Method for generating a code, method for verifying a code, document comprising the code, and use of the code

Abstract

The invention relates to a method for generating a code 3 which is based on first biometric characteristics 2', to a method for verifying a code 3, to a document 4 comprising the code 3, and to the use of the code 3.

Description

[0001] Method for generating a code, a method for verifying a code and a document comprising the code and the use of the code

[0002] The invention relates to a method for generating a code based on first biometric characteristics, a method for verifying a code and a document comprising the code and the use of the code.

[0003] It is well known that graphical, machine-readable codes are used to store information and read it using a device. These offer the advantage of being easy to create and read by devices, even those without significant computing power. This makes the information contained in the codes quickly and easily available.

[0004] The amount of information that can be stored in the code is, however, limited. The amount of information that can be stored in the code depends on the size of the code, for example on a document. The space available, for example on the document, limits the size of the code and thus the amount of information. The area of ​​application for these codes is therefore restricted to simple applications, such as storing an internet address or other simple alphanumeric information in a relatively small amount. The size of the code or the available space is to be understood as a two-dimensional graphic image size or two-dimensional dimension, in particular a two-dimensional print or a two-dimensional image on a substrate, for example on a document, with the code components, in particular the modules, having a two-dimensional minimum size orMinimum dimensions to provide the desired machine readability.

[0005] Personal data can be more complex. More complex personal data is biometric data, such as biometric characteristics. These are suitable for identifying individuals due to their difficult-to-manipulate properties, but also require more storage capacity due to their complexity.

[0006] When larger or more complex quantities of a person's personal data need to be stored to ensure secure verification and authenticity of a document, the storage capacity of a graphic, machine-readable code, such as a 1D barcode or QR code, is easily exceeded. Currently, memory chips are used to store personal data embedded in the document to be verified. However, memory chips are significantly more complex to manufacture and require more expensive materials than machine-readable graphic codes. Furthermore, recycling these materials also presents a challenge.

[0007] Thus, there is no practical solution for providing biometric data, such as biometric characteristics, cost-effectively and securely so that they can be easily verified using simple electronic devices. Furthermore, there is currently no standard specification for machine-readable, graphical codes based on biometric data. The object of the invention is to provide an improved code for verifying a document, wherein the code is in the form of a geometric pattern and is based on biometric characteristics.

[0008] The object is achieved by a method, in particular by a method according to claim 1, for generating a code based on first biometric characteristics, wherein the method comprises at least the following steps: a) capturing at least one piece of biometric information from a real person, a photo and / or a video, b) extracting first biometric characteristics from the at least one piece of biometric information, in particular by means of a neural network, in the form of a first data set, c) compressing the first data set to obtain a compressed first data set, d) adding a second data set with payload data to the compressed first data set, e) generating a first hash value via the compressed first data set or via the compressed first data set and the second data set, f) generating a signature for the first hash value, g) generating the code in the form of a geometric pattern,comprising the first compressed data set based on the first biometric characteristics, the first hash value and the signature.

[0009] Preferred embodiments of the method according to claim 1 are described in subclaims 2 to 24.

[0010] Furthermore, the object is achieved by a method, preferably by a method according to claim 25, for verifying a code, in particular a code according to claim 1, wherein the method comprises at least one of the following steps A) or B), A) creating a compressed third data set, wherein the compressed third data set is based on second biometric characteristics, wherein step A) comprises at least the following sub-steps:

[0011] - Capturing at least one piece of biometric information from a real person, a photo or a video,

[0012] - Extracting second biometric characteristics, in particular by means of a neural network, from the at least one biometric information in the form of a third data set,

[0013] - Compressing the third data set to obtain a compressed third data set, or

[0014] B) opening a database comprising at least one compressed third data set, wherein the compressed third data set is based on second biometric characteristics, the method further comprising at least the following steps: h) reading a compressed first data set from the code, wherein the compressed first data set is based on first biometric characteristics, and comparing the compressed first data set with the compressed third data set, i) generating an evaluation parameter that enables the evaluation of a match between the compressed first data set and the compressed third data set, j) outputting an evaluation based on the generated evaluation parameter.

[0015] Preferred embodiments of the method according to claim 25 are described in subclaims 26 to 31.

[0016] Furthermore, the object is achieved by a document, in particular a document according to claim 32, which comprises a code generated according to claim 1. Preferred embodiments of the document according to claim 32 are described in subclaims 33 to 39.

[0017] Furthermore, the object is achieved by the use, in particular by the use according to claim 40, of the code generated according to claim 1 in a document, wherein the document is selected from the group consisting of certificate, deed, ticket, preferably airline ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, an official document, preferably visa, residence permit, proof of arrival, temporary ID, residence permit or combinations thereof.

[0018] The present invention makes it possible to combine biometric information, preferably biometric characteristics, with machine-readable graphic codes. This allows an object, for example a document, to be linked to a person, which is one of the main objectives for documents issued by authorities or organizers. Since the generation of a code in the form of graphic patterns entails negligible additional material effort, biometric information, preferably biometric characteristics, can be incorporated into documents cost-effectively. In addition, it is possible to easily verify the information, preferably biometric characteristics, using simple devices, for example, conventional smartphones.Nevertheless, the invention can ensure that the stored information, preferably biometric characteristics, is securely protected against forgery or manipulation due to asymmetric cryptography.

[0019] Further advantageous embodiments of the invention are defined in the subclaims. In step a), at least one piece of biometric information is captured. Step a) is preferably the first step of the method for generating the code and is performed before step b).

[0020] Biometric information is primarily understood as the physiological characteristics of a body or body part. Biometric information that is present in all individuals within a group to be distinguished, but is unique to a single individual, is preferably unique to a person, is constant regardless of age or time of measurement, can be captured using a defined metric or measurement method, and can be captured in essentially all individuals.

[0021] Preferably, at least the biometric information of a face, a fingerprint, hand veins, and / or an eye, preferably an iris and / or a retina, is captured. Particularly preferably, the biometric information of a face is captured.

[0022] It is possible that the at least one piece of biometric information is present in two-dimensional and / or three-dimensional form and / or that the at least one piece of biometric information is captured in two-dimensional and / or three-dimensional form.

[0023] Parts or configurations of at least one piece of biometric information that are characteristic of an individual person, in particular that enable differentiation from other persons, are understood as biometric characteristics. They are unique for an individual person compared to a group of people to be differentiated. Biometric characteristics are preferably comprised by or derived from at least one piece of biometric information. The biometric characteristics are preferably dependent on a person's physiology. Biometric characteristics that are recorded to generate the code can be referred to as first biometric characteristics. Biometric characteristics that are additionally recorded to verify the code can be referred to as second biometric characteristics.If the first and second biometric characteristics are captured from the same person, the second biometric characteristics may correspond to the first biometric characteristics or include them completely or partially. In particular, the second biometric characteristics have the same embodiments as the first biometric characteristics of the first data set. This enables a comparison of the compressed first data set with the compressed third data set.

[0024] It is possible for the at least one piece of biometric information to be captured using a device selected from the group consisting of a smartphone, camera, tablet, PC, terminal, IoT devices, or combinations thereof. The device comprises at least one device for capturing images and a processor for executing at least one further step of the method.

[0025] This offers the advantage that no expensive special equipment has to be used, but that the procedure can also be carried out with common equipment.

[0026] It is possible that the device is a client device and the at least one piece of biometric information is sent over a network to an external server where one or more or all of the following steps are performed.

[0027] Alternatively, it is possible for all steps of the method to be performed locally on the device, preferably without a network connection, in particular a wireless network connection, to an external server. In step b), first biometric characteristics are extracted from at least one piece of biometric information, in particular, they are recognized and extracted. Step b) is preferably performed after step a) and before step c).

[0028] It is possible that one or more biometric characteristics, in particular the first and / or second biometric characteristics, are selected from the group consisting of the tip of the nose, the corner of the eye, the eyebrow, the tip of the chin, the center of the nose, the center of the eyes, the contour of the eyes, the contour of the lips, the contour of the nose, loops, arches, swirls and / or minutiae of a fingerprint, veins of a hand and combinations thereof.

[0029] It is possible for the number of biometric characteristics, in particular the first and / or second biometric characteristics, to be selected from the range of 50 to 200, preferably from 60 to 100, more preferably from 65 to 70. Even more preferably, the number of biometric characteristics selected is 68. In particular, all extracted first biometric characteristics are included in the first data set. In other words, in step b), a first data set is obtained that is based on all extracted first biometric characteristics.

[0030] This makes it possible to collect a sufficient number of biometric characteristics to allow for a sufficiently high degree of matching and thus identification of a person. An even greater number of biometric characteristics appears unnecessary, as the increase in accuracy would be negligible, but significantly larger data storage capacities would be required.

[0031] It is possible for the biometric characteristics, in particular the first biometric characteristics in the first data set or the second biometric characteristics in the third data set, to be described and extracted as vectors and / or as coordinates in a coordinate system. Furthermore, it is also possible for the biometric characteristics, in particular the first and / or second biometric characteristics, to be assigned a type or a class, for example the basic pattern of a fingerprint. It is also possible for the biometric characteristics, in particular the first and / or second biometric characteristics, to be described and extracted as binary data points, preferably of an iris scan. It is possible for the biometric characteristics, in particular the first and / or second biometric characteristics, to be described and extracted as a

[0032] Combination of at least two of the above variants can be described and extracted.

[0033] This makes it possible to describe the position or configuration of the biometric characteristics, in particular the first and / or second biometric characteristics, absolutely and / or relative to other biometric characteristics and to each other.

[0034] It is possible that step b) for extracting the first biometric characteristics comprises at least the following sub-steps:

[0035] - Identifying and defining a first area to be used for extraction,

[0036] - Recognition of the first biometric characteristics in the first area,

[0037] - Extracting the first biometric characteristics in the first area.

[0038] In particular, the first area is defined by the bounding rectangle method. Here, a rectangle is placed over the at least one piece of biometric information, for example, a face. This is done in such a way that the first biometric characteristics preferably lie at least 90% within the rectangle, wherein the rectangle preferably forms the smallest circumscribing rectangle of the at least one piece of biometric information. It is possible for the extraction of the first biometric characteristics in the first area to be carried out by a method selected from the group consisting of a holistic method, a method with a restricted local model, a regression-based method, or a combination method of deep learning models and 3D shape models, preferably FacemarkLBF, FacemarkKazemi, FacemarkAAM, or combinations thereof.Preferably, the extraction of the first biometric characteristics in the first area is carried out by a combination method of deep learning models and 3D shape models, preferably by FacemarkLBF.

