DEVICE FOR PERSONALIZING A PERSON'S IDENTITY DOCUMENT

Description

TECHNICAL AREA

[0001] The present invention relates to the field of testing a reference facial image of a person, which is to be stored in an identity document of the person. TECHNICAL BACKGROUND

[0002] The publication "Anonymous: "Technical guideline for production data collection, quality control and transmission for passports", 2010 (2010-01-01), XP055676218, Found on the Internet: URL:https: / / www.bsi.bund.de / SharedDocs / Downloads / DE / BSI / Publikationen / TechnischeRichtlinien / TR03104 / BSI-TR_031_04_A1_V2_1_5.pdf;jsessionid=68F0241576AA1E810DB8C9097105B9D7.1_cid341?_blob=publicationFile&v=1 [accessed on 2020-03-12]" is a technical guideline for production data collection, quality control and transmission for passports.

[0003] The publication "Anonymous: "Technical Guideline 03146 ​​Electronic Image Transmission for Applying for Official Documents (E-Image hD)", October 1, 2014 (2014-10-01), XP055676232, Found on the Internet: URL:https: / / www.bsi.bund.de / SharedDocs / Downloads / DE / BSI / Publikationen / TechnischeRichtlinien / TR03146 / TR03146.pdf?_blob=publicationFile&v=1 [accessed on 2020-03-12]" is a technical guideline regarding electronic image transmission for applying for official documents.

[0004] The publication "SCHERHAG ULRICH ET AL: "Towards Detection of Morphed Face Images in Electronic Travel Documents", 2018 13TH IAPR INTERNATIONAL WORKSHOP ON DOCUMENT ANALYSIS SYSTEMS (DAS), IEEE, April 24, 2018 (2018-04-24), pages 187-192, DOI: 10.1109 / DAS.2018.11, [accessed on 2018-06-22]", discloses a method in which face morphing attacks are detected in the context of using an identity document.

[0005] When verifying a person's identity, a comparison is typically made between a reference facial image of the person, which is printed on the person's identity document or stored digitally in the identity document, and an actual face or facial image of the person.

[0006] Recent studies suggest that it may be possible to create a single digital reference facial image from digital reference images of individuals who appear nearly identical, using image manipulation. This image could then match all individuals during manual inspection by border control officers or forensic experts, as well as during automated inspection by access control systems. Such image manipulation could, for example, involve morphing, where an intermediate transition between multiple digital reference facial images is calculated. This would make it possible, for instance, to apply for an identity document that could be used by several people, as the single digital reference facial image would biometrically match all of them. This poses particular challenges to the reliability of automated access control systems.

[0007] A classic approach to solving this problem is to technically or organizationally prevent manipulated digital reference facial images from being accepted and incorporated into an identity document during the application process. However, in the case of passports as identity documents, this approach could only be implemented if all issuing countries or authorities worldwide adopted this procedure. In practice, this is difficult to ensure. Furthermore, an electronic signature attached to a data set by a manufacturer typically provides no indication of how the digital reference facial image was originally created or whether it has been subsequently manipulated.

[0008] The aforementioned challenges affect both countries or authorities where applicants provide reference facial images in the form of printed passport photos, which are then scanned, and countries or authorities where applicants receive both printed passport photos and references or links to digitally stored images from the photographer, which can then be retrieved from servers. DESCRIPTION OF THE INVENTION

[0009] The purpose of the present invention is to create an efficient concept for personalizing a person's identity document.

[0010] This problem is solved by the features of the independent claims and thus defined by the attached independent claims. Advantageous embodiments of the invention are the subject of the description, the drawings, and the dependent claims.

[0011] The present invention is based on the finding that the above problem can be solved by capturing a high-quality digital reference facial image of the person, in which a biometric reference feature is captured. This biometric reference feature is then compared with a biometric comparison reference feature to obtain a manipulation indicator, which represents the degree of image manipulation in the digital reference facial image. The manipulation indicator can, for example, represent a confidence level or probability that image manipulation has occurred. If no image manipulation is detected, the digital reference facial image of the person can be appropriately adapted for the identity document, for example, by converting it into a digital reference facial image with a lower reference image resolution.The digital reference facial image can then be stored in the identity document. The converted digital reference facial image can have a reference image resolution according to established standards, such as those of the ICAO.

