outline of a character to be inspected formed in a predetermined area of the identification card with the outline of a genuine character. wherein the authenticity determination means calculates a vector from an origin to a contour pixel in an image of a character region of a positive character, detects a corresponding pixel at the end of the vector starting from the origin in an image of the character region of the character to be inspected, and if the corresponding pixel is not a contour pixel, calculates a difference in position between the detected contour pixel and the corresponding pixel by searching for pixels in a range outside the corresponding pixel as a difference vector, and determines the authenticity of the identification card using a determination value based on the difference vector. is. [Effects of the Invention]

[0017] According to the present invention, it is possible to provide an authenticity determination device and the like that can suitably determine the authenticity of an identification card. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 shows an authenticity determination system 1. [Figure 2] FIG. 2 is a diagram showing the hardware configuration of the authenticity determination device 3. [Figure 3] FIG. 2 is a diagram showing the hardware configuration of a user terminal 5. [Figure 4] FIG. 2 is a block diagram showing the functions of the authenticity determination device 3. [Figure 5] 1 is a diagram showing an ID card 10. FIG. [Figure 6] 3 is a flowchart showing an outline of the processing of the authenticity determination system 1. [Figure 7] An example of the ID card 10 photo capture screen. [Figure 8] 10 is a flowchart showing an authenticity determination method. [Figure 9] FIG. 10 is a diagram for explaining extraction of a character range 200. [Figure 10] 3A to 3C are diagrams for explaining the creation of a character image 300. [Figure 11] 3A and 3B are diagrams for explaining binarization of a character image 300. [Figure 12] 4A to 4C are diagrams for explaining notch removal from a binary image 400. FIG. [Figure 13] 5A to 5C are diagrams for explaining the creation of an inspection image 500. [Figure 14] FIG. 10 is a diagram illustrating calculation of a difference vector Vec. [Figure 15] FIG. 10 is a diagram illustrating correction of a vector v. [Figure 16] 10 is a flowchart showing the procedure for determining authenticity. [Figure 17] FIG. 1 is a diagram illustrating reflection of a hologram. DETAILED DESCRIPTION OF THE INVENTION

[0019] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

[0020] (1. Authenticity Determination System 1) 1 is a diagram showing an authenticity determination system 1 having an authenticity determination device 3 according to an embodiment of the present invention. The authenticity determination system 1 is configured by connecting the authenticity determination device 3 and a user terminal 5 via a network so that they can communicate with each other. In the authenticity determination system 1, a captured image of a user's ID card 10 is sent from the user terminal 5 to the authenticity determination device 3, and the authenticity determination device 3 determines the authenticity of the ID card 10 from the captured image.

[0021] 2 is a diagram showing the hardware configuration of the authenticity determination device 3. The authenticity determination device 3 can be realized by a computer configured by connecting a control unit 31, a storage unit 32, a communication unit 33, etc. via a bus or the like. However, the authenticity determination device 3 is not limited to this and various configurations can be used as appropriate, and it is also possible to realize the authenticity determination device 3 using multiple computers.

[0022] The control unit 31 is composed of a CPU, ROM, RAM, etc. The CPU loads programs related to the processing of the authenticity determination device 3 stored in storage media such as the memory unit 32 and ROM into a work area on the RAM and executes them. The ROM is a non-volatile memory that permanently stores programs such as the boot program and BIOS, as well as data. The RAM is a volatile memory that temporarily stores programs and data loaded from the memory unit 32, ROM, etc., and also has a work area that the control unit 31 uses to perform various processes.

[0023] The storage unit 32 is a hard disk drive, a solid state drive, a flash memory, or the like, and stores programs executed by the authenticity determination device 3 in the processes described below, data required for program execution, an OS, and the like.

[0024] The communication unit 33 is a communication interface that mediates communication via a network, and communicates with the user terminal 5 .

[0025] The user terminal 5 is a terminal carried by the user, which takes a picture of the ID card 10 and transmits the photographed image to the authenticity determination device 3. As the user terminal 5, for example, a mobile terminal such as a smartphone or a tablet terminal is used.

[0026] Fig. 3 is a diagram showing the hardware configuration of the user terminal 5. As shown in Fig. 3, the user terminal 5 is configured by connecting a control unit 51, a storage unit 52, a display unit 53, an input unit 54, a communication unit 55, a camera 56, an audio input / output unit 57, etc. via a bus or the like. However, the configuration is not limited to this, and various other configurations may be used as appropriate.