[0039] In particular, a method is used that can recognize and track at least one piece of biometric information and / or a first biometric characteristic, for example from a real person, a photograph, a video, and / or a sequence of images. It is possible for the contours and / or centers of eyes, lips, or nose to be recognized and / or tracked. It is also possible for loops, arches, swirls, and / or minutiae of a fingerprint to be recognized and / or tracked. Furthermore, it is also possible for the veins of a hand to be recognized and / or tracked. Preferably, a method is used that uses at least one neural network. Neural networks are preferably trained using large data sets of at least one piece of biometric information, preferably facial images, in particular in which biometric characteristics have been manually marked.Preferably, they are trained to recognize patterns in data and use these patterns to identify and track biometric characteristics. Specifically, the method uses a cascade regression model, particularly one that includes multiple layers of neural networks. Holistic methods are methods that explicitly model the holistic appearance of the face and global patterns of facial shape, such as the external facial shape.

[0040] Constrained Local Model (CLM) methods are methods that rely on the explicit local appearance of the face and explicit global patterns of the face shape.

[0041] Regression-based methods are methods that use holistic or local appearance information and implicitly embed global facial shape patterns for joint landmark detection.

[0042] FacemarkLBF describes a model for recognizing and tracking biometric information, in particular face recognition and face tracking. FacemarkLBF can be trained to recognize and track biometric information of a face, a fingerprint, palm veins, and / or an eye, preferably an iris and / or a retina. FacemarkLBF divides a piece of biometric information, for example, a face, into sub-regions. It then recognizes and tracks features, in particular biometric characteristics, within these sub-regions. The FacemarkLBF model uses an algorithm that recognizes and tracks these features, in particular biometric characteristics, in each sub-region of the biometric information, for example, a face. A prediction is made about where these features, in particular biometric characteristics, will be located in the next image.The advantage of this is that movements and changes in biometric information, for example, in the face, can be tracked even when the biometric information is moving or the lighting changes. The FacemarkLBF model uses a cascade regression model.

[0043] The cascade regression model is a machine learning method that comprises multiple layers of neural networks. Preferably, each neural network in a layer in the cascade regression model is trained to detect and track specific features, particularly biometric characteristics, in a piece of biometric information, such as a face.

[0044] The output of a first-layer neural network is preferably used as input for a second-layer network, which in turn detects and tracks further features, particularly biometric characteristics. In this way, the tracking of biometric information and / or characteristics becomes increasingly precise and accurate.

[0045] To train the cascade regression model, datasets comprising a large number of images are provided. The images contain biometric information and biometric characteristics. The datasets allow the cascade regression model to be calibrated with respect to capturing the represented biometric characteristics. The cascade regression model learns how to recognize and track the biometric characteristics to which it is or should be calibrated by making a series of binary decisions, particularly through classifiers. The cascade regression model consists of several layers of classifiers that are applied sequentially to recognize and / or track a specific feature or property, particularly a biometric characteristic. Each classifier of the cascade regression model is a binary classifier.This means that the classifier either affirms or denies the presence of a particular feature or characteristic, especially a biometric characteristic. In the context of the cascade regression model, multiple layers of neural networks can be used to detect complex patterns in the input data and classify them into different categories. Each layer of the neural network can be thought of as a classifier, classifying a subset of the input data and passing the output to the next layer to detect even more complex patterns. Cascading allows for gradual processing of the data, enabling the model to achieve greater accuracy or performance.

[0046] The classifiers are applied to an image one after the other, and only if a classifier provides a positive result, particularly in terms of detecting a characteristic, preferably at a specific position, is the next classifier applied to the image, particularly within the bounding rectangle.

[0047] The use of cascade regression models offers the advantage of effectively and efficiently detecting and tracking complex features such as biometric characteristics, especially first and / or second biometrics. This is possible even when they occur at different angles or lighting conditions.

[0048] It is possible to use a method that can detect and track at least one piece of biometric information and / or at least one first biometric characteristic, the method using a convolutional neural network as the neural network.

[0049] A convolutional neural network (abbreviated CNN) is a neural network that is preferably used for processing images and / or other multidimensional data. It can be used in image recognition and image classification, recognition of biometric information, preferably facial recognition and / or speech recognition, and other applications or combinations thereof, particularly when large amounts of complex data and data sets need to be processed. A convolutional neural network preferably comprises at least two layers, wherein the at least two layers each perform different functions. The convolutional neural network preferably comprises at least one convolutional layer as the first layer and at least one pooling layer as the second layer. The layers are executed sequentially, starting from the first layer.

[0050] In the convolutional layer, a set of filters is convolved over the input image to generate one or more feature maps. These feature maps detect various features in an image, such as edges, textures, and / or shapes.

[0051] The convolutional layer is followed, in particular, by a pooling layer, which reduces the size of the one or more feature maps and extracts the information most relevant for further processing. The pooling layer can, in particular, be designed as a max-pooling layer. Preferably, the max-pooling layer extracts the maximum of each sub-region of the one or more feature maps.

[0052] The convolutional neural network can have additional layers.

[0053] Preferably, the convolutional neural network has one or more fully connected layers following the first and second layers. The one or more fully connected layers combine features from the first and / or second layers or from preceding further fully connected layers and each outputs a prediction. The prediction may, for example, include whether an input image represents a certain class of object. A convolutional neural network is trained in the various layers by optimizing weights that best represent the features of the input image and provide the highest accuracy in classifying data.

[0054] The biometric characteristics, in particular the first and / or second biometric characteristics, can be recognized and marked, for example, as points and / or coordinates in at least one piece of biometric information, for example, a face. These biometric characteristics, in particular the first and / or second biometric characteristics, can be used to identify and track the at least one piece of biometric information independently of various conditions such as position and lighting. If the at least one piece of biometric information is a face, it can be identified and tracked independently of various facial expressions, eye movements, gaze direction, facial hair, glasses, and head posture.It is possible that algorithms for recognizing biometric characteristics, in particular the first and / or second biometric characteristics, are used in the automatic identification of key points in an image and / or video.

[0055] In step c), the first data set based on the extracted biometric characteristics, in particular the first biometric characteristics, is compressed. Step c) is preferably performed after step b) and before step d).

[0056] The first data set based on the extracted first biometric characteristics is preferably an array of 128 floating-point numbers, particularly at the beginning of step c). Each floating-point number preferably comprises 4 bytes. In other words, the first data set comprises a data volume of 128 x 4 bytes, i.e., 512 bytes. Preferably, in step c), the data volume of each floating-point number, preferably of the array of 128 floating-point numbers, is reduced. In particular, the data volume of the first data set is reduced by a factor of 4. The data volume of each floating-point number is compressed such that it preferably comprises 1 byte. In step c), a compressed first data set is obtained, which preferably comprises 128 x 1 bytes, i.e., 128 bytes.

[0057] It is possible that the quantization method is used for compression. In particular, each floating-point number is reduced to a certain number of discrete values. The reduction of floating-point numbers is referred to as rounding. Quantization preferably results in an integer for each floating-point number. In particular, each floating-point number is a 4-byte floating-point number, and each integer is a 1-byte integer.

[0058] This offers the advantage of reducing the amount of data in the initial data set and thus the amount of data that must be included in the generated code. This allows the method to be executed on mobile devices or IoT devices that have limited resources, such as memory and computing power. By reducing the size of the method parameters, the resource requirements can be reduced without significantly compromising the method, for example, its accuracy.

[0059] In particular, embedding numbers are used to reduce the floating-point numbers and for compression in step c) of the first data set. Using the embedding numbers, the floating-point number can be scaled or normalized to a single byte. The embedding numbers are obtained, in particular, using a deep learning method. LFW images can be used as training data or sample data to determine the embedding numbers.

[0060] LFW images are a collection of facial images collected as part of the "Labeled Faces in the Wild" (LFW) project. They serve as a benchmark for facial recognition to test the performance of various facial recognition algorithms. Specifically, the LFW images consist of publicly available images and include a large number of people of different ethnicities, ages, and genders, as well as in different positions and expressions. The images are of varying quality, particularly in resolution and lighting, and vary in size. The LFW images have been contributed to the development of new facial recognition algorithms and technologies.

[0061] It is possible to determine embedding numbers for biometric characteristics, in particular for the first and / or second biometric characteristics, from image collections adapted to the specific biometric characteristics, in particular to the first and / or second biometric characteristics. This can be, for example, a collection of fingerprints or close-ups of eyes, in particular irises and / or retinas.

[0062] The embedding numbers represent typical, learned deviations in the training data. Using these typical, learned deviations and the number of possible states in a byte, 4-byte floating-point numbers of the first biometric characteristics of the first data set or the second biometric characteristics of the third data set can be compressed into 1-byte integers of the compressed first or third data set. Preferably, a byte has 255 possible states. The training data is preferably selected depending on the respective application, in particular depending on the first biometric information. In other words, the training data includes faces, for example, if the first data set and / or third data set includes or is intended to include biometric characteristics of faces. Thus, embedding numbers are obtained depending on the respective application.Preferably, the embedding width, in particular the embedding numbers, are determined as a function of the first data set and / or as a function of the training data. The embedding width, in particular the embedding numbers, are determined in such a way that 1-byte integers are obtained using the 4-byte floating-point numbers of the first biometric characteristics of the first data set or the second biometric characteristics of the third data set and the method according to the invention, in particular since the first and / or third data set has a direct and / or indirect relationship with the training data and / or the embedding width.

[0063] Advantageously, the compression applies to all biometric characteristics, especially the first and / or second biometric characteristics, so that the data remains internally consistent. Thus, each 1-byte integer contains the typical learned deviation.

[0064] To map a floating-point number to the range of embedding numbers, it is quantized. Quantization is the conversion to a discrete value within the range of embedding numbers.

[0065] Step c) comprises in particular the following two sub-steps:

[0066] - Calculation of the embedding width, in particular using embedding numbers,

[0067] - Quantizing a 4-byte floating-point number of the first data set, in particular using the embedding width and the embedding numbers, to obtain a 1-byte integer. The embedding width, which is calculated in the first substep of step c), is in particular the difference between the highest and the lowest embedding number.

[0068] It is possible to calculate the embedding width using the following equation (1 ):

[0069] Embedding width = max(embedding numbers) — min(embedding numbers) (1 )

[0070] The second substep of step c), in which the 4-byte floating-point number is quantized, in particular mapped to the range of embedding numbers, can be described as follows:

[0071] - Subtracting the lowest embedding number from the 4-byte floating-point number, dividing the result by the embedding width to obtain a value in 4-byte format within the embedding range,

[0072] - Dividing the previously obtained result by the 4-byte floating point number,

[0073] - Multiplying the previous result by the number of possible embedding values,

[0074] - Round the previous result to the nearest integer, yielding a 1-byte integer.

[0075] In particular, the second sub-step of step b) may comprise the following equations (2) to (5): x0— Min(embedding number) (2)

[0076] 1 Embedding width x3= x2■ n (4) x4= Round(x3) (5)

[0077] Here, x0 corresponds to the 4-byte floating-point number and x4 to the 1-byte integer. xi to x3 correspond to the intermediate results of equations (2) to (4). The number of possible embedding values ​​is abbreviated as n, where n is preferably 255.

[0078] In particular, the second sub-step of step c), i.e. the quantization of the 4-byte floating-point number, may comprise one or more rounding steps between the arithmetic operations and / or the equations (1) to (5).

