Computer-implemented method, data processing apparatus, and computer program for copy protection

Abstract

A computer-implemented method for preventing fraudulent processing of a digital representation of at least a portion of a document is provided. A computer-implemented method, data processing apparatus, and computer program product for preventing fraudulent processing of a digital representation of at least a portion of a document are provided, particularly where the document is a banknote. The method includes providing data, the data being based on a digital representation of at least a portion of a test element. The digital representation may be an image file corresponding to at least one portion of the test element. The method also includes analyzing the data with respect to data representing at least one characteristic property of the at least one portion of the document. The method further includes activating a prohibiting means if the data based on the digital representation of the at least one portion of the test element is similar to the data representing the at least one characteristic property. The prohibiting means prohibits further processing of the data based on the digital representation of the at least one portion of the test element, particularly including copying, transmitting, printing, and / or reproducing the data. Alternatively, the prohibiting means modifies the data to prevent the data from being transmitted, printed, reproduced, and / or further modified by the data processing apparatus.

Description

Technical Field

[0001] The present invention relates to a computer-implemented method, a data processing apparatus, and a computer program for copy protection. Mu It relates to.

Background Art

[0002] Security-related documents such as airline tickets and banknotes are often targets of forgery. One approach to countering forged documents relates to the assessment of the authenticity of suspicious documents. However, this approach is a downstream activity in the sense that the original document has already been copied by the time the authentication process is executed on the suspicious document. Therefore, authentication-related countermeasures are not very desirable.

[0003] Copying of the original document itself can be performed by a scanner, a printer, and / or a copying device. In this regard, another approach relates to a method of recognizing security-related documents and prohibiting their copying before it is executed.

[0004] For the purpose of preventing the above-described acts, various security mechanisms exist. They may be, for example, printed graphic design elements recognized by a special detector within the aforementioned device. The detector can then trigger a response that prevents the desired act, such as a denial of processing or the printing of a highly degraded image. Such graphic elements may be designed to have an appearance such that they are part of the pattern of the security document. Examples of the use of such elements can be found in U.S. Patent No. 5,845,008. In other cases, special signals that are hardly perceptible visually can be added to the design to be printed, and as a result, the design can be recognized by a special detector within the aforementioned device and then trigger a response as described above. Examples of the use of such elements can be found in U.S. Patent No. 6,449,377.

[0005] However, these security mechanisms have inherent vulnerabilities. Graphic design elements can often be easily recognized for their intended purposes, even when attempts are made to make them look like part of a pattern. As a result, it is possible to slightly modify the graphic design elements so that special detectors can no longer identify them, and thus disrupt the desired actions of counterfeiters. Graphic design elements can also be misused by applying them to other documents that legitimate users do not intend to protect, resulting in the inability of people to complete actions such as scanning, copying, or printing the said documents.

[0006] Special signals such as digital watermarks can also have undesirable characteristics that make printed documents look distorted. In the case of banknote designs, this can be particularly undesirable. Distortion may be reduced, sometimes at the expense of signal strength, and a compromise is usually required.

[0007] Artificial intelligence (AI) combined with machine learning is increasingly being used in applications such as face recognition and other object identification. In such applications, there can be an infinite number of potential images that need to be robustly recognized. For example, an application trained to recognize images on a gray list may encounter any one of several individual characteristics such as the size, pose, age, color shade, lighting, or some other of the gray list. Applications designed to reliably recognize an individual's face must face similar variations, increasing at least the computational complexity and computational resource needs of the application.

Summary of the Invention

Problems to be Solved by the Invention

[0008] The object of the present invention is to overcome or at least reduce the drawbacks of known methods for copy protection according to the prior art, respective data processing devices and respective computer program products.

[0009] The object technical problem is solved according to the subject matter of the independent claims. The advantages (preferred embodiments) are described in the forms for carrying out the following invention and / or the attached drawings and the dependent claims.

Means for Solving the Problem

[0010] According to one aspect, a computer-implemented method for preventing unauthorized processing of at least a part of a digital representation of a document is provided. According to the method, data is provided, the data being based on at least a part of a digital representation of a test element. Data based on the digital representation of at least one part of the test element is analyzed with respect to data representing at least one characteristic property of at least one part of the document. If data based on the digital representation of at least one part of the test element is similar to data representing at least one characteristic property, inhibitory means are activated. Further processing of data based on the digital representation of at least one part of the test element is then prohibited by the inhibitory means. Alternatively, data based on the digital representation of at least one part of the test element is corrected by the inhibitory means so as to prevent the data from being transmitted and / or printed and / or copied and / or further corrected by a data processing device.

[0011] In the context of the present method, the document may be one of a security-related document, in particular a banknote, a check, a bill, a ticket, a passport or a boarding pass. In these document forms, unauthorized processing of the respective digital representation of a test element similar to the document poses a significant risk for both economic and security-related reasons. These risks are avoided or at least reduced by the method for preventing unauthorized processing.

[0012] Within the context of this method, a test element refers to an item that could potentially be considered by an objective (unbiased) observer as being one of the above types of documents. In other words, a test element (sample, probe) within this context may be something that is sufficiently similar to the original document such that an objective observer could mistakenly consider the test element to be the original document. For example, a counterfeit banknote could potentially be considered as the real banknote. There could be a significant deviation between the test element and the original, yet there is a possibility of misinterpretation by an objective observer. In this regard, when the document and the test element are directly compared to each other, such a deviation may be easily recognized. However, the memory of an objective observer is limited. Therefore, when an objective observer examines only the test element without immediate access to the original document, the observer may consider the test element to be the (original) document. An objective observer does not necessarily have to be an expert in the field, but could be considered as someone who generally uses the original document. A part of the test element may in particular be a one - dimensional or two - dimensional part of the test element.

[0013] Within the context of this method, a digital representation refers to digital code that is usually written in a computer language and is thus computer - readable code. Therefore, a digital representation of a part of an item (original document, test element or training document) can be a data file representing a part of the item. The data file may be suitable for describing the nature of the item by digital code.