[0027] The control unit 51, the storage unit 52, and the communication unit 55 are substantially the same as the above-described control unit 31, the storage unit 32, and the communication unit 33. The display unit 53 has a display device such as a liquid crystal panel, and is provided with a touch panel as an input unit 54 for inputting information to the user terminal 5. The audio input / output unit 57 has a microphone and a speaker used for audio input and output.

[0028] Camera 56 is an area camera that includes an optical lens, an imaging element such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal-Oxide Semiconductor), an A / D (Analog / Digital) converter, etc. Camera 56 photoelectrically converts the subject image input through the optical lens using the imaging element to generate an analog image signal. The A / D converter then converts the analog image signal into digital image data.

[0029] 4 is a block diagram showing the functions of the authenticity determination device 3. The authenticity determination device 3 includes a receiving means 301, an authenticity determination means 302, and the like.

[0030] The receiving means 301 is a means by which the control unit 31 of the authenticity determination device 3 receives a photographed image of the ID card 10 from the user terminal 5 via the communication unit 33 .

[0031] The authenticity determining means 302 determines the authenticity of the ID card 10 by using the photographed image of the ID card 10 received by the control unit 31 of the authenticity determining device 3 from the user terminal 5. Details of this authenticity determination will be described later.

[0032] (2. ID Card 10) In this embodiment, the ID card 10 is an identification card that is the subject of authentication determination. An identification card is a medium used for identity verification by financial institutions, mobile phone carriers, various government agencies, etc., such as a residence card, driver's license, My Number card, or special permanent resident card. In this embodiment, a hologram is formed on the ID card 10 as an anti-counterfeit measure, but the present invention is not limited to this.

[0033] FIG. 5 is a diagram showing an outline of an ID card 10 (residence card), and shows a schematic diagram in which some of the face information of the ID card 10 is omitted.

[0034] Various information is printed on the ID card 10. For example, for a residence card, examples of the information on the face of the card include the country of issue 11, residence number 12, date of birth of the card holder 13, facial image 14, residence status 15, whether or not there are employment restrictions 16, date of issuance of the card 17, and expiration date 18.

[0035] Among these, residence number 12, date of birth 13, facial image 14, residence status 15, presence or absence of employment restrictions 16, date of issue 17, expiration date 18, etc. are on-demand information that differs depending on each individual ID card 10.

[0036] Meanwhile, characters such as "number," "date of birth," "residence status," "whether or not there is a work restriction," "date of issue," and "this card is..." that indicate the entry fields for residence number 12, date of birth 13, residence status 15, whether or not there is a work restriction 16, date of issue 17, and expiration date 18, as well as the card issuing country 11 and the name of ID card 10 ("residence card"), are used in common on each ID card 10. These characters are formed in advance in a predetermined position on the card substrate before the on-demand information is printed, and are called pre-printed portions.

[0037] The ID card 10 is manufactured by printing the on-demand information on a card substrate on which the pre-printed portion has been formed in advance. The configuration of the on-demand information and pre-printed portion varies depending on the type of ID card 10, such as a residence card, driver's license, My Number card, or special permanent resident card.

[0038] (3. Overview of the processing of the authenticity determination system 1) Fig. 6 is a flowchart showing an outline of the processing executed by the authenticity determination system 1. S1 to S2 and S7 to S8 in Fig. 6 are processing executed by the control unit 51 of the user terminal 5 by controlling each unit of the user terminal 5, and S3 to S6 are processing executed by the control unit 31 of the authenticity determination device 3 by controlling each unit of the authenticity determination device 3.

[0039] In this embodiment, the user first operates the user terminal 5 to launch a dedicated application program pre-stored in the storage unit 52. The user terminal 5 then displays a menu screen (not shown) on the display unit 53, and activates the camera 56 in accordance with the user's selection on the menu screen to photograph the face of the ID card 10 (S1).

[0040] 7 shows an example of the screen for capturing the ID card 10. In this embodiment, when the camera is started, instructions are displayed on the display unit 53 of the user terminal 5 to prompt the user to place the ID card 10 within a predetermined frame 531, and then the ID card 10 is captured.

[0041] When the user terminal 5 photographs the ID card 10, it transmits the photographed image to the authenticity determination device 3 (S2). The photographed image is an RGB image in which the pixel colors are determined by the gradation values of the R (red), G (green), and B (blue) components. When the authenticity determination device 3 receives the photographed image of the ID card 10 (S3), it determines the authenticity of the ID card 10 using the photographed image (S4).

[0042] Authenticity determination in this embodiment is performed based on the fact that a predetermined font is used for the characters on the face of the genuine ID card 10. Details of the authenticity determination method will be described later.