[0079] Advantageously, the use of the reduced, scaled integers does not result in a significant reduction in the accuracy of detecting a match between two data sets based on biometric characteristics, in particular first and / or second biometric characteristics, for example in document verification, compared to using the original floating-point numbers. The method according to the invention thus has a high overall accuracy compared to the use of an uncompressed data set, or leads to comparable results in use. The term "overall accuracy" is preferably understood to mean the percentage of identical results of a system that uses a compressed data set compared to a system that uses an uncompressed data set, for example in a code verification method.

[0080] The overall accuracy of the method according to the invention is preferably at least 95%, more preferably at least 97.5%, even more preferably at least 98.9%, particularly with respect to a method using an uncompressed first data set. The overall accuracy is expressed as a percentage and lies between 0% and 100%, with 100% meaning complete agreement between the results of a system using a compressed data set and a system using an uncompressed data set.

[0081] The positive accuracy is preferably at least 80%, more preferably at least 90%, even more preferably at least 98.2%. The negative accuracy is preferably at most 2%, more preferably at most 1%, even more preferably at most 0.5%.

[0082] The term "positive accuracy" describes the ability of a system to accurately detect and locate the actual, correct position of the biometric characteristics, in particular the first and / or second biometric characteristics, on the biometric information, for example, on the face. A system with high positive accuracy has a higher probability of correctly / positively detecting the biometric characteristics, in particular the first and / or second biometric characteristics. A system with low positive accuracy has a higher probability of incorrectly detecting or locating the biometric characteristics, in particular the first and / or second biometric characteristics.

[0083] The term "negative accuracy" describes the error rate or failure rate in the recognition and / or processing of biometric information and / or biometric characteristics, particularly facial recognition or facial processing. It indicates the rate of incorrectly recognized or incorrectly located biometric characteristics, particularly the first and / or second biometric characteristics. A high "negative accuracy" means that the system has a high error rate and has difficulty accurately recognizing and locating biometric characteristics. Positive accuracy and negative accuracy are each expressed as a percentage and range from 0% to 100%.

[0084] In step d), payload data is provided in at least one second data set. The second data set is added to the first data set. In particular, after the addition, the first data set comprises the at least one second data set. Step d) is preferably performed after step c) and before step e).

[0085] The payload may include or be data that specifies the document in which the code is located and / or the person to whom the document is issued. The combination of biometric characteristics and document data and / or personal data guarantees the authenticity and integrity of a document and its connection to its owner.

[0086] It is possible that the payload and / or the second data set contains document data selected from the group consisting of personal data, document type, country of issue, issuing authority, signer identifier, certification number, serial number, check digit, image data, document issue date, signature date, validity period, machine-readable zone (MRZ), URL (Uniform Resource Locator) or combinations thereof.

[0087] It is possible that the user data and / or the second data set contain one or more personal data selected from the group consisting of name, nationality, gender, eye color, height, signature, date of birth, place of birth or combinations thereof.

[0088] It is possible that the payload and / or the second data set is provided as a data type selected from the group consisting of alphanum, string, multistring, date, float, integer, binary or combinations thereof.

[0089] In step e), a first hash value is generated using the compressed first data set or the compressed first data set and the second data set. Step e) is preferably performed after step d) and before step f).

[0090] A hash value, in particular the first and / or second hash value, is obtained using a hash function. One or more data records are entered into the hash function as input. A unique cryptographic hash value, in particular the first and / or second hash value, is generated as output. In particular, the hash function is irreversible. In other words, it is not possible to reconstruct and obtain from the hash value the one or more data records that were provided and entered for calculating the hash value, in particular the first and / or second hash value. Hash values ​​offer the advantage that they guarantee the integrity of data records, in particular by ensuring that the data records have not been and / or cannot be manipulated and / or that no faulty data records exist.

[0091] In particular, a signature is created using the first hash value and a signature algorithm, in particular in step f).

[0092] The one or more data sets for the input can preferably be based on biometric characteristics, in particular on first and / or second biometric characteristics, and / or on user data, more preferably comprising at least the compressed first data set or additionally the second data set.

[0093] Step e) preferably comprises the following sub-steps, more preferably in the specified order: e1) Providing at least the compressed first data set or the compressed first data set and the second data set, and providing a hash function, e2) Input of at least the compressed first data set or the compressed first data set and the second data set into the hash function, e3) Calculating the first hash value from at least the compressed first data set or from the compressed first data set and the second data set, e4) Output of the first hash value.

[0094] The hash value, in particular the first hash value and / or the second hash value, preferably has a unique and / or fixed length of a character sequence of hexadecimal numbers. A SHA-1 function or a SHA-2 function, preferably a SHA-256 function, or a SHA-3 function is preferably provided as the hash function. In particular, a SHA-256 function is provided. A 256-bit hash value is preferably obtained as the hash value, in particular the first hash value and / or the second hash value. The hash value, in particular the first hash value and / or the second hash value, is preferably a hexadecimal number with 64 characters.

[0095] In step f), a signature is generated for the first hash value obtained in step e). Step f) is preferably performed after step e) and before step g).

[0096] It is possible for the first hash value, particularly in step f), to be encrypted asymmetrically. Preferably, a signature is generated using a signature algorithm, particularly using a Digital Signature Algorithm (abbreviated DSA). It is possible for the signature algorithm to use elliptic curve cryptography. Particularly preferably, an Elliptic Curve Digital Signature Algorithm (abbreviated ECDSA) is used as the signature algorithm to generate the signature.

[0097] The Elliptic Curve Digital Signature Algorithm is a cryptographic method for signing data that is based on elliptic curves and generates a signature that includes two signature parameters.

[0098] Preferably, the elliptic curve designated "brainpool256r1" is used to calculate the signature. In particular, the signature comprises a first signature parameter r and a second signature parameter s. The first signature parameter r and the second signature parameter s are preferably each 256 bits long. The data volume for the first and second signature parameters is preferably 512 bits or 64 bytes.

[0099] Preferably, step f) comprises the following sub-steps, which are preferably carried out in the specified order:

[0100] - Selection of an elliptic curve and a base point on the elliptic curve, in particular the curve and the base point being publicly known,

[0101] - Selection of a random, secret integer called the private key, using the private key to calculate a public key, the public key lying on the chosen elliptic curve,

[0102] - Providing the first hash value generated in step e),

[0103] - Selection of a random number as signature key,

[0104] - Calculation of the first signature parameter r, where the first signature parameter r corresponds to the x-coordinate value of the point obtained by multiplying the base point by the signature key,

[0105] - Calculation of the second signature parameter s, whereby the second signature parameter s is obtained by multiplying the inverse of the signature key by the first hash value and subsequently adding the private key, modulating it with the order of the elliptic curve.

[0106] Preferably, the validity of the signature is checked, particularly in a step of the verification process. A valid signature serves as proof of the integrity of the data. The data has then not been tampered with.

[0107] To check the validity of the signature, at least the following sub-steps are preferably carried out, preferably in the specified order:

[0108] - Calculating a second hash value from the compressed first data set or from the compressed first data set and the second data set,

[0109] - Decryption of the signature by using the sender's public key, obtaining the first hash value,

[0110] - Comparison of the newly calculated second hash value and the first hash value.

[0111] "Sender" refers to the original creator of the code and signature, i.e., the person who performed the process, in particular step f) or the substeps of step f), to generate the code. The public key is provided, preferably in verification software or on a server.

[0112] In particular, if the original hash value, in particular the first hash value, matches the newly calculated hash value, in particular the second hash value, the data records have not been tampered with (integrity). If the signature cannot or could not be successfully verified using the sender's public key, the signature is invalid or the wrong public key was used. The hash value for checking the validity of the signature, in particular the second hash value, is preferably calculated using a hash function. In particular, the same hash function as described in step e) is used. The hash function is preferably an SHA-256 function. In particular, the calculation of the second hash value comprises the substeps of step e), whereby the second hash value is obtained instead of the first hash value.

[0113] Preferably, a signature, preferably an ECDSA 256-bit signature, more preferably when using the SHA-256 function, in particular using the curve brainpool256r1 , has a size of 64 bytes.

[0114] In step g), a code is obtained based on the compressed first data set. This code thus enables the biometric data extracted in step b) to be retrieved, evaluated, and verified. Step g) is preferably performed after step f). The code can be generated using a state-of-the-art algorithm.

[0115] The code is preferably present as a geometric and graphic pattern. It is possible for the code to be generated as a machine-readable 2D code, preferably a matrix code, QR code, micro-QR code, DataMatrix code, MaxiCode, Aztec code, JAB code, Han Xin code, a dot code, a stacked code, preferably Codablock, Code 49, PDF417, or a combination thereof. Preferably, the 2D code can be combined with a 1D code. In particular, the code is not a 1D code and / or is not human-readable.

[0116] The code preferably has dimensions of 20 mm x 20 mm.

[0117] The code preferably has 80 modules x 80 modules. A module is the smallest binary graphic unit that makes up a code. For example, in a QR code, a module is a white or black pixel and / or a dot.

[0118] It is possible that the code contains a bug fix.

[0119] Error correction is preferably understood to mean that it is possible to detect and correct errors during the transmission and / or reading of data records from a code. This is achieved by adding redundant data information to the first data record or the compressed first data record. This allows a data record and / or code to contain up to a defined percentage of redundant data. In other words, some information or data elements of the first compressed data record and / or a header and / or the first hash value and / or the signature are present at least in duplicate. The redundant data information can be used by a reading device to detect errors and / or correct them during reading.

[0120] Preferably, the error correction comprises a Reed-Solomon code, which makes it possible to implement a Reed-Solomon error correction method. A Reed-Solomon error correction method makes it possible to detect and / or correct erroneous and / or missing data elements during readout.

[0121] This provides the advantage that the data records contained in the code can be read even if there are errors of up to 7%, 15%, 25% or 30% of the data elements of the data record, depending on the level of error correction.

[0122] There are different levels of error correction, which differ in the percentage of redundant data. Preferably, the code has error correction levels L, M, Q, or H, preferably L, M, or Q. In particular, the code has error correction levels L.

[0123] Level L error correction can correct up to 7% of data elements.

[0124] In other words, up to 7% of the data elements can be missing and / or misread, while the code is still readable, especially 100%.

[0125] Error correction M can correct up to 15% of the data elements. This means that up to 15% of the data elements can be missing and / or misread, while the code is still readable, specifically 100%.

[0126] Level Q error correction can correct up to 25% of data elements, and level H error correction can correct up to 30% of data elements. In other words, up to 25% and / or 30% of data elements can be missing and / or misread, respectively, while the code is still readable, specifically 100%.

[0127] The code comprises at least the compressed first data set or additionally the second data set with the payload data, the first hash value over the compressed first data set or additionally also over the second data set with the payload data, a signature and preferably a header and preferably an error correction. It is possible for the header to be part of the payload data. This makes it possible for the header to be and / or be included in the hash value, in particular the first and / or second hash value. It is possible for the error correction not to be and / or be included in the hash value, in particular the first and / or second hash value.