[0014] Within the context of this method, the data provided does not have to be a complete data file. The data provided may be a fragment of a data file, for example, an image file fragment, a text file fragment or a pdf file fragment representing a part of the test element. Usually, the data of the test element is at least part of an image file. Therefore, the data may be a string of data when transmitted from a first data - processing device to another data - processing device.

[0015] In particular, the data file of the item (original document, test element or training document) may be an image file, text file or pdf file of the document (test element or training document). The data file may be an image file having a resolution within the range of 50 dpi to 2000 dpi, in particular within the range of 100 dpi to 1000 dpi, more particularly within the range of 200 dpi to 600 dpi, and even more particularly within the range of 300 dpi to 400 dpi.

[0016] In the context of this method, the characteristic property of the document may be the physical property of the document or the substance applied to the document, based on which an objective observer may consider the test element to be the (original) document. This characteristic property will be explained in more detail later.

[0017] Within the context of this method, data similarity does not require that the data representing the test element and the document be equal to each other. The data according to this method is basically related to the nature of each item for which the data is in digital representation. However, these natures can of course be described by various data due to, for example, various available data file formats or computer languages. Thus, data related to various items can be considered similar to each other if the data can describe the nature of the test element that may be substantially similar to the characteristic property. For example, the characteristic property of a document may be two blue lines oriented parallel to each other. Then, the data of the test element may describe two blue lines parallel to each other. However, the data related to the test element may be written in different computer languages. Alternatively, different color codes may be used for the test element to describe the color of the line (e.g., RGB vs. CYMK). Independent of these deviations regarding the underlying code, the characteristics of the test element may be different from the characteristic property but can still be approved as being similar. In this regard, the similarity of characteristics relates to the evaluation of similarity by an objective observer. This method may also take into account the pixelation of the digital representation of at least one part of the test element when the similarity with the data of the document is determined. Furthermore, this method may take into account the resolution and / or color distribution and / or contrast distribution and / or luminance distribution characteristics of the test element. Thus, this method may be configured to interpret the data of the test element and recognize the nature of the item to which the data relates. If that nature is similar to the characteristic property, each data can be considered similar to each other within the scope of the meaning of this method. Note that the evaluation regarding the similarity of characteristics is different from the evaluation regarding authenticity. This method does not authenticate the characteristics of the test element related to the document or the test element itself. This will be described later.

[0018] In the context of this method, the prohibiting means may be a software or hardware implementation structure. The prohibiting means may be configured to determine whether an action is applicable to some data. In this regard, the prohibiting means may be configured to issue instructions to a bus structure, an interface, a data processing unit, a memory, etc. Therefore, the prohibiting means may have a master function such that the remaining components of the data processing device are slaves compared to the prohibiting means.

[0019] In the context of this method, further processing of the data may include copying and / or transmitting and / or printing and / or replicating the data. Therefore, all such actions applied to the data can be prohibited by the prohibiting means. Thus, further processing refers not only to the transmission of a data file but also to any action applied to a part of such a data file, for example, the replication of each data string.

[0020] In the context of this method, correcting the data may include correcting the data to include marks / attributes / flags, and the marks / attributes / flags prevent the data from being transmitted and / or printed and / or replicated and / or further corrected by a data processing device. Therefore, the prohibiting means may be configured to correct the data so that other data processing devices can immediately recognize that the processing of the data is prohibited. Alternatively, the prohibiting means can also be configured to delete the data.

[0021] Therefore, a method is provided for advantageously prohibiting further processing of data based on at least a partial digital representation of a test element when the data is similar to data representing at least one characteristic property of a document. In other words, when the test element is sufficiently similar to the document, further processing of its digital representation is advantageously prohibited.

[0022] In contrast to the aforementioned problems of prior art face recognition and other object recognition using AI assistance (where countless potential images would be recognized), in the field of security documents such as banknotes, it can be expected that the number of images to be recognized will be much smaller. In particular, it can be expected that there are a finite number, perhaps in the hundreds, of things to be recognized. A basic image, for example, means a flat and evenly illuminated image of the denomination and side (front or back) of a given series of banknotes. Naturally, variations such as individual serial numbers and other identifiers, as well as minor variations in print quality and registration and image resolution, are expected.

[0023] The object of the invention described herein is to provide a means for recognizing security document images in a form that is considered by an average person to represent a security document (without authentication by a security mechanism), so it is not necessary to recognize all possible variations of the document. The reason is that according to the present invention, there is no need to prevent the digital image from being replicated if it is discovered by an average person that a given security document representation is too distorted to be acceptable as a security document in the case of replication. Therefore, it can be expected that the training set of the present invention and the corresponding computational resources required for the detection process will be significantly smaller than, for example, a typical training set and computational resources for an AI application in a face recognition procedure.

[0024] A further advantage of the present invention is that the method of the present invention does not authenticate the security mechanism of the digital document image, but simply recognizes that the image is sufficiently similar to the basic image in the training set that an average person can recognize as an actual security document image. For example, when a counterfeit banknote is carefully inspected, the difference from the genuine banknote will surely become apparent, but the method of the present invention does not perform a detailed inspection. Nevertheless, the average person does not recognize the difference, and thus thinks that the replicated document represents a security document, and there are sufficiently small differences such as those related to forgery that are not recognized as forgery typically in the short time when currency exchange is carried out. In contrast, when AI is used for, for example, human face recognition, the error in "authenticating" a given face can have serious legal consequences.

[0025] The present invention does not detect a special signal of the document image, and thus there is no need to add an additional mechanism such as a copy code to the security document, so there is no obvious advantage for those who access the detection code of the present invention.

[0026] In fact, the fact that there is no need to use a special signal, copy code or graphic design element is a great advantage of the present invention. As described above, the use of these technologies may visually change the security document in a way that is aesthetically unfavorable and at the same time creates security vulnerabilities. In the method of the present invention, the designer of the security document does not need to be involved in the addition of special mechanisms that are aesthetically destructive and / or create vulnerabilities because the image is globally recognized by the detector.

