Identity verification documents, methods for authenticating such documents, and methods for manufacturing such documents

Abstract

This document provides identification documents, as well as methods for authenticating and manufacturing such documents. [Solution] A method for manufacturing an identity verification document (1), comprising the steps of: printing a background pattern (10a, 10b) on the identity verification document, the background pattern including a motif that provides contrast to at least partially cover a text field (Tf1) of the identity verification document (1) (S1); printing identification information, which includes one or more characters, on the text field (Tf1) (S2); selecting a pattern portion (11a, 11b) of an image of the identity verification document (1), the pattern portion including at least one character of the identification information and a portion of the background pattern that covers at least one character (S3); encoding the selected pattern portion (11a, 12a) into a machine-readable code (S4); and printing the machine-readable code (13a) on the identity verification document (1) (S5).

Description

Technical Field

[0001] The present invention relates to a document used for identifying an individual or a product.

Background Art

[0002] Document authentication technology is important in ensuring the validity and integrity of documents across various fields, including law, finance, and the political sector. These technologies encompass various methods designed to verify the authenticity of both physical and digital documents and protect them from forgery, tampering, and fraud. They not only contribute to the protection of confidential information but also promote credibility and reliability in document transactions and communications.

[0003] As a conventional method of performing such authentication, for example, registering security features known from the design layout of a document, such as a watermark, hologram, or special ink, may be done.

[0004] Such approaches have been used for a long time to ensure the security of physical documents. They are often general-purpose for a given document type, meaning that any authentication given to a certain type of document will test the same features. Thus, if a fraudster creates a fraudulent presentation from a known genuine sample, the fake document may inherit many - all - of the features used in these authentication operations.

[0005] On the other hand, with the advent of digital technology, new solutions such as digital signatures, blockchain, and biometrics have been introduced. However, these solutions also have drawbacks such as complex and expensive implementation, vulnerability to cyberattacks, or privacy concerns. Furthermore, integrating new digital authentication technologies with existing systems can be highly difficult. Considerable effort and resources are required to ensure compatibility and interoperability between different platforms and devices.

[0006] Therefore, there is a need for authentication technologies for existing documents that enhance security and fraud prevention. [Overview of the project] [Means for solving the problem]

[0007] This invention aims to improve the reliability of fraud detection in secure credential authentication systems using computer vision analysis.

[0008] Therefore, a method for manufacturing identity verification documents, - A background pattern which includes a motif that creates a contrast that at least partially covers the text fields of the identification document, is printed on the identification document. - The step of printing identification information, which includes one or more characters, into a text field. - A step of selecting a pattern portion of an image of an identity verification document, which includes at least one character of the identification information and a portion of the background pattern that covers this at least one character. - The step of encoding the selected pattern portion into a machine-readable code, - The step of printing a machine-readable code on the identification document, A method including this is provided.

[0009] This identification can be used to securely authenticate individuals or objects. In fact, a machine-readable code is printed during the manufacturing of identity verification documents. The printed machine-readable code encodes the pattern portion corresponding to an image containing characters and an identifying background pattern. During document authentication, if the pattern portion of a document does not match the encoded pattern portion, the document may be rejected. This further increases the likelihood of detecting altered or forged documents. In fact, if the identification information in a text field is changed, the characters and background pattern should be different from those in the encoded pattern portion, resulting in a different image when authenticating the document. This solution can also be added to existing document designs to enhance the security of fraud detection.

[0010] Other preferred embodiments of the present invention, though not limited to these, are as follows, either individually or in technically feasible combinations: - The background pattern completely covers the text field. - The background pattern is designed to produce predictable binary results when the threshold determination method proposed by Nobuyuki Otsu is applied. - The contrasting motifs in the background pattern do not periodically cover the text field. - At least one character is the last character of the identification information, and the selected pattern portion includes a portion of the text field that follows the last character. - Encoding the selected pattern portion involves binarizing the selected pattern portion using a binarization threshold. - Encoding the selected pattern portion includes applying a compression algorithm to the selected pattern portion. - The method further includes the step of encoding metadata information, which includes at least one of the location information of a selected pattern portion in the identification document and the security threshold for authentication of the identification document, into a machine-readable code before printing the machine-readable code on the identification document. - The method further includes the step of signing a machine-readable code using cryptography, and then printing the signed machine-readable code onto an identification document. - The machine-readable code will be printed on the identification document at an image quality between 200 DPI and 400 DPI. - The machine-readable code is printed, at least partially, with ultraviolet ink, preferably on the ghost portrait of the identification document. - The machine-readable code is a QR code.

[0011] The present invention also relates to an identity verification document which includes at least a text field on which identification information, comprising one or more characters, is printed, and a background pattern which includes a contrasting motif that at least partially covers the text field, and an identity verification document which includes a machine-readable code which encodes a pattern portion of the image of the identity verification document containing at least one character of the identification information, and a portion of the background pattern that covers this at least one character, preferably overlapping with the at least one character.