[0128] A header contains standard information required for a code. This standard information preferably includes information about the document on which the code is located, more preferably the type, version, and country of issue of the document.

[0129] The code can preferably comprise a data volume selected from the range of 1 byte to 520 bytes, more preferably from 32 bytes to 412 bytes. In particular, the data volume includes at least the compressed first data set, the first hash value, the signature, and optionally the header.

[0130] It is possible for the code, especially before it is placed on a document, to be embedded in a two-dimensional discrete complex function G(fx, fy) with an fx frequency coordinate and an fy frequency coordinate, the two-dimensional discrete complex function G(fx, fy) being Fourier transformed, and binarized into a two-dimensional image. The resulting two-dimensional image is placed on the document. A method for verifying such a code includes a further retransformation step, which is performed before the compressed first data set is read from the code.

[0131] This offers the advantage of further increasing the security of the code against forgery, since decrypting the code requires the further step of retransformation before the code can be verified.

[0132] It is possible for the code to be and / or be placed on or in a document. The code is preferably configured such that it can be placed and / or used in a document.

[0133] The code may be provided, provided, and / or used digitally, for example, as a PDF or image file. The code may be provided and / or used digitally on an end-user device, preferably a smartphone, tablet, and / or computer. Alternatively, the code may be provided or used as an embossed and / or printed layer.

[0134] It is possible that the document is or is provided as a certificate, deed, ticket, preferably airline ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, or an official document, preferably visa, residence permit, proof of arrival, temporary ID, residence permit.

[0135] In particular, the document contains one or more personal data selected from the group consisting of name, nationality, gender, photograph, signature, date of birth, or combinations thereof. In particular, the personal information is human-readable and / or machine-readable.

[0136] It is possible that the document contains the payload that is and / or will be included in the code in a human-readable and / or machine-readable format.

[0137] It is possible that the document is provided and / or exists digitally, preferably as a PDF or image file, or physically.

[0138] If the document is physical, it is possible that the document contains one or more security elements selected from the group consisting of security printing, metal layer, RFID chip, optically variable security element, machine-readable element or combinations thereof.

[0139] Preferably, the optically variable security element has an optically active relief structure selected from the group consisting of diffraction grating, hologram, zero-order diffraction structure, blaze grating, macrostructure, in particular lens structure or microprism structure, mirror surface, matt structure, in particular anisotropic or isotropic matt structure.It is possible for the optically variable security element to comprise elements, individually or in combination, selected from the group consisting of a metallic reflection layer, a reflection layer with a high refractive index (HRI layer), a volume hologram, a thin film structure with a color change effect, in particular a Fabry-Perot three-layer thin film structure, liquid crystal material, preferably nematic or cholesteric liquid crystal material, dyes and / or pigments, preferably phosphorescent pigments, luminescent pigments, thermochromic pigments and / or optically variable pigments.

[0140] Alternatively, the document, particularly if it is in digital form, may have at least one or more security elements selected from the group consisting of digital watermarks, electronic and / or digital signatures, asymmetric and / or symmetric encryption, digital signatures, password protection or combinations thereof.

[0141] The code may partially or completely overlap or be overlapped by a security element, particularly when viewed perpendicular to the plane spanned by the document. It is possible that the code is partially or completely enclosed by a decorative element of the document. The code may partially or completely overlap or be overlapped by the decorative element.

[0142] The method, in particular one, two or more steps of the method, for verifying the code can be carried out on an external server and / or locally on a device. It is possible for the device to have no network connection. In particular, the method is configured for verifying the code according to the invention. Preferably, one, two or more steps of the method for verifying the code can be carried out on a device selected from the group consisting of smartphone, camera, tablet, PC, terminal, IoT devices, or combinations thereof. The device has at least one device for image capture and a processor for executing at least one further step of the method.

[0143] In step A), a compressed third data set is created, the compressed third data set being based on second biometric characteristics. The biometric characteristics captured and extracted in step A) are referred to as second biometric characteristics. Preferably, in step A), data, such as a photo or a video, is sent to or created by a server, a client device, or a device.

[0144] The sent or created data is optionally checked in step A). ​​In particular, the image quality, orientation, cropping, and / or sufficient lighting can be checked.

[0145] In particular, it is possible that in step A) at least the following points are checked on the server, the client device or the device:

[0146] - Checking whether at least one biometric information is available,

[0147] - Checking whether a single piece of biometric information or two or more pieces of biometric information are present,

[0148] - Checking whether at least one piece of biometric information is present with dimensions selected from the range of at least 100x100 pixels, preferably at least 80x80 pixels.

[0149] It is possible that the orientation of the biometric information is optionally checked on the server, the client device or the device. In doing so, the orientation is checked in at least one x, y and z direction, whereby in particular the roll-pitch-yaw angles are used. Roll refers to the rotation in the longitudinal direction of the at least one piece of biometric information x-axis or longitudinal axis. Pitch refers to the rotation around the y-axis or transverse axis of the at least one piece of biometric information. Yaw refers to the rotation around the z-axis of the reference system or vertical axis.

[0150] For example, if a face is present as biometric information, the longitudinal axis can be understood as an imaginary axis that enters at the front of the head and exits at the back of the head. The y-axis corresponds to an imaginary axis that runs from one ear to the ear on the opposite side of the head. The z-axis is perpendicular to the x-axis and the y-axis and runs from the top of the head toward the larynx. The roll-pitch-yaw angles can assume values ​​between 0° and 360°. The x-, y-, and z-axes are perpendicular to each other.

[0151] The yaw angle is preferably selected from the range of 324° to 36°, more preferably from 336° to 24°, even more preferably from 348° to 12°. In other words, the deviation from 0° is preferably up to 36°, more preferably up to 24°, even more preferably up to 12°.

[0152] The roll angle is preferably selected from the range of 221° to 299°, preferably from 234° to 286°, more preferably from 247° to 273°. In other words, the deviation from 260° is preferably up to 39°, more preferably up to 26°, even more preferably up to 13°.

[0153] It is possible not to evaluate the pitch angle or to only evaluate the yaw angle and the roll angle.

[0154] The image quality check is considered successful if at least one piece of biometric information is recognized and is present with the minimum number of pixels described above and the orientation of the biometric information is arranged within the angles described above, in particular yaw angle and roll angle.

[0155] If the verification is deemed unsuccessful, i.e., the image quality is not deemed sufficient for successful completion of the subsequent procedural steps as described above, a signal is generated and, preferably, a request is made to re-enter biometric information and / or repeat the verification. The signal can be a visual or acoustic signal and / or a display text. If the verification is deemed successful, the next step or sub-step follows.

[0156] Preferably, in step A), the second biometric characteristics are extracted, in particular using a neural network, in particular wherein the neural network is preferably a convolutional neural network. It is possible for the substep for extracting the second biometric characteristics from step A) to be carried out as described above in step b).

[0157] For example, DeepFace or FaceNet can be used to extract the biometric data, especially in step A), which include convolutional neural networks.

[0158] The biometric information and / or second biometric characteristics can be compressed as already described above, in particular analogously to step c). The third data set is preferably quantized. A compressed third data set is preferably obtained, wherein said third data set comprises 1-byte integers in which the second biometric characteristics are stored. The third compressed data set is based on the second biometric characteristics extracted in step A) from at least one piece of biometric information. As an alternative to step A), the third compressed data set can be provided in step B) by a database, wherein the database comprises at least one third data set. The third data set is preferably based on second biometric characteristics. The third compressed data set preferably comprises 1-byte integers. The third data set is preferably base64 encoded.The database is preferably provided by accessing an external server or is available locally on the device. The compressed third data set preferably has no signature and / or hash value.

[0159] In step h), a code based on first biometric characteristics is preferably provided. The code is preferably generated according to the method for generating a code according to the invention. In particular, the first biometric characteristics correspond to the biometric characteristics captured and extracted in steps a) and b). In step h), the compressed first data set is first read from the code. Subsequently, the compressed first data set based on the first biometric characteristics is compared with the compressed third data set based on the second biometric characteristics.

[0160] Subsequently, in step i), an evaluation parameter for the difference between the compressed first data set and the compressed third data set is selected and calculated. This evaluation parameter makes it possible to assess whether the compressed first data set and the compressed second data set are based on the same person's biometric characteristics.

[0161] The evaluation parameter can, for example, be a distance, preferably in a 2D image space, between the compressed first data set and the compressed third data set. A triplet loss algorithm is preferably used to determine the evaluation parameter.

[0162] The triplet loss algorithm is a machine learning method that preferentially generates embeddings using a specific type of loss function in conjunction with neural networks. Embeddings that reflect semantic similarity between different input examples are preferentially generated. The triplet loss algorithm enables differentiation between similar objects assigned to different so-called classes. The triplet loss algorithm aims to maximize the distance between positive and negative examples in an embedding space, while simultaneously minimizing the distances between examples within the same class.

[0163] Essentially, three images are input to the algorithm: an "anchor" (A), a "positive example" (P), and a "negative example" (N). The algorithm learns to minimize the distance between the anchor and the positive example, since they belong to the same class. On the other hand, the distance between the anchor and the negative example is maximized, since they belong to different classes. The loss of the triplet loss algorithm is the distance between the anchor and the positive example minus the distance between the anchor and the negative example plus a margin. The margin is a constant.

[0164] In particular, the loss of the triplet loss algorithm can be determined by the following equation (6):

[0165] Loss = Distance (A and P) - Distance (A and N) + Margin (6)

[0166] The anchor (A) and the positive example (P) are preferably both images of the same person, and the negative example (N) is another image of a different person. A neural network trained with the triplet loss algorithm causes the neural network to represent faces in a specific way in a 2D image space (“embedding space”). In this 2D image space, faces belonging to the same person are positioned close together, i.e., with a small distance between them. Faces of different people are represented further apart, i.e., with a greater distance between them.

[0167] The triplet loss algorithm is used during neural network training to optimize the embeddings so they are useful for face recognition. The trained neural network is then used to extract the embeddings. The embeddings can then be compared using a Euclidean distance. If the measured distance between the two embeddings is small, this suggests that they are the same person. Similarly, if the distance is large, they are likely different people.

[0168] Embedding is understood to mean a compact representation of data, preferably a clear representation of biometric characteristics, in particular first and / or second biometric characteristics.

[0169] In step j), an assessment is made as to whether the obtained evaluation parameters lie within a specified order of magnitude or value range or do not exceed a specified threshold. If the evaluation parameter is within the specified order of magnitude or value range or below the threshold, a positive assessment is issued. If the evaluation parameter lies outside the specified order of magnitude or value range or above the threshold, a negative assessment is issued.

[0170] A positive assessment means that the first and second biometric characteristics on which the compressed first and compressed third data sets are based match, and it has been verified that they originate from the same person.

[0171] The output of the rating is preferably an optical or acoustic signal and / or a display text, whereby the output differs between positive and negative ratings.

[0172] The method for verifying a code preferably comprises the further step:

[0173] - Checking the validity of the signature.