[0027] Another advantage of the present invention is that any security document can be used in the training set regardless of whether the security document was created before or after the implementation of the detector. In contrast, if copy prevention relies on the use of special signals, copy codes, or graphic design elements created only after the implementation of detection, there is no possibility of detecting security documents containing new codes. Conversely, if the detector does not detect old copy codes, it is not possible to detect documents that may still be in use, and thus it is not possible to prevent their copying using prior art authentication methods for copy prevention.

[0028] According to another aspect, data based on the digital representation of at least one part of the test element may be recorded by an inspection device. The inspection device can be an appearance inspection device and may be based on image recognition or computer vision technology. The inspection device may in particular be a scanning device (pick-up head) and / or an imaging device and / or a camera. The inspection device may be configured to provide a data file (document, test element or training document) of the item that can be an image file, text file or pdf file of the item. The inspection device may be particularly configured to provide an image file having a resolution within the range of 50 dpi to 2000 dpi, in particular within the range of 100 dpi to 1000 dpi, more particularly within the range of 200 dpi to 600 dpi, and even more particularly within the range of 300 dpi to 400 dpi. Thus, if a test element is present, the method may advantageously include creating data based on the digital representation of at least one part of the test element by the inspection device. Thus, the method does not depend on external resources that provide each data so that the risk of manipulation of the data representing a part of the test element is reduced.

[0029] According to another aspect, at least a part of the training document may be provided. Data representing the digital representation of at least one part of the training document may be recorded by an inspection device. The inspection device may in particular be one of the aforementioned types. Data representing the characteristic properties of at least one part of the document can be identified within the data representing the digital representation of at least one part of the training document by artificial intelligence (AI) and / or machine learning (ML).

[0030] In the context of the present method, artificial intelligence (AI) refers to software or hardware-based technologies, such as algorithms, configured to elicit decisions. AI can also be configured to automatically utilize the data provided with respect to the intended purpose and automatically provide each result to the user.

[0031] In the context of the present method, machine learning refers to software or hardware-based technologies that may include AI capabilities. ML may be configured to include multiple inputs in order to improve the process of eliciting decisions. In other words, ML may be configured to recognize several similar inputs in order to improve the prospect of the accuracy of eliciting decisions as compared to the accuracy of eliciting decisions based on a single input. For example, if several similar training documents are provided, ML can be configured to identify the characteristic properties that all of these training documents have in common. Furthermore, ML can also be configured to identify the characteristic properties within a set of training documents even if the individual training documents of the set, such as banknotes with different values, are different.

[0032] In the context of this method, the training document can be a document used to identify at least one characteristic feature. Based on the characteristic feature, an objective observer can evaluate the similarity of the test element compared to the (original) document. In other words, using the training document, methods such as AI and / or ML (or related software or hardware-based means) can be trained for the set of qualities by which the test element is analyzed. In this regard, data representing the digital representation of at least one part of the training document is provided to the AI and / or ML. Then, the AI and / or ML may be configured to identify the qualities representing the characteristic feature that determines what the test element is analyzed for within these data.

[0033] Accordingly, the method advantageously includes a training process that defines what the test element is analyzed for and automatically evaluates which qualities included in the training document can function as appropriate characteristic features to evaluate the similarity between the test element and the document. A method designed in this way is independent of a given input regarding the qualities by which the test element is analyzed.

[0034] According to another aspect, the data representing the characteristic feature and determined based on the training document may be stored in a memory. In particular, the data may be stored encrypted and / or error-coded. Furthermore, the data may also be stored in a local manner (within the same data processing structure) or a non-local manner (on a server and / or external memory). The data may be stored in the memory to build a database. Thus, the availability of the data representing the characteristic feature can be advantageously retained so that the data can be used when analyzing the test element.

[0035] According to another aspect, the inspection device may be configured to record data representing a digital representation of at least one portion of a test element or a training document, substantially independently of at least one of the angular direction of an item related to the inspection device, the item inspected by the inspection device that has been trimmed and / or mutilated, the resolution provided by the inspection device, the distortion of an item related to the inspection device, the scaling effect applied to the item inspected by the inspection device, and combinations thereof. The item inspected by the inspection device can be the subject of various actions by the user. Such actions can generally affect the recording of data. A method of using an inspection device designed in such a way is advantageously robust against external influences.

[0036] According to another aspect, the inspection device may operate in a reflection mode and / or a transmission mode. The inspection device may comprise a detector and a radiation emission source. The detector can be configured to detect radiation emitted by the radiation emission source and reflected by or transmitted through the item being inspected. The item inspected by the inspection device may include characteristics that can be identifiable only in one of the reflection mode and the transmission mode. Thus, a method having an inspection device configured in such a way meets all requirements independent of the mode in which the characteristics can be identified.

[0037] According to another aspect, AI and / or ML may be configured to perform an analysis regarding the similarity of data based on the digital representation of at least one part of a test element as compared to data representing at least one characteristic property of at least one part of a document. If AI and / or ML discovers that data based on the digital representation of at least one part of a test element is similar to data representing at least one characteristic property, AI and / or ML may also subsequently be configured to activate a prohibiting means. The method so configured includes the same AI and / or ML used to determine the data representing the characteristic property. Thus, advantageously, the same and / or ML (criteria) is uniformly applied throughout the method so that potential variability is reduced.

[0038] According to another aspect, data based on the digital representation of at least one part of a test element may be analyzed by AI and / or ML regarding data representing characteristic properties, and the data representing the characteristic properties is stored in a memory. The memory provides the possibility of storing data regarding characteristic properties for analysis so that steps can be advantageously performed independently of each other in time.

