Determination system, determination method and program

Abstract

To provide a determination system capable of improving accuracy of thickness determination of a card-like object.SOLUTION: A determination system includes first recognition means, image rotation means, thickness determination means, second recognition means, and collation means. The first recognition means recognizes first information from a first image obtained by imaging the front face of a card-like object from the front side. The image rotation means rotates a second image obtained by imaging the front face from an oblique direction so that the front face is turned over and the lower surface of the object is exposed, by 180 degrees. The thickness determination means detects the length in the width direction of the lower surface from the second image rotated by 180 degrees, and determines whether or not the thickness of the object is appropriate. The second recognition means recognizes the first information from the second image rotated by 180 degrees. The collation means collates the first information recognized by the first recognition means with the first information recognized by the second recognition means. The determination system determines that the detected length is the length in the width direction of the lower surface, when the first information is in agreement.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to a determination system, a determination method, and a program.

Background Art

[0002] In recent years, systems that capture personal identification documents such as driver's licenses and My Number cards using smartphones or tablet PCs (Personal Computers) with a camera function and import them as images have been widely used.

[0003] Also, recently, due to the Anti-Money Laundering Act, it has been required to confirm the thickness of the personal identification document in the captured image of the personal identification document. For example, Patent Document 1 discloses an example in which a driver's license is imaged obliquely so as to appear trapezoidal so that the thickness of the lower surface (the surface sandwiched between the two lower sides of the front and back of the document surface; also referred to as the bottom surface) can be seen.

[0004] Furthermore, when imaging a personal identification document obliquely so as to appear trapezoidal, if the personal identification document is placed on a table or the like, a shadow is formed near the lower surface, and in order to prevent misdetection by taking the length including the width of the shadow as the thickness of the lower surface, a method of imaging while tilting the document so that the upper surface (the surface opposite to the lower surface; also referred to as the top surface) is exposed (so that it cannot be placed on a table or the like) is also conceivable.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, depending on the identity verification document, when imaging is performed while tilting the document so that it appears trapezoidal and the thickness of the upper surface can be seen, the information described in the identity verification document to be extracted and recognized from the captured image is located far from the two lines representing the upper side, resulting in problems such as a decrease in recognition accuracy.

[0007] Embodiments of the present invention provide a determination system, a determination method, and a program that can improve the accuracy of determining the thickness of a card-shaped object.

Means for Solving the Problems

[0008] According to the embodiment, the determination system includes a first recognition unit, an image rotation unit, a thickness determination unit, a second recognition unit, and a verification unit. The first recognition unit recognizes first information from a first image obtained by imaging the front side of a coupon surface on which first information of a card-shaped object is described. The image rotation unit rotates a second image obtained by imaging the coupon surface obliquely so that the coupon surface is inverted vertically and the lower surface of the object is exposed, by 180 degrees. The thickness determination unit detects the length in the width direction of the lower surface from the second image rotated by 180 degrees by the image rotation unit, and determines whether the thickness of the object is appropriate. The second recognition unit recognizes the first information from the second image rotated by 180 degrees by the image rotation unit. The verification unit verifies the first information recognized by the first recognition unit and the first information recognized by the second recognition unit. When the verification unit confirms the match of the first information, the determination system determines that the length detected by the thickness determination unit is the length in the width direction of the lower surface.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Embodiments for Carrying Out the Invention

[0010] Hereinafter, embodiments will be described with reference to the drawings.

[0011] (First Embodiment) First, the first embodiment will be described.

[0012] FIG. 1 is a diagram showing a configuration example of the personal identification system 1 of the first embodiment.

[0013] The personal identification system 1 is a system that captures a document for personal identification, such as a driver's license or a My Number card, as an image. The personal identification system 1 can be realized by an electronic device such as a smartphone or a tablet PC.

[0014] The personal identification system 1 includes a camera 11 and a personal identification document thickness determination unit (determination system) 12.

[0015] The camera 11 is a device that forms an image of light from a subject input to a lens as a digital image using an imaging element such as a CMOS image sensor. The personal identification system 1 captures an image of a document for personal identification with the camera 11. The image of the captured document for personal identification is input to the personal identification document thickness determination unit 12.

[0016] The document thickness determination unit 12 for identity verification is a module that determines whether the thickness of the identity verification document in the input image is appropriate. The document thickness determination unit 12 for identity verification is realized by a processor executing a program in which a procedure for determining whether the thickness of the identity verification document in the image is appropriate is described. Note that the document thickness determination unit 12 for identity verification may be realized as hardware such as an electric circuit or the like.

[0017] FIG. 2 is a diagram showing a configuration example of the document thickness determination unit 12 for identity verification.