[0012] Identity verification documents include, for example, an ID card, a driver's license, or a passport.

[0013] Identity verification documents are, for example, documents that identify a product or animal.

[0014] In another embodiment, a method for authenticating the aforementioned identity verification documents, - The step of obtaining authentication information corresponding to the text field portion of the identity verification document by decrypting the machine-readable code, - A step in which input information is extracted from an image of the identity verification document, and the extracted input information includes the text field portion of the identity verification document. - A step to compare the extracted input information with the authentication information and obtain a similarity match score as a result of the comparison, - providing an authentication result based on a similarity match score; A method including the above has been proposed.

[0015] Typically, when the similarity match score is greater than a security threshold, the authentication result is positive, and when the similarity match score is lower than the security threshold, the authentication result is negative.

[0016] Non-limiting, other suitable embodiments of the authentication method are as follows, either alone or in technically feasible combinations. - The method further includes obtaining location information by reading a machine-readable code, and the input information is extracted based on this location information. - The authentication information is a grayscale image, and the method further includes converting the extracted input information into a grayscale image before comparison. - The authentication information is a black-and-white image, and the method further includes obtaining a binarization threshold by reading a machine-readable code, and converting the extracted input information into a black-and-white image based on the binarization threshold before the comparison step. - The similarity match score is one of cross-correlation, correlation coefficient, and sum of squared differences. - The machine-readable code is encrypted using a digital signature, and the method further includes authenticating the machine-readable code by verifying the digital signature before comparison. If the digital signature is not verified, the authentication result is negative.

[0017] In another aspect, an authentication system including a processor configured to perform the above-described authentication method has been proposed.

[0018] When executed by the aforementioned authentication system, a computer program including instructions to cause the authentication system to perform the authentication method described herein is also contemplated.

Brief Description of the Drawings

[0019] <Represents a common personal identification document of the card type. [Figure 2] It is a flowchart showing a method for manufacturing a personal identification document according to one aspect of the present invention. [Figure 3] It is a schematic diagram of the personal identification document of FIG. 1 after the background pattern is printed in two embodiments. [Figure 4] These are two examples of the background pattern. [Figure 5] It is a schematic diagram of the personal identification document of FIG. 3, and the pattern portion is selected. [Figure 6] It represents the selected pattern portion after binarization of one of the personal identification documents of FIG. 5. [Figure 7] In one embodiment, it is a schematic diagram encoding the selected pattern portion. [Figure 8] It represents the personal identification document of FIG. 3 after the machine-readable code is printed. [Figure 9] It is a flowchart showing a method for authenticating a personal identification document according to another aspect of the present invention. [Figure 10] In another embodiment, it is a flowchart showing a method for authenticating a personal identification document. [Figure 11] It is a schematic diagram of the forged personal identification document of FIG. 8. [Figure 12] It is a schematic diagram authenticating the forged personal identification document of FIG. 11 using the authentication information of the machine-readable code.

Mode for Carrying Out the Invention

[0020] According to the first aspect, a personal identification document is described. Hereinafter, the personal identification document may sometimes be simply referred to as a "document". The personal identification document can be any personal identification document such as an identity card, a driver's license, or a passport. Such a personal identification document can be used daily to confirm the identity of an individual called the owner.

[0021] <000,0125> Figure 1 illustrates a typical national ID card. The format of identification documents is not limited to cards and could potentially extend to passports or other documents attached to personal identification information. While this specification describes personal identification documents, more generally, identification documents can be used to identify products, animals, or any other objects.

[0022] In the traditional method, Identity Document 1 includes a photo field Pf1 where the owner's photograph can be printed. Identity Document 1 also includes various text fields where general information can be printed. "General information" is understood to be information that is not related to the owner, or is not directly and explicitly related to the owner. For example, the text field Atf at the top of Identity Document 1 prints the name of the competent authority or agency that serves the Identity Document. This text field cannot be individually changed in Identity Document 1 within documents of the same category. For example, the text field Tf3 may print other general information, such as a postal code or the date of issue or expiration of the document. This information can be identical for each Identity Document for all owners living in the same area, for example.

[0023] Identity Verification Document 1 includes several text fields Tf1, Tf2 on which identification information is printed or will be printed. This identification information may be personal identification information (PII), which is any information about the owner maintained by an institution and can be used to identify or track the owner's identity, such as name, social security number, date and place of birth, address, etc. The length of the identification information may vary and may contain multiple words separated by spaces. The text may include uppercase or lowercase letters and numbers.

[0024] Identity verification documents may be issued by government agencies or by non-governmental organizations such as contracting agencies of government agencies. The production of identity verification documents is typically done using a printer. The printer is equipped with a processor configured to receive information from an interface. The printer may have memory for storing information or instructions. The memory can be accessed by the processor. The interface typically includes a scanner for scanning the printed identity verification document and a screen for displaying the information. The interface can be used to input information and may include, for example, a keyboard.