[0039] According to another aspect, the characteristic feature may be at least one of a single or a plurality of specific distributions of contrast levels and / or colors and / or marks arranged and / or printed and / or applied on the surface of the item and / or included within the item. The characteristic feature may also be at least one of a single or a plurality of shapes of marks arranged and / or printed on the surface of the item and / or included within the item. Further, the characteristic feature may be at least one of a single or a plurality of moiré patterns, microstructures, microtexts, digital security marking mechanisms invisible to the naked eye, gyoshé, rainbow, concave lenses, optically variable elements, holograms, optical lenses, watermarks, QR codes and fingerprints. Further, the characteristic feature may be at least one of a single or a plurality of specific materials arranged on the surface of the item and / or included within the item, in particular, the specific material includes at least one of fabrics such as paper, polymers and cotton. Also, the characteristic feature may be at least one of a single or a plurality of security mechanisms arranged on the surface of the item and / or included within the item. The security mechanism may include at least one of a hologram, a microlens, an embedded security thread, a security foil, a transparent or translucent window, a label and a symbol. Further, the characteristic feature may be any combination of the foregoing. The method configured in such a manner is advantageously improved with respect to security. Since the characteristic feature is not determined a priori, it is generally unknown according to which criteria the test element is analyzed. Therefore, the test element cannot be designed according to a predetermined known criterion. Further, since the possible features are extensive, each feature can generally be evaluated individually with respect to its suitability to act as a characteristic feature. Therefore, the method is also advantageously improved with respect to its reliability.

[0040] According to another aspect, the characteristic features identified within the data representing the training document may be independent of at least one of the angular orientation of the item with respect to the inspection device, trimming and / or cutting of the item inspected by the inspection device, the resolution provided by the inspection device, the distortion of the item with respect to the inspection device, the scaling effect applied to the item inspected by the inspection device, and combinations thereof. Thus, the characteristic features can be robust with respect to various external influences that can affect the document (training document) during normal use.

[0041] According to another aspect, the method may include a step of determining a reference value. The reference value may be based on the probability that data representing at least one characteristic feature of at least one part of the document is similar to data based on the digital representation of at least one part of the test element. The reference value can be considered true if the reference value is greater than a predetermined threshold. The step of determining the reference value may be included within the step of analyzing data based on the digital representation of at least one part of the test element with respect to data representing at least one characteristic feature of at least one part of the document. Further, the activation of the prohibiting means may depend on whether the reference value is true. A reference value indicating whether the data representing the test element is similar to the data representing the characteristic feature may be determined based on each probability. Thus, a method configured in this way advantageously takes into account a certain degree of deviation between the respective data, but the method can still reliably determine the similarity of the respective data.

[0042] According to another aspect, the method may include an operation step on the data before analyzing whether data based on a digital representation of at least one part of the test element is similar to data representing at least one characteristic property of the document. In particular, the data may be transformed and / or filtered for analyzing the data with respect to special properties, for example, a 1D or 2D Fourier transform, a logarithmic transform, or a Laplacian filter may be applied to the data. According to the method configured in such a way, the special properties of the data can also be advantageously analyzed.

[0043] According to another aspect, the method can be configured to provide a false positive rate and / or a false negative rate of one in ten million or better with respect to an evaluation of whether data based on a digital representation of at least one portion of a test element is similar to data representing at least one characteristic property of a document. Generally, an event in which an objective observer believes that a portion of a test element is similar to at least a portion of a document is assigned a positive event. Conversely, an event in which an objective observer believes that a portion of a test element is not similar to a document is assigned a negative event. In this regard, the similarity between the test element and the document can still exist even if the test element is modified with respect to the pixelation effect and / or respective resolution and / or color distribution and / or contrast distribution and / or luminance distribution described by the data. A false positive event indicates that the similarity of the data has been approved even though a portion of the element is not similar to the document. A false negative event indicates that the data has been approved as dissimilar to each other even though a portion of the element is actually similar to a portion of the document. The false positive rate and the false negative rate basically indicate the reliability of a method that depends on a statistical test. The ratio is empirically tested based on a sufficiently large data set to achieve the desired performance. The method can also be configured such that it has a high tolerance for false positives or false negatives and / or uses a secondary and more thorough determination process for cases that trigger a positive but inconclusive answer. A method configured in this way advantageously determines the similarity of the data with an acceptable fault tolerance.

[0044] According to another aspect, the method may take into account the pixelation of data based on a digital representation of at least one portion of a test element when a reference value is determined. When determining the reference value, the method may further take into account the resolution and / or color distribution and / or contrast distribution and / or luminance distribution of the data based on a digital representation of at least one portion of the test element. In a method configured in this way, since various effects are taken into account, the accuracy of the reference value is improved.

[0045] According to another aspect, the method may be configured to be locally executable within a first data processing device. The first data processing device may comprise a storage memory capable of storing each code of the method. Alternatively, the method may also be configured to be remotely executable. According to an alternative, the first data processing device may be connected to a second data processing device via a data connection. Then, the method may be configured to be executable on the second data processing device via the data connection. A method configured in such a manner is executable almost independently of the data processing structure. Advantageously, it can also be executed based on a server-client system or, for example, in a local manner if a network connection is not available.

[0046] According to another aspect, the method may be based on code, and each code of the method may have a binary size of 100 kB to 50 MB, particularly 200 kB to 10 MB, and more particularly 500 kB to 1 MB. Since the code has a similarly small size, the code can be advantageously implemented even in non-high-end data processing devices such as scanning devices, printers, copying devices, etc.

[0047] According to another aspect, data based on the digital representation of at least one part of a test element may be provided based on the transmission of data from a first data processing device to a second data processing device. In other words, the method may be configured to act on the data transmitted between data processing devices. Further transmission can be prohibited if the data is similar to data representing at least one characteristic property of at least one part of a document. The digital representation of at least one part of an element may represent a one-dimensional or two-dimensional part of the test element. For example, a line-by-line scan of an element may be transmitted from a scanning device to a data processing unit. The method may advantageously be configured to recognize this data and analyze the data with respect to data representing at least one characteristic property of at least one part of a document.

[0048] According to another aspect, the method may be configured to be executable within a period of less than 60 seconds, particularly within a period of 100 milliseconds to 30 seconds, especially within a period of less than 1 second. The method configured in such a way can advantageously be applied with an acceptable time consumption even during normal real-time data processing procedures such as printing or scanning of elements.

[0049] According to another aspect, the method may be configured to avoid authenticating data representing at least one part of the digital representation of a test element with respect to the digital representation of at least one part of a document. The purpose of the method can be to determine the similarity between these data. The method is not intended to authenticate elements with respect to the document. The authentication process is a very important process that is stored in a certified facility for security reasons. Since the method is generally configured to be implemented on common hardware or software (copying devices and / or printers and / or scanning devices) that are also available to ordinary customers, the method advantageously does not include an authentication function, and as a result, the details of the authentication process are kept confidential.