[0018] The document thickness determination unit 12 for identity verification includes an edge detection unit [1] 21, a character recognition unit [1] 22, an image 180-degree rotation unit 23, an edge detection unit [2] 24, a thickness determination unit 25, a character recognition unit [2] 26, and a collation unit 27. Here, for the sake of easy understanding of the explanation, the edge detection unit [1] 21 and the edge detection unit [2] 24 are shown as different modules, but the edge detection unit [1] 21 and the edge detection unit [2] 24 may be the same module that operates as the edge detection unit [1] 21 at a certain timing and operates as the edge detection unit [2] 24 at another timing. Similarly, the character recognition unit [1] 22 and the character recognition unit [2] 26 may be the same module.

[0019] The document thickness determination unit 12 for identity verification in the first embodiment inputs two images: a front image (first image) and a thickness determination image (second image). In other words, in the identity verification system 1 of the first embodiment, the imaging of the identity verification document by the camera 11 is performed twice: the imaging of the front image and the imaging of the thickness determination image. The front image is an image obtained by imaging the document surface on which the recognition target characters (first information) of the identity verification document are described from the front. On the other hand, the thickness determination image is an image obtained by imaging the document surface obliquely so that the document surface is inverted vertically and the lower surface of the identity verification document is exposed. The point of imaging the thickness determination image so that the document surface is inverted vertically is an original idea unique to the identity verification system 1 of the first embodiment, and the details of this point will be described later.

[0020] FIG. 3 is a diagram showing an example of a front image of an identity document and an image for thickness determination in the identity verification system 1 of the first embodiment. Here, an example is shown in which a My Number Card is used as the identity document. FIG. 3(a) is the front image, and FIG. 3(b) is the image for thickness determination. When the My Number Card is used as the identity document, the recognition target characters to be described later are, for example, four-digit numbers (in FIG. 3, "0458") near the lower side of the ticket surface.

[0021] Regarding the imaging of the image for thickness determination shown in FIG. 3(b), in which the ticket surface is upside down and the lower surface of the identity document is exposed, for example, as shown in FIG. 4, a guide frame a1 with an animation representing the state of holding the identity document by hand can be superimposed on the live view image displayed when the camera 11 is activated on a smartphone or a tablet PC, etc., to support it.

[0022] Here, as a comparative example, with reference to FIGS. 5 to 9, points that were problematic when capturing the identity document as an image will be described.

[0023] FIG. 5 is a diagram showing an example of a front image of a driver's license (identity document) and an image for thickness determination in a comparative example. FIG. 5(a) is the front image, and FIG. 5(b) is the image for thickness determination. As shown in FIG. 5(b), in the comparative example, the image for thickness determination has the ticket surface not upside down.

[0024] Using an image of the identity document captured in a trapezoidal shape as shown in FIG. 5(b), by using an existing edge detection method, two lines representing the lower side are found from the image, and the thickness (the length in the width direction of the lower surface) of the captured identity document is measured by measuring the interval between them. However, if an attempt is made to determine the thickness by this method, two lines included in a pattern in the background, for example, which have nothing to do with the lower side, may be detected from the image, and correct determination may not be possible.

[0025] Therefore, based on the two detected lines, for example, the driver's license number (b2) can be cut out and recognized, and then compared with the driver's license number (b1) included in the front image to confirm whether the lower side has been correctly found. That is, the characters (in the case of a driver's license, the driver's license number) used to check whether the thickness determination is correctly performed are the characters to be recognized. Note that the extraction of the driver's license number (b1) from the front image is performed, for example, based on the four lines representing the four sides of the driver's license.

[0026] Figure 6 is a diagram showing an example of the front image and the image for thickness determination of the My Number card (identity document) in the comparative example. Fig. 6(a) is the front image, and Fig. 6(b) is the image for thickness determination. Also in the case of the My Number card, as in the case of the driver's license, as shown in Fig. 6(b), the image for thickness determination in the comparative example has the ticket surface not inverted vertically.

[0027] In the case of the My Number card, for example, by recognizing and comparing the four-digit numbers (c1, c2) described in the lower left of the ticket surface, it is possible to confirm whether the lower side has been correctly found. That is, in the case of the My Number card, the four-digit numbers (c1, c2) are the characters to be authenticated.

[0028] By the way, when performing the thickness determination of the lower side, if the identity document is placed on a table or the like and imaged, a shadow may be formed around the lower side, and there is a possibility that the thickness may be determined to be thicker than the original value. Figure 7 is a diagram showing an example of an image of the identity document imaged with a shadow formed around the lower side of the identity document placed on a table or the like.

[0029] FIG. 7 shows an example in which a health insurance card is used as an identity document. In the example of FIG. 7, a shadow 150 is formed near the lower edge of the health insurance card 100 on the table 200 on which the health insurance card 100 is placed. In this case, although the length between the two lower edges of the front (front and back) of the health insurance card 100 should be calculated as the thickness, there was a possibility that the length obtained by adding the width of the shadow 150 was calculated as the thickness. As a result, a health insurance card that is considerably thinner than a driver's license is imaged with a shadow and has the same thickness as a driver's license.