[0174] This step is preferably performed during or after step A) or B) and before step h). The test is performed as above and may include the substeps listed there.

[0175] In particular, if the verification reveals that the signature is invalid, i.e., that the data has been tampered with, a signal is issued and the process is preferably terminated. The signal can be, for example, a visual or acoustic signal and / or a display text.

[0176] A preferred method for generating a code comprises at least the following steps a), b), c), d), e), f) and g), preferably in the specified order. A preferred method for verifying a code comprises at least the following steps A), h), i), and j) or B), h), i), and j), preferably in the specified order. It is also possible for the method steps to be carried out once or more than once in the method. In particular, method steps can be repeated in the method and / or the method steps of the method can comprise sub-steps. Of course, the above-mentioned material features can also be used equivalently in a method or the above-mentioned method features can be used in the product.

[0177] The invention is explained below using several exemplary embodiments with the aid of the accompanying drawings. The exemplary embodiments shown are therefore not to be understood as limiting.

[0178] Fig. 1 shows a schematic representation of a process of

[0179] Method for generating a code having a compressed first data set based on first biometric characteristics.

[0180] Fig. 2 shows a further schematic representation of a process flow of a method for verifying a code having a compressed first data set based on first biometric characteristics

[0181] Fig. 3 shows a schematic representation of first and / or second biometric characteristics.

[0182] Fig. 4 shows a schematic representation of a document.

[0183] Fig. 1 shows a schematic sequence of the method for generating a code 3, which comprises a compressed first data set based on first biometric characteristics 2'. The code 3 generated in the method according to Fig. 1 can be verified by the method schematically described in Fig. 2, in which second biometric characteristics 2" are extracted and a compressed third data set is created which is based on the second biometric characteristics 2". For verification, the compressed first and third data sets are compared with one another. The method for generating the code 3 has at least the following steps, which are particularly shown in Fig.1: a) capturing at least one piece of biometric information 1 from a real person, a photo and / or a video, b) extracting first biometric characteristics 2' from the at least one piece of biometric information 1, in particular by means of a neural network, in the form of a first data set, c) compressing the first data set to obtain a compressed first data set, d) adding a second data set with payload data to the compressed first data set, e) generating a first hash value via the compressed first data set or via the compressed first data set and the second data set, f) generating a signature for the first hash value, g) generating the code 3 in the form of a geometric pattern, comprising the first compressed data set based on the first biometric characteristics 2', the first hash value and the signature.

[0184] Step a) is preferably the first step of the process and is carried out before step b).

[0185] Fig. 3 shows a schematic representation of a face whose biometric information 1 is being captured. It is also possible for the biometric information 1 to be captured from a fingerprint, hand veins, and / or an eye, preferably an iris and / or a retina. It is possible for the at least one piece of biometric information 1 to be present and / or captured in two-dimensional and / or three-dimensional form. It is possible for the at least one piece of biometric information 1 to be captured by means of a device selected from the group of smartphones, cameras, tablets, PCs, terminals, IoT devices, or combinations thereof. The device has at least one device for capturing images and a processor for carrying out at least one further step of the method.

[0186] It is possible for the device to be a client device, and for the at least one piece of biometric information 1 to be sent over a network to an external server, where one or more or all of the subsequent steps are performed. Alternatively, it is possible for all steps of the method to be performed locally on the device, preferably without a network connection to an external server.

[0187] In step b), biometric characteristics 2' are extracted from at least one piece of biometric information 1, in particular, they are recognized and extracted. Step b) is preferably performed after step a) and before step c).

[0188] It is possible that, as schematically shown in Fig. 3, one or more of the biometric characteristics 2, in particular the first and / or second biometric characteristics 2', 2", are selected from the group consisting of the tip of the nose, the corner of the eye, the eyebrow and the tip of the chin, the contour of the eyes, the contour of the lips, the contour of the nose, loops, arches, swirls and / or minutiae of a fingerprint, veins of a hand and combinations thereof.

[0189] It is possible that the number of biometric characteristics 2, in particular of the first and / or second biometric characteristics 2', 2", is selected from the range from 50 to 200, preferably from 60 to 100, more preferably from 65 to 70. Even more preferably, as shown in Fig. 3, the number 68 is selected for the number of biometric characteristics 2, in particular of the first and / or second biometric characteristics 2', 2". In particular, in step b), a first data set is obtained which is based on all extracted first biometric characteristics 2'.

[0190] It is possible for the biometric characteristics 2, in particular the first biometric characteristics 2' in the first data set or the second biometric characteristics 2" in the third data set, to be described and extracted as vectors and / or as coordinates in a coordinate system. Furthermore, it is also possible for the biometric characteristics, in particular the first and / or second biometric characteristics 2', 2", to be assigned a type or class, for example, the basic pattern of a fingerprint. It is also possible for the biometric characteristics 2, in particular the first and / or second biometric characteristics 2', 2", to be described and extracted as binary data points, preferably from an iris scan.It is possible that the biometric characteristics 2, in particular the first and / or second biometric characteristics 2', 2", are described and extracted as a combination of at least two of the above variants.

[0191] It is possible that step b) for extracting the first biometric characteristics 2' comprises at least the following sub-steps:

[0192] - Identifying and defining a first area to be used for extraction,

[0193] - Recognition of the first biometric characteristics 2' in the first area,

[0194] - Extracting the first biometric characteristics 2' in the first area.

[0195] In particular, the first area is defined using the bounding rectangle method. Here, a rectangle is placed over the at least one piece of biometric information 1, for example, a face, with the first biometric characteristics 2', preferably at least 90%, lying within the rectangle. The rectangle is preferably the smallest circumscribing rectangle of the at least one piece of biometric information 1.

[0196] It is possible for the extraction of the first characteristic 2' in the first area to be performed using a method selected from the group consisting of a holistic method, a method with a restricted local model, a regression-based method, or a combination method of deep learning models and 3D shape models, preferably FacemarkLBF, FacemarkKazemi, FacemarkAAM, or combinations thereof. Preferably, the extraction of the first biometric characteristic 2' in the first area is performed using a combination method of deep learning models and 3D shape models, preferably FacemarkLBF.

[0197] In particular, a method is used that can recognize and track at least one piece of biometric information 1 and / or a first biometric characteristic 2', for example, from a real person, a photograph, a video, and / or a sequence of images. It is possible for the contours and / or centers of eyes, lips, or nose to be recognized and / or tracked. It is also possible for loops, arches, swirls, and / or minutiae of a fingerprint to be recognized and / or tracked. Furthermore, it is also possible for the veins of a hand to be recognized and / or tracked. Preferably, a method is used that utilizes at least one neural network.

[0198] The biometric characteristics 2, in particular the first and / or second biometric characteristics 2', 2", can be recognized and marked, for example, as points and / or coordinates in a piece of biometric information 1, for example a face. In step c), the first data set based on the extracted biometric characteristics 2, in particular the first biometric characteristics 2, 2", is compressed. Step c) is preferably performed after step b) and before step d).

[0199] The first data set based on the extracted first biometric characteristics 2' is preferably an array of 128 floating-point numbers, particularly at the beginning of step c). Each floating-point number preferably comprises 4 bytes. In other words, the first data set comprises a data volume of 128 x 4 bytes, i.e., 512 bytes.

[0200] Preferably, in step c), the data volume of each floating-point number, preferably of the array of 128 floating-point numbers, is reduced. In particular, the data volume of the first data set is reduced by a factor of 4. The data volume of each floating-point number is compressed such that it preferably comprises 1 byte. A compressed first data set is obtained, which preferably comprises 128 x 1 bytes, i.e., 128 bytes.

[0201] It is possible to use the quantization method for compression. Specifically, each floating-point number is reduced to a certain number of discrete values. Quantization preferably results in an integer for each floating-point number.

[0202] In particular, embedding numbers are used to reduce the floating-point numbers and for compression in step c) of the first data set. The embedding numbers are obtained using a deep learning method. LFW images can be used as training data or sample data to determine the embedding numbers.

[0203] It is possible to determine embedding numbers for biometric characteristics 2, in particular for the first and / or second biometric characteristics 2', 2", from image collections adapted to the specific biometric characteristics 2, in particular to the first and / or second biometric characteristics 2', 2". This can be, for example, a collection of fingerprints or close-ups of eyes, in particular irises and / or retinas.

[0204] Using the embedding numbers, 4-byte floating-point numbers of the first biometric characteristics 2' of the first data set or the second biometric characteristics 2" of the third data set can be compressed into 1-byte integers of the compressed first or third data set, respectively. Preferably, one byte has 255 possible states.

[0205] Advantageously, the compression acts on all biometric characteristics 2, in particular for the first and / or second biometric characteristics 2', 2", so that the data remains consistent.

[0206] Step c) comprises in particular the following two sub-steps:

[0207] - Calculation of the embedding width, in particular using embedding numbers,

[0208] - Quantizing a 4-byte floating-point number of the first data set, in particular using the embedding width and the embedding numbers, to obtain a 1-byte integer.

[0209] For example, 1000 LFW images could be quantized with 4 bytes to 1 byte to obtain embedding numbers ranging from -0.52077 to 0.58290. If other LFW images or a different number of LFW images are used, different embedding numbers may be obtained. For the following example, it can be assumed that a first data set is provided, in particular one that contains 4-byte floating-point numbers from a range of 10,000 to 511,000.

[0210] The embedding width can be calculated by the equation (1 ) described above as follows: Embedding width = 0.58290 — (—0.52077) = 1.10367 (1 )

[0211] The second substep of step c) in which the 4-byte floating-point number is quantized, in particular mapped to the range of embedding numbers, can be described as follows:

[0212] - Subtracting the lowest embedding number from the 4-byte floating-point number, dividing the result by the embedding width to obtain a value in 4-byte format within the embedding range,

[0213] - Dividing the previously obtained result by the 4-byte floating point number,

[0214] - Multiplying the previous result by the number of possible embedding values,

[0215] - Round the previous result to the nearest integer, yielding a 1-byte integer.

[0216] In particular, the second sub-step of step b) can be performed using equations (2) to (5) described above. In particular, the second sub-step of step c), i.e., quantizing the 4-byte floating-point number, can include one or more rounding steps between the arithmetic operations and / or equations (1) to (5).

[0217] For example, if the floating point number 10,000 is to be converted into an integer using the exemplary embedding numbers and the determined embedding width, the calculation operation based on equations (2) to (5) described above would be as follows:

[0218] 10,000 - (-0.52077)

[0219] Xi = 9.532532369277048

[0220] 1.10367

[0221] 9.532532369277048 = 0.9532532369277048

[0222] 10,000 x3= 0.9532532369277048- 255 = 243.0795754165647 x4= Rounds(243, 0795754165647) = 243

[0223] The value quantized from the 4-byte floating-point number 10,000 is therefore the 1-byte integer 243.