[0050] According to another aspect, an apparatus for data processing is provided. The apparatus may comprise means for executing the above-described method. Further, the apparatus may also comprise means such that a modification of the above-described method can be executed. In particular, the apparatus for data processing may comprise a data processing unit such as a CPU. The data processing unit may be configured to analyze and / or process and / or correct data. The data processing apparatus may further comprise at least one of a storage memory for storing data, an interface for communicating with other data structures, and a data bus for transmitting data between different components. Further, the data processing apparatus may be a component of a scanning device (pickup head), a printer or a copying device. Thus, a simple and effective data processing apparatus is provided that is advantageously configured to execute the shown method.

[0051] According to another aspect, the apparatus for data processing may comprise a lower ARM type multi-core CPU or a similar CPU commonly used in mobile devices. The apparatus may further comprise a main memory within the range of 4 MB to 8 GB, more particularly within the range of 16 MB to 2 GB, more particularly within the range of 64 MB to 512 MB, more particularly within the range of 128 MB to 256 MB. The method may be configured to use a main memory of the indicated size and be executable on the indicated CPU type in a local or remote manner.

[0052] According to another aspect, a computer program product is provided. The computer program product may include instructions that, when executed by a data processing apparatus, cause the apparatus to perform the above-described method. The instructions may be the result of the code of the computer program product. The computer program product may be written in a computer language suitable for implementation in a scanning device (pickup head), a printer, or a copying device. The computer program product may also be configured to be stored in a volatile or non-volatile memory, a hard disk drive, or a computer-readable medium such as a USB memory, a CD, etc. The computer program product may be configured to be executable from an external or internal memory of the data processing unit. Thus, a computer program product is provided that is advantageously configured to perform the method shown in various configurations such as external or internal memory.

[0053] According to another aspect, the code of the computer program product may be configured to be stored in an encrypted and / or error-coded manner. Some of the underlying technologies and instructions should be kept secret for security reasons. Thus, if the code is stored encrypted, the underlying technologies and instructions may advantageously be prevented from being made public.

Brief Description of the Drawings

[0054] Further aspects and features of the present invention will be derived from the following description of the preferred embodiments of the present invention with reference to the accompanying drawings.

[0055]

Figure 1

Figure 2

Figure 3

Mode for Carrying Out the Invention

[0056] FIG. 1 is a simplified schematic flowchart according to method 100 for preventing improper processing of a digital representation of at least a part of a document. Method 100 includes several steps indicated by boxes. Boxes with dashed lines indicate steps that are optional, either alone or in any combination. Thus, the method includes at least steps 110, 120, and 130. Regarding steps 132 and 134, only one of the two needs to be present. Steps 132 and 134 are alternatives to each other.

[0057] In step 110, data is provided, and the data is based on a digital representation of at least a part of the test element. In this regard, providing the data may depend on further optional steps 140, 190, and 182. In step 140, the test element (or a part thereof) is arranged so as to be able to record data representing the nature of the test element. For this purpose, an inspection device may be used. The inspection device can be an appearance inspection device such as a scan unit or a camera, and may be based on image recognition or computer vision technology. The inspection device may be configured to generate data describing the test element. For example, the generated data may include information regarding the contrast level and / or color and / or distribution of marks arranged and / or printed on and / or included in the surface of the test element. In particular, the recorded data may be an image file representing the test element. Thus, the data may be part of an image file created by the inspection device. In this regard, an image file is a normal method for describing the (graphic) nature of an item. The inspection device may be configured to directly provide the recorded data. Thus, step 140 may directly provide the data to step 110. Alternatively, the data is provided according to step 190. In step 190, the data is transmitted from a first data processing device to a second data processing device, for example as a string. The first data processing device may be, for example, a normal CPU, and the second data processing device may be implemented, for example, in a network printer. The CPU and the network printer may be connected by a data bus, for example a network connection. The method may be configured such that, in step 190, the transmitted data is recognized by appropriate means of the method. The method may be configured such that the data is intercepted and / or a copy thereof is created and provided. Thus, the data may be directly provided from steps 140 and 190 to step 110. Alternatively, the data may be stored in a storage memory in step 182 before being provided to step 110.The memory may be an external memory or an internal memory. The memory may be a computer-readable medium such as a main memory, a hard disk drive, a USB memory, or a CD. By storing data in the memory in step 182, the temporal independence of the method is provided.

[0058] In step 120, data based on the digital representation of at least one part of the test element is analyzed with respect to data representing at least one characteristic property of at least one part of the document. The characteristic property of the document may be a graphic property or a structural property of the document such as a special barcode. Thus, the method may be configured to interpret the data and analyze whether the data represents features similar to the characteristic properties of the document. Usually, the data of the test element is at least part of an image file. If this image file contains characteristics similar to an exemplary special barcode, the method determines that the respective data are similar to each other.

[0059] Data representing characteristic features may also be provided to the method. In this regard, in step 150, training documents may be provided. In step 160, the training documents may be inspected so that data representing the nature of the training documents can be recorded. The recording of the data may be performed by an inspection device, in particular an appearance inspection device as described above with respect to the test element. In particular, the recorded data may be an image file representing the training document. In step 170, at least one characteristic feature may be identified within the data recorded with respect to the training document. For this purpose, artificial intelligence (AI) and / or machine learning (ML) techniques may be applied. Thus, the data (usually an image file) may be inspected by AI and / or ML for characteristics representing appropriate criteria for evaluating the similarity with the test element. The characteristic feature selected in step 170 is a feature that, if realized in the test element in a similar way, an objective observer can consider the test element and the training document to be similar to each other. Alternatively, steps 150 and 160 may be repeated several times, which means that several training documents are provided. Thus, at least one characteristic feature identified in step 170 by AI and / or ML may also depend on the set of data recorded with respect to multiple training documents. Once the characteristic feature is identified, the data representing the feature may be provided to step 120 for evaluation of data similarity. Alternatively, in step 184, the data representing the characteristic feature may be stored in an external or internal memory. Thus, the method can be executed in a timely manner independently of steps 150, 160 and 170. Further, in step 184, a database may be constructed to have various conceivable characteristic features. This is particularly interesting when the method is implemented in a server-client based structure.

