Method, computer readable medium, and control device

Abstract

According to various aspects, a method (300) comprises: determining (201) two or more graphic patterns of a graphic reference layout (102) based on a plurality of print renditions (114) of the graphic reference layout (102), the two or more graphic patterns satisfying criteria for printing related position changes (d) relative to each other; determining (205) a model (250) of the graphic reference layout (102) based on a result of the determining (201) of the two or more graphic patterns, where the model implements a representation (250a, 250b, 250c) of the graphic reference layout (102) for each of the two or more graphic patterns, at least a portion of the graphical pattern is masked or at least marked by the representation.

Description

Technical Field

[0001] Each aspect generally involves a method, a computer-readable medium, and a control device. Background Technology

[0002] Banknotes are typically used for financial transactions worldwide and thus circulate through various instances such as bank terminals, shops, safes, and wallets. During circulation, the condition of banknotes can deteriorate (e.g., due to mechanical stress and / or contamination). Therefore, some instances are used to check whether banknotes are suitable for further use and, if necessary, remove them from circulation. One challenge in this context includes distinguishing various forms of contamination, such as stains and graffiti, from the potentially intricate designs of banknotes. Attached Figure Description

[0003] Throughout the accompanying drawings, it should be noted that similar reference numerals are used to describe the same or similar elements, features, and structures. The drawings are not necessarily drawn to scale; rather, the emphasis is generally placed on illustrating aspects of this disclosure. In the following description, some aspects of this disclosure are described with reference to the following drawings, in which: Figure 1 A schematic diagram of the printing process is shown, based on various aspects. Figure 2 , Figure 3 and Figure 4 Each shows schematic diagrams illustrating various methods based on different aspects; and Figure 5 A schematic diagram of a bank terminal is shown, based on various aspects. Detailed Implementation

[0004] The following detailed description refers to the accompanying drawings, which illustrate, by way of illustration, specific details and aspects in which the present disclosure may be practiced. One or more aspects are described in sufficient detail to enable those skilled in the art to practice the present disclosure. Other aspects may be utilized, and structural, logical, and electrical changes may be made, without departing from the scope of the present disclosure. The aspects are not necessarily mutually exclusive, as some aspects may be combined with one or more other aspects to form new aspects. The aspects are described in terms of combination methods, and the aspects are described in terms of combination apparatus. However, it is understood that the aspects described in terms of combination methods can be similarly applied to apparatus, and vice versa. Throughout the drawings, it should be noted that similar reference numerals are used to depict the same or similar elements, features, and structures.

[0005] This disclosure may include various processes (e.g., methods and functions thereof). In some embodiments, the processes may be executed by hardware components or may be embodied in computer-readable instructions that can be used to execute the processes by a general-purpose or special-purpose processor or logic circuit programmed with the instructions. Alternatively, the processes may be executed by a combination of hardware and software. Depending on various aspects, one or more processes executed by one or more processors may be illustratively implemented as counterparts by code segments stored in memory, wherein, if executed by one or more processors, the code segments cause one or more processors to perform the processes (e.g., functions and methods). For example, code segments provided as part of software may be updated on demand via (e.g., mobile) networks.

[0006] As an example, the term "processor" as used herein can be understood as any type of entity capable of processing data, signals, etc. As an example, data or signals can be processed according to one or more specific functions performed by a processor. Therefore, as an example, a processor can be or include analog circuits, digital circuits, mixed-signal circuits, logic circuits, processors, microprocessors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), field-programmable gate arrays (FPGAs), integrated circuits, application-specific integrated circuits (ASICs), or any combination thereof. Any other type of implementation of the corresponding functions, which will be further described in detail below, can also be understood as a processor or logic circuit. It is understood that any two (or more) processors or logic circuits detailed herein can be implemented as a single entity with equivalent functionality, and conversely, any single processor or logic circuit detailed herein can be implemented as two (or more) separate entities with equivalent functionality. It is understood that one or more of the method steps detailed herein can be performed (e.g., implemented) by a processor, or can be performed (e.g., implemented) by one or more specific functions performed by a processor.

[0007] The term "control device" (also known as a controller) may be understood herein to refer to any type of logical implementation entity that can be implemented by one or more processors (e.g., including dedicated circuit systems) executing software stored in memory, firmware, or any combination thereof. Thus, a "control device" may include hardwired or programmable logic circuitry, such as a programmable processor, for example a microprocessor (e.g., a Complex Instruction Set Computer (CISC) processor or a Reduced Instruction Set Computer (RISC) processor). A "control device" may, for example, include one or more processors executing software or at least segments of code (e.g., any type of computer program, such as a computer program using virtual machine code, such as Java).

[0008] The control device may optionally include memory, for example, storing code segments representing processes provided by the control device (e.g., controlling one or more operational functions). Additionally or alternatively, the memory may store one or more standards, rules, and algorithms, as detailed herein, for example.

[0009] As used herein, “memory” can be understood as a non-transitory computer-readable medium in which data or information can be stored for retrieval. Therefore, references to “memory” included herein can be understood to refer to volatile or non-volatile memory, including random access memory (“RAM”), read-only memory (“ROM”), flash memory, solid-state storage devices, magnetic tape, hard disk drives, optical drives (by way of example), or any combination thereof. Furthermore, it is understood that, by way of example, registers, shift registers, processor registers, and data buffers are also covered by the term memory herein. It is understood that a single component referred to as “memory” or “a memory” can consist of more than one different type of memory, and therefore can refer to a collection of components including one or more types of memory. It is readily understood that any single memory component can be separated into multiple commonly equivalent memory components, and vice versa. Furthermore, while memory can be depicted as separate from one or more other components (as in the figures), it is understood that memory can be integrated within another component, such as on a common integrated chip.

