Methods for determining the originality and classification of a product

Abstract

Method for determining the authenticity of a product, comprising the steps of: - Providing a dynamically modifiable three-dimensional (3D) security feature and a product-specific machine-readable two-dimensional (2D) code on the product or product packaging, wherein the surface of the 3D security feature has unique, characteristic, optically detectable security features resulting from the surface texture, the distribution of colors, blurring, and / or contrasts on the surface; - Reading 100 to 10,000 surface features from several selected unique minutiae and / or characteristics of the surface of the dynamically modifiable 3D security feature using an optical detection device in an area no longer perceptible to the unarmed human eye for characterizing the surface.- Formation of at least one individual feature matrix characteristic of the product from the read surface features of the unique minutiae and / or characteristics of the surface of the dynamically changeable 3D security feature using an algorithm or rule; - Storing the at least one feature matrix characteristic of the product directly or indirectly in a database or as a code applied to the product or the product packaging, whereby the following steps are carried out during the identification of the product: - Reading the surface of the dynamically changeable 3D security feature and its minutiae and / or surface characteristics with any optical detection device and forming at least one feature matrix with the same algorithm or rule as for the stored feature matrix.- Verifying the surface by comparing at least a part of the feature matrix of the read-in dynamically modifiable 3D security feature with at least a part of the individual feature matrix stored as a code in the database or the product or product packaging, whereby the verification is positive if the compared feature matrices agree at least partially within a predefined tolerance range, where the degree of agreement of the compared feature matrices is at least 90%, - only with a positive verification is the 3D security feature stored in the database supplemented or replaced by the read-in dynamically modifiable 3D security feature.so that the read 3D security feature corresponds to the current security feature in the database for later verification and the machine-readable two-dimensional (2D) code characteristic of the product is read from the product or the product packaging.

Description

[0001] The present invention relates to an improved method for determining the originality of a product in order to avoid product counterfeiting or product manipulation.

[0002] Product counterfeiting and imitations of genuine products pose a global problem. According to a study by the trade organization OECD and the European Union Intellectual Property Office (EUIPO), counterfeit goods worth at least €460 billion are traded worldwide, with approximately 25% of this occurring within the EU. Not only consumer goods such as handbags and watches are copied, but also high-tech products like automotive parts, pharmaceuticals, and payment methods. With high-priced watches, not only the timepiece itself is counterfeited, but often the accompanying certificates as well. Counterfeit electronic devices or safety-relevant automotive components such as brake pads rarely meet safety standards. In particular, counterfeit pharmaceuticals and food products are a very serious global problem. In the worst-case scenario, counterfeit products threaten the health or even the life of the user.

[0003] Not only are the products themselves counterfeited, but also serial numbers and two-dimensional codes, such as QR codes. A QR code consists of a square matrix of black and white squares, which represent the encoded data in a binary system. A square marker in three of the four corners indicates the orientation of the QR code. To read a QR code, a digital image of the encoded data is first created, and then the encoded data contained in the image is converted into text. Nowadays, QR codes can be conveniently read with a smartphone. However, QR codes themselves are also counterfeited and are frequently the target of cyberattacks, as they can be easily copied or replaced with malicious code.For example, users are not directed to the original website, but to a website that allows phishing or contains malware that can potentially be installed on the user's device.

[0004] There is still no reliable system to ensure the authenticity of a QR code before it is scanned. The damage caused by counterfeit QR codes is immense. The same applies to other two-dimensional machine-readable codes affixed to a product or product packaging for tracking or product description purposes.

[0005] Dynamic QR codes carry the additional risk that they can change over time, for example, to redirect from one URL (Uniform Resource Locator; internet address) to a new one. Therefore, the user doesn't know whether the current dynamic QR code actually corresponds to the original. Overall, a technology offering an additional security hierarchy to protect machine-readable two-dimensional codes is lacking.

[0006] Counterfeiters often swap genuine goods for counterfeit ones within the supply chain. Manufacturers lack the ability to control the entire supply chain and make it verifiable for the customer. Trade in counterfeit products from non-EU countries into the EU accounts for over 6.8% of imports. While static and dynamic security features are known for authentication, these are not designed for two-dimensional codes, such as those found on product labels or packaging.