[0025] A method for manufacturing secure identification documents is described according to the first embodiment illustrated in Figure 2.

[0026] First, a blank identification document is obtained. During step S1 of the manufacturing method, the printer prints background patterns 10a and 10b onto identification document 1. During step S2, the printer prints identification information into the text fields of identification document 1. In the illustrated example, the text field Tf1 corresponds to the full name of the document owner, so the identification information is "John Doe". The identification information includes five lowercase letters and two uppercase letters.

[0027] Alternatively, or in combination, the identification information may correspond to any other information that distinguishes the owner of a document and includes at least one character, such as a last name, first name, document number, address, date of birth, or password. Therefore, the content of the corresponding text fields will differ significantly among owners.

[0028] The text fields in question may or may not be the same across multiple identification documents within the same category. The text fields in question may or may not be the same across identification documents across multiple categories.

[0029] Background patterns 10a and 10b contain a contrasting motif that at least partially covers the text field. "Covering" is understood to mean that the background pattern extends (at least partially) across the text field, typically over a rectangular box, but may also overlap or be superimposed on the printed characters in text field Tf1.

[0030] Background patterns 10a and 10b can be distinguished by the fact that they differ from the background of the remaining parts of identity verification documents 1a and 1b. A "contrasting motif" is understood to mean that the motif is not blurred and does not resemble the background of the remaining parts of identity verification documents 1a and 1b.

[0031] Background patterns 10a and 10b have high contrast. For example, the contrast ratio, which is the ratio between the maximum and minimum intensity, is better than 2:1. The pattern contrast ratio R is defined as follows: R=(max(L1,L2)+0.05) / (min(L1,L2)+0.05) In the formula, L1 and L2 are L channel values, i.e., luminance, obtained after converting a color RGB image to an HSL (hue-saturation-luminance) image for the foreground (1), e.g., a motif, and the background (2), which is the rest of the background pattern. Preferably, the pattern contrast ratio R is greater than 2.0:1, and more preferably greater than 3.0:1.

[0032] Other measures could include root mean square (RMS) contrast, i.e., the standard deviation of pixel intensity normalized by the mean intensity. In another example, the histogram of pixel intensity is not homogeneous, meaning that background patterns 10a and 10b contain pixels with high intensity variability.

[0033] Preferably, the background pattern is designed to produce a predictable binary result when the threshold selection method proposed by Nobuyuki Otsu in "A threshold selection method from gray-level histograms (Threshold selection method from density distribution). IEEE Trans.Sys.Man.Cyber.9(1):62-66." is applied. In the example illustrated in Figure 4, background pattern 10c is not considered to have high contrast. In fact, the boundaries of the motif are blurred, and the binarized pattern 12c is unpredictable. On the other hand, background pattern 10d is usable for its intended purpose. That is, the foreground and background pixels are easily distinguishable, and the binarized pattern 12d can be clearly determined.

[0034] Advantageously, contrasting motifs can be easily distinguished using copies with image quality below 400 DPI. This allows the authentication device to reliably match identity documents 1a and 1b with the mobile capture even if the mobile capture is of poor quality, for example, blurry.

[0035] The background pattern may be the same across multiple identification documents. The background pattern may differ between identification documents within the same category, or between identification documents spanning multiple categories.

[0036] For example, the exemplary document 1a at the top of Figure 3 has a background pattern 10a with a clear outline and higher contrast than the background pattern 10b of document 1b at the bottom of Figure 3.

[0037] Preferably, the background patterns 10a and 10b are not periodic; that is, contrasting motifs are not identically reproduced at different positions in the text field Tf1. This enhances the reliability of the authentication method for documents obtained by the proposed manufacturing method, as will be described later.

[0038] The contrast between the background patterns 10a and 10b and the printed characters must be sufficient to ensure that i) the identification information is legible to humans, and ii) that optical character recognition can be performed correctly on the images of the identification documents 1a and 1b. Typically, the identification information is printed on the background patterns 10a and 10b in a dark color near RGB=(0,0,0). Preferably, the printed characters of the identification information have a contrast-to-noise ratio with respect to the background patterns 10a and 10b greater than 4.5:1.

[0039] The pattern contrast ratio R between L1 corresponding to the characters and L2 corresponding to the background pattern, as defined above, is greater than 1.5:1, preferably greater than 2.0:1, and more preferably greater than 2.5:1.

[0040] Background patterns 10a and 10b can be printed before or after printing the identification information. Preferably, step S1 is performed before step S2, i.e., the identification background pattern is added behind the area of ​​the text field under consideration where the identification information may be printed. In alternative embodiments where background patterns 10a and 10b are printed on top of the identification information, background patterns 10a and 10b must be sufficiently transparent so that the printed identification information is legible to humans. In particular, the transparency must be sufficient so that the contrast-to-noise ratio of the printed characters compared to the printed background is greater than 4.5:1, or the pattern contrast ratio of the printed characters compared to the printed background is greater than 1.5:1, preferably greater than 2.0:1, and more preferably greater than 2.5:1.