[0060] The evaluation of the similarity between the data representing the test element and the characteristic features of the document is performed in step 120. In step 130, if the data based on the digital representation of at least one part of the test element is similar to the data representing at least one characteristic feature, the prohibiting means is activated. Thus, step 130 explains the technical effect of the method when each data is considered by the method to be similar to each other. The prohibiting means may be configured to cause at least one technical effect described with respect to steps 132 and 134. According to step 132, the data regarding the test element is prohibited from further processing by the prohibiting means. The prohibiting means may have a master function. The prohibiting means may then issue an instruction to prohibit further processing of each data. Alternatively, according to step 134, the data is corrected by the prohibiting means so that the data is prevented from being transmitted and / or printed and / or copied and / or further corrected by other data processing devices. In this regard, a special flag and / or attribute and / or mark indicating that further processing of the data is prohibited may be assigned to the data.

[0061] Alternatively, the method may be configured such that steps 120 and 130 can also be performed by AI and / or ML. Thus, it is also possible to apply AI and / or ML to determine the similarity between the data of the test element and the characteristic features.

[0062] FIG. 2 is a simplified schematic diagram of a data processing apparatus 200 that utilizes an inspection apparatus for data recording. The data processing apparatus 200 may be configured to execute the method 100. Further, the data processing apparatus 200 is connected to the external inspection apparatus described above with respect to steps 140 and 160. The inspection apparatus according to the present embodiment includes a radiation emission source 220 and detectors 232, 234. Here, the radiation emission source 220 and the detectors 232, 234 are external components connected to the data processing apparatus 200. Alternatively, the radiation emission source 220 and the detectors 232, 234 may be internal components of the data processing apparatus 200. The radiation emission source 220 and the detectors 232, 234 may be combined with the data processing apparatus 200 in a scanning apparatus, a printer, or a copying apparatus.

[0063] An item 210 (document, test element, or training document) is arranged such that the radiation emission source 220 can emit radiation toward the item 210. The radiation emitted by the radiation emission source 220 may have any wavelength suitable for investigating the item 210. In particular, the radiation may have a wavelength corresponding to UV light and / or visible light and / or IR radiation. The operation of the radiation emission source 220 may depend on each command issued by the data processing apparatus 200. Next, the inspection apparatus may be configured such that the detector 232 can detect the radiation reflected by the item 232, and the reflection of the radiation depends on the physical properties of the item 210. Accordingly, information regarding the physical properties of the item 210 can be obtained, and a digital representation of the item 210 can be generated. The digital representation may be generated by the detector 232 and communicated to the data processing apparatus 200. Alternatively, the data processing apparatus 200 generates a digital representation of the item 210 based on the information received from the detector 232. Further, the inspection apparatus may be configured such that the detector 234 can detect the radiation that has passed through the item 210. Different physical properties of the item can be evaluated depending on whether the radiation is detected in the reflection mode or the transmission mode.

[0064] FIG. 3 is a simplified schematic diagram of a training document 310 and a test element 320. The training document 310 may be substantially similar to a document whose digital representation is to be prohibited from unauthorized processing. The training document 310 includes various characteristics 332, 334, 336, 338, each surrounded by a dashed box. These characteristics may be, for example, the distribution of marks such as the first characteristic 332. The second characteristic 334 includes several concentric circles. Thus, the second characteristic 334 is particularly invariant with respect to a linear inspection of the training document 310 as long as the linear inspection coincides with the center of the concentric circles. The third characteristic 336 includes two individual lines oriented parallel to each other. The fourth characteristic 338 includes a character-based code printed on the training document 310. The character-based code may be the serial number of the training document 310. The characteristics may vary with respect to several properties such as the position, color, contour, etc. within the training document 310. Of course, additional characteristics, such as those described with respect to characteristic characteristics within the context of the present method, may be part of the training document 310. All of the characteristics included in the training document 310 can serve as characteristic characteristics for evaluating the similarity between the document and the test element 320. Different training documents 310 may have some of the characteristics shown in common with the training document 310 shown in FIG. 3. However, the training documents may also differ from each other with respect to some characteristics.

[0065] Test element 320 also exists. Test element 320 includes several characteristics 333, 335, 337, 339. These characteristics deviate from the features of the training document 310 with respect to several properties such as their position, distribution, data format, number, shape, etc. By directly comparing the training document 310 and the test element 320, the deviation between these items can be appropriately identified. However, when the test element 320 is inspected independently of the training document 320, a person may consider the test element 320 to be the (original) document. This misunderstanding may occur because the memory of the exact nature of the characteristics of the training document 310 or the (original) document is limited when the document does not currently exist. Therefore, although an objective observer may approve that the test element 320 is similar to the training document 310, upon detailed comparison of these items, there may be significant differences between these items. As a result, the objective observer may identify that the data based on the digital representation of the characteristics 333, 335, 337, 339 of the test element 320 is similar to the data representing the characteristic features 332, 334, 336, 338 of the training document 310. Therefore, the processing of the data representing the test element 320 is prohibited, or the data is corrected by prohibited means so that the data is not processed further.

[0066] Although the present invention has been described above with reference to specific embodiments, the present invention is not limited to these embodiments, and it is doubtless that further alternatives within the scope of the claimed invention will occur to those skilled in the art.

[0067] Embodiment 1. A computer-implemented method for preventing unauthorized processing of a digital representation of at least a part of a document, in particular a part of a banknote, comprising: a) providing data, wherein the data is based on the digital representation of at least a part of a test element, in particular the digital representation of the at least one part of the test element is an image file corresponding to the at least one part of the test element, the step of providing; b) analyzing the data based on the digital representation of the at least one part of the test element with respect to data representing at least one characteristic property of the at least one part of the document; c) activating a prohibiting means if the data based on the digital representation of the at least one part of the test element is similar to the data representing the at least one characteristic property; comprising; further processing of the data based on the digital representation of the at least one part of the test element is prohibited by the prohibiting means and / or the data based on the digital representation of the at least one part of the test element is corrected by the prohibiting means so as to prevent further processing of the data. A computer-implemented method.