[0043] If the authenticity determination device 3 determines that the ID card 10 is not genuine (S5; NO), it transmits a message to the user terminal 5 indicating that the ID card 10 is not genuine (S6). When the user terminal 5 receives a message indicating that the ID card 10 is not genuine (S7), it displays this message on the display unit 53 (S8) and ends the process.

[0044] On the other hand, if the authenticity determination device 3 determines that the ID card 10 is authentic (S5; YES), it completes the authenticity determination process. For example, when the above authenticity determination is performed as identity verification when opening an account at a financial institution on the Internet, after the authenticity determination process, user information (such as name and date of birth) on the face of the ID card 10 can be obtained from a photographed image of the ID card 10 as information required for opening the account, and the user information can be registered and stored in the memory unit 52.

[0045] (3-1. Authenticity determination based on character shape) As described above, in this embodiment, the authenticity of the genuine ID card 10 is determined in step S4 (see FIG. 6) based on the shape of the characters, which are printed on the face of the genuine ID card 10 using a predetermined font.

[0046] That is, the characters in the pre-printed portion and on-demand information on the ID card 10 are written in a predetermined font within a predetermined range. Therefore, by comparing the shape of the characters in the predetermined range extracted from the photographed image of the ID card 10 with the correct characters used in the predetermined range on the genuine ID card 10, it is possible to identify a fake ID card 10 that uses a different font from that of the correct ID card 10.

[0047] FIG. 8 is a flowchart showing the steps of the authenticity determination method in S4 (see FIG. 6).

[0048] In this embodiment, in S4, a predetermined range of the captured image (RGB image) of the ID card 10 is first converted to grayscale to create a grayscale image. The B and R components of a predetermined range of the captured image are then extracted, and an image based on the gradation values of the B component (hereinafter referred to as the B image) and an image based on the gradation values of the R component (hereinafter referred to as the R image) are created (S401). The predetermined range can be set in advance as the range in which the characters to be inspected exist, and in this embodiment, it is the range in which the characters in the pre-printed portion described above are formed.

[0049] Next, the authenticity determination device 3 binarizes the grayscale image (S402). A grayscale image is an image in which pixel brightness is expressed by a gradation value of 0 to 255, with a gradation value of 0 being black and a gradation value of 255 being white. A binary image is an image in which the character portion is represented by black pixels (pixels of a first gradation value) and the background portion is represented by white pixels (pixels of a second gradation value). This is also true for other binary images described later. However, the specifications of grayscale images and the representation of binary images are not limited to those described above.

[0050] Next, the authenticity determination device 3 cuts out the range of characters to be inspected (referred to as a character range) from the binarized image (S403).

[0051] While the extraction method is not particularly limited, in this embodiment, first, columns 101 and rows 102 containing black pixels are detected in a binarized image 100, as shown in Fig. 9(a). Then, as shown in Fig. 9(b), a region where a predetermined number or more of columns 101 containing black pixels are consecutive and located near the center of the image 100 is designated as 101a, and a region where a predetermined number or more of rows 102 containing black pixels are consecutive and located similarly near the center of the image 100 is designated as 102a, and a range 100a consisting of regions 101a and 102a is extracted.

[0052] The authenticity determination device 3 performs the processes shown in Figures 9(a) and 9(b) on the extracted range 100a as necessary, and determines the finally extracted range 200 as the character range as shown in Figure 9(c). The above-mentioned predetermined number is, for example, 10 dots, but is not limited to this. Figure 10(a) shows an example in which the character to be inspected is the character "tome" (return) in the pre-printed portion of the ID card 10, and the character range 200 of "tome" is extracted by the processes of S402 and S403.

[0053] As shown in FIG. 10(b), the authenticity determination device 3 converts the resolution (dpi) of the original grayscale image 20 from which the character range 200 was cut out so that the character range 200 has a predetermined size (S404).

[0054] In this embodiment, the predetermined size is set to multiple numbers of dots in both the vertical and horizontal directions (three numbers: 89, 90, and 91 in this embodiment), and the processes in steps S405 to S410, which will be described later, are performed on multiple grayscale images 20' converted so that the character range 200 has these sizes. This is to reduce the influence of noise in the captured image and improve the accuracy of determination. However, the predetermined size may be one, and there are no particular limitations on the specific size. The magnification during resolution conversion is a magnification that adds the larger of the number of dots in both the vertical and horizontal directions of the character range 200 in the original grayscale image 20 to the number of dots described above, and the resolution conversion can be performed using, for example, the bicubic method.