[0030] To solve this problem, it is also conceivable to adopt a method of imaging the upper edge instead of the lower edge. FIG. 8 is a diagram showing an example of a front image of a driver's license (identity document) and a thickness determination image captured to expose the upper surface in a comparative example.

[0031] FIG. 8(a) is a front image. And FIG. 8(b) is a thickness determination image captured to measure the thickness of the upper surface of the driver's license. When imaging is performed so as to know the thickness of the upper surface as shown in FIG. 8(b), since the driver's license cannot be placed on the table for imaging, a case where a shadow is formed can be avoided. Therefore, there is an advantage that the thickness can be measured more accurately.

[0032] Further, FIG. 9 is a diagram showing an example of a front image of a My Number card (identity document) and a thickness determination image captured to expose the upper surface in a comparative example. Also in the case of the My Number card, as in the case of the driver's license, when imaging is performed so as to know the thickness of the upper surface in FIG. 9(b), since the driver's license cannot be placed on the table for imaging, a case where a shadow is formed can be avoided.

[0033] However, in this method, there is no problem in the case of a driver's license, but in the case of a My Number card, since the four lower left digits, which are the characters to be recognized, are very far from the upper edge, the focus is blurred and the accuracy of character recognition decreases, making it difficult to use for verification.

[0034] On the one hand, the date of birth near the upper edge of the My Number Card has small characters, which also causes a problem of reduced recognition accuracy. Although it is conceivable to use the name for verification, since the name is in Chinese characters, there is also a problem that the recognition accuracy is lower compared to numbers.

[0035] For example, during the imaging of the images in FIG. 8(b) and FIG. 9(b), if a mechanism is provided such that the shutter is released when the verification is successful, and the user moves the smartphone little by little towards the identification document until the shutter is released for the purpose of identity verification, there is a problem that if the recognition accuracy of the characters to be recognized decreases, the phenomenon that the shutter does not release easily occurs.

[0036] Based on the above, returning to FIG. 2, the description of the configuration of the identification document thickness determination unit 12 of the identity verification system 1 of the first embodiment will be continued.

[0037] As described above, a front image such as FIG. 3(a) and a thickness determination image such as FIG. 3(b) are input to the identification document thickness determination unit 12.

[0038] The edge detection unit [1] 21 detects four lines representing the four sides of the identification document from the input front image. The character recognition unit [1] 22 cuts out the characters to be recognized from the front image based on the four lines detected by the edge detection unit [1] 21 and performs character recognition. As described above, the identification document in FIG. 3 is a My Number Card, and the character recognition unit [1] 22 cuts out and recognizes, for example, four digits ("0458") near the lower edge of the coupon surface.

[0039] On the other hand, when the thickness determination image is input to the identification document thickness determination unit 12, it is supplied to the image 180-degree rotation unit 23, which is a component unique to the identification document thickness determination unit 12 of the first embodiment provided in front of the edge detection unit [2] 24. The image 180-degree rotation unit 23 vertically flips the input thickness determination image. The edge detection unit [2] 24 detects two lines representing the lower side of the identification document from the thickness determination image vertically flipped by the image 180-degree rotation unit 23.

[0040] The thickness determination unit 25 calculates the length between the two lines detected by the edge detection unit [2] 24 and determines whether it is equal to or greater than a threshold value.

[0041] Also, the character recognition unit [2] 26 also cuts out the character to be recognized from the thickness determination image based on the two lines detected by the edge detection unit [2] 24 and performs character recognition. In the thickness determination image inverted vertically by the image 180-degree rotation unit 23, the four-digit number that is the character to be recognized is located near the lower surface (lower side) exposed for thickness calculation. Therefore, compared with the case of a general thickness determination image in which the four-digit number that is the character to be recognized is located far from the upper surface (upper side) exposed for thickness calculation shown in FIG. 9(b), the accuracy of character recognition of the character to be recognized is improved.

[0042] The collation unit 27 collates the character to be recognized on the identity document in the front image character-recognized by the character recognition unit [1] 22 with the character to be recognized on the identity document in the thickness determination image character-recognized by the character recognition unit [2] 26. When the two character recognition results match, the collation unit 27 determines that the thickness determination of the identity document by the thickness determination unit 25 is based on the two lines representing the lower side correctly detected from the identity document by the edge detection unit [2] 24.

[0043] That is, the identity document thickness determination unit 12 of the first embodiment is equipped with a function (image 180-degree rotation unit 23) for rotating the image by 180 degrees, so that it is possible to adopt an imaging method in which the lower side of the My Number Card is imaged facing upward. As a result, imaging while placed on a table can be almost prevented, and thickness determination can be performed with high accuracy. Also, since there are numbers with high character recognition accuracy near the lower side shown in the upper part of the image in FIG. 3(b), by recognizing these characters, it has the feature that collation with the character recognition result of a photo taken from the front with high accuracy can be performed. Of course, for a driver's license, it may also be imaged in such a way that the lower side is facing upward so that the thickness of the lower surface can be known. Since the driver's license number is also shown larger, the collation accuracy can be improved.