[0041] Preferably, the background patterns 10a and 10b cover the entire area of ​​the text field Tf1. “Entire area” is understood to mean that the written identification information is contained within the background patterns 10a and 10b. In the illustrated embodiment, the background patterns 10a and 10b completely cover the text field Tf1 containing the full name. The shape of the background patterns 10a and 10b is rectangular, and their width and length are determined by the selected text field Tf1.

[0042] The printer interface acquires images of identity verification documents 1a and 1b with background patterns 10a and 10b and identification information printed on them. The images can be scanned images of identity verification documents 1a and 1b that include at least the target text field Tf1. Preferably, the images are acquired at a resolution of 400 DPI or higher.

[0043] With respect to Figure 5, during step S3, the processor selects a pattern portion of the image of the identification document 1. The selected pattern portions 11a and 11b include at least one character of the identification information and a portion of the background pattern covering this at least one character, that is, the background pattern can be superimposed either below or above the at least one character. In other words, the selected pattern portions 11a and 11b include both the printed character or number and several contrasting motifs surrounding the printed identification information, for example, motifs next to the character or above or below the character, depending on the size of the motifs and the line spacing.

[0044] Typically, the selected pattern portions 11a and 11b are rectangular patches that cover the last character of the identification information and a portion of the background pattern surrounding the last character, called trailing whitespace.

[0045] Alternatively, or in combination, the selected pattern portion includes randomly selected characters of the printed identification information and portions of the background patterns 10a, 10b above and below the selected characters. This is particularly suitable when there is sufficient line spacing and the contrasting motifs of the background pattern are sufficiently small, and the selected pattern portions 11a, 11b are guaranteed to contain contrasting motifs, for example, as evaluated by Otsu's threshold determination method described above. Preferably, the selected pattern portions 11a, 11b are contained within a text field covered by the background patterns 10a, 10b, meaning that they do not contain portions of the generic background of the identification documents 1a, 1b.

[0046] Alternatively, or in combination, the selected pattern portion includes two or more characters from the identification information. Preferably, the pattern portion 10a includes multiple characters from the identification information. The pattern portion 10a may include multiple adjacent characters, or it may consist of different portions containing non-adjacent characters. In particular, the number of characters in the selected pattern portion is determined by the available memory or data storage space. For example, it may store multiple characters from the entire identification information in addition to the last character and trailing whitespace.

[0047] In another embodiment, multiple background patterns are printed on multiple text fields, and a pattern portion is selected from each text field covered by each background pattern.

[0048] During step S4, the processor encodes the selected pattern portion 11a into machine-readable code. The selected pattern portion 11a can be encoded as a representation of itself, for example, as a pixel matrix of binarized values. Preferably, the selected pattern portion 11a is encoded as a feature vector. This alternative is known to be more color-friendly in one embodiment, as it does not require binarization and the selected pattern portion is not converted to grayscale, and the required memory space is reduced to well a few hundred bytes. Features can be extracted using various methods. For example, Scale-Invariant Feature Transform (SIFT), Speeded Up Robust Feature (SURF), or KAZE generate string-based descriptors, while Oriented FAST and Rotated BRIEF (ORB), binary robust invariant scalable keypoints (BRISK), or Accelerated KAZE (AKAZE) generate binary descriptors. Next, feature point matching can be performed, for example, using the L1 or L2 norm for string-based descriptors, or the Hamming distance for binary descriptors. Other methods of feature point matching include approximate nearest neighbor (FLANN) search, commonly known as the brute-force (BF) method.

[0049] Machine-readable codes may include barcodes, quick-response (QR) codes, or any other symbolic codes. Machine-readable codes can encode the personal identification information of the owner of the identification document 1a. Barcodes are one-dimensional, meaning that the information can only be scanned horizontally. QR codes are two-dimensional, meaning that the information can be read both horizontally and vertically, allowing for the storage of more data.

[0050] Typically, the pattern portion 11a is extracted from the image of the identification document 1a.

[0051] Optionally, with respect to Figure 6, the selected pattern portion 11a is binarized to obtain the binarized pattern portion 12a. "Binarization" means that after binarization, the pixels of the selected pattern portion 11a are set to two different values, e.g., 0 or 1. For example, if the selected pattern portion 11a is color, the selected pattern portion 11a is first converted to grayscale. Each pixel of the selected pattern portion 11a has a value between 0 (black) and 255 (white). Next, a binarization threshold is selected. During binarization, pixel values ​​above the binarization threshold are set to the first value (usually 255 or 1 for white), and values ​​below that are set to 0 (black). Using binarization makes it possible to save memory space for later encoding into machine-readable code. Preferably, the background pattern 10a is designed to produce a predictable binary result when Otsu's thresholding method is applied, as described above.