[0055] Next, the authenticity determination device 3 again extracts the character to be inspected from the grayscale image 20' to create a character image (S405). Here, as shown in Fig. 10(c), the range of the character image 300 is set to a range whose center C coincides with the center of the character range 200 in the grayscale image 20' and whose size is larger than the predetermined size (for example, 104 dots vertically and horizontally), but is not limited to this.

[0056] Next, the authenticity determination device 3 performs noise removal processing (S406) on the character image 300. In this embodiment, low-pass filtering and expansion / contraction processing are performed as noise removal processing, thereby removing high-frequency tone value changes (noise).

[0057] In this embodiment, in S406, first, low-pass filtering is performed on the character image 300. Low-pass filtering is a known technique, but in this embodiment, for example, as shown in the following equation (1), the gradation value pn of each pixel in a certain range centered on a pixel of interest in the character image 300 is multiplied by a coefficient kn (weighting coefficient) corresponding to the position of each pixel to calculate the sum Σ(pn×kn), and this sum Σ(pn×kn) is then divided by the sum Σkn of the coefficients kn to obtain the gradation value p0′ of the pixel of interest, and this process is performed with each pixel of the character image 300 as the pixel of interest. p0' = Σ(pn × kn) / Σkn…(1)

[0058] However, in this embodiment, the above sums Σ(pn×kn) and Σkn are calculated for pixels within the above fixed range whose gradation value pn differs from the original gradation value p0 of the pixel of interest by less than a predetermined value. By ignoring pixels whose gradation values differ greatly from the pixel of interest when calculating the sums Σ(pn×kn) and Σkn, low-pass filtering can be performed to prevent blurring of the edges around the characters. The above predetermined value is, for example, 30, but is not limited to this.

[0059] In S406, the character image 300 after the low-pass filter process is further subjected to expansion / contraction processing of the character portion. Expansion / contraction processing is known, but in this embodiment, for example, a process is performed in which the gradation value of a pixel of interest is replaced with the largest value among the gradation values of pixels in a certain range centered on the pixel of interest, with each pixel of the character image 300 being treated as the pixel of interest. After the character portion is contracted in this way, the reverse process is performed in which the gradation value of the pixel of interest is replaced with the smallest value among the gradation values of pixels in a certain range centered on the pixel of interest, with each pixel of the character image 300 being treated as the pixel of interest. By expanding the character portion after contraction in this way, the outline of the character portion is shaped and noise is reduced.

[0060] Finally, the process of S406 is terminated by performing a trimming process on the character image 300. The part to be trimmed is a part of a predetermined width on the outer edge of the character image 300, and the width is determined according to the size of each fixed range set in the process of S407.

[0061] Thereafter, the authenticity determination device 3 performs a binarization process on the character image 300 (S407).

[0062] In this embodiment, as the positive data to be used for authenticity determination, a positive binary image of the character to be inspected (in this embodiment, "tame") is registered in advance in the storage unit 32 of the authenticity determination device 3. In S407, in order to perform optimal binarization after comparison with the positive data, the threshold value for binarization is determined by comparison with the positive data.

[0063] More specifically, a histogram showing the number of pixels for each gradation value is first created for the character image 300. Then, as shown in FIG. 11(a), the range between the minimum and maximum gradation values V appearing in the character image 300 is equally divided, and the division values Dn and the center value Cn of each division interval D are set. There are various methods for setting this, but in this example, the minimum gradation value V within the division interval D is set as the division value Dn, and the maximum value of the division value Dn is 1 greater than the maximum gradation value V appearing in the character image 300. The center value Cn can also be calculated by averaging the upper and lower division values Dn. In reality, the number of division intervals D is approximately 16, but for the sake of explanation, the number is set to 3 in the example of FIG. 11.

[0064] Next, for each divided section D, the center of gravity value Mn is calculated using the following formula (2) using each gradation value V within the divided section D and the number of pixels S for each gradation value V. Then, as shown in Fig. 11(b), a new divided value Dn is calculated by averaging the center of gravity values Mn of the upper and lower divided sections D, and the divided value Dn (excluding the maximum and minimum values of the divided value Dn) is updated accordingly. Mn=Σ(V·S) / ΣS…(2)

[0065] Then, as shown in Figure 11(c), a new center value Cn is calculated in the same manner as above using the new updated divided value Dn, and the center value Cn is updated accordingly. The processes of Figures 11(a) to 11(c) are repeated until there is no difference between the new center value Cn and the center value Cn before the update, and the divided value Dn at the end of the process (excluding the minimum value of the divided value Dn) is set as the threshold. This results in multiple thresholds. Note that each value other than the center of gravity value Mn is an integer value, and in the above process, the calculated value is converted to an integer by rounding, etc., as necessary.