[0224] For example, if the 4-byte floating point number 511,000 is to be converted into the format of a 1-byte integer, the following results according to equations (2) to (5):

[0225] 511,000 - (-0.52077) = 463.4725687932081

[0226] 1.10367

[0227] 463.4725687932081

[0228] *2 = 0.9069913283624425

[0229] 511,000 x3= 0.9069913283624425 ■ 255 = 231.2827887324228 x4= Rounds(231, 2827887324228) = 231

[0230] The 231 corresponds to the 511,000, but now as a 1-byte integer instead of a 4-byte floating point number.

[0231] The following shows an example of a first data set, as it might appear before step c). It contains 128 floating-point numbers in float format, each consisting of 4 bytes. This first data set therefore comprises 512 bytes.

[0232] -0.05925874 -0.01128396 0.09105307 -0.03127807 -

[0233] 0.06801271 0.00182864

[0234] 0.02190954 -0.03301878 0.1174328 -0.12124632 0.2046176

[0235] -0.03022786

[0236] -0.1650396 -0.03561758 0.02826818 0.04809535 -

[0237] 0.08440463 -0.13857858 -0.04214942 0.12700349 0.0040139 0 .0755067

[0238] 0.04215983 0 02811784

[0239] -0.18864852 0.32442543 -0.10503909 -0.12915151

[0240] 0.04021796 - 0.09991094

[0241] -0.02277122 0.12424234 -0.07866608 -0.02084278

[0242] 0.04711079 0.13625742

[0243] -0.084192 -0.06349151 0.16590628 0 .0514999 -0.1250737

[0244] 0.05737659

[0245] -0.02666266 0.39832145 0.19206464 0 .02395936 0.0366288

[0246] 0.01073211

[0247] 0.14245601 -0.28562331 0.14108598 0.12950768 0.1153863

[0248] 0.03340553

[0249] 0.13990811 -0.05962767 -0.05303854 0.09821685

[0250] 0.12946202 0.14813364

[0251] 0.073308 -0.10102654 0.00718761 -0.13756338

[0252] 0.13108446 0.04731053

[0253] -0.1450232 -0.03344052 0.10073769 -0.18781862

[0254] 0.09244437 0.04796299

[0255] -0.14790541 -0.18098995 -0.24642134 0.1148652

[0256] 0.37924495 0.20492479

[0257] -0.14766243 0.01187875 0.01043329 0 .02256429

[0258] 0.12094933 0.12964407

[0259] -0.01805194 -0.02777201 -0.0671116 0 .06098519

[0260] 0.20639569 - 0.02059942

[0261] -0.08582722 0.20526092 -0.05562493 0.03028861 0.0125723

[0262] 0.00467008

[0263] -0.06393289 -0.0395912 -0.09014382 0.01064388

[0264] 0.04050598 - 0.07019708

[0265] -0.02232645 0.11580256 -0.15094413 0.19078945

[0266] 0.05291561 0 02259036 0.02638829 -0.0248037 -0.12892082 0.01049134

[0267] 0.16260809 -0.25910816

[0268] 0.07229158 0.18393528 0.03622625 0.13120936

[0269] 0.06309938 -0.02336612

[0270] 0.02109955 -0.04417466 -0.20827481 -0.07516982

[0271] 0.03874544 0.03451699

[0272] 0.00411214 0.01026958

[0273] The above first data set is compressed as described in step c), and the floating-point numbers are converted to integers, in particular, with an embedding width specific to the exemplary first data set being determined. The following shows how the above first data set could be present as a compressed first data set after step c), now comprising 128 integers of 1 byte each. This compressed first data set thus comprises 128 bytes.

[0274] 103 114 139 109 101 117 122 109 145 88 166 110 77 108 124

[0275] 128 97 84

[0276] 107 86 118 135 107 124 72 39 92 86 127 93 111 147 98

[0277] 112 128 150

[0278] 97 102 157 129 87 131 111 212 163 123 126 120 151 48 151

[0279] 148 145 125

[0280] 150 103 104 140 86 152 135 93 119 84 148 128 82 109 141

[0281] 72 95 128

[0282] 81 74 58 144 208 166 82 120 119 122 146 148 113 110 101

[0283] 132 166 112

[0284] 96 166 104 124 120 118 102 107 95 119 127 100 112 145 81

[0285] 163 104 122 123 111 86 119 156 55 134 161 126 148 132 111 122 106 67

[0286] 99 126 125

[0287] 118 119

[0288] The first record or the compressed first record is shown below as base64 encoded, with both records being present as float.

[0289] This is intended to further illustrate the difference achieved by compression. For example, a base64-encoded data set from an array of 1284-byte floating-point numbers looks like this: <h2 style=";text-align:left;direction:ltr">

[0290] <h2 style=";text-align:left;direction:ltr"> S71yvVzgOLwIero9bxOAvT9K173 6ru86o3uzPLM+B72cgPA9 / k / 4vUeHUT5 roPe8 JAApvr3 j Eb2 skuc8of 9EPVbcrL2L5w2 + 36QsvTQNAr4 Jh4M7QqOaPc mvLLl iV+Y8 271S 88 / d8RPj g9kr7YeBA+qZ 0EpqdP7D091Ag9EkQPPiU 8dL3xPlm96iX JPbGRBL5YsBc+gSKWPQHnzrOLhus7atOMvgE7Bj 6wyEE99o AUvu34CL2QT849hlNAvnlTvb3XdEQ9hHQXvm5VOb7gVXy+cj 7rPWUswj 701 1E+ 0 j QXvhefQj xk8co8v9i4PEi 09z lqwQQ+qeGTvCCC47 zOcYm9m8t5PWRZ uz4dwK181savvesvUj 731mO90h / 4Paz 8TTx4B5k7P+ + CvWIqIrl Tnbi 9q2M uPJnpJT19w4+ 99uW2vOQp7TOZkRq+T15DPgq+WLlrD7 k8QS zYPCExy7 zSAw S + 3+MrPLyC Jj 7TqYS+nA2UPYhZPD76YRQ9vl sGP j 86gT10ar+ 899isPH3wN L3+RVW+ovKzvYqzHj 2wYQ09Ib+GO75BKDw=<h2 style=";text-align:left;direction:ltr"> <h2 style=";text-align:left;direction:ltr">

[0291] Compared to the above exemplary data set, a compressed first data set comprising 1-byte integers base64 encoded as a float can be exemplary as follows. The data volume and thus the required storage space is reduced by a factor of 4 to 172 characters: Z3KLbWVlem2RWKZuTWx8gGFUal Z2h2t8SCdcVn9db5NicICWYWadgVeDb9S je 35 1 zCXl JF9 Imdo j FaYhl 13V JSAUm2NS F+AUUo 6 kNCmUnh3 epKUcW5 IhK ZwYKZofHh2 Zmt f d39kc JFRo2h6e29Wd5w3hqF+l IRvempDY359dnc=

[0292] Advantageously, the use of the reduced, scaled integers does not result in a significant reduction in the accuracy of detecting a match between two data sets based on biometric characteristics 2 and / or two biometric characteristics 2, for example, in document verification, compared to using the original floating-point numbers. The overall accuracy of the method according to the invention is preferably at least 95%, more preferably at least 97.5%, even more preferably at least 98.9%, particularly with respect to a method with an uncompressed first data set.

[0293] The positive accuracy is preferably at least 80%, more preferably at least 90%, even more preferably at least 98.2%. The negative accuracy is preferably at most 2%, more preferably at most 1%, even more preferably at most 0.5%.

[0294] In step d), payload data is provided in at least one second data set. The second data set is added to the first data set. In particular, after the addition, the first data set comprises the at least one second data set. Step d) is preferably performed after step c) and before step e).

[0295] The payload may include or be data that specifies the document 4 in which the code 3 is arranged and / or a person to whom the document 4 is issued. An exemplary document 4 is shown in Fig. 4. It is possible for the payload and / or the second data set to contain document data selected from the group consisting of personal data, document type, country of issue, issuing authority, signer identifier, certification number, serial number, check digit, image data, document issue date, signature date, validity period, machine-readable zone (MRZ), URL (Uniform Resource Locator), or combinations thereof.

[0296] It is possible that the user data and / or the second data set contain one or more personal data 5 selected from the group consisting of name, nationality, gender, eye color, height, signature, date of birth, place of birth or combinations thereof.

[0297] In step e), a first hash value is generated using the compressed first data set or the compressed first data set and the second data set. Step e) is preferably performed after step d) and before step f).

[0298] The one or more data sets that are entered into the hash function to create the hash value can preferably be based on biometric characteristics 2 and / or user data, more preferably comprising at least the compressed first data set.

[0299] A SHA-1 function or a SHA-2 function, preferably a SHA-256 function, or a SHA-3 function is preferably provided as the hash function. In particular, a SHA-256 function is provided. A 256-bit hash value is preferably obtained as the hash value, in particular the first hash value and / or the second hash value. The hash value, in particular the first hash value and / or the second hash value, is preferably a hexadecimal number with 64 characters. In step f), a signature is generated for the first hash value obtained in step e). Step f) is preferably carried out after step e) and before step g).

[0300] It is possible that the signature algorithm uses an Elliptic Curve Digital Signature Algorithm (ECDSA).

[0301] Preferably, a signature, preferably an ECDSA 256-bit signature, more preferably when using the SHA-256 function, in particular using the curve brainpool256r1 , has a size of 64 bytes.

[0302] In step g), a code 3 is obtained based on the compressed first data set. Code 3 thus enables the biometric information 1 and / or biometric characteristics 2 extracted in step b) to be retrieved, evaluated, and verified. Step g) is preferably performed after step f).

[0303] Code 3 is preferably present as a geometric and graphic pattern. Code 3 can be generated as a machine-readable 2D code, preferably a matrix code, QR code, micro-QR code, DataMatrix code, MaxiCode, Aztec code, JAB code, Han Xin code, a dot code, a stacked code, preferably Codablock, Code 49, PDF417, or a combination thereof. Preferably, the 2D code can be combined with a 1D code. In particular, Code 3 is not a 1D code and / or is not human-readable.

[0304] Code 3 preferably has dimensions of 20 mm x 20 mm. Code 3 preferably has 80 modules x 80 modules. Code 3 may include error correction. Code 3 comprises at least the compressed first data set or additionally the second data set with the payload data, the first hash value over the compressed first data set or additionally also over the second data set with the payload data, a signature, and preferably a header, and preferably error correction. The header may be part of the payload data. This makes it possible for the header to be and / or be included in the hash value, in particular the first and / or second hash value. Error correction may not be and / or be included in the hash value, in particular the first and / or second hash value.

[0305] Code 3 can preferably comprise a data volume selected from the range of 1 byte to 520 bytes, more preferably from 32 bytes to 412 bytes. In particular, the data volume includes at least the compressed first data set, the first hash value, the signature, and optionally the header.

[0306] It is possible for the code 3, particularly before it is placed on a document 4, to be embedded in a two-dimensional discrete complex function G(fx, fy) with an fx frequency coordinate and an fy frequency coordinate, the two-dimensional discrete complex function G(fx, fy) to be Fourier transformed, and binarized into a two-dimensional image. The resulting two-dimensional image is placed on the document 4. A method for verifying such a code 3 comprises a further retransformation step, which is performed before the compressed first data set is read from the code 3.