[0012] Furthermore, additional approaches can be used to determine further indicators of manipulation, whereby the manipulation indicator and the other manipulation indicators can be evaluated together to detect image manipulation. This joint evaluation can be carried out, for example, using a statistical approach or a neural network.

[0013] By using a high-quality digital reference facial image for image manipulation detection, the efficiency and reliability of the detection process can be significantly increased. Furthermore, the security and reliability of access control systems that use identity documents can be enhanced. This increased efficiency and reliability of manipulation detection is also expected to result in fewer individuals submitting manipulated reference facial images.

[0014] According to a first aspect, the invention relates to a device for personalizing a person's identity document. The device comprises an image acquisition device configured to acquire a first digital reference facial image of the person, wherein the first digital reference facial image of the person has a first reference image resolution. The device further comprises a tamper detection device configured to acquire a biometric reference feature in the first digital reference facial image of the person, to compare the biometric reference feature with a biometric comparison reference feature in order to obtain a tamper indicator, and to detect image manipulation in the first digital reference facial image based on the tamper indicator.The device further comprises an image generation unit configured to generate a second digital reference facial image of the person based on the first digital reference facial image if no image manipulation is detected in the first digital reference facial image, wherein the second digital reference facial image of the person has a second reference image resolution. The device further comprises a personalization unit configured to store the second digital reference facial image of the person in the person's identity document in order to personalize the person's identity document.

[0015] Image manipulation can include, for example, morphing, where an intermediate transition is calculated between different digital reference images of faces. Morphing can encompass techniques such as warping, tweening, and / or cross-dissolving. However, image manipulation can also include any other form of image manipulation, such as geometric image distortion or texture alteration.

[0016] The identity document can be, for example, a passport, national identity card, visa, access control card, authorization card, company ID card, tax stamp, ticket, birth certificate, driver's license, vehicle registration document, means of payment, bank card, or credit card. The identity document can be, for example, a machine-readable travel document (MRTD) or an electronic machine-readable travel document (eMRTD). The identity document can include electronic storage in which the second digital reference facial image is to be stored. The second digital reference facial image can be assigned to a predefined data group, for example, Data Group 2 (DG2).

[0017] According to one embodiment, the image acquisition device includes a communication interface configured to retrieve and capture the person's first digital reference facial image. This first digital reference facial image can, for example, be stored on a server. This offers the advantage of efficient provision of the first digital reference facial image.

[0018] According to one embodiment, the image acquisition device comprises a communication interface configured to receive a reference to the person's first digital reference facial image and to retrieve the person's first digital reference facial image using the reference. The reference can, for example, be a link. The first digital reference facial image can, for example, be stored on a server. This achieves the advantage of efficiently providing the first digital reference facial image.

[0019] According to one embodiment, the image acquisition device comprises an image scanner, wherein the image scanner is configured to optically scan a photographic print of a person's face in order to capture the person's first digital reference facial image. This achieves the advantage of efficiently providing the first digital reference facial image.

[0020] According to one embodiment, the device includes a database in which the biometric comparison reference feature is pre-stored. This offers the advantage that the biometric comparison reference feature can be provided efficiently.

[0021] According to one embodiment, the device comprises an image sensor configured to capture a digital facial image of the person, wherein the manipulation detection device is configured to capture the biometric comparison reference feature in the digital facial image of the person. This achieves the advantage that the biometric comparison reference feature can be provided efficiently.

[0022] According to one embodiment, the manipulation detection device is further configured to capture another biometric reference feature in the first digital reference facial image of the person, to capture another biometric comparison reference feature in the digital facial image of the person, to compare the second biometric reference feature with the second biometric comparison reference feature to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference facial image based on the second manipulation indicator. This achieves the advantage that a further manipulation indicator can be taken into account for detecting the image manipulation.

[0023] According to one embodiment, the biometric reference feature and / or the further biometric reference feature is one of the following biometric reference features: a reference eye spacing of the person, a reference eye color of the person, a reference eye width of the person, a reference eye height of the person, a reference nose height of the person, a reference nose width of the person, a reference ear spacing of the person, a reference ear width of the person, a reference ear height of the person, a reference mouth height of the person, a reference forehead height of the person, a reference forehead width of the person, or a reference cheek spacing of the person. This achieves the advantage that characteristic features of human faces can be taken into account.