[0066] The authenticity determination device 3 binarizes the character image 300 using each threshold value, compares the number of black pixels in the binarized character image with the number of black pixels in the positive data, and determines the threshold value that minimizes the difference as the final threshold value. In S407, the character image 300 is binarized using this threshold value, thereby achieving optimal binarization in comparison with the positive data.

[0067] FIG. 12(a) is an example of a binary image 400 obtained by the process of S407, and the authenticity determination device 3 performs notch deletion processing on this binary image 400 (S408).

[0068] The notches to be removed are minute (for example, one-dot) irregularities, and in this embodiment, when a pattern with a notch is detected within a certain range centered on a pixel of interest in the binary image 400, the pattern is replaced with a pattern without a notch, as shown in Fig. 12(b), and this process is performed with each pixel of the binary image 400 as the pixel of interest. There is a one-to-one correspondence between patterns with notches and patterns without notches, and a plurality of patterns other than the example shown in Fig. 12(b) are prepared for each of the notches for white pixels and the notches for black pixels.

[0069] In this embodiment, the same replacement process is performed again as necessary on the binary image 400 after the above replacement process has been performed, thereby reliably removing small irregularities present in the binary image 400 and improving the accuracy of authenticity determination.

[0070] Next, the authenticity determination device 3 performs character outline extraction processing on the binary image 400 after notch deletion to create an inspection image (S409).

[0071] In this embodiment, as shown in Fig. 13(a), for a pixel of interest in binary image 400, if the pixel of interest is black (first gradation value) and at least one of the eight pixels surrounding the pixel of interest is white (second gradation value), the pixel of interest is set to white, and otherwise the pixel of interest is set to black. This process is performed on each pixel of binary image 400 as the pixel of interest. This creates an inspection image 500 in which the outline pixels of characters are shown in white, as shown in Fig. 13(b).

[0072] On the other hand, in this embodiment, as shown in Fig. 13(c), a similar contour extraction process is performed on positive data to obtain a positive contour image. However, the conditions for contour extraction are different from those described above. For a pixel of interest in positive data (binary image), if the pixel of interest is black (first gradation value) and at least one of the four pixels above, below, left, and right of the pixel of interest is white (second gradation value), the pixel of interest is set to white; otherwise, the pixel of interest is set to black. As a result, a positive contour image R is created for positive data as well, as shown in Fig. 13(d), in which the contour pixels of characters are shown in white.

[0073] As can be seen from a comparison of Figures 13(a) and 13(c), the conditions for contour extraction are stricter for the positive data, which searches for a white pixel using only the four pixels above, below, left, and right of the pixel of interest, than for binary image 400, which searches for a white pixel using the eight pixels surrounding the pixel of interest.As a result, the positive contour image R has fewer contour pixels than the inspection image 500.

[0074] The authenticity determination device 3 then compares the contour portion of the inspection image 500 with the contour portion of the correct contour image R, and calculates a determination value to be used for authenticity determination based on the difference between these contour portions (S410).

[0075] 14(a), in this embodiment, rectangular regions circumscribing the contour pixels are extracted as character regions t and r for each of the inspection image 500 and the positive contour image R. The character regions t and r are regions whose horizontal range extends between the leftmost and rightmost contour pixels in the image (inspection image 500, positive contour image R) and whose vertical range extends between the topmost and bottommost contour pixels in the image.

[0076] As shown in Fig. 14(b), the authenticity determination device 3 calculates a vector v(xr, yr) from the origin o(Xr0, Yr0) of the character region r for the contour pixels of the character region r in the positive contour image R. Then, as shown in Fig. 14(c), it detects a corresponding pixel 502 at the end of the vector v whose starting point is the origin 501(Xt0, Yt0) of the character region t in the inspection image 500.

[0077] If the corresponding pixel 502 is not a contour pixel, pixels are searched for in a certain range of the character region t centered on the corresponding pixel 502 (hereinafter sometimes referred to as the search range) in order from the vicinity of the corresponding pixel 502 toward the outside. The specific search order is not particularly limited. Then, the difference in position between the first detected contour pixel and the corresponding pixel 502 is recorded in the storage unit 32 as a difference vector Vec(x,y). If the corresponding pixel 502 is a contour pixel, (0,0) is recorded as the difference vector Vec(x,y). Note that the origin o(Xr0,Yr0) and the origin 501(Xt0,Yt0) are pixels in the upper left corners of the character regions r and t, respectively, but are not limited to this.