[0068] Embodiment 2. d) recording the data based on the digital representation of the at least one part of the test element by an inspection device, in particular by an appearance inspection device such as a scanning device and / or a camera, wherein in particular the inspection device is configured to provide a data file of the at least one part of the test element having a resolution in the range of 50 dpi to 2000 dpi, in particular in the range of 100 dpi to 1000 dpi, more particularly in the range of 200 dpi to 600 dpi, even more particularly in the range of 300 dpi to 400 dpi; The computer-implemented method according to Embodiment 1, further comprising.

[0069] Embodiment 3. e) providing at least a part of a training document; f) A step of recording data representing a digital representation of at least one part of the training document by an inspection device, particularly an appearance inspection device such as a scanning device and / or a camera, wherein, in particular, the inspection device is configured to provide a data file of at least one part of the training document having a resolution in the range of 50 dpi to 2000 dpi, particularly in the range of 100 dpi to 1000 dpi, more particularly in the range of 200 dpi to 600 dpi, and even more particularly in the range of 300 dpi to 400 dpi, the recording step; g) A step of identifying data representing characteristic features in the data representing the digital representation of at least one part of the training document by artificial intelligence and / or machine learning; A computer-implemented method according to any of the preceding embodiments, further comprising.

[0070] Embodiment 4. h) A step of storing the data identified in step g) in a storage memory, wherein, in particular, the data is stored encrypted and / or error-coded, the storing step A computer-implemented method according to Embodiment 3, further comprising.

[0071] Embodiment 5. The inspection device records the data in step d) and / or in - the angular direction of the item with respect to the inspection device, - the item inspected by the inspection device that has been trimmed and / or cut, - the resolution provided by the inspection device, - the distortion of the item with respect to the inspection device, and - the scaling effect applied to the item inspected by the inspection device substantially independently of at least one of, a computer-implemented method according to any one of Embodiments 2 to 4.

[0072] Embodiment 6. The inspection device operates in a reflection mode and / or a transition mode. In particular, the inspection device comprises a detector and a radiation emission source, and is a computer-implemented method according to any one of Embodiments 2 to 5.

[0073] Embodiment 7. Steps b) and c) are performed by the artificial intelligence and / or machine learning. In particular, in step c), the artificial intelligence and / or machine learning determines whether the data based on the digital representation of the at least one part of the test element is similar to the data representing the at least one characteristic feature of the at least one part of the document, and is a computer-implemented method according to any one of Embodiments 3 to 6.

[0074] Embodiment 8. The characteristic features are - the single or multiple specific distributions of contrast levels and / or colors and / or marks arranged and / or printed on the surface of the item and / or contained within the item, and - the single or multiple shapes of marks arranged and / or printed on the surface of the item and / or contained within the item, and - single or multiple moiré patterns, microstructures, microtexts, cryptographs, gyoshes, rainbows, concave lenses, optical elements, holograms, kinegrams, optical lenses, watermarks, QR codes and fingerprints, - single or multiple specific materials arranged on the surface of the item and / or contained within the item, in particular, the specific materials include at least one of fabrics such as paper, polymers and cotton, - single or multiple security mechanisms arranged on the surface of the item and / or contained within the item, in particular, the security mechanisms include at least one of holograms, microlenses, embedded security threads, windows, labelings and symbols, and - combinations thereof Any computer-implemented method of the previous embodiments, which is at least one of them.

[0075] Embodiment 9. Step b) includes determining a reference value, and the reference value is based on the probability that the data representing the at least one characteristic property of the at least one part of the document is similar to the data based on the digital representation of the at least one part of the test element. The reference value is true when the reference value is greater than a predetermined threshold value. Step c) includes activating the prohibiting means when the reference value is true. Any computer-implemented method of the previous embodiments.

[0076] Embodiment 10. Determining the reference value takes into account the pixelation of the data based on the digital representation of the at least one part of the test element. In particular, determining the reference value also takes into account the resolution and / or color distribution and / or contrast distribution and / or luminance distribution of the data based on the digital representation of the at least one part of the test element. The computer-implemented method of Embodiment 9.

[0077] Embodiment 11. The method is configured to be executed locally or remotely within a first data processing device comprising a storage memory in which the respective codes of the method are stored. The first data processing device comprises a storage memory in which the respective codes of the method are stored. The first data processing device is connected via a data connection to a second data processing device. The method is executed on the second data processing device via the data connection and / or The data based on the digital representation of the at least one part of the test element in step a) is provided based on the transmission of the data from the first data processing device to the second data processing device. In particular, the digital representation of the at least one part represents a one-dimensional or two-dimensional part of the test element and / or The method is executed within a period of less than 60 seconds, in particular within a period between 100 milliseconds and 30 seconds, in particular within a period of less than 1 second. A computer-implemented method according to any one of the preceding embodiments.

[0078] Embodiment 12. A computer-implemented method according to any one of the preceding embodiments, wherein for the digital representation of the at least one part of the document, the data representing the at least one part of the digital representation of the test element is not authenticated.

[0079] Embodiment 13. An apparatus for data processing, comprising means for performing the method according to any one of Embodiments 1 to 12.

[0080] Embodiment 14. A computer program product comprising instructions, which, when executed by an apparatus for data processing, cause the apparatus for data processing to perform the method according to any one of Embodiments 1 to 12.

[0081] Embodiment 15. The computer program product according to Embodiment 14, wherein the code of the computer program product is stored encrypted and / or error-coded.