[0044] In addition, in such an identity verification procedure, there may be cases of sophisticated fraud where the identity verification procedure is bypassed by presenting pre-prepared identity verification documents or still images of the face. There may also be cases of fraud where the machine automatically presents an image to bypass identity verification. In such cases, when capturing the images in FIGS. 3(a) and 3(b), instead of capturing still images, capturing them as a series of moving images makes it possible to determine that a human is actually operating rather than showing a pre-prepared photo. Also, a method of capturing photos or moving images during the process of turning the document over may be adopted.

[0045] FIG. 10 is a flowchart showing the processing flow of the identity document thickness determination unit 12 of the identity verification system 1 according to the first embodiment.

[0046] The identity document thickness determination unit 12 first acquires a front image (S101). The edge detection unit [1] 21 detects four lines (edges) representing the four sides of the identity document from the front image (S102). The character recognition unit [1] 22 cuts out and recognizes the characters to be recognized on the identity document from the front image based on the edges detected by the edge detection unit [1] 21 (S103).

[0047] Subsequently, the identity document thickness determination unit 12 acquires a thickness determination image (S104). The image 180-degree rotation unit 23 flips the thickness determination image vertically (S105). The edge detection unit [2] 24 detects two lines (edges) representing the lower side of the identity document from the vertically flipped thickness determination image (S106). The thickness determination unit 25 determines the thickness of the identity document in the thickness determination image based on the edges detected by the edge detection unit [2] 24 (S107).

[0048] The identity document thickness determination unit 12 determines whether the thickness of the identity document determined by the thickness determination unit 25 is equal to or greater than the threshold value (S108). If it is less than the threshold value (S108: NO), the identity document thickness determination unit 12 abnormally terminates the process of determining the thickness of the identity document.

[0049] When the thickness is equal to or greater than the threshold value (S108: YES), next, the character recognition unit [2] 26 cuts out and recognizes the character to be recognized on the personal identification document from the thickness determination image based on the edges detected by the edge detection unit [2] 24 (S109). The collation unit 27 collates the characters recognized from the front image with the characters recognized from the thickness determination image (S110). When the two characters do not match (S111: NO), the personal identification document thickness determination unit 12 determines that the determination result of the thickness determination unit 25 is not based on the two lines (edges) representing the lower side of the personal identification document, and abnormally terminates the process of determining the thickness of the personal identification document. When the two characters match (S111: YES), the personal identification document thickness determination unit 12 determines that the determination result of the thickness determination unit 25 is based on the two lines (edges) representing the lower side of the personal identification document and is correctly determined, and normally terminates the process of determining the thickness of the personal identification document.

[0050] As described above, the personal identification document thickness determination unit (determination system) 12 of the first embodiment has a unique configuration (image 180-degree rotation unit 23) of inverting the thickness determination image vertically, so that the accuracy of determining the thickness of the personal identification document can be improved.

[0051] (Second Embodiment) Next, the second embodiment will be described.

[0052] FIG. 11 is a diagram showing a configuration example of the personal identification document thickness determination unit 12-2 in the personal identification system 1 of the second embodiment.

[0053] The personal identification document thickness determination unit 12-2 of the second embodiment has a face image extraction unit [1] 28 and a face image extraction unit [2] 29 instead of the character recognition unit [1] 22 and the character recognition unit [2] 26 of the personal identification document thickness determination unit 12 of the first embodiment. Further, instead of the image 180-degree rotation unit 23, a face trapezoid correction unit 30 is installed after the face image extraction unit [2] 29 and before the collation unit 27. That is, it is installed so as to be interposed between the face image extraction unit [2] 29 and the collation unit 27. Note that the face image extraction unit [1] 28 and the face image extraction unit [2] 29 may be the same module that operates as the face image extraction unit [1] 28 at a certain timing and operates as the face image extraction unit [2] 29 at another timing.

[0054] In the personal identification system 1 of the second embodiment, as in the comparative example, a front image and a thickness determination image are captured as shown in FIGS. 5(a) and 5(b).

[0055] In the personal identification system 1 of the first embodiment, the recognition target characters of the personal identification document in the front image and the thickness determination image are character-recognized, and the characters recognized from the front image are collated with the characters recognized from the thickness determination image. In contrast, in the personal identification system 1 of the second embodiment, a face collation technique is used to collate the face image of the personal identification document in the front image with the face image of the personal identification document in the thickness determination image.

[0056] The face image extraction unit [1] 28 extracts a face image from the front image based on four lines representing the four sides of the personal identification document detected by the edge detection unit [1] 21. The face image extraction unit [2] 28 extracts a face image from the thickness determination image based on two lines representing the lower side of the personal identification document detected by the edge detection unit [2] 24. The thickness determination image is an image captured of the personal identification document obliquely rather than from the front. Therefore, the face image extracted from the thickness determination image is also an image captured obliquely. The face trapezoid correction unit 30 corrects the face image extracted by the face image extraction unit [2] 28 into an image captured from the front.