[0024] According to one embodiment, the biometric reference feature and / or the further biometric reference feature is one of the following biometric reference features: a reference eye spacing of the person, a reference eye color of the person, a reference eye width of the person, a reference eye height of the person, a reference nose height of the person, a reference nose width of the person, a reference ear spacing of the person, a reference ear width of the person, a reference ear height of the person, a reference mouth height of the person, a reference mouth width of the person, a reference forehead height of the person, a reference forehead width of the person, or a reference cheek spacing of the person. This achieves the advantage that characteristic features of human faces can be taken into account.

[0025] According to one embodiment, the manipulation detection device is further configured to transform the first digital reference face image into a transformation space, in particular a frequency space, in order to obtain a first transformed digital reference face image, to transform the digital face image into the transformation space, in particular the frequency space, in order to obtain a transformed digital face image, to compare the first transformed digital reference face image with the transformed digital face image in order to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference face image on the basis of the further manipulation indicator.This offers the advantage of being able to compare properties of the first digital reference face image and the digital face image, which become more apparent in the transformation space, particularly the frequency space. For example, regular surface structures in the frequency space can be compared more efficiently.

[0026] According to one embodiment, the manipulation detection device is further configured to detect a reference image feature in the first digital reference face image, to detect an image feature in the digital face image, to determine a difference between a reference position of the reference image feature in the first digital reference face image and a position of the image feature in the digital face image in order to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on the further manipulation indicator. For example, a feature extraction algorithm can be used to determine the reference image feature and / or the image feature. This achieves the advantage that characteristic properties of the first digital reference face image and the digital face image can be taken into account.

[0027] According to one embodiment, the image feature is one of the following: a corner in the digital face image, or an edge in the digital face image. This offers the advantage of taking characteristic properties of digital images into account. However, the image feature can also be any other suitable image feature that can be provided by a feature extraction algorithm.

[0028] According to one embodiment, the reference image feature is one of the following: a reference corner in the first digital reference face image, or a reference edge in the first digital reference face image. This achieves the advantage that characteristic properties of digital images can be taken into account. However, the reference image feature can also be any other suitable reference image feature that can be provided by a feature extraction algorithm.

[0029] According to one embodiment, the manipulation detection device is further configured to perform image manipulation using the first digital reference face image and the digital face image to obtain a manipulated digital reference face image, to compare the first digital reference face image with the manipulated digital reference face image to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on this further manipulation indicator. This achieves the advantage of being able to take into account the effects of image manipulation.

[0030] According to one embodiment, the manipulation detection device is further configured to detect the degree of a distortion artifact from image manipulation, in particular duplication or smearing, in the first digital reference face image in order to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on this further manipulation indicator. This achieves the advantage that typical distortion artifacts from image manipulation can be taken into account.

[0031] According to one embodiment, the manipulation detection device is configured to detect a statistical reference feature in the first digital reference face image, to detect a statistical feature in the digital face image of the person, to compare the statistical reference feature with the statistical feature to obtain a further manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on the further manipulation indicator. This achieves the advantage that statistical properties of both the first digital reference face image and the digital face image can be taken into account.

[0032] According to one embodiment, the manipulation detection device is further configured to determine the difference between the manipulation indicator and a predetermined manipulation reference value to obtain a difference value; to determine the difference between at least one further manipulation indicator and at least one further predetermined manipulation reference value to obtain at least one further difference value; to determine the sum of the difference value and the at least one further difference value to obtain a sum value; and to detect image manipulation in the first digital reference face image if the sum value exceeds a predetermined sum threshold. The manipulation indicators can each represent a confidence level or a probability of image manipulation.This offers the advantage that a joint evaluation of the manipulation indicators can be carried out using a statistical approach.

[0033] According to one embodiment, the manipulation detection device is further configured to detect image manipulation in the first digital reference face image using a neural network. The neural network comprises an input layer with one input and at least one further input, and an output layer with one output. The manipulation indicator is assigned to the input, the at least one further manipulation indicator is assigned to the at least one further input, and the output indicates the image manipulation in the first digital reference face image. The manipulation indicators can each represent a confidence level or a probability for the presence of image manipulation. This offers the advantage that the manipulation indicators can be evaluated jointly using a neural network.

[0034] The neural network can comprise multiple neurons, with each neuron within that network assigned a weight value. The neural network can be pre-trained, for example using a deep learning approach, to determine these weight values. This training can be based on multiple previously known digital reference images of faces with deliberate image manipulations.