[0007] EP 1 982 296 B1 describes a method for reading information from a multilayer body with an optically machine-readable identifier, in which an electronic image is generated from a relief structure using an optoelectronic reading device to increase security. The relief structures have different areas that are provided with an optically effective reflective layer or separating layer.

[0008] EP 2 039 527 B1 describes a security element and a method for verifying such a security element, comprising an optically readable, machine-readable code that is not visible to the unarmed eye and is introduced by laser engraving by removing the reflective layer in the area of ​​the code.

[0009] German patent DE 10 2016 218 545 A1 describes an identification feature for identifying an object with a random structure formed from fragments of a colorant or from particles of at least one dye. For example, within a QR code, the random structure is formed from pigments as the identification element. For optical perception, the identification element must be activated by an energy input, for example, by UV light.

[0010] A similar method is described in DE 10 2015 219 400 B4, in which an identification feature is also formed within a QR code, but this feature must also be made visible for reading, for example, using fluorescent or phosphorescent dyes. Alternatively, various dispersing agents can be used, as described, for example, in DE 10 2015 219 399 B4. DE 10 2015 219 396 B4 describes the use of dye particles whose surface consists of a polymer, such as chitosan alginate, cellulose, resin, polynucleic acid, polysaccharide, or polystyrene. DE 10 2015 219 395 B4 describes how the random structure of the identification element can be created by coagulation of neighboring pixels.

[0011] German patent DE 10 2008 016 435 A1 describes an adhesive label with an area that has a unique surface structure, wherein at least one product code and a digital signature are affixed to the adhesive label. The method provides that the surface structure of the area is first captured, and then a digital signature is generated based on a security feature derived from the captured surface structure, a product code assigned to the object, and a secret signature key. Subsequently, the product code and the digital signature are applied to the adhesive label, and the adhesive label is then affixed to the object. The security features can be applied to the adhesive label as a machine-readable code.

[0012] However, these solutions are not satisfactory for use in a product line or supply chain. To verify a product's authenticity, it is necessary that all participants in the supply chain, right up to the end customer, have the ability to check the product's authenticity. In the simplest case, this should be done via smartphone, i.e., through optical scanning of security-relevant features and subsequent data comparison to verify the security feature.

[0013] Against this background, the object of the present invention is to provide an improved method that makes it possible to determine the originality of a product within a supply chain cost-effectively and safely for everyone at any time.

[0014] This problem is solved by a method having the features of claim 1. Preferred embodiments are found in the dependent claims.

[0015] The method according to the invention is based on security features that are not detectable by the naked eye and are essentially attributable to the individual surface structures of a security element. "Naked eye" refers to the ability to detect a security feature without aids, i.e., detectability with the naked eye. Possible aids include known optical detection or reading devices, such as cameras or other optics. The invention takes advantage of the fact that every surface is unique and can be assigned a distinct identity. Even if a label or product packaging is printed on a printing line by a specific printer, each individual printed product differs due to minute macro-features on its surface that can be detected by an optical detection device, such as a camera.According to the invention, the term "product" includes any kind of goods, documents, certificates, printed matter, means of payment, valuables or things.

[0016] The detection of security features preferably takes place in an area imperceptible to the naked human eye, preferably in the macro to micro range, ideally in a range of 1 mm to 1 µm, and requires that the surface to be detected possesses unique, characteristic, optically detectable security features resulting from the surface texture, the distribution of colors, blurring, and / or contrasts on the surface. Each individual character generated contains minute inaccuracies at the pixel level, which manifest as noise or blurring. The characters can be generated, for example, by printing, engraving, machining, or demixing. Engraving a 2D code, such as a serial number, can be done, for example, with a laser. Machining includes, for example, grinding the surface.One example is characteristic machining marks on a metallic surface. This separation can occur, for instance, during a drying process. In addition to color printing, this also includes the 3D printing of security features and their associated characteristics.

[0017] The invention enables both the characterization of products and the classification of machines, such as all possible types of manufacturing or processing machines. In characterization, the characteristics of each product are recorded individually, so that even counterfeits printed on the same printing press can be detected. In classification, different characteristics within a machine are assigned to the same machine. This means that products manufactured or processed on other machines are identified as counterfeits. This is relevant, for example, for printing presses, where the characteristics of each individual printed product (e.g., a label) are determined, so that even counterfeits printed on the same printing press can be detected during the characterization process. Classification, in turn, makes it possible to determine whether a printed product was printed on the same printing press.Such differences are due to the automated production process. This means that, based on product-specific surface characteristics, it is possible to determine on which machine a product or product line was manufactured, printed, or otherwise processed.