[0057] In the verification unit 27 of the personal identification document thickness determination unit 12-2 of the second embodiment, the face image extracted from the front image by the face image extraction unit [1] 28 is compared with the face image after the face trapezoid correction unit 30 corrects the thickness determination image by the face image extraction unit [2] 29. When the two face images match, the verification unit 27 determines that the thickness determination of the personal identification document by the thickness determination unit 25 is based on the two lines representing the lower side correctly detected from the personal identification document by the edge detection unit [2] 24.

[0058] Since the personal identification document thickness determination unit 12 of the first embodiment has a face trapezoid correction function (face trapezoid correction unit 30) and a face verification function (verification unit 27), face verification can be used for the verification of the front image and the thickness determination image. Since face verification can be performed at a relatively low image resolution compared to character recognition, the thickness determination can be performed with higher accuracy.

[0059] FIG. 12 is a flowchart showing the processing flow of the personal identification document thickness determination unit 12-2 of the personal identification system 1 of the second embodiment.

[0060] The personal identification document thickness determination unit 12-2 first acquires a front image (S201). The edge detection unit [1] 21 detects four lines (edges) representing the four sides of the personal identification document from the front image (S202). The face image extraction unit [1] 28 extracts the face image on the personal identification document from the front image based on the edges detected by the edge detection unit [1] 21 (S203).

[0061] Also, the personal identification document thickness determination unit 12-2 acquires a thickness determination image (S204). The edge detection unit [2] 24 detects two lines (edges) representing the lower side of the personal identification document from the thickness determination image (S205). The thickness determination unit 25 determines the thickness of the personal identification document in the thickness determination image based on the edges detected by the edge detection unit [2] 24 (S206).

[0062] The document thickness determination unit 12-2 for identity verification determines whether the thickness of the identity verification document determined by the thickness determination unit 25 is equal to or greater than the threshold value (S207). If it is less than the threshold value (S207: NO), the document thickness determination unit 12-2 for identity verification abnormally terminates the process of determining the thickness of the identity verification document.

[0063] If the thickness is equal to or greater than the threshold value (S108: YES), then the face image extraction unit [2] 29 extracts a face image on the identity verification document from the thickness determination image based on the edges detected by the edge detection unit [2] 24 (S208). The face trapezoid correction unit 30 corrects the face image extracted by the face image extraction unit [2] 29 into an image captured from the front (S209).

[0064] The matching unit 27 matches the face image extracted from the front image with the face image extracted from the thickness determination image (S210). If the two face images do not match (S211: NO), the document thickness determination unit 12-2 for identity verification determines that the determination result of the thickness determination unit 25 is not based on the two lines (edges) representing the lower side of the identity verification document, and abnormally terminates the process of determining the thickness of the identity verification document. If the two face images match (S211: YES), the document thickness determination unit 12-2 for identity verification determines that the determination result of the thickness determination unit 25 is based on the two lines (edges) representing the lower side of the identity verification document and is correctly determined, and normally terminates the process of determining the thickness of the identity verification document.

[0065] As described above, the document thickness determination unit (determination system) 12-2 for identity verification in the second embodiment can improve the accuracy of determining the thickness of the identity verification document by performing matching using the face image of the identity verification document.

[0066] (Third Embodiment) Next, the third embodiment will be described.

[0067] FIG. 13 is a diagram showing a configuration example of the document thickness determination unit 12-3 in the identity verification system 1 of the third embodiment.

[0068] The document thickness determination unit 12-2 of the third embodiment has an imaging orientation random selection unit 31 instead of the image 180-degree rotation unit 23 of the document thickness determination unit 12 of the first embodiment.

[0069] When the personal identification system 1 captures an image for thickness determination using the camera 11, the imaging orientation random selection unit 31 randomly selects one of, for example, four types of orientations each time to determine the orientation in which the personal identification document should be imaged. The personal identification system 1 captures an image for thickness determination using the camera 11 with the personal identification document in the orientation selected by the imaging orientation random selection unit 31. The orientation selected by the imaging orientation random selection unit 31 is also notified to the character recognition unit [2]26.

[0070] FIG. 14 is a diagram illustrating four types of imaging orientations of the personal identification document that can be selected by the imaging orientation random selection unit 31. In FIG. 14, the imaging orientation of the personal identification document is indicated by a guide frame displayed to support the capture of an image for thickness determination.