[0078] The authenticity determination device 3 calculates the difference vector Vec(x,y) for each contour pixel in the character region r of the positive contour image R using the above procedure, calculates the sum ΣVec(x,y), and calculates the average value avgVec(x,y) of Vec(x,y) using the following equation (3): N is the total number of difference vectors Vec, i.e., the number of contour pixels in the character region r. avgVec(x,y)=(1 / N)ΣVec(x,y)…(3)

[0079] If the absolute values of both elements x and y of avgVec are less than 0.5, the authenticity determination device 3 terminates the process, and determines the judgment value Result as the average value of the lengths of the difference vector Vec(x, y) when the avgVec was obtained multiplied by 100, and calculates the judgment value Result using the following equation (4). Result=(100 / N)Σ(Vec(x) 2 +Vec(y) 2 ) 1 / 2 …(4)

[0080] On the other hand, if the absolute value of either or both of the x and y elements of avgVec is 0.5 or greater, the authenticity determination device 3 corrects (offsets) the origin 501 of the inspection image 500 by the vector IntavgVec(x,y), which is obtained by rounding off each x and y element of avgVec(x,y) to an integer, as shown in FIG. 14(d), and recalculates the difference vector Vec(x,y) for each contour pixel in the character region r of the positive contour image R using the procedure described above. Then, the determination value Result is calculated from these difference vectors Vec(x,y) using equation (4). Note that, although equation (4) multiplies the average length of the difference vector Vec(x,y) by 100, a value without multiplication by 100 can also be used as the determination value Result.

[0081] The processes of S405 to S410 are performed on the three grayscale images 20' (see FIG. 10(b)) obtained by the resolution conversion of S404, and the authenticity determination device 3 determines the smallest value of the determination values Result obtained for each grayscale image 20' as the final determination value, and performs authenticity determination using the determination value Result (S411). The procedure for authenticity determination will be described later.

[0082] The basic calculation procedure for the judgment value Result is as described above. However, in S410, it is also desirable to correct the vector v and the position (Xt0, Yt0) of the origin 501, taking into account the difference in size between the character region r in the positive contour image R and the character region t in the inspection image 500.

[0083] For example, to correct the vector v, the difference in size between the two character regions r and t is calculated as the horizontal magnification magX and the vertical magnification magY using the following equations (5) and (6), and the one of magX and magY closest to 1 is determined as the comparison magnification mag. magX = (horizontal size of character area t) / (horizontal size of character area r) ... (5) magY = (vertical size of character area t) / (vertical size of character area r) ... (6)

[0084] 15, vector v(xr, yr) used to detect corresponding pixel 502 within character region t of inspection image 500 is corrected as vector v·mag by multiplying each element xr, yr by comparison magnification mag. Note that each element (xr·mag, yr·mag) of vector v·mag is converted to an integer by rounding or the like. The position of origin 501 of character region t of inspection image 500 can also be corrected taking into account the above-mentioned comparison magnification mag. For example, the position of origin 501 is offset so that it approaches or moves away from the center of character region t depending on the value of comparison magnification mag.

[0085] Additionally, with regard to the search for pixels in the character region t (see FIG. 14(c)), the number of times each contour pixel in the character region t has been detected (the number of times the difference vector Vec has been calculated) may be recorded, and the search may be performed under the condition that the pixel being searched for is a contour pixel and that the number of times the pixel has been detected is equal to or less than a predetermined value. For example, if the predetermined value is 2, a pixel that has been detected 3 or more times is not detected as a contour pixel, even if it is a contour pixel, and the next pixel is searched for in the search order described above.

[0086] Furthermore, if the contour pixel for which the difference vector Vec is to be calculated is not detected even after the entire search range has been searched, the difference vector Vec(x,y) can be recorded as (8,8) if the number of times the contour pixel was not detected, including the search, is even, or (-8,-8) if the number of times is odd. By recording values with the same absolute value and opposite sign according to whether the number is even or odd, the influence on the calculation of the sum ΣVec(x,y) of the difference vectors Vec can be reduced. Note that the absolute value can be determined taking into account the size of the search range, etc., and is not limited to 8 as described above. The size and shape of the search range are also not limited. For example, the search range may be a square or a circular area.

[0087] (3-2. Procedure for authenticity determination using the judgment value Result) As described above, in S411 (see FIG. 8), authenticity is determined using the determination value Result. This determination value Result is the final value obtained from the grayscale image 20 created in S401, but in this embodiment, the R image and B image created in S401 are also used for determination as needed, and the procedure for doing so will be described below.