[0018] The invention is based on the realization that the process product is different and therefore unique in every manufacturing or processing operation, even when using the same machine.

[0019] Each generated 3D security feature is distinguished by characteristic features attributable to the specific product. According to the invention, these structural features are referred to as "minutiae" at the macro level, meaning they are the finest detectable features of the surface. Furthermore, the detection of the security element's surface also includes surface characteristics; that is, the invention also takes into account the diffraction or reflection behavior of light, or any coatings or artifacts that may be present due to the detection optics but can be detected by the algorithm used in the inventive method.

[0020] The surface can refer, for example, to an area of ​​a printed product (e.g., label, stamp, foil, adhesive tape, packaging, banknote, or poster) or to areas of the product itself that are highly characteristic of the product. Even the three-dimensional structure of an engraving, such as an engraved serial number, exhibits a three-dimensional pattern that can be optically detected at the micro or macro level and can serve as a security feature. Products that have undergone a manufacturing or processing procedure also differ individually based on minute structural features or surface properties that can be detected using appropriate optics, such as those built into cameras or screening systems. Optical detection devices, such as those already present in standard smartphones or other mobile communication devices (tablets, laptops, etc.), are preferably used.

[0021] The method according to the invention comprises, in a first step, providing a dynamically modifiable three-dimensional (3D) security feature and a machine-readable two-dimensional (2D) code characteristic of the product on the product or the product packaging. The 2D code can be applied directly to the product itself or indirectly, for example, in the form of a label. Preferably, the code is engraved, printed, or otherwise applied directly to the product or the product packaging. The surface of the 3D security feature exhibits unique, characteristic, optically detectable security features resulting from the surface texture, the distribution of colors, blurring, and / or contrasts on the surface.

[0022] In a further process step, unique minutiae and / or characteristics of the surface of the dynamically changeable security feature in one or more areas are read using an optical detection device in the macro range. Preferably, between 100 and 10,000 surface features, more preferably between 500 and 2000 features, are read.

[0023] The macro range refers to a resolution of 1 mm to 1 µm, i.e., a range where not all details are visible to the human eye without aids. Depending on the application, this range may also fall within the micro range; therefore, the term "macro range," as used in the present invention, also includes the micro range or parts thereof. The detectability of details to the human eye decreases with increasing distance; that is, the characteristics of a distant poster are not readily visible without aids. When using an aid, such as optics, such details become detectable, allowing minutiae or characteristics to be defined. Accordingly, the technology can also be used to detect minutiae or other characteristics of a poster or other surface from a greater distance (e.g.,between 20m and 100m) to read in order to generate a feature matrix.

[0024] According to the invention, “minutiae” refers to the finest features or structural elements on the surface, the surface itself, or other features that characterize the surface.

[0025] A "dynamically modifiable 3D security feature" refers to a security feature that has a three-dimensional structure and can change over time. Such changes can occur, for example, in the form of color changes, structural changes (e.g., creases, bumps, etc.), or through overpainting. Color changes, fading, or changes in color contrast also fall under this category.

[0026] In a preferred implementation, the 3D security feature undergoes a change in its surface between two separate points in time; that is, the minutiae or characteristics of the surface of the 3D security feature are partially changed, while other areas remain unchanged. The modified security feature can then be stored in the database as a new, current security feature.

[0027] After the unique, individual minutiae and / or surface characteristics selected for each product have been generated, at least one individual feature matrix characteristic of the product is created, preferably one or more feature vectors. According to the invention, the creation of at least one individual feature matrix characteristic of the product, or preferably one feature vector, from the selected unique minutiae and / or surface characteristics of the read surface of the dynamically modifiable 3D security element is carried out using an algorithm or a rule. This includes the use of a code, for example, a hash code. Subsequently, the at least one feature matrix characteristic of the product is stored in data form or other form in a database or as a code directly on the product or the product packaging.The code is preferably applied by further printing on the product or the product packaging. The printed image from this printing process can also be used as a 3D security feature and processed in accordance with the invention.

[0028] In a preferred variant, the feature matrix is ​​not stored directly in a database or integrated into code, but indirectly, for example, in the form of an image, preferably an image file. Furthermore, it can be provided that the stored data is kept in encrypted form in the database or the code. The feature matrix can then be extracted from the stored image during a query.