[0071] FIG. 14(a) shows an example of the upper side upward orientation in which the personal identification document is imaged obliquely from the side so that the upper surface is exposed without inverting the document vertically. FIG. 14(b) shows an example of the lower side upward orientation in which the personal identification document is imaged obliquely from the side with the document inverted vertically so that the lower surface is exposed. FIG. 14(c) shows an example of the lower side downward orientation in which the personal identification document is imaged obliquely from the side without inverting the document vertically so that the lower surface is exposed. FIG. 14(d) shows an example of the upper side downward orientation in which the personal identification document is imaged obliquely from the side with the document inverted vertically so that the upper surface is exposed.

[0072] In the personal identification system 1 of the third embodiment, according to these guide displays, the user is asked to capture an image of the personal identification document. Since it is difficult for a machine to take a photo according to a randomly selected guide, by providing a function (imaging orientation random selection unit 31) to randomly select a guide for capturing an image for thickness determination, it is possible to distinguish whether a human is actually operating or a machine is operating.

[0073] FIG. 15 is a flowchart showing the processing flow of the personal identification document thickness determination unit 12-3 of the personal identification system 1 according to the third embodiment.

[0074] The personal identification document thickness determination unit 12-3 first acquires a front image (S301). The edge detection unit [1] 21 detects four lines (edges) representing the four sides of the personal identification document from the front image (S302). Further, the character recognition unit [1] 22 cuts out and recognizes the characters to be recognized on the personal identification document from the front image based on the edges detected by the edge detection unit [1] 21 (S303).

[0075] Next, the imaging orientation random selection unit 31 selects the imaging orientation of the thickness determination image (S304). Thereafter, the personal identification document thickness determination unit 12-32 acquires the thickness determination image captured in the orientation selected by the imaging orientation random selection unit 31 (S305). The edge detection unit [2] 24 detects two lines (edges) representing the upper side or the lower side of the personal identification document from the thickness determination image (S306). The thickness determination unit 25 determines the thickness of the personal identification document in the thickness determination image based on the edges detected by the edge detection unit [2] 24 (S307).

[0076] The personal identification document thickness determination unit 12-3 determines whether the thickness of the personal identification document determined by the thickness determination unit 25 is equal to or greater than the threshold value (S308). If it is less than the threshold value (S308: NO), the personal identification document thickness determination unit 12-3 abnormally terminates the process of determining the thickness of the personal identification document.

[0077] When the thickness is equal to or greater than the threshold value (S308: YES), this time, the character recognition unit [2] 26 cuts out and recognizes the character to be recognized on the personal identification document from the thickness determination image based on the edges detected by the edge detection unit [2] 24 and the imaging orientation randomly selected by the imaging orientation random selection unit 31 (S309). The collation unit 27 collates the characters recognized from the front image with the characters recognized from the thickness determination image (S310). When the two characters do not match (S311: NO), the personal identification document thickness determination unit 12 determines that the determination result of the thickness determination unit 25 is not based on the two lines (edges) representing the upper or lower side of the personal identification document, and abnormally terminates the process of determining the thickness of the personal identification document. When the two characters match (S311: NO), the personal identification document thickness determination unit 12 determines that the determination result of the thickness determination unit 25 is based on the two lines (edges) representing the upper or lower side of the personal identification document and is correctly determined, and normally terminates the process of determining the thickness of the personal identification document.

[0078] As described above, the personal identification document thickness determination unit (determination system) 12-3 of the third embodiment has a unique configuration (imaging orientation random selection unit 31) that randomly selects the imaging orientation of the thickness determination image, thereby eliminating machine-based fraud and improving the accuracy of determining the thickness of the personal identification document.

[0079] (Fourth Embodiment) Next, the fourth embodiment will be described.

[0080] In the fourth embodiment, it is assumed that the my number card is used as the personal identification document. And in the personal identification system 1 of the fourth embodiment, when imaging the thickness determination image, the left side of the my number card is photographed with the bottom or top facing up.

[0081] FIG. 16 is a diagram showing an example of a thickness determination image of the my number card in the personal identification system 1 of the fourth embodiment.

[0082] FIG. 16(a) is an example of the left side facing down of the My Number Card. FIG. 16(b) is an example of the left side facing up of the My Number Card.

[0083] Since the four-digit number and the photo of the My Number Card are both near the left side, both the character recognition of the four-digit number and the face verification can be performed with high accuracy. Note that the personal identification system 1 of the fourth embodiment can be realized by utilizing a part of the configuration of the personal identification document thickness determination unit 12 of the first embodiment or the configuration of the personal identification document thickness determination unit 12-2 of the second embodiment.

[0084] Also, since the vertical width of the My Number Card is smaller than the horizontal width, when photographing in the orientation shown in FIG. 16, the thickness can be relatively enlarged in the photograph, so the thickness can be measured relatively accurately. In addition, since the screen of a smartphone is often a vertical screen, it is also suitable for taking photos with a smartphone.

[0085] As described above, the personal identification system 1 of the fourth embodiment can improve the accuracy of determining the thickness of the personal identification document.

[0086] (Fifth Embodiment) Next, the fifth embodiment will be described.