[0052] Preferably, the processor compresses the selected pattern portions 11a and 12a. This makes it possible to limit the memory required to store the selected pattern portions 11a and 12a. Preferably, the selected pattern portions 11a and 12a are highly compressed, i.e., the memory space required to store the compressed pattern portions 11a and 12a is less than one-tenth. Preferably, the memory space required to store the compressed pattern portions 11a and 12a is less than 1 kilobyte.

[0053] In one embodiment, the processor applies a lossy compression algorithm, such as JPEG compression, to the selected pattern portions 11a, 12a. However, because the selected pattern portions 11a, 12a contain sharp edges associated with text characters, the lossy compression algorithm may cause visual distortion known as ringing. Preferably, the processor applies a lossless compression algorithm, such as PNG compression, to the selected pattern portions 11a, 12a. A lossless compression algorithm is particularly advantageous when the background pattern 10a is selected so that a binary representation 12a of the text and background can be reliably created by simple preprocessing. For example, a high-contrast grayscale background pattern with a size of 400x100 can be stored in PNG format in 2446 bytes. Using 1-byte / pixel representation compression on the grayscale bitmap pattern portion results in 40,000 bytes. In this context, compression is necessary, for example, using a 1-bit / pixel binary representation supported by PNG to reduce the image to 1216 bytes. Another alternative is to crop these 1-byte / pixel representations to 100x100 and save them as binary PNGs, reducing the required memory space to 437 bytes.

[0054] Compression is particularly interesting in embodiments where multiple background patterns are used in different text fields. In fact, the storage space of machine-readable codes is limited. For example, a very high-density QR code has a storage capacity of up to 1700 bytes, but requires a dedicated document scanner to decrypt. On the other hand, a low-density QR code can be read by other media such as smartphones, but has a storage capacity of less than 1000 bytes.

[0055] Preferably, with respect to Figure 7, the processor 2 further encodes the metadata information Data1 and Data2 into a machine-readable code 13a. In other words, the payload of the machine-readable code 13a makes it possible to reconstruct the appropriate metadata and binarized pattern portion 12a encoded in the QR code 13a. For example, the metadata information Data1 and Data2 include a binarization threshold, an indicator that one or more text fields are selected, location information such as the index of each character selected in the identification information, the dimensions of the selected pattern portion 11a, or the position of the edges of the selected pattern portion. Advantageously, the metadata information can include location information. For example, location information is stored as a label or index of the text field in question, if a number is printed next to the text field in question and includes the start and end ranges of the characters. If the pattern portion contains trailing whitespace, an incremented end range (end range + 1) may correspond to the trailing whitespace. Alternatively, location information is stored as a bounding box definition in units of 400 dpi pixels that defines the region of interest, i.e., the outline of the pattern portion within the identification document.

[0056] As will be described later, metadata information may also include information useful for document authentication, such as recommended score thresholds or security thresholds, or recommended matching methods.

[0057] Preferably, the machine-readable code is digitally signed. As previously described, the selected pattern portion 11a can be binarized before the digital signature is created, or the selected pattern portion 11a can be binarized and further compressed before the digital signature is created.

[0058] Typically, a single digital signature is calculated for all the data to be stored, including, for example, multiple selectable pattern components and metadata as described above. The digital signature is then added to the payload before a machine-readable code is generated and printed on the identification document.

[0059] Adding a digitally signed signature to a machine-readable code enhances the security of identity verification document 1a. In fact, any attempt to forge a document by altering the machine-readable code will be detected. Only verification authorities with a secure key can accurately verify the information encoded in the machine-readable code by matching it with the signing key.

[0060] Regarding Figure 8, the printer then prints a machine-readable code on the identification document 1a. Here, the machine-readable code 13a is printed on the lower right of the identification document 1a.

[0061] The machine-readable code can be printed on any suitable location on the identification document 1a, including on various sides or pages of the document. Here, the machine-readable code is shown on the front side of document 1a, but if it is too large to be displayed on the front side of the document in card format, it may also be shown on the back. Alternatively, the printer may print the machine-readable code on a separate substrate and then securely attach it to the identification document 1a, for example, by adhesive.

[0062] Preferably, the machine-readable code is printed with UV ink, at least partially or entirely. This makes the machine-readable code less conspicuous on the identification document, making it difficult to forge. Thus, the security level is improved. While machine-readable codes printed with UV ink can be read by scanners with multispectral illumination, they may not be readable by mobile devices. It also becomes possible to print the machine-readable code in a larger format without masking the text fields of the document. Machine-readable codes printed with UV ink can be printed across the entire document. Larger codes are easier to read with scanners with multispectral illumination. It also suggests that altering part of the document may, in some cases, break the machine-readable code, leading to negative results during authentication. Instead, the machine-readable code is printed with UV ink as a ghost portrait of the identification document, i.e., a smaller version of the original photographic image on the document printed in semi-transparent ink.