[0307] It is possible for the code 3 to be arranged on or in a document 4. The code 3 is preferably configured such that it can be arranged and / or used in a document 4.

[0308] It is possible that the Code 3 is available digitally, for example as a PDF or as

[0309] The Code 3 can be present and / or used digitally on an end-user device, preferably a smartphone and / or tablet and / or computer. Alternatively, the Code 3 can be present and / or used as an embossed and / or printed layer.

[0310] Fig. 4 shows an exemplary document 4 on which the code 3 is arranged. Furthermore, the document contains additional personal data 5 and security elements 6.

[0311] It is possible that the document 4 is or is provided as a certificate, deed, ticket, preferably airline ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, or an official document, preferably visa, residence permit, proof of arrival, temporary identity card, residence permit.

[0312] The personal data 5 shown in the document 4 according to Fig. 4 can be selected from the group consisting of name, nationality, gender, photograph, signature, date of birth, or combinations thereof. In particular, the personal data 5 are human-readable and machine-readable.

[0313] It is possible that Document 4 contains the payload data encompassed by Code 3 in a human-readable and / or machine-readable format. It is possible that Document 4 is provided and / or exists digitally, preferably as a PDF or image file, or physically.

[0314] If the document 4 is physically present, it is possible for the document 4 to have one or more security elements 6 selected from the group consisting of security printing, metal layer, RFID chip, optically variable security element, machine-readable element, or combinations thereof. Preferably, the optically variable security element has an optically active relief structure selected from the group consisting of diffraction grating, hologram, zero-order diffraction structure, blaze grating, macrostructure, in particular lens structure or microprism structure, mirror surface, matte structure, in particular anisotropic or isotropic matte structure.It is possible for the optically variable security element to comprise elements, individually or in combination, selected from the group consisting of a metallic reflection layer, a reflection layer with a high refractive index (HRI layer), a volume hologram, a thin film structure with a color change effect, in particular a Fabry-Perot three-layer thin film structure, liquid crystal material, preferably nematic or cholesteric liquid crystal material, dyes and / or pigments, preferably phosphorescent pigments, luminescent pigments, thermochromic pigments and / or optically variable pigments.

[0315] Alternatively, the document 4, particularly if it is in digital form, may have at least one or more security elements 6 selected from the group consisting of digital watermarks, electronic and / or digital signatures, asymmetric and / or symmetric encryptions, digital signatures, password protection or combinations thereof.

[0316] Code 3 may partially or completely overlap or be overlapped by a security element 6, particularly when viewed perpendicular to the plane spanned by document 4. It is possible for code 3 to be partially or completely enclosed by a decorative element of the document. Code 3 may partially or completely overlap or be overlapped by the decorative element.

[0317] In Fig. 2, a schematic representation of a sequence of a method for verifying a code 3, in particular the code 3 is generated according to the method shown in Fig. 1, wherein the code 3 has a first data set based on first biometric characteristics 2'.

[0318] The method shown in Fig. 2 comprises at least one of the following

[0319] Follow steps A) or B):

[0320] A) Creating a compressed third data set, wherein the compressed third data set is based on second biometric characteristics 2", wherein step A) comprises at least the following sub-steps:

[0321] - Capturing at least one piece of biometric information 1 of a real person, a photo or a video,

[0322] - Extracting second biometric characteristics 2", in particular by means of a neural network, from the at least one biometric information 1 in the form of a third data set,

[0323] - Compressing the third data set to obtain a compressed third data set, or

[0324] B) Opening a database comprising at least one compressed third data set, wherein the compressed third data set is based on second biometric characteristics 2".

[0325] Furthermore, the method shown in Fig. 2 further comprises at least the following steps: h) reading out a compressed first data set from the code 3, wherein the compressed first data set is based on first biometric characteristics 2' and comparing the compressed first data set with the compressed third data set, i) generating an evaluation parameter which enables the evaluation of a match between the compressed first data set and the compressed third data set, j) outputting an evaluation based on the generated evaluation parameter. The method according to Fig. 2, in particular one, two or more steps of the method, for verifying the code 3 can be carried out on an external server and / or carried out locally on a device. It is possible that the device does not have a network connection. In particular, the method is configured for verifying the code 3 according to the invention.

[0326] Preferably, one, two, or more steps of the method for verifying the code 3 can be performed on a device selected from the group consisting of a smartphone, camera, tablet, PC, terminal, IoT device, or combinations thereof. The device has at least one device for image capture and a processor for executing at least one further step of the method.

[0327] In step A), a compressed third data set is created, whereby the compressed third data set is based on second biometric characteristics 2". Data sent to or created by the server, the client device, or the device is optionally checked in step A). ​​In particular, the image quality, orientation, cropping, and / or sufficient lighting can be checked.

[0328] In particular, it is possible that in step A) at least the following points are checked on the server, the client device or the device:

[0329] - Check whether at least one biometric information 1 is available,

[0330] - Checking whether a single piece of biometric information 1 or two or more pieces of biometric information 1 are present,

[0331] - Checking whether the at least one piece of biometric information 1 has dimensions selected from the range of at least 100x100 pixels, preferably at least 80x80 pixels. Optionally, the orientation of the biometric information 1 on the server, the client device, or the device can be checked. The orientation is checked along an x, y, and z direction, with particular reference to the roll, pitch, and yaw angles.

[0332] The yaw angle is preferably selected from the range from 324° to 36°, more preferably from 336° to 24°, even more preferably from 348° to 12°. In other words, the deviation from 0° is preferably up to 36°, more preferably up to 24°, even more preferably up to 12°. The roll angle is preferably selected from the range from 221° to 299°, preferably from 234° to 286°, even more preferably from 247° to 273°. In other words, the deviation from 260° is preferably up to 39°, more preferably up to 26°, even more preferably up to 13°. It is possible not to evaluate the pitch angle or to evaluate only the yaw angle and the roll angle.

[0333] The image quality check is assessed as successful if at least one piece of biometric information 1 is recognized and this is present with the minimum number of pixels described above and the orientation of the biometric information 1 is arranged within the angles described above, in particular yaw angle and roll angle.

[0334] If the verification is deemed unsuccessful, i.e., the image quality is deemed insufficient for successful completion of the subsequent procedural steps, a signal is generated and, preferably, a request is made to re-enter biometric information 1 and / or repeat the verification. The signal can be a visual or acoustic signal and / or a display text. If the verification is deemed successful, the next step or sub-step follows.

[0335] Preferably, in step A) of the method according to Fig. 2, the second biometric characteristics 2" are extracted, in particular by means of a neural network, wherein the neural network is preferably a convolutional neural network. It is possible for the substep for extracting the second biometric characteristics from step A) to be carried out as described above in step b).

[0336] For example, DeepFace or FaceNet can be used to extract the biometric information 1 and / or the second biometric characteristics 2", in particular in step A), wherein DeepFace or FaceNet comprise convolutional neural networks.

[0337] The biometric information 1 and / or biometric second characteristics 2" can be compressed in the method according to Fig. 2 as already described above, in particular analogously to step c) of the method according to Fig. 1. Preferably, the third data set is quantized. A compressed third data set is preferably obtained, said third data set comprising 1-byte integers in which the second biometric characteristics 2" are stored. The third compressed data set is based on the second biometric characteristics 2" extracted in step A) from at least one piece of biometric information 1.

[0338] In the method according to Fig. 2, as an alternative to step A), in step B) the third compressed data set can be provided by a database, wherein the database comprises at least one third data set. The third data set is preferably based on second biometric characteristics 2'. The third compressed data set preferably comprises 1-byte integers. The third data set is preferably base64 encoded. The database is preferably provided by accessing an external server or is present locally on the device. Preferably, no signature and / or hash value is present for the compressed third data set. In step h) of the method according to Fig. 2, preferably a code 3 is provided which is based on first biometric characteristics 2'. The code 3 has preferably been generated according to the inventive method for generating a code 3.In particular, the first biometric characteristics 2' correspond to the biometric characteristics 2 that were captured and extracted in steps a) and b) of the method according to Fig. 1. In step h), the compressed first data set is first read from the code 3. Subsequently, the compressed first data set based on the first biometric characteristics 2' is compared with the compressed third data set based on the second biometric characteristics 2".

[0339] Subsequently, in step i), an evaluation parameter for the difference between the compressed first data set and the compressed third data set is selected and calculated. This evaluation parameter makes it possible to assess whether the compressed first data set and the compressed second data set are based on the same person's biometric characteristics 2.

[0340] The evaluation parameter can, for example, be a distance, preferably in a 2D image space, between the compressed first data set and the compressed third data set. A triplet loss algorithm is preferably used to determine the evaluation parameter.

[0341] In step j), an assessment is made as to whether the obtained assessment parameters lie within a predefined order of magnitude or value range or do not exceed a predefined threshold. If the assessment parameter is within the assigned order of magnitude or value range or below the threshold, a positive assessment is issued. If the assessment parameter lies outside the assigned order of magnitude or value range or above the threshold, a negative assessment is issued. A positive assessment is understood to mean that the biometric characteristics 2, in particular the first and second characteristics 2', 2", on which the compressed first and compressed third data sets are based, match, and it has been verified that these originate from the same person.

[0342] The output of the rating is preferably an optical or acoustic signal and / or a display text, whereby the output differs between positive and negative ratings.

[0343] The method for verifying a code 3 preferably comprises the further step:

[0344] - Checking the validity of the signature.

[0345] This step is preferably performed during or after step A) or B) and before step h). The test is performed as above and may include the substeps listed there.

[0346] The hash value for verifying the validity of the signature, in particular the second hash value, is preferably calculated using a hash function. In particular, the same hash function as described in step e) is used. The hash function is preferably an SHA-256 function. In particular, the calculation of the second hash value comprises the substeps of step e), whereby the second hash value is obtained instead of the first hash value.

[0347] In particular, if the check shows that the signature is not valid, i.e. that the data has been manipulated, a signal is output and the method is preferably terminated. The signal can be, for example, an optical or acoustic signal and / or a display text. A preferred method for generating a code 3 has at least the following steps a), b), c), d), e), f) and g), in particular in the specified order. A preferred method for verifying a code 3 has at least the following steps A), h), i), and j) or B), h), i), and j), in particular in the specified order. It is furthermore possible for the method steps to be carried out once or more than once in the method. In particular, method steps can be repeated within the method and / or the method steps of the method can have sub-steps.

[0348] The data sets shown as examples are preferably not to be understood as restrictive, but serve only for illustration purposes.