[0035] According to one embodiment, the tamper detection device is further configured to scale the first digital reference face image of the person and the digital face image of the person to a common scale, and / or to rotate the first digital reference face image of the person and the digital face image of the person to a common alignment axis. This achieves the advantage that the respective tamper indicators can be determined more efficiently.

[0036] According to one embodiment, the tamper detection device is further configured to crop the first digital reference facial image of the person and / or the digital facial image of the person to a predetermined image section. The predetermined image section can, for example, include an area around the person's eyes, nose (especially the tip of the nose), mouth, ears, beard, hair, glasses (especially the frame), or the edge of the face. This achieves the advantage of enabling more efficient determination of the respective tamper indicators.

[0037] According to one embodiment, the image generation device is configured to convert the first digital reference face image with the first reference image resolution into the second digital reference face image with the second reference image resolution. This achieves the advantage that the second digital reference face image can be efficiently generated based on the first digital reference face image.

[0038] According to one embodiment, the first reference image resolution is higher than the second reference image resolution. This offers the advantage that, for example, a high-resolution first digital reference facial image can be used as the basis for detecting image manipulation, while the lower-resolution second digital reference facial image can be stored in the person's identity document.

[0039] According to a second aspect, the invention relates to a method for personalizing a person's identity document. The method comprises capturing a first digital reference facial image of the person by an image capture device, wherein the first digital reference facial image of the person has a first reference image resolution, capturing a biometric reference feature in the first digital reference facial image of the person by a tamper detection device, comparing the biometric reference feature with a biometric comparison reference feature by the tamper detection device to obtain a tamper indicator, and detecting image manipulation in the first digital reference facial image based on the tamper indicator by the tamper detection device if no image manipulation is detected in the first digital reference facial image.a generating of a second digital reference facial image of the person based on the first digital reference facial image of the person by an image generation device, wherein the second digital reference facial image of the person has a second reference image resolution, and a storing of the second digital reference facial image of the person in the identity document of the person by a personalization device in order to personalize the identity document of the person.

[0040] The method can be carried out by the device according to the first aspect of the invention. Further features of the method result directly from the features or functionality of the device according to the first aspect of the invention.

[0041] According to a third aspect, the invention relates to a computer program with program code for carrying out the method according to the second aspect of the invention when the program code is executed by a device. The device can be programmed to execute the program code.

[0042] The invention can be implemented in hardware and / or in software. DESCRIPTION OF THE FIGURES

[0043] Further examples are explained in more detail with reference to the accompanying drawings. These show: Fig. 1 a schematic diagram of a device for personalizing a person's identity document; and Fig. 2 A schematic diagram of a procedure for personalizing a person's identity document. DETAILED DESCRIPTION OF THE FIGURES

[0044] Fig. 1Figure 100 shows a schematic diagram of a device for personalizing a person's identity document. Device 100 splits the conventional image capture process into several sub-steps relating to image capture, tamper detection, and image generation.

[0045] The device 100 comprises an image acquisition device 101 configured to acquire a first digital reference facial image of the person, wherein the first digital reference facial image of the person has a first reference image resolution. The image acquisition device 101 may include a communication interface configured to retrieve the first digital reference facial image of the person in order to acquire it. The communication interface may additionally or alternatively be configured to receive a reference to the first digital reference facial image of the person and to retrieve the first digital reference facial image of the person using the reference in order to acquire it. The reference may, for example, be a link. The first digital reference facial image may, for example, be stored on a server.Alternatively, the image acquisition device 101 can include an image scanner, wherein the image scanner is configured to optically scan a photographic print of the person's face in order to capture the person's first digital reference facial image. Consequently, a high-resolution first digital reference facial image of the person can be provided.

[0046] The first digital reference facial image is therefore digitally captured by the image capture device 101. Unlike conventional image capture, the first digital reference facial image is typically not a portrait suitable for identity documents according to data group DG2, with an ICAO-recommended Inter-Eye Distance (IED) of 120 pixels, but rather a high-resolution image that is as unprocessed and uncompressed as possible. The first digital reference facial image can be as high-resolution as the photographic equipment allows. For example, the IED can be in the range of >400 pixels or higher, which is easily achievable with modern cameras. The first digital reference facial image can be compressed almost without loss. Almost without loss because most cameras can perform initial JPEG compression.During the development of raw images in the camera, certain processing steps can always take place, which can lead to a slightly reduced resolution.