Claims

1. A computer-implemented method for preventing unauthorized processing of digital representations of at least a portion of a test element, wherein the test element deviates from an original security document and a training document, but is similar to the original security document to such an extent that an objective observer might erroneously judge the test element to be the original security document, a) providing at least a portion of the training document; b) recording, by an inspection device, an image file of the training document representing a digital representation of at least a portion of the training document, wherein the inspection device is configured to provide a data file of at least a portion of the training document having a resolution within a range of 50 dpi to 2000 dpi, the recording step; c) identifying, by artificial intelligence, data representing characteristic features within the image file of the training document, wherein the image file of the training document is inspected by artificial intelligence to identify a set of traits representing appropriate criteria for evaluating similarity to the test element, and the traits of the characteristic features are such that, when realized in the test element as well, the objective observer might erroneously judge the test element to be the original security document, and the characteristic features by which the test element is analyzed are independent of a given input, the identifying step; d) providing an image file of the test element, wherein the image file of the test element is based on a digital representation of at least a portion of the test element, and the test element deviates from the original security document and the training document, the providing step; e) analyzing, by the artificial intelligence, the image file of the test element with respect to a identified trait representing at least one characteristic feature of at least one part of the training document within the image file of the training document, wherein the trait, if realized similarly in the test element, may cause an objective observer to erroneously judge the test element as the original security document; the analyzing step; f) activating a prohibiting means if the image file of the test element is similar to the image file representing the at least one characteristic feature; comprising; further processing of the image file of the test element is prohibited by the prohibiting means and / or the image file of the test element is corrected by the prohibiting means so as to prevent further processing of the image file of the test element; the similarity between the image file of the test element and the image file representing the at least one characteristic feature is related to an incorrect evaluation of similarity by an objective observer; the computer-implemented method does not authenticate the image file of the test element with respect to a digital representation of an image file of a training document including at least a part of the training document; A computer-implemented method.

2. g) recording, by an inspection device, in particular by an appearance inspection device of a scanning device and / or a camera, the image file of the test element, wherein in particular the inspection device is configured to provide a data file of at least one part of the test element having a resolution within a range of 50 dpi to 2000 dpi, in particular within a range of 100 dpi to 1000 dpi, more particularly within a range of 200 dpi to 600 dpi, even more particularly within a range of 300 dpi to 400 dpi; the recording step; The computer-implemented method according to claim 1, further comprising.

3. The computer-implemented method according to claim 1 or 2, wherein step c) is performed by the artificial intelligence including machine learning.

4. h) A step of storing an image file of the test element identified in step c) in a storage memory, in particular, the image file of the test element is stored in an encrypted and / or error-coded manner, the storing step. The computer-implemented method according to claim 3, further comprising.

5. In step b) and / or g), the inspection device - The angular direction of the original document and / or the test element and / or the training document regarding the inspection device, - The original document and / or the test element and / or the training document inspected by the inspection device that has been trimmed and / or cut, - The resolution provided by the inspection device, - The distortion of the original document and / or the test element and / or the training document regarding the inspection device, and - The scaling effect applied to the original document and / or the test element and / or the training document inspected by the inspection device The computer-implemented method according to any one of claims 2 to 4, which is configured to record an image file of the test element substantially independently of at least one of.

6. The inspection device operates in a reflection mode and / or a transmission mode, in particular, the inspection device includes a detector and a radiation emission source. The computer-implemented method according to any one of claims 2 to 5.

7. Steps e) and f) are performed by the artificial intelligence including machine learning. The computer-implemented method according to any one of claims 3 to 6.

8. The characteristic feature is - A single or multiple specific distributions of contrast levels and / or colors and / or marks arranged and / or printed on the surface of the original document and / or the test element and / or the training document, and / or included in the original document and / or the test element and / or the training document, - A single or multiple shapes of marks arranged and / or printed on the surface of the original document and / or the test element and / or the training document, and / or included in the original document and / or the test element and / or the training document, - One or more moiré patterns, microstructures, microtexts, digital security marking mechanisms invisible to the naked eye, gyo-sha, rainbows, concave lenses, optical variable elements, holograms, optical lenses, watermarks, QR codes, and fingerprints, - One or more specific materials arranged on the surface of the original document and / or the test element and / or the training document, and / or included in the original document and / or the test element and / or the training document, in particular, the specific material includes at least one of paper, polymer, and fabric, - One or more security mechanisms arranged on the surface of the original document and / or the test element and / or the training document, and / or included in the original document and / or the test element and / or the training document, in particular, the security mechanism includes at least one of holograms, microlenses, embedded security threads, windows, labels, and symbols, and - Combinations thereof The computer-implemented method according to any one of claims 1 to 7, which is at least one of the above.

9. Step e) includes determining a reference value, the reference value being based on the probability that the image file of the test element is similar to the image file of the test element, the reference value being true when the reference value is greater than a predetermined threshold value, and step f) includes activating the prohibiting means when the reference value is true, the computer-implemented method according to any one of claims 1 to 8.

10. Determining the reference value takes into account the pixelation of the image file of the test element. In particular, determining the reference value also takes into account the resolution and / or color distribution and / or contrast distribution and / or luminance distribution of the image file of the test element, the computer-implemented method according to claim 9.

11. The computer-implemented method is configured to be executed locally or remotely within a first data processing device comprising a storage memory in which the respective codes of the computer-implemented method are stored, the first data processing device comprising a storage memory in which the respective codes of the computer-implemented method are stored, the first data processing device being connected via a data connection to a second data processing device, and the computer-implemented method being executed on the second data processing device via the data connection, the computer-implemented method according to any one of claims 1 to 10.

12. The image file of the test element in step d) is provided based on the transmission of the image file of the test element from the first data processing device to the second data processing device. In particular, at least a partial digital representation of the test element represents a one-dimensional or two-dimensional portion of the test element, the computer-implemented method according to any one of claims 1 to 11.

13. The computer-implemented method is executed within a period of less than 60 seconds, in particular within a period of 100 milliseconds to 30 seconds, in particular within a period of less than 1 second, the computer-implemented method according to any one of claims 1 to 12.

14. An apparatus for data processing, comprising means for performing the computer-implemented method according to any one of claims 1 to 13. **Claim 15** A computer program comprising instructions which, when the computer program is executed by an apparatus for data processing, cause the apparatus for data processing to perform the computer-implemented method according to any one of claims 1 to 13. **Claim 16** The computer program according to claim 15, wherein the code of the computer program is stored encrypted and / or error-coded.

Images