[0088] 16 is a flowchart showing the procedure for determining authenticity in S411. In this embodiment, the determination value Result is compared with a predetermined reference value, and if the determination value Result is smaller than the reference value (S4110; YES), the ID card 10 is determined to be authentic (S4111). The reference value can be determined appropriately taking into consideration the determination accuracy, etc.

[0089] On the other hand, if the determination value Resut is equal to or greater than the reference value (S4110; NO), the ID card 10 is not determined to be authentic. However, in this embodiment, as shown in Fig. 17, taking into consideration the possibility that the characters in the binary image 400 are missing significantly due to the hologram being reflected in the captured image of the ID card 10, causing an erroneous calculation of the determination value Result, if the determination value Result is equal to or greater than a certain value (S4112; YES), the processes of S402 to S410 are performed on image B to calculate the determination value Result (S4113).

[0090] This is because, unlike grayscale image 20, which is a grayscale version of the RGB components of the captured image of ID card 10, the color of the hologram does not appear when viewing the individual color components of the captured image of ID card 10, particularly the R and B components, and the effect on the text may be minimal. For example, if the hologram color in the captured image is red, viewing the B component of the captured image will have little effect on the text. The above fixed value is set to a value greater than the reference value, and the specific value can be determined as appropriate taking into account factors such as judgment accuracy.

[0091] If the judgment value Result obtained by S4113 is not greater than the above-mentioned certain value (S4114; NO), the authenticity determination device 3 compares the judgment value Result with a predetermined reference value, and if the judgment value Result is smaller than the reference value (S4115; YES), it determines that the ID card 10 is authentic (S4116).

[0092] If the determination value Result is smaller than a certain value in S4112 (S4112; NO) or if the determination value Result is equal to or greater than the reference value in S4115 (S4115; NO), the ID card 10 is determined to be false (S4117).

[0093] On the other hand, if the judgment value Result is equal to or greater than a certain value in S4114 (S4114; YES), the processes of S402 to S410 are further performed on the R image to calculate the judgment value Result (S4118), and the smallest of the judgment values Result obtained from the grayscale image 20, the B image, and the R image is determined to be the final judgment value (S4119).

[0094] If the determination value Result is smaller than the reference value (S4115; YES), the ID card 10 is determined to be true (S4116), and if the determination value Result is equal to or greater than the reference value (S4115; NO), the ID card 10 is determined to be false (S4117).

[0095] As described above, in this embodiment, the font of the characters formed on the ID card 10 is predetermined, and authenticity is determined by comparing the outline of the character to be inspected formed in a predetermined area on the ID card 10 with the outline of the correct character, thereby making it possible to suitably distinguish between a genuine ID card 10 in which the characters to be inspected are formed in a predetermined font and a counterfeit ID card 10 in which the characters to be inspected are formed in a font other than the predetermined font. The authentication method of this embodiment is particularly effective for ID cards 10 that are difficult to simply copy because they have anti-counterfeiting measures such as holograms, and can suitably distinguish counterfeit ID cards 10 that have been altered, such as by creating the card face from scratch.

[0096] In this embodiment, the characters to be inspected are characters in the pre-printed portion. For characters in the pre-printed portion, the range in which each character will be printed is predetermined, so the correct characters to be compared are known in advance, and only a small number of correct data items are required. Furthermore, by comparing the shapes, a counterfeit ID card 10 can be detected not only when the font is different, but also when the characters themselves are different.

[0097] In addition, in this embodiment, the difference vector Vec calculated by the method described in Figure 14, etc., can be used to appropriately evaluate the difference between the outline of the character being inspected and the outline of the correct character, allowing for appropriate authenticity determination.

[0098] 14 and other figures, the method of calculating the difference vector Vec searches for pixels in the character region t to be inspected that correspond to the contour pixels of the positive character, so the number of pixel searches corresponds to the number of contour pixels of the positive character, and the more contour pixels there are in the character to be inspected, the sooner a search will be completed. Therefore, as explained in FIG. 13 and other figures, by using strict criteria to reduce the number of contour pixels of the positive character and using looser criteria to increase the number of contour pixels of the character to be inspected, it is possible to quickly determine authenticity and prevent a genuine ID card 10 from being mistakenly determined to be fake.

[0099] However, the present invention is not limited to the above-described embodiment. For example, in the present embodiment, authenticity determination is performed as identity verification when opening an account, but the application scene or purpose of the authenticity determination of the present embodiment is not particularly limited.