[0047] The device 100 further comprises a manipulation detection device 103, which is configured to detect a biometric reference feature in the first digital reference facial image of the person, to compare the biometric reference feature with a biometric comparison reference feature in order to obtain a first manipulation indicator, and to detect image manipulation in the first digital reference facial image on the basis of the first manipulation indicator.

[0048] The device 100, in particular the manipulation detection device 103, can include a database in which the biometric comparison reference feature is pre-stored. In summary, a biometric comparison or biometric thresholding can be performed against the pre-stored biometric comparison reference feature. The biometric comparison reference feature can represent a typical biometric characteristic of human faces. The greater the difference, the higher the confidence or probability of image manipulation. The pre-stored biometric comparison feature can, for example, be taken from a previous identity document of the person.

[0049] The device 100, in particular the image acquisition device 101 or the tamper detection device 103, can include an image sensor configured to capture a digital facial image of the person, wherein the tamper detection device 103 is configured to capture the biometric comparison reference feature in the digital facial image of the person. The image sensor can, for example, be a still camera or a video camera. The video camera can provide a video with a plurality of image frames, wherein the digital facial image can be formed by or contained within an image frame. The described concept can therefore be repeated for several image frames. Consequently, the biometric comparison reference feature can be determined live. The quality of the captured digital facial image may not need to be particularly high.In some cases, a self-portrait or selfie may suffice.

[0050] The manipulation detection device 103 can further be configured to capture an additional biometric reference feature in the first digital reference facial image of the person, to capture an additional biometric comparison reference feature in the digital facial image of the person, to compare the additional biometric reference feature with the additional biometric comparison reference feature to obtain a second manipulation indicator, and to further detect the image manipulation in the first digital reference facial image based on the second manipulation indicator. The biometric reference feature and the additional biometric reference feature can be different. The biometric comparison reference feature and the additional biometric comparison reference feature can also be different.

[0051] The manipulation detection device 103 can further be configured to transform the first digital reference face image into a transformation space, in particular a frequency space, in order to obtain a first transformed digital reference face image, to transform the digital face image into the transformation space, in particular the frequency space, in order to obtain a transformed digital face image, to compare the first transformed digital reference face image with the transformed digital face image in order to obtain a third manipulation indicator, and to further detect the image manipulation in the first digital reference face image on the basis of the third manipulation indicator.For example, it can be examined whether the first digital reference face image and the digital face image have a similarly homogeneous structure and whether the respective structural properties in the various areas, such as the background, foreground, face, or selected facial areas, correspond to predetermined criteria. In particular, the image areas around the eyes, nose, and / or mouth of the person, as well as the edges of the person's face, can be considered.

[0052] The manipulation detection device 103 can further be configured to detect a reference image feature in the first digital reference face image, to detect an image feature in the digital face image, to determine a difference between a reference position of the reference image feature in the first digital reference face image and a position of the image feature in the digital face image in order to obtain a fourth manipulation indicator, and furthermore to detect the image manipulation in the first digital reference face image based on the fourth manipulation indicator. For example, a feature extraction algorithm can be used to determine the reference image feature and / or the image feature. In summary, characteristic properties of the first digital reference face image and the digital face image can be examined.

[0053] Furthermore, a plurality of reference image features can be captured in the first digital reference face image and a plurality of image features in the digital face image. After filtering, it can be examined whether the respective number of reference image features or image features, and / or the differences in the reference positions or positions, correspond. It can also be examined whether the distribution of correlation components differs between the first digital reference face image and the digital face image.

[0054] The manipulation detection device 103 can further be configured to perform image manipulation using the first digital reference face image and the digital face image to obtain a manipulated digital reference face image, to compare the first digital reference face image with the manipulated digital reference face image to obtain a fifth manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on the fifth manipulation indicator. In summary, it can be investigated how the first digital reference face image would change if it were manipulated using the digital face image.

[0055] The manipulation detection device 103 can further be configured to detect a degree of distortion artifact of image manipulation, in particular a duplication or smearing, in the first digital reference face image in order to obtain a sixth manipulation indicator, and furthermore to detect the image manipulation in the first digital reference face image on the basis of the sixth manipulation indicator. Such distortion artifacts can be detected in particular in the area of ​​the irises, the rims of the eyes, the eyebrows, the eyelashes, the hairlines, the lips, the corners of the mouth, the nostrils, or the flexures at the edge of a pair of glasses.