[0087] In the fifth embodiment, it is assumed that the My Number Card is used as a personal identification document. And in the personal identification system 1 of the fifth embodiment, when imaging the thickness determination image, the back surface of the My Number Card is imaged with its top and bottom reversed.

[0088] FIG. 17 is a diagram showing an example of the thickness determination image of the My Number Card in the personal identification system 1 of the fifth embodiment.

[0089] On the back of the My Number Card, there is a QR code (registered trademark) and a check digit. Therefore, by searching for and recognizing the QR code based on the side where the thickness was determined, and comparing the recognition result with an image taken of the back of the QR code from the front, verification can be performed with extremely high accuracy. Note that the personal identification system 1 of the fifth embodiment can be realized by utilizing a part of the configuration of the personal identification document thickness determination unit 12 of the first embodiment.

[0090] Since the QR code is also located near the lower side, the recognition accuracy can be improved by imaging it with the image flipped vertically. Also, since the My Number included in the QR code has a check digit, even without performing collation with the image taken from the front, if the check digit of the My Number is the correct value, it can be said that the side being focused on is indeed the lower side of the My Number.

[0091] As described above, the personal identification system 1 of the fifth embodiment can improve the accuracy of thickness determination of the personal identification document.

[0092] (Sixth Embodiment) Next, the sixth embodiment will be described.

[0093] In the sixth embodiment, when imaging the thickness determination image, the vicinity of the upper right corner or the lower right corner of the personal identification document is photographed.

[0094] FIG. 18 is a diagram showing an example of the thickness determination image of the personal identification document in the personal identification system 1 of the sixth embodiment.

[0095] FIG. 18(a) is an example of imaging the lower left corner of the personal identification document. FIG. 18(b) is an example of imaging the upper left corner of the personal identification document.

[0096] When capturing an identity document as shown in FIGS. 18(a) and 18(b), compared with the case of capturing the identity document so as to include the entire ticket surface, the identity document can be captured larger, and both the left side and the bottom surface (FIG. 18(a)) or the left side and the top surface (FIG. 18(b)) can be exposed, and either can be used to perform thickness determination. Therefore, both thickness determination and verification can be made more accurate.

[0097] Also, in the case of the identity verification system 1 of the sixth embodiment, when capturing an image for thickness determination, as shown in FIGS. 19(a) to 19(c), the guide display may be displayed while randomly changing the oblique angle little by little. In this case, since the oblique angle can be randomly changed little by little to instruct the user to align with the frame, it becomes easier to distinguish between the machine and the human. Also, since a face photo can be captured, it can also be used for face verification purposes.

[0098] As described above, the identity verification system 1 of the sixth embodiment can improve the accuracy of thickness determination of the identity document.

[0099] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0100] 1... Identity verification system, 11... Camera, 12... Identity document thickness determination unit, 21... Edge detection unit [1], 22... Character recognition unit [1], 23... Image 180-degree rotation unit, 24... Edge detection unit [2], 25... Thickness determination unit, 26... Character recognition unit [2], 27... Verification unit, 28... Face image extraction unit [1], 29... Face image extraction unit [2], 30... Face trapezoid correction unit, 31... Imaging orientation random selection unit.

Claims

1. First recognition means for recognizing the first information from a first image obtained by imaging, from the front, a face of a card-shaped object on which the first information is described; Image rotation means for rotating by 180 degrees a second image obtained by imaging, obliquely from the side, the face so that the face is inverted vertically and the lower surface of the object is exposed; Thickness determination means for detecting the length in the width direction of the lower surface from the second image rotated by 180 degrees by the image rotation means and determining whether the thickness of the object is appropriate; Second recognition means for recognizing the first information from the second image rotated by 180 degrees by the image rotation means; Verification means for verifying the first information recognized by the first recognition means and the first information recognized by the second recognition means; Comprising: A determination system that, when the verification means confirms the coincidence of the first information, determines that the length detected by the thickness determination means is the length in the width direction of the lower surface.

2. First recognition means for extracting a face image from a first image obtained by imaging, from the front, a face of a card-shaped object on which the face image is described; Thickness determination means for detecting the length in the width direction of the upper surface from a second image obtained by imaging, obliquely from the side, the face so that the upper surface of the object is exposed and determining whether the thickness of the object is appropriate; Second recognition means for extracting the face image from the second image; Face image correction means for correcting the face image extracted by the second recognition means into a face image obtained by imaging from the front; Verification means for verifying the face image recognized by the first recognition means and the face image extracted by the second recognition means and corrected by the face image correction means; Comprising: A determination system that, when the verification means confirms the coincidence of the face images, determines that the length detected by the thickness determination means is the length in the width direction of the upper surface.