[0063] Preferably, machine-readable codes are printed on identification documents at a resolution between 200 and 400 DPI. This resolution level is standard and can be manufactured using any printer for identification documents. Depending on the size of the compressed pattern payload, high-density machine-readable codes may be required, which may necessitate the highest resolution in this range or higher. For example, high-density machine-readable codes are printed at a resolution between 400 and 600 DPI. As mentioned above, high-density machine-readable codes offer greater storage capacity but require a specific scanner for reading.

[0064] As will be explained later, the selected image quality level is sufficient for certification.

[0065] In another embodiment, identity documents can be authenticated to identify the owner's identity. The authentication method is performed by an authentication system including a processor. The authentication system includes a memory that stores a security key used to digitally sign data, such as machine-readable codes or metadata information used to manufacture identity documents. The memory is accessible from the processor. The memory stores a computer program that includes instructions for performing the authentication method described herein.

[0066] For example, the authentication system could be, for instance, a scanning device in border security screening. Identification documents could be received by the scanning device before boarding an international flight. The scanning device could include a scanner having a photosensitive camera that scans the front or back of the identification document under various lighting conditions, such as visible light, infrared, and ultraviolet light. The photosensitive camera could include, for example, a charge-coupled element.

[0067] Alternatively, the authentication system may consist of a mobile device, such as a camera-equipped cell phone, owned by a security inspector or officer. Owner authentication may be performed to control access to buildings where security is guaranteed, or during routine tasks such as traffic control.

[0068] The identification documents obtained using the proposed manufacturing method can then be authenticated using the authentication methods illustrated in Figures 9 and 10.

[0069] The authentication system first obtains an image representing the identity verification document to be authenticated. The image of the identity verification document may include both the target text field and the machine-readable code. Alternatively, multiple images of the identity verification document may be obtained, for example, a first image containing the machine-readable code and a second image containing the target text field. Here, the input document image includes at least a QR code 13a printed on the document and a text field Tf1. The input document image can be obtained using a scanner or camera.

[0070] During step A1, the authentication system obtains authentication information corresponding to the text fields on the identification document by decrypting the machine-readable code. In other words, it can scan the machine-readable code to extract the encoded personal information of the owner of the identification document.

[0071] Preferably, during preliminary step A0, the processor authenticates the machine-readable code by verifying the digital signature used to encrypt the machine-readable code. This speeds up the authentication process by directly outputting a negative authentication result if the digital signature cannot be verified. The processor can verify the scanned machine-readable code using the encryption key stored in memory.

[0072] During step A2, the processor extracts input information from the image of the identification document. The extracted input information includes the text field portion of the identification document.

[0073] The extracted input information may correspond to a portion of the input document image. This portion typically corresponds to a slightly larger rectangle of the input document image and includes the last character of the target text field and trailing whitespace, including the background pattern. Preferably, the extracted input information also includes any additionally stored characters. For example, the processor can implement optical character recognition (OCR). After OCR processing of the input information, the positions of the last character of the identification information in the target text field and any additionally stored characters can be determined.

[0074] The authentication system's memory can store information about the location of the pattern portion to be authenticated. Alternatively, the authentication system can store this information in memory accessible to it. This information may include, for example, the dimensions of the target text field or the pattern portion used when manufacturing the identification document.

[0075] In addition, or instead, the information is embedded in a machine-readable code printed on the identification document for authentication. The information is then recovered by directly reading the machine-readable code. This is particularly advantageous in one embodiment where the target text field or selection pattern portion varies considerably depending on the category of the identification document.

[0076] During step A3, the authentication system's processor compares the extracted input information with the authentication information.

[0077] The comparison results in a similarity match score (Sms). The similarity match score (Sms) is calculated by the processor from the extracted input information and the authentication information obtained through template matching. For example, the similarity match score (Sms) is one of the following: correlation coefficient, sum of squared differences, or cross-correlation. Some template matching methods, such as the sum of squared differences, can inherently present numerically smaller scores the better the match. In such cases, it is preferable to convert the similarity match score from a distance score to a similarity score by, for example, using a normalized version of the scale to obtain a normalized difference score in the range [0.0, 1.0], and then calculating the difference between 1.0 and the obtained normalized difference score.

[0078] In embodiments where the features of the selected pattern portion are encoded into machine-readable code, instead of performing template matching, similar features can be extracted from the input image during step A3 and feature point matching can be performed.

[0079] Preferably, the processor implements a scaling algorithm before comparison to ensure that the extracted input information is scaled to the same effective DPI as the authentication information, i.e., the pattern portion encoded in machine-readable code. This makes it possible to improve the reliability of the similarity match score (SMS).

[0080] Preferably, if preprocessing, such as threshold determination for binarization, is applied to the text field portion during the extraction of input information, the same operation is applied to the input information.