[0056] The manipulation detection device 103 can further be configured to detect a statistical reference feature in the first digital reference face image, to detect a statistical feature in the digital face image of the person, to compare the statistical reference feature with the statistical feature to obtain a seventh manipulation indicator, and to further detect the image manipulation in the first digital reference face image based on the seventh manipulation indicator. The statistical reference feature and the statistical feature can represent noise in the first digital reference face image and in the digital face image, respectively.

[0057] The tamper detection device 103 can further be configured to scale the person's first digital reference face image and the person's digital face image to a common scale, and / or to rotate the person's first digital reference face image and the person's digital face image to a common alignment axis. The tamper detection device 103 can further be configured to crop the person's first digital reference face image and / or the person's digital face image to a predetermined image area. The predetermined image area can, for example, include an area around the person's eyes, nose (in particular the tip of the nose), mouth, ears, beard, hair, glasses (in particular the rim of the glasses), or the edge of the person's face.

[0058] The aforementioned approaches for determining the first, second, third, fourth, fifth, sixth, and seventh manipulation indicators can be applied in any combination. For example, only the first and third manipulation indicators, or only the first, second, fifth, and seventh manipulation indicators, can be determined. However, all of the aforementioned manipulation indicators can also be determined.

[0059] A joint evaluation of the manipulation indicators used can then be carried out, for example, using a statistical approach or a neural network. The joint evaluation will be described below using examples of a first, a second, and a third manipulation indicator.

[0060] In the statistical approach, the manipulation detection device 103 is configured, for example, to determine a difference between the first manipulation indicator and a predetermined first manipulation reference value to obtain a first difference value, to determine a difference between the second manipulation indicator and a predetermined second manipulation reference value to obtain a second difference value, to determine a difference between the third manipulation indicator and a predetermined third manipulation reference value to obtain a third difference value, to determine a sum of the first difference value, the second difference value, and the third difference value to obtain a sum value, and to detect image manipulation in the first digital reference face image if the sum value exceeds a predetermined sum threshold.

[0061] Alternatively, the manipulation detection device 103 is configured, for example, to detect image manipulation in the first digital reference face image using the neural network, wherein the neural network has an input layer with a first input, a second input, and a third input, and wherein the neural network has an output layer with an output, wherein the first manipulation indicator is assigned to the first input, the second manipulation indicator is assigned to the second input, the third manipulation indicator is assigned to the third input, and wherein the output indicates the image manipulation in the first digital reference face image. The neural network is thus used to investigate whether the first digital reference face image has been manipulated.

[0062] In summary, a multiple manipulation detectors can be used on a high-resolution, low-loss first digital reference face image, and their results can be appropriately processed or fused. For example, an analysis for morphing artifacts can be performed. Furthermore, an analysis for presentation-attack artifacts can be performed. The manipulation detection device 103 can, for example, determine a confidence score indicating how likely an image manipulation has been detected.

[0063] The device 100 further comprises an image generation unit 105, which is configured, if no image manipulation is detected in the first digital reference face image, to generate a second digital reference face image of the person based on the first digital reference face image of the person, wherein the second digital reference face image of the person has a second reference image resolution. The image generation unit 105 can be configured to convert the first digital reference face image with the first reference image resolution into the second digital reference face image with the second reference image resolution. The first reference image resolution can be higher than the second reference image resolution.

[0064] In summary, the first digital reference face image is further processed. The image generation unit 105 can generate all images required for personalizing the identity document. The generated second digital reference face image can be provided, for example, in JPEG or JPEG2000 format for Data Group 2 (DG2), for the image to be printed on a data page of the identity document, as well as for holographic and other secondary images. By independently generating the second digital reference face image from a high-quality first digital reference face image, the advantage is achieved that an optimal image can be provided depending on the personalization technology used for the respective image. Furthermore, multiple lossy compressions can be avoided.Any limitations that may arise from a previous resolution limit, for example an IED of 90 pixels, can therefore be eliminated.

[0065] The device 100 further comprises a personalization unit 107, which is configured to store the second digital reference facial image of the person in the person's identity document in order to personalize the person's identity document. The personalization unit 107 can communicate with the person's identity document via a wired or wireless connection.

[0066] Device 100 can automatically detect image manipulation and can be used, for example, as a support tool for a user. This support can be provided using software and / or hardware-based tools. For instance, the user can be guided to an area where Device 100 has detected image manipulation. Furthermore, the device can zoom in on this area. Such areas could include, for example, an eye, nostrils, an ear, or the mouth of the person. Depending on the application, this allows both experienced and inexperienced users to manually examine the initial digital reference facial image for image manipulation or anomalies.