[0029] According to the invention, the following steps are performed to identify the product within or at the end of a supply chain. For verification, the surface of the dynamically modifiable 3D security feature, including its minutiae and / or surface characteristics, is first scanned using any optical scanning device, and at least one feature vector is generated. The same algorithm or rule is applied as described above for the initial mapping of the dynamically modifiable 3D security feature. The optical scanning device can be any smartphone, tablet, or other camera.

[0030] The surface is then verified by comparing the at least one feature vector of the read-in, dynamically modifiable 3D security feature with the individual, at least one feature vector stored in the database. Verification is successful if the compared feature vectors match within a predefined tolerance range. This tolerance range defines the degree of match and is preferably at least 90%, preferably between 95% and 99%.

[0031] Preferably, access to a data room is only granted after successful verification. For example, redirection to a URL via a QR code could only occur after successful verification, i.e., after confirmation that it is the original. The data room preferably contains product-related data, such as its serial number, manufacturing date, quality assurance information, etc.

[0032] To enhance security, the invention not only analyzes the surface at the macro or micro level, but in a preferred embodiment, it also employs dynamic recognition. This means that upon successful verification, the 3D security feature stored in the database is replaced or supplemented for subsequent verification by the read-in, dynamically modifiable 3D security feature. Dynamic security can also incorporate other factors, such as a geotag, a time factor, lighting and shadow conditions during the recording, environmental features, or weather data. The use of different scanning optics (e.g., from different smartphones) can also generate dynamic security features.

[0033] The method according to the invention makes it possible to analyze an object based on its unique surface structures and surface characteristics, so that a unique identity is calculated from the acquired surface information using the algorithms or rules employed. Each product and each resulting feature matrix is ​​thus unique, similar to a fingerprint. For verification, the read surface information or the feature matrix generated from it is compared with a database in which the surface information or the resulting feature matrix is ​​stored.

[0034] A key feature of the dynamic security component is that while optical alterations within a predefined tolerance range are permitted, the security check is performed dynamically. This means that as soon as a comparison or verification is successful, the old entry in the database is supplemented or replaced by the new one that has just been read. Even if a counterfeiter were in possession of the original security feature, it would be of little use to them, as the feature currently stored in the database has already changed. The result is a multidimensional, dynamic identity.

[0035] This makes the invention far more secure than conventional static methods and also distinguishes it through the robustness of the system. Due to the multidimensional dynamic identity used in the present invention, it is possible to generate individual product identifiers and verify them during database queries. It is impossible for a counterfeiter to imitate the minutiae and characteristics of the surface or the feature matrices used, as such an undertaking would only be possible with considerable, highly technical, and time-consuming effort. Counterfeiting attempts are therefore no longer worthwhile. The counterfeiter must assume that the currently valid security feature has changed in the meantime. Verification with an old feature would then yield no positive result, meaning the counterfeit would be detected during the database query.

[0036] Preferably, the machine-readable 2D code characteristic of the product, located on the product or its packaging, is only read after successful verification. This means that before a user is redirected to a URL or data room via a machine-readable code (e.g., a QR code), verification must first take place, and only then is the machine-readable code read. This prevents, for example, unintentional redirection to a data room from which malware is being transmitted.

[0037] As already explained at the outset, every machine-generated result (e.g., the printed image from a printing press) is unique. The printing process, for example, results in the surface of the dynamically modifiable 3D security feature having an inhomogeneous ink distribution. According to the invention, this is exploited to generate the minutiae and thus the derived feature vectors. This allows the inventive method and its algorithms to be used for the classification of a machine, such as a printing press or processing machine (e.g., grinding machine, engraving machine, laser). That is, it can be determined beyond doubt on which machine a product (e.g., a printed product) was manufactured. The invention can therefore also be used in the field of solar cells or high-quality watches to determine whether the ID numbers engraved (e.g., with a laser) are genuine.

[0038] When a sufficiently large number of products within a product line are analyzed, it is preferably provided that the feature matrices or feature vectors are also adapted or optimized to ensure the individuality of each product and its security features. This adaptation can be achieved using artificial intelligence (AI), for example, by applying a machine learning method. This allows the system to optimize itself because, within a product line, unique minutiae and / or surface characteristics within the 3D security feature are generated for each individual product, resulting in a unique feature matrix characteristic of that product. Preferably, the AI ​​can be combined with a rule, i.e., the rule can specify certain conditions for the adaptation or optimization process.