[0081] Preferably, the similarity match score is calculated using grayscale or black and white, i.e., binarized images. This improves the reliability of the similarity match score, because varying lighting conditions in mobile capture environments can lead to inconsistent color representations of digital images of documents, potentially reducing the reliability of color pattern matching.

[0082] For example, in one embodiment where the authentication information is a grayscale image, the processor converts the extracted input information into a grayscale image before comparison.

[0083] For example, in one embodiment where the authentication information is a grayscale image, the authentication system first obtains a binarization threshold. The binarization threshold can be reconstructed by reading a machine-readable code or stored in the authentication system's memory. Next, the processor converts the extracted input information into a grayscale image based on the binarization threshold before comparison.

[0084] In one embodiment, where multiple background patterns are used in different text fields, and multiple pattern portions are selected and stored in machine-readable code, the previous step is repeated for each stored pattern portion. In other words, each stored pattern portion is individually matched with its corresponding search area.

[0085] Preferably, if the quality of the extracted input image falls below a quality threshold, the authentication system provides a warning to the interface indicating that the comparison cannot be performed and reports an inconclusive authentication. In this case, the operator running the authentication system can provide a new input image to extract higher quality input information.

[0086] During step A4, the authentication system provides authentication results based on the similarity match score (SMS).

[0087] For example, if the similarity match score Sms is greater than the security threshold St, the authentication result is positive, meaning the input document passes the authentication test; or, if the similarity match score Sms is less than the security threshold St, the authentication result is negative, meaning the input document is not authenticated. The security threshold can be stored in the authentication system's memory. Alternatively, the security threshold can be encoded within the machine-readable code as part of the metadata information. In this case, the security threshold is obtained following step A1 when the machine-readable code is read.

[0088] The threshold may vary depending on the card design. The threshold may be fixed or vary depending on the category of the identification document. Preferably, the security threshold may vary depending on the nature of the machine-readable code. For example, in one embodiment, where the machine-readable code is a 2D barcode with limited memory capacity, the authentication information has low resolution due to high compression before encoding. In this context, the security threshold will be lower than the security threshold that can be used in one embodiment where the machine-readable code is a QR code with higher memory capacity and better resolution authentication information.

[0089] In embodiments where multiple background patterns are used, preferably, in order to output a positive authentication result that constitutes successful authentication of the document, all similarity match scores must be above the corresponding security threshold.

[0090] As illustrated in Figures 11 and 12, one of the advantages of the proposed method is that, while it is based on a document design common to all owners, the selection pattern portion is unique to most owners based on the length of the owner's identification information and the characters selected in the target text field.

[0091] Figures 11 and 12 illustrate the reliability of the proposed authentication method when a fraudster attempts to obtain Document 1, which is obtained using a manufacturing method starting with the general document shown in Figure 1 and Document 1a, which is guaranteed to be secure in Figure 8.

[0092] To obtain the forged document 1c shown in Figure 11, the fraudster here erased the letters of the original name "John Doe" so they were illegible, and then replaced them with the longer name "Johnny Deer." The result is a visually plausible identification document 1c, which means that the forged document 1c may pass checks that rely solely on static patterns, and even tests based on data cross-checking if the fraudster also knows how to falsify other relevant data sources such as the machine-readable zone (MRZ).

[0093] Figure 12 illustrates the various steps of the authentication method in this case.

[0094] Processor 3 reads the QR code 13a to restore the binarized pattern portion 12a and metadata information Data1 and Data2. Preferably, the digital signature of the QR code 13a is verified beforehand. If the verification of the QR code 13a fails, the document 1c cannot be authenticated. An error or negative authentication result is output to the monitor. Otherwise, if the verification of the QR code 13a is successful, the restored initial selected pattern portion is considered secure. Advantageously, QR code authentication is performed before further processing. This buys time and eliminates the need to calculate a similarity match score.

[0095] Using metadata information Data1, Data2, and an image of the forged document 1c, processor 3 extracts the corresponding pattern portion and applies binarization to obtain input information 12c. Next, the restored pattern portion is template-matched with a selected portion of the input document image. Processor 3 compares the authentication information 12a corresponding to the initial binarized selected pattern portion with the input information 12c corresponding to the binarized selected pattern portion of the forged document 1c.

[0096] In the illustrated case, the new name is longer, so the newly selected pattern area contains different characters. Therefore, the position of the last character in the name text field has been changed. Using the proposed authentication method, the authentication result is negative because the similarity match score between the selected patterns is low. In fact, the two pattern parts do not match. Processor 3 outputs a negative authentication result to the monitor, and the forged document 1c is not authenticated.

[0097] Processor 3 uses OCR to detect the position of the last character, and in another case where the selected pattern portion includes the last character and trailing spaces, the overall length of the identification information changes, and therefore the background pattern is also different. As a result, the two pattern portions do not match, and the document is not authenticated.