[0100] In this embodiment, the characters to be inspected are characters in the pre-printed portion, but they may also be characters in on-demand information. In this case, a known OCR (Optical Character Recognition) process is performed on a predetermined area of the captured image containing the characters to be inspected to recognize the characters, and the positive data corresponding to the characters is used to determine the authenticity.

[0101] However, for on-demand information, the amount of positive data that needs to be prepared is large because the characters formed within a predetermined range of the captured image differ for each individual ID card 10. However, as for on-demand information, there are ranges that consist only of numbers 0 to 9, such as the date of birth, and ranges that consist only of letters A to Z, so when characters within these ranges are used to determine authenticity, a relatively small amount of positive data is sufficient.

[0102] Furthermore, in this embodiment, when binarizing the character image 300 in S407 (see FIG. 8), the final threshold value is determined from among a plurality of threshold values by comparing the number of black pixels at the time of binarization with the number of black pixels in the positive data. However, for each binary image that has been binarized using all threshold values or a plurality of threshold values among which the number of black pixels at the time of binarization is close to the number of black pixels in the positive data, the above-described procedure may be followed up to the calculation of the judgment value Result using equation (4), and the smallest judgment value Result may be determined as the final threshold value.

[0103] In addition, although the present embodiment tests for one character, the ID card 10 may be ultimately determined to be genuine if the ID card 10 is determined to be genuine for all characters in the above-described authenticity determination procedure, or if the number of characters that determine the ID card 10 to be genuine is a predetermined number or more.

[0104] While the preferred embodiments of the present invention have been described above with reference to the accompanying drawings, the present invention is not limited to these examples. It is clear that those skilled in the art can conceive of various modifications or alterations within the scope of the technical ideas disclosed herein, and it is understood that these modifications also fall within the technical scope of the present invention. [Explanation of symbols]

[0105] 1: Authenticity determination system 3: Authenticity determination device 5: User terminal 10: ID card 301: Receiving method 302: Authenticity determination means

Claims

1. A device for determining the authenticity of a personal identification card, an authenticity determination means for determining the authenticity of the identification card by using a photographed image of the identification card and comparing the outline of a character to be inspected formed in a predetermined area of the identification card with the outline of a genuine character; The authenticity determining means In the image of the character region of the positive character, a vector from the origin to the contour pixel is calculated; In an image of a character region of the character to be inspected, a corresponding pixel at the end of the vector starting from the origin is detected; If the corresponding pixel is not a contour pixel, a difference between the position of the contour pixel detected by searching for pixels in a range outside the corresponding pixel and the position of the corresponding pixel is calculated as a difference vector; The authenticity of the personal identification card is determined based on a determination value based on the difference vector. An authenticity determination device characterized by:

2. 2. The authenticity determination device according to claim 1, wherein the characters to be inspected are characters on a pre-printed portion of the personal identification card.

3. 2. The authenticity determination device according to claim 1, wherein the predetermined range is a range in which numbers or alphabets are formed as on-demand information of the personal identification card.

4. The contour pixels of the character to be inspected are In a binary image in which the character has a first gradation value and the background has a second gradation value, if the pixel of interest has the first gradation value and at least one of the eight pixels surrounding the pixel of interest has the second gradation value, the pixel of interest is set to the second gradation value, and in other cases, the pixel of interest is set to the first gradation value, thereby obtaining a pixel of the second gradation value; The positive character contour pixels are An authenticity determination device according to any one of claims 1 to 3, characterized in that in a binary image in which the character is a first gradation value and the background is a second gradation value, if the pixel of interest has the first gradation value and at least one of the four pixels above, below, left, and right of the pixel of interest has the second gradation value, the pixel of interest is set to the second gradation value, and in all other cases, the pixel of interest is set to the first gradation value, thereby obtaining a pixel of the second gradation value.

5. 5. The authenticity determining device according to claim 1, further comprising a receiving means for receiving the photographed image from a user terminal.

6. Computer, A device for determining the authenticity of a personal identification card, A program for causing the device to function as an authenticity determination device having an authenticity determination means for determining the authenticity of the identification card by using a photographed image of the identification card and comparing the outline of a character to be inspected formed in a predetermined area of the identification card with the outline of a correct character, The authenticity determining means In the image of the character region of the positive character, a vector from the origin to the contour pixel is calculated; In an image of a character region of the character to be inspected, a corresponding pixel at the end of the vector starting from the origin is detected; If the corresponding pixel is not a contour pixel, a difference between the position of the contour pixel detected by searching for pixels in a range outside the corresponding pixel and the position of the corresponding pixel is calculated as a difference vector; The authenticity of the personal identification card is determined based on a determination value based on the difference vector. A program characterized by:

Images