[0039] At any point in the supply chain, it is possible to have the product verified, for example, at various logistics stations, at the retailer, or at the end customer's location. Therefore, in a preferred embodiment, the inventive method comprises multiple readings, verifications, and, if necessary, replacements or additions to the dynamically modifiable 3D security feature characteristic of the product within the database within a supply chain or lifecycle. Security is particularly high when dynamic integrity is applied, i.e., when the 3D security feature is replaced during the reading process within a supply chain by a modified security feature that has changed slightly in some areas while remaining unchanged in others.Preferably, upon successful verification, the 3D security feature stored in the database is supplemented or replaced by the dynamically modifiable 3D security feature that has been read in, so that the read 3D security feature corresponds to the current security feature in the database for later verification. The authorized replacement of a previous security feature could also be included in this dynamic process.

[0040] Preferably, the machine-readable 2D code is a Data Matrix code, a QR code, an RFID tag, an NFC tag, or a barcode. The 3D security feature can be located either outside or inside the machine-readable 2D code. Either only areas of the 3D security feature can be read to generate the minutiae and / or surface characteristics, or a combination of the 3D security feature and areas of the machine-readable 2D code can be used. Preferably, the dynamically modifiable 3D security feature and the product-specific machine-readable 2D code are provided on a printed label. To increase security, the data stored in the database for the dynamically modifiable 3D security feature can also be secured via a blockchain.

[0041] The invention is explained in more detail in the following drawings. Fig. Figure 1 shows the application of the inventive method using the example of a postage stamp equipped with a security feature according to the invention. Fig. Figure 2 shows another embodiment in which the labeling of a product package is used for reading, Fig. Figure 3 shows a QR code with an integrated security feature. Fig. Figure 4 shows the formation of minutiae and their conversion into a feature vector. Fig. Figure 5 shows a supply chain with the possibility of reading the security feature at different stages of the supply chain. Fig. Figure 6 shows a schematic example of pressure tolerances with visible distinguishing features for the formation of a feature matrix.

[0042] Fig. Figure 1 shows a postage stamp with a standard Data Matrix code as a 2D code. Alternatively, it could of course be a QR code or other machine-readable 2D code. Historically, a postage stamp has served as both a means of payment and a receipt. The amount of money on the stamp is intended to cover the costs of transporting the letter or goods. With digitalization, it is necessary to connect a physical postage stamp with the digital world to create added digital value. Therefore, counterfeit protection must be ensured and illegal reuse prevented.

[0043] As part of the printing process, each individual stamp is unique. Even slight blurring and color variations are enough for each stamp to be an individual carrier of distinctive security features. In the version shown, the 2D code affixed to the stamp serves as a security feature, which is read by an optical recognition device, in this case a smartphone. Due to the printing-related variations, each 2D code can be individually identified as a three-dimensional security feature, even if the stamps were printed on the same printing press with the same ink. This not only prevents the stamp from being used twice or counterfeited, but also gives it its own physical identity. This physical identity is a key to the digital identity and the associated digital content.This is because the contents can only be read if the original stamp is also physically present.

[0044] The original is only considered to exist if the read dynamic security feature matches the security feature originally stored in the database within a tolerance range. The tolerance range score is individually configurable and preferably lies between 95 and 100%. In the implementation shown, upon successful verification, the user gains access to a data room, i.e., to data or information that may be related to the stamp.

[0045] Due to the subtle surface differences, it is practically impossible for a forger to produce a duplicate of this stamp given the time and resources available to them. The generation of the minutiae and / or surface characteristics is based on unique, individual features and the application of an algorithm, resulting in a unique feature matrix for each stamp. Only a match with this feature vector, or parts thereof, provides positive verification. This ensures the user that the stamp is not a forgery.

[0046] The algorithms according to the invention can be used in conjunction with conventional smartphone cameras, allowing each stamp to be individually scanned and analyzed at macro or micro levels. The invention enables the unambiguous characterization of stamp surfaces, and depending on the application, the scanning process can be automated, making mass-market applications feasible. This allows, for example, the efficient scanning of large quantities of stamps (e.g., 100,000 stamps). Ideally, this is done after the printing process or following the cutting process, depending on what makes more sense in a production line.