[0349] Of course, the listed variants can be combined with each other as desired and do not represent any limitation

[0350] List of reference symbols

[0351] 1 biometric information

[0352] 2, 2', 2“ biometric characteristic 3 code

[0353] 4 Document

[0354] 5 Personal data

[0355] 6 Security element

Claims

Patent claims 1. A method for generating a code (3) based on biometric characteristics (2), characterized in that the method comprises at least the following steps: a) capturing at least one piece of biometric information (1) from a real person, a photo, and / or a video, b) extracting first biometric characteristics (2') from the at least one piece of biometric information (1), in particular by means of a neural network, in the form of a first data set, c) compressing the first data set to obtain a compressed first data set, d) adding a second data set with user data to the compressed first data set, e) generating a first hash value using the compressed first data set or using the compressed first data set and the second data set, f) generating a signature for the first hash value, g) generating the code (3) in the form of a geometric pattern comprising the first compressed data set,which is based on the first biometric characteristics (2'), the first hash value and the signature., 2. Method according to claim 1, characterized in that the method of quantization is used for compression in step c).

3. Method according to one of the preceding claims, characterized in that embedding numbers are used for the compression in step c), in particular wherein the embedding numbers are obtained with the aid of a deep learning method.

4. Method according to one of the preceding claims, characterized in that step c) comprises the following two sub-steps: - Calculation of an embedding width, in particular using the embedding numbers, - Quantizing a 4-byte floating-point number of the first data set, in particular using the embedding width and the embedding numbers, to obtain a 1-byte integer.

5. Method according to claim 4, characterized in that the sub-step of step c) in which the 4-byte floating-point number is quantized is described as follows: - Subtracting the lowest embedding number from the 4-byte floating point number, dividing the result by the embedding width to obtain a value in 4-byte format within the embedding range, - Dividing the previously obtained result by the 4-byte floating point number, - Multiplying the previous result by the number of possible embedding values, - Round the previous result to the nearest integer, yielding a 1-byte integer.

6. Method according to one of the preceding claims, characterized in that the first data set at the beginning of step c) is an array of 128 floating-point numbers, each floating-point number comprising 4 bytes, and the compressed first data set, in particular in step c), is an array of 128 integers, each integer comprising 1 byte.

7. Method according to one of the preceding claims, characterized in that the size of the compressed first data set is reduced in step c) by a factor of 4 compared to the first data set.

8. Method according to one of the preceding claims, characterized in that the at least one piece of biometric information (1) of a face, a fingerprint, the hand veins and / or an eye, preferably an iris and / or a retina, is recorded.

9. Method according to one of the preceding claims, characterized in that the at least one item of biometric information (1) is captured by means of a device selected from the group of smartphone, camera, tablet, PC, terminal, IoT devices or combinations thereof.

10. Method according to claim 9, characterized in that that the device is a client device and the at least one piece of biometric information (1 ) is sent via a network to an external server on which one or more or all of the following steps are carried out.

11. Method according to claim 9, characterized in that all steps of the method are carried out locally on the device, preferably without a network connection, in particular a wireless network connection, to an external server.

12. Method according to one of the preceding claims, characterized in that the number of first biometric characteristics (2') is selected from the range from 50 to 200, preferably from 60 to 100, more preferably from 65 to 70, even more preferably that the number 68 is selected for the number of first biometric characteristics (2').

13. Method according to one of the preceding claims, characterized in that the first biometric characteristics (2'), in particular in the first data set, are described as vectors and / or as coordinates in a coordinate system, as a type or class, as binary data points and / or as a combination thereof.

14. Method according to one of the preceding claims, characterized in that step b) for extracting the first biometric Characteristics (2') comprises at least the following sub-steps: - Identifying and defining a first area to be used for extraction, - recognition of the first biometric characteristics (2') in the first area, - Extracting the first biometric characteristics (2') in the first area.

15. Method according to one of the preceding claims, characterized in that the overall accuracy is at least 95%, more preferably at least 97.5%, even more preferably at least 98.9%, in particular with respect to a method with an uncompressed first data set.

16. Method according to one of the preceding claims, characterized in that the positive accuracy of the method is at least 80%, preferably at least 90%, more preferably at least 98.2%, and / or that the negative accuracy of the method is at most 2%, preferably at most 1%, more preferably at most 0.5%.

17. Method according to one of the preceding claims, characterized in that the payload data contain document data selected from the group consisting of personal data, document type, country of issue, issuing authority, signer identifier, certification number, serial number, check digit, image data, document issue date, signature date, validity period, machine-readable zone, URL or combinations thereof, and / or that the payload data and / or the second data record contain one or more personal data (5) selected from the group consisting of name, nationality, gender, eye color, height, signature, date of birth, place of birth or combinations thereof.

18. Method according to one of the preceding claims, characterized in that the first hash value is obtained by a hash function, wherein a SHA-1 function or a SHA-2 function, preferably a SHA-256 function, or a SHA-3 function is provided as the hash function.

19. Method according to one of the preceding claims, characterized in that step e) comprises the following sub-steps, preferably in the specified order: e1) providing at least the compressed first data set and providing the hash function, e2) inputting at least the compressed first data set into the hash function, e3) calculating the first hash value from at least the first compressed data set, e4) outputting the first hash value.

20. Method according to one of the preceding claims, characterized in that the signature is generated by an elliptic curve digital signature algorithm. 21 . Method according to one of the preceding claims, characterized in that the code (3) is a machine-readable 2D code, preferably a matrix code, QR code, Micro-QR code, DataMatrix code, MaxiCode, Aztec code, JAB code, Han-Xin code, a dot code, a stacked code, preferably Codablock, Code 49, PDF417, or a combination thereof, and in particular combined with a 1D code.

22. Method according to one of the preceding claims, characterized in that the code (3) contains a data set selected from the range of 1 byte to 520 bytes, preferably from 32 bytes to 412 bytes.

23. Method according to one of the preceding claims, characterized in that the code (3) is embedded in a two-dimensional discrete, complex function G(fx, fy) with an fx frequency coordinate and an fy frequency coordinate, the two-dimensional, discrete, complex function G(fx, fy) is Fourier transformed, and binarized to a two-dimensional image.

24. Method according to one of the preceding claims, characterized in that the code (3) is arranged on or in a document (4), wherein the document (4) is provided as a certificate, deed, ticket, preferably airline ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, an official document, preferably visa, residence permit, proof of arrival, temporary ID, residence permit.

25. Method for verifying a code (3), in particular a code (3) generated according to claim 1, characterized in that the method comprises at least one of the following steps A) or B), A) Creating a compressed third data set, wherein the compressed third data set is based on second biometric characteristics (2"), wherein step A) comprises at least the following sub-steps: - capturing at least one piece of biometric information (1 ) of a real person, a photo or a video, - Extracting second biometric characteristics (2"), in particular by means of a neural network, from the at least one biometric information (1) in the form of a third data set, - Compressing the third data set to obtain a compressed third data set, or B) Opening a database comprising at least one compressed third data set, wherein the compressed third data set is based on second biometric characteristics (2"), the method further comprising at least the following steps: h) Reading a compressed first data set from the code (3), wherein the compressed first data set is based on first biometric characteristics (2') and comparing the compressed first data set with the compressed third data set, i) Generating an evaluation parameter that enables the evaluation of a match between the compressed first data set and the compressed third data set, j) Outputting an evaluation based on the generated evaluation parameter.

26. Method according to claim 25, characterized in that in step A) at least the following points are checked on the server, the client device or the device: - Checking whether at least one biometric information (1 ) is available, - Checking whether a single piece of biometric information (1) or two or more pieces of biometric information (1 ) are present, - Checking whether the at least one piece of biometric information (1) is present with dimensions selected from the range of at least 100x100 pixels, preferably at least 80x80 pixels.

27. Method according to one of claims 25 to 26, characterized in that the orientation of the at least one item of biometric information (1) is checked on the server, the client device or the device, wherein in particular the orientation about at least one x, y and z direction is checked.

28. Method according to one of claims 25 to 27, characterized in that the roll-pitch-yaw angles are used for the test, in particular wherein the yaw angle is selected from the range from 324° to 36°, more preferably from 336° to 24°, even more preferably from 348° to 12°, or in particular wherein the roll angle is selected from the range from 221° to 299°, preferably from 234° to 286°, more preferably from 247° to 273°.

29. Method according to one of claims 25 to 28, characterized in that in step A) the second biometric characteristics (2") are extracted by means of a neural network, wherein the neural network is preferably a convolutional neural network.

30. Method according to one of claims 25 to 29, characterized in that that the evaluation parameter is a distance, preferably in a 2D image space, between the compressed first and compressed third data sets.

31. Method according to one of claims 25 to 30, characterized in that the method comprises the further step: - Checking the validity of the signature.

32. Document (4) comprising a code (3) generated according to claim 1.

33. Document (4) according to claim 32, characterized in that the document (4) is present digitally, preferably as a PDF or as an image file, or physically, in particular as an embossing and / or as a printing layer.

34. Document (4) according to one of claims 32 to 33, characterized in that the document (4) is a certificate, document, ticket, preferably a flight ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, an official document, preferably a visa, residence permit, proof of arrival, temporary ID card and / or residence permit.

35. Document (4) according to one of claims 32 to 34, characterized in that the document (4) comprises one or more personal data (5) selected from the group consisting of name, nationality, gender, eye colour, height, photograph, signature, date of birth, place of birth or combinations thereof, in particular wherein the personal data (5) are human-readable and / or machine-readable.

36. Document (4) according to one of claims 32 to 35, characterized in that the document (4) has the useful data comprised by the code (3) in a human-readable and / or machine-readable manner.

37. Document (4) according to one of claims 32 to 36, characterized in that the document (4) has one or more security elements (6) selected from the group consisting of security printing, metal layer, RFID chip, optically variable security element, machine-readable element or combinations thereof, when it is physically present, or that the document (4) has at least one or more security elements (6) selected from the group consisting of digital watermarks, electronic and / or digital signatures, asymmetric and / or symmetric encryptions, digital signatures, password protection or combinations thereof, when it is digitally present.

38. Document (4) according to one of claims 32 to 37, characterized in that the optically variable security element (6) has an optically active relief structure selected from the group consisting of diffraction grating, hologram, zero-order diffraction structure, blaze grating, macrostructure, in particular lens structure or microprism structure, mirror surface, matt structure, in particular anisotropic or isotropic matt structure, and / or that the optically variable security element (6) comprises elements, individually or in combination, selected from the group consisting of metallic reflection layer, reflection layer with high refractive index, volume hologram, thin film structure with color change effect, in particular Fabry-Perot three-layer thin film structure, liquid crystal material, preferably nematic or cholesteric liquid crystal material, dyes and / or pigments, preferably phosphorescent pigments, luminescent pigments, thermochromic pigments and / or optically variable pigments.

39. Document (4) according to one of claims 32 to 38, characterized in that the code (3) partially or completely overlaps with or is overlapped by a security element (6), in particular when viewed perpendicular to the plane spanned by the document (4) and / or that the code (3) is partially or completely enclosed by a decorative element.

40. Use of the code (3) generated according to claim 1 in a document (4), wherein the document is selected from the group consisting of certificate, deed, ticket, preferably airline ticket, train ticket, event ticket, concert ticket, conference ticket, admission ticket, an access control, an official document, preferably visa, residence permit, proof of arrival, temporary ID, residence permit or combinations thereof.

Images