[0067] A reverse approach is also possible, in which the user marks a suspicious area in the first digital reference facial image beforehand, which is then examined by the device 100, or the user excludes certain areas in advance. This can be advantageous, for example, if it is already known that certain areas have been intentionally and lawfully altered.

[0068] All of the aforementioned components of the device 100, in particular the image acquisition device 101, the tamper detection device 103, the image sensor, the image generation device 105, and the personalization device 107, can be arranged spatially adjacent to one another or separately and communicate with each other, for example, via communication links over a communication network. For example, the image acquisition device 101, the tamper detection device 103, and the image sensor can be located in a passport office, whereas the image generation device 105 and the personalization device 107 can be located at a manufacturer of the identity document. Alternatively, the tamper detection device 103 can also be located at the manufacturer of the identity document.

[0069] Device 100 thus provides a means of determining whether a passport photo submitted by an applicant has been manipulated, in particular morphed, or otherwise impermissibly modified. It can also determine whether further alterations, such as cosmetic enhancements, have taken place. For example, blemishes might have been removed, wrinkles smoothed, or the nose reshaped. This may also contribute to improved image quality and can be a beneficial side effect. Device 100 can be used by authorities, particularly in the context of passport application procedures. Device 100 can be implemented decentrally or centrally, or on a server.

[0070] Fig. 2 shows a schematic diagram of a procedure 200 for personalizing a person's identity document.

[0071] The method 200 comprises capturing 201 a first digital reference facial image of the person by an image capture device, wherein the first digital reference facial image of the person has a first reference image resolution, capturing 203 a biometric reference feature in the first digital reference facial image of the person by a tamper detection device, comparing 205 the biometric reference feature with a biometric comparison reference feature by the tamper detection device to obtain a tamper indicator, and detecting 207 image manipulation in the first digital reference facial image on the basis of the tamper indicator by the tamper detection device if no image manipulation is detected in the first digital reference facial image.a generating 209 of a second digital reference facial image of the person based on the first digital reference facial image of the person by an image generation device, wherein the second digital reference facial image of the person has a second reference image resolution, and a storing 211 of the second digital reference facial image of the person in the identity document of the person by a personalization device in order to personalize the identity document of the person.

[0072] Method 200 can be carried out using a computer program with program code. The device can be programmed to execute the program code.

[0073] All features shown or described in connection with individual embodiments of the invention can be provided in any combination in the object according to the invention in order to simultaneously realize their advantageous effects. REFERENCE MARK LIST

[0074] 100 Device 101 Image acquisition device 103 Tamper detection device 105 Image generation device 107 Personalization device 200Process 201Capture 203Capture 205Compare 207Recognize 209Generate 211Save

Claims

1. A device (100) for personalizing an identity document of a person, comprising: an image capture device (101), which is configured to capture a first digital reference facial image of the person, wherein the first digital reference facial image of the person has a first reference image resolution; a manipulation detection device (103), which is configured to detect a biometric reference feature in the first digital reference facial image of the person, to compare the biometric reference feature with a biometric comparison reference feature in order to obtain a manipulation indicator, and to detect an image manipulation in the first digital reference facial image on the basis of the manipulation indicator; wherein the manipulation indicator represents a confidence for the presence of an image manipulation, wherein the greater the difference, the higher the confidence for an image manipulation, an image generation device (105), which is configured, if no image manipulation is detected in the first digital reference facial image, to generate a second digital reference facial image of the person on the basis of the first digital reference facial image of the person, wherein the second digital reference facial image of the person has a second reference image resolution; and a personalization device (107), which is configured to store the second digital reference facial image of the person in the identity document of the person in order to personalize the identity document of the person, wherein the first reference image resolution is greater than the second reference image resolution, wherein the image generation device (105) is configured to convert the first digital reference facial image having the first reference image resolution into the second digital reference facial image having the second reference image resolution.

2. The device (100) according to claim 1, wherein the image capture device (101) comprises a communication interface, which is configured to receive a reference to the first digital reference facial image of the person and to retrieve the first digital reference facial image of the person using the reference in order to capture the first digital reference facial image of the person.