[0047] The security feature can be printed on a white background. However, various colors or color schemes can also be used to improve counterfeit protection. Equipping a machine-readable 2D code with a security feature creates a secure code. The functionality of the machine-readable 2D code remains intact, while simultaneously allowing verification that it is an original code. This also protects the code itself from copying or manipulation.

[0048] In Fig. Figure 2 shows another application example. In this case, the labels are paper labels from a roll, each bearing a product name. Enlarging a letter on one of the labels reveals that the printed image is highly inconsistent and characterized by blurring, even when using the same printer and ink. Thus, each individual printed product is unique; that is, each label and each character on the paper roll is individual and can be identified by its minutiae and surface characteristics. In the close-up of the selected character "P," it is evident that the ink distribution is inhomogeneous due to machine tolerances. This pattern can therefore be used to create minutiae and / or surface characteristics, ultimately resulting in feature matrices that are unique to each individual label.These feature matrices are then stored in the database for each individual label. Every single square millimeter of a surface possesses unique surface structures and properties. When creating the minutiae and / or characteristics of the surface, such fine features are captured, analyzed, and ultimately stored as a feature matrix in data form in a database. Alternatively, the data can also be stored directly on the product or product packaging itself, for example, as a subsequently printed code.

[0049] Every surface inevitably undergoes changes over time, representing a dynamic process. The degree of permissible change can be defined in a rule, so that when the security feature is read, a score is determined that indicates whether it is the original product or not. The score calculation involves a metric evaluation that incorporates an actual value, a target value, and a tolerance indicator. Additional parameters can also be included in the evaluation and score calculation, such as their relationship to other factors, for example, time elapsed, geographical distance, brightness, vibrations, and changes in temperature and / or humidity.

[0050] To increase counterfeit protection, the dynamic changes to the surface can be incorporated by recording the changing security feature as a new security feature in the database or adding to the existing one at the time of querying. This allows changes to the security feature to be documented. Additionally, the system is capable of learning, meaning that a machine learning method can be used to optimize the recognition process and the generation of feature vectors. This involves the application of technologies such as AI, hashing, PPK encryption, blockchain technologies, robotics, and IoT.

[0051] To improve data integrity, a hybrid blockchain approach is preferably used to generate mathematical certificates for each individual data packet. These certificates are preferably protected by a blockchain.

[0052] Using multiple public blockchains can improve data integrity, and this approach also saves a considerable amount of energy compared to solutions that directly utilize blockchain technology.

[0053] In Fig. Figure 3 shows a QR code in which a three-dimensional dynamic security feature according to the invention is integrated at its center. The QR code is a widely used information carrier that is machine-readable. However, it is easily forged. Using the dynamic surface recognition according to the invention, counterfeit protection for the QR code is achieved while simultaneously maintaining machine readability. This means that when the user scans the QR code with a smartphone, they gain access to a data space, for example, by being redirected to a URL, i.e., an internet address. To verify whether the QR code is genuine, the security feature is scanned by the smartphone camera and compared with a database entry. Key factors in this comparison include the surface texture and the ink distribution, which are unique for each printed product.This means that, on a macroscopic level, it can be ruled out that the code is forged, as this would require an unreasonable effort. For example, the forger would have to reliably replicate surface features, blurring, and contrasts using a computer, which is virtually impossible. The forger could only create a duplicate that would be visible to the naked eye; however, the forgery would be immediately detected during verification after the security feature is read.

[0054] In Fig. Figure 4 illustrates the creation of a feature matrix or feature vector. From the scanned surface of a security feature, so-called minutiae and / or surface characteristics are generated—that is, the finest feature elements, which are combined to form one or more feature vectors. Each feature matrix obtained from the individual features represents a unique identifier for each product. The feature matrix is ​​stored in a database as data or affixed to a product as a code. The feature matrix is ​​thus comparable to a fingerprint; that is, it exhibits an individual pattern that can be assigned to a specific product. During verification, the calculated feature matrix, or parts thereof, are compared with the matrix or parts thereof stored in the database.