[0098] This authentication method is particularly robust against changes in identifying information that shorten or lengthen the printed text of the target text field. In both cases, even if the new text ends with the same character as the original text, sampling at different locations within the target text field will capture different segments of the background pattern. Attempts to maintain the same length with a different number of characters using an alternative font or a different font size should also be detected using this approach, as the shape of the last character of the text field will no longer match what is stored in the initially selected pattern segment. [Explanation of Symbols]

[0099] 1. Identity verification documents 1a Identity verification documents 1b Identity verification documents 1c Forged document 2 processors 3 processors 10a Background Pattern 10b Background Pattern 10c Background Pattern 10d background pattern 11a Selection pattern section 11b Selection pattern section 11c Newly selected pattern region 12a Binarized pattern portion 12c Binarized pattern portion 12d binarized pattern 13a Machine-readable code Atf text field A0 Preliminary Step A1 Step A2 Step A3 Step A4 Step Data1 Metadata Information Data2 Metadata Information Pf1 Photo Field SMS Similarity Matching Score St Security Threshold S1 Step S2 Step S3 Step S4 Step S5 Step Tf1 Text Field Tf2 Text Field Tf3 Text Field

Claims

1. A method for manufacturing an identity verification document (1), - Step (S1) of printing a background pattern (10a, 10b) on the identity verification document, the background pattern including a motif that provides contrast to at least partially cover the text field (Tf1) of the identity verification document (1), - The step (S2) of printing the identification information, which includes one or more characters, into the text field (Tf1), - A step (S3) of selecting a pattern portion (11a, 11b) of the image of the identity verification document (1) which includes at least one character of the identification information and a portion of the background pattern that covers the at least one character, - A step (S4) of encoding the selected pattern portion (11a, 12a) into a machine-readable code, - Step (S5) of printing the machine-readable code (13a) onto the identity verification document (1), A method that includes this.

2. A method for manufacturing an identity verification document (1) according to claim 1, wherein the contrasting motifs of the background patterns (10a, 10b) do not periodically cover the text field (Tf1).

3. A method for producing an identity verification document (1) according to claim 1 or 2, wherein the at least one character is the last character of the identification information, and the selection pattern portion (11a, 11b) includes a portion of the text field (Tf1) that follows the last character.

4. A method for producing an identity verification document (1) according to claim 1 or 2, wherein the step of encoding the selected pattern portion (11a, 12a) includes binarizing the selected pattern portion (11a) using a binarization threshold.

5. A method for manufacturing an identity document (1) according to claim 1 or 2, further comprising the step of encoding metadata information, which includes at least one of the location information of the selected pattern portion in the identity document and the security threshold for authentication of the identity document, into the machine-readable code (13a) before printing the machine-readable code (13a) onto the identity document (1).

6. A method for manufacturing an identity document according to claim 1 or 2, further comprising the step of signing the machine-readable code (13a) using cryptography, and then printing the signed machine-readable code (13a) onto the identity document (1).

7. A method for manufacturing an identity verification document according to claim 1 or 2, wherein the machine-readable code (13a) is printed on the identity verification document (1) at an image quality between 200 DPI and 400 DPI.

8. A method for manufacturing the identity document (1) according to claim 1 or 2, wherein the machine-readable code (13a) is printed at least partially with ultraviolet ink, preferably on the ghost portrait of the identity document (1).

9. A method for manufacturing an identity verification document according to claim 1 or 2, wherein the machine-readable code (13a) is a QR code.

10. Identity verification document (1) comprising at least a text field (Tf1) on which identification information, comprising one or more characters, is printed, and a background pattern (10a, 10b) comprising a contrasting motif that at least partially covers the text field (Tf1), and an identity verification document (1) comprising a machine-readable code (13a) which encodes a pattern portion of the image of the identity verification document (1) comprising at least one character of the identification information, and a portion of the background pattern that covers the at least one character, is printed.

11. The identity verification document (1) according to claim 10, which is a personal identification document that is one of the following: an ID card, a driver's license, or a passport.

12. The identity verification document (1) according to claim 10, which is a document that certifies a product or animal.

13. A method for authenticating an identity verification document as described in any one of claims 10 to 12, - Step (A1) of obtaining authentication information corresponding to the text field portion of the identity verification document (1) by decrypting the machine-readable code (13a), - Step (A2) involves extracting input information from the image of the identity verification document (1), and ensuring that the extracted input information includes the text field portion of the identity verification document (1). - Step (A3) of comparing the extracted input information with the authentication information and obtaining a similarity match score (Sms) as a result of the comparison, - Step (A4) of providing authentication results based on the similarity matching score (Sms), A method that includes this.

14. The authentication method according to claim 13, further comprising obtaining location information by reading the machine-readable code (13a), and the input information being extracted based on the location information.

15. The authentication method according to claim 13, further comprising converting the extracted input information into a grayscale image before comparison, wherein the authentication information is a grayscale image.

16. An authentication system including a processor configured to perform the authentication method described in claim 13.

Images