3. The device (100) according to claim 1, wherein the image capture device (101) comprises an image scanner, wherein the image scanner is configured to optically scan a photographic print of a face of the person in order to capture the first digital reference facial image of the person.

4. The device (100) according to one of the preceding claims, comprising: an image sensor, which is configured to capture a digital facial image of the person; wherein the manipulation detection device (103) is configured to detect the biometric comparison reference feature in the digital facial image of the person.

5. The device (100) according to claim 4, wherein the manipulation detection device (103) is further configured to detect a further biometric reference feature in the first digital reference facial image of the person, to detect a further biometric comparison reference feature in the digital facial image of the person, to compare the further biometric reference feature with the further biometric comparison reference feature in order to obtain a further manipulation indicator, and to detect the image manipulation in the first digital reference facial image further based on the further manipulation indicator.

6. The device (100) according to one of claims 4 or 5, wherein the manipulation detection device (103) is further configured to transform the first digital reference facial image into a transformation space, in particular a frequency space, in order to obtain a first transformed digital reference facial image, to transform the digital facial image into the transformation space, in particular the frequency space, in order to obtain a transformed digital facial image, to compare the first transformed digital reference facial image with the transformed digital facial image in order to obtain a further manipulation indicator, and to detect the image manipulation in the first digital reference facial image further on the basis of the further manipulation indicator.

7. The device (100) according to one of claims 4 to 6, wherein the manipulation detection device (103) is further configured to detect a reference image feature in the first digital reference facial image, detect an image feature in the digital facial image, determine a difference between a reference position of the reference image feature in the first digital reference facial image and a position of the image feature in the digital facial image to obtain a further manipulation indicator, and detect the image manipulation in the first digital reference facial image further based on the further manipulation indicator.

8. The device (100) of claim 7, wherein the image feature is one of the following image features: a corner in the digital facial image, or an edge in the digital facial image.

9. The device (100) according to one of claims 4 to 8, wherein the manipulation detection device (103) is further configured to perform an image manipulation using the first digital reference facial image and the digital facial image in order to obtain a manipulated digital reference facial image, to compare the first digital reference facial image with the manipulated digital reference facial image in order to obtain a further manipulation indicator, and to detect the image manipulation in the first digital reference facial image further based on the further manipulation indicator.

10. The device (100) according to one of claims 5 to 9, wherein the manipulation detection device (103) is further configured to determine a difference between the manipulation indicator and a predetermined manipulation reference value in order to obtain a difference value, to determine a difference between a further manipulation indicator and a further predetermined manipulation reference value in order to obtain a further difference value, to determine a sum of the difference value and the further difference value in order to obtain a sum value, and to detect the image manipulation in the first digital reference facial image if the sum value exceeds a predetermined sum threshold value.

11. The device (100) according to one of claims 5 to 9, wherein the manipulation detection device (103) is further configured to detect the image manipulation in the first digital reference facial image using a neural network, wherein the neural network has an input layer with an input and a further input, wherein the neural network has an output layer with an output, wherein the manipulation indicator is associated with the input, wherein the further manipulation indicator is associated with the further input, and wherein the output indicates the image manipulation in the first digital reference facial image.

12. A method (200) for personalizing an identity document of a person, comprising: capturing (201) a first digital reference facial image of the person by an image capturing device (101), wherein the first digital reference facial image of the person has a first reference image resolution; detecting (203) a biometric reference feature in the first digital reference facial image of the person by a manipulation detection device (103); comparing (205) the biometric reference feature with a biometric comparison reference feature by the manipulation detection device (103) in order to obtain a manipulation indicator; wherein the manipulation indicator represents a confidence for the presence of image manipulation, wherein the greater the difference, the higher the confidence for an image manipulation, detecting (207) an image manipulation in the first digital reference facial image on the basis of the manipulation indicator by the manipulation detection device (103); if no image manipulation is detected in the first digital reference facial image, generating (209) a second digital reference facial image of the person on the basis of the first digital reference facial image of the person by an image generation device (105), wherein the second digital reference facial image of the person has a second reference image resolution; and storing (211) the second digital reference facial image of the person in the identity document of the person by a personalization device (107) in order to personalize the identity document of the person, wherein the first reference image resolution is greater than the second reference image resolution, wherein the first digital reference facial image having the first reference image resolution is converted into the second digital reference facial image having the second reference image resolution.

13. A computer program comprising a program code for performing the method (200) according to claim 12, when the program code is executed by a device (100).