[0055] In Fig. Figure 5 shows a supply chain. The security feature, in this case a QR code, is read using an optical recognition device, i.e., a camera. To ensure that only the genuine product is present in the supply chain, the scan can be performed at various points, such as when loading a pallet, after transport, or upon delivery to the point of sale. The recipient can thus verify that the product is genuine and not a counterfeit. If desired, the existing security feature can also be supplemented after the scan.

[0056] In Fig.Figure 6 shows two different print results where the security-relevant features are located in different positions. Three nested circles or dots of varying sizes are shown, arranged differently at specific positions (compare the circles / dots at the 4 o'clock, 5 o'clock, and 6 o'clock positions). Based on the extracted security features, different characteristic feature matrices are created for each product.

[0057] This is a schematic example of printing tolerances, such as those that typically occur in a similar form during a printing process. However, the technology according to the invention can also be used to individually characterize and identify, for example, posters viewed from a great distance. The details are then no longer discernible to the naked human eye from such a distance without aids (such as an optical magnifying device, a camera, or a lens system). QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] EP 1982296 B1

[0007] EP 2039527 B1

[0008] DE 102016218545 A1

[0009] DE 102015219400 B4

[0010] DE 102015219399 B4

[0010] DE 102015219396 B4

[0010] DE 102015219395 B4

[0010] DE 102008016435 A1

[0011]

Claims

[1] Method for determining the originality of a product, comprising the steps: - Providing a dynamically modifiable three-dimensional (3D) security feature and a product-specific machine-readable two-dimensional (2D) code on the product or product packaging, wherein the surface of the 3D security feature has unique, characteristic, optically detectable security features resulting from the surface texture, the distribution of colors, blurring and / or contrasts on the surface, - Reading several selected unique minutiae and / or characteristics of the surface of the dynamically changing 3D security feature with an optical detection device in an area no longer perceptible to the unarmed human eye, - Formation of at least one individual feature matrix characteristic of the product from the selected unique minutiae of the read surface of the dynamically changeable 3D security feature using an algorithm or rule, - Storing at least one characteristic matrix for the product directly or indirectly in a database or as a code applied to the product or product packaging, whereby the following steps are carried out in the course of identifying the product: - Reading the surface of the dynamically modifiable 3D security feature and its minutiae and / or surface characteristics using any optical scanning device and generating at least one feature matrix according to the algorithm or rule applied for the storage, - Verifying the surface by comparing at least a part of the feature matrix of the read-in dynamically changeable 3D security feature with at least a part of the individual feature matrix stored as a code in the database or the product or product packaging, whereby the verification is positive if the compared feature matrices agree at least partially within a specified tolerance range. [2] Method according to claim 1, characterized by , that only after a positive verification will the machine-readable 2D code characteristic of the product be displayed or executed on the product or the product packaging. [3] Method according to claim 1 or claim 2, characterized by, that a data room containing data about the product is provided, whereby access to the data room only occurs after the verification of the interface has been successful. [4] Method according to any one of claims 1 to 3, characterized by , that within a product line, unique minutiae and / or surface characteristics are formed for each individual product of the product line within the 3D security feature, leading to an individual feature vector characteristic of the product, using a machine learning method for adaptation or optimization. [5] Method according to any one of claims 1 to 4, characterized by, that upon positive verification, the 3D security feature stored in the database is supplemented or replaced by the read-in dynamically changeable 3D security feature, so that the read-in 3D security feature corresponds to the current security feature in the database for later verification. [6] Method according to any one of claims 1 to 5, characterized by , that multiple reading, verification, supplementation or replacement of the dynamically changeable 3D security feature characteristic of the product takes place in the database within a supply chain or life cycle. [7] Method according to any one of claims 1 to 6, characterized by that the machine-readable 2D code is a Data Matrix code, QR code, RFID tag, NFC tag or barcode. [8] Method according to claim 7, characterized bythat the 3D security feature is located outside or inside the machine-readable 2D code, or overlaps with it. [9] Method according to claim 8, characterized by , that for the formation of the minutiae and / or characteristics of the surface, both areas of the dynamically changeable 3D security feature and areas of the machine-readable 2D code are selected. [10] Method according to any one of claims 1 to 9, characterized by , that the provision of the dynamically modifiable 3D security feature and the machine-readable 2D code characteristic of the product is carried out on a printed label. [11] Method according to any one of claims 1 to 10, characterized by , that the data stored in the database for the dynamically changeable 3D security feature is additionally secured via a blockchain.

Images