3. First recognition means for recognizing the first information from a first image obtained by imaging, from the front, a face of a card-shaped object on which the first information is described; Selection means for randomly selecting the orientation of the object at the time of imaging a second image to be imaged obliquely from the side; Thickness determination means for detecting the length in the width direction of the upper surface or the lower surface of the object from the second image imaged in the orientation selected by the selection means and determining whether the thickness of the object is appropriate; Second recognition means for recognizing the first information from the second image imaged in the orientation selected by the selection means; A collating means for collating the first information recognized by the first recognition means with the first information recognized by the second recognition means; comprising; A determination system that, when the collating means confirms the coincidence of the first information, determines that the length detected by the thickness determination means is the length in the width direction of the upper surface or the lower surface.

4. The selection means a first orientation in which the coupon surface is imaged so that the coupon surface is not inverted vertically and the upper surface is exposed; a second orientation in which the coupon surface is imaged so that the coupon surface is inverted vertically and the lower surface is exposed; a third orientation in which the coupon surface is imaged so that the coupon surface is not inverted vertically and the lower surface is exposed; a fourth orientation in which the coupon surface is imaged so that the coupon surface is inverted vertically and the upper surface is exposed; The determination system according to claim 3, which randomly selects one from at least two of the four orientations.

5. A first recognition means for recognizing the first information from a first image obtained by imaging a coupon surface on which the first information of a horizontally long card-shaped object is described from the front; a thickness determination means for detecting the length in the width direction of one side surface from a second image obtained by obliquely imaging the coupon surface so that one side surface of the object is exposed with the width direction facing the vertical direction, and determining whether the thickness of the object is appropriate; a second recognition means for extracting the first image from the second image; a collating means for collating the first information recognized by the first recognition means with the first information recognized by the second recognition means; comprising; A determination system that, when the collating means confirms the coincidence of the first information, determines that the length detected by the thickness determination means is the length in the width direction of one side surface.

6. The determination system according to claim 5, wherein the first information is a number or a face image.

7. A first recognition means for recognizing the first information from a first image obtained by imaging a first coupon surface on which the first information of a card-shaped object is described from the front; a thickness determination means for detecting the length in the width direction of the lower surface from a second image obtained by imaging the second coupon surface on which a two-dimensional code including a check digit corresponding to the first information and facing the first coupon surface is inverted vertically and the lower surface of the object is exposed, and determining whether the thickness of the object is appropriate; a second recognition means for recognizing the two-dimensional code from the second image; Verification means for verifying whether the check digit included in the two-dimensional code recognized by the second recognition means is the correct value corresponding to the first information recognized by the first recognition means; comprising; A determination system that, when the validity of the check digit is confirmed by the verification means, determines that the length detected by the thickness determination means is the length in the width direction of the lower surface. **Claim 8** First recognition means for recognizing the first information from a first image obtained by imaging a coupon surface on which the first information of a card-shaped object is described from the front; Thickness determination means for detecting the length in the width direction of at least one of the two surfaces from a second image obtained by partially imaging a corner of the coupon surface obliquely so that the two surfaces including the area where the first information is described and one side surface orthogonal to the upper surface of the object or the lower surface of the object and the one side surface are exposed, and determining whether the thickness of the object is appropriate; Second recognition means for recognizing the first information from the second image; Verification means for verifying the first information recognized by the first recognition means and the first information recognized by the second recognition means; comprising; A determination system that, when the coincidence of the first information is confirmed by the verification means, determines that the length detected by the thickness determination means is the length in the width direction of the lower surface. **Claim 9** The determination system according to claim 8, further comprising selection means for randomly selecting an angle for tilting the object when imaging the second image. **Claim 10** A determination method executed by an electronic device capable of imaging or inputting an image, comprising: recognizing the first information from a first image obtained by imaging a coupon surface on which the first information of a card-shaped object is described from the front; rotating a second image obtained by imaging the coupon surface obliquely so that the coupon surface is inverted vertically and the lower surface of the object is exposed, by 180 degrees; detecting the length in the width direction of the lower surface from the rotated second image by 180 degrees, and determining whether the thickness of the object is appropriate; recognizing the first information from the rotated second image by 180 degrees; verifying the first information recognized from the first image and the first information recognized from the second image; and determining that the detected length is the length in the width direction of the lower surface when the coincidence of the first information is confirmed. A determination method comprising the steps. **Claim 11** A computer, First recognition means for recognizing the first information from a first image obtained by imaging the face of a card-shaped object on which the first information is described from the front; Image rotation means for rotating a second image obtained by imaging the face obliquely so that the face is inverted vertically and the lower surface of the object is exposed, by 180 degrees; Thickness determination means for detecting the length in the width direction of the lower surface from the second image rotated by 180 degrees by the image rotation means, and determining whether the thickness of the object is appropriate; Second recognition means for recognizing the first information from the second image rotated by 180 degrees by the image rotation means; Verification means for verifying the first information recognized by the first recognition means and the first information recognized by the second recognition means; Functioning as; A program for determining that the length detected by the thickness determination means is the length in the width direction of the lower surface when the coincidence of the first information is confirmed by the verification means.

Images