[0010] Refer to “image data” and the processing of image data. Image data can be the result of optical sensing (e.g., of an object) of reality (also known as an imaging process or simply imaging), and can therefore be considered a digital representation of reality, such as its current state at the time the image data is acquired. For example, an imaging process may include (e.g., by means of an optical system, such as including one or more lenses and / or one or more mirrors) projecting light onto the surface of an image sensor (e.g., a Bayer sensor) and reading out the image sensor. The image data read out from the image sensor may be raw image data (also known as RAW), including information about the light sensed by the image sensor, such as per pixel (also known as raster graphics). Image data may optionally (e.g., before or during processing) be converted to an image format type different from RAW, such as converting to another type of raster graphics or vector graphics, so that further processing can be based on that image format, or can be arbitrarily converted between different image formats. The process may optionally include interpolation (e.g., using de-mosaic) of the data read out from the image sensor, for example, to obtain full multicolor information for each pixel or to use less memory or processing power. Image data may be optionally compressed (e.g., to require less storage space or computing power) or uncompressed (e.g., to avoid distortion). The corresponding image format providing the image data may also define a color model, according to which the color information of the image data is specified. For example, image data may be represented as coordinates in a color space based on the color model.

[0011] The simplest case is the binary color space, according to which each pixel provides a bit value (e.g., representing a black or white value). In more complex color spaces (e.g., grayscale color models), the color of the color space is identified along with its intensity levels, for example, providing multiple levels between black and white for each pixel (also called grayscale values). However, color models can also be defined by multiple (e.g., two or more) basic colors (such as red, green, and blue). For example, wavelength-sensitive image sensors can be used to acquire image data that includes multicolor information. The corresponding light properties can be expressed according to the color model used. It is understood that references to image data or color models herein can be similarly applied to multicolor (also called polychromatic) or monochromatic color models. Display devices can be used for the graphical reproduction of image data on the display device. For such graphical reproduction, the image data is converted by a graphics processor into control signals for controlling the display device. For ease of understanding, image data as mentioned herein is depicted as a graphical reproduction of image data, but can be processed into any image format type, for example, as a bit string. For example, image data stored in a storage medium may include, for example, bit strings stored in a file (also called a digital image or image file) according to a specific image format type.

[0012] An imaging apparatus may be understood herein as a device through which an imaging process can be performed. For example, an imaging apparatus may include an image sensor, optionally an optical system, and a data interface for providing readout image data. In the context of an apparatus (e.g., an imaging apparatus), the term “type” may be understood herein to refer to construction (also known as construction type) that is identical for all apparatuses of the same type (e.g., the same production series). For example, two imaging apparatuses of the same type match in their construction properties (e.g., their architecture), such as the construction properties of their image sensors (e.g., their pixel density, their total number of pixels, their size, technology, etc.), their optical systems (also known as photographic objectives), and / or their data interfaces. For example, two imaging apparatuses of the same type match in one or more (e.g., each) of the following sensory properties, such as sensory nonuniformity (light response nonuniformity), signal-to-noise ratio, resolution, pixel density (e.g., pixels per inch), etc. Similarly, image data sensed by two imaging devices of the same type (and the processing results based thereon) match one or more of the following image properties (e.g., each): for example, sensory nonuniformity (light response nonuniformity), signal-to-noise ratio, resolution, pixel density (e.g., pixels per inch).

[0013] The term "banking terminal" may be understood herein to mean any device that includes an interface (e.g., a human-machine interface) for interacting with a person and capable of performing one or more banking transactions in response to such interaction (e.g., in response to human instruction). Examples of banking transactions may include: receiving cash from a person, distributing cash to a person, and financial transfers from or to a person's account (e.g., account-to-account payments or financial transfers).

[0014] The term “cash” as used herein may be understood to mean any (e.g., physical) medium of financial exchange, such as banknotes and coins. The term “banknote” as used herein may be understood to mean any negotiable promissory note (e.g., certificate), produced by a bank or other licensed institution, and payable on demand to the holder. In particular, this document refers to banknotes in printed form (e.g., sheet material), such as printed paper, printed fabric, or printed polymer sheet. The term “money” as used herein may be understood to refer to the financial system and its measurement of cash as a medium of financial exchange.

[0015] The term "printing" as used herein can be understood to refer to any process (also known as a printing process) that reproduces a graphic reference layout by transferring a colorant (also called a dye) onto a surface to be printed (such as paper, fabric, or polymer sheet). The term "colorant" as used herein can be understood to refer to any substance that, when applied to a surface, can (e.g., permanently) change the color of the surface (e.g., color). Examples of colorants include inks, dyes, paints, pigments, etc. Examples of printing processes include gravure printing, offset printing, inkjet printing, laser printing, letterpress printing, screen printing, etc.

[0016] A graphic reference layout (also known as a printing template in the context of printing) can be provided in any digital or analog form and can include one or more graphic patterns (also known as layout components), for example, in the form of printed glyphs (e.g., one or more numerals and / or one or more letters) and / or decorative elements (e.g., ornaments and / or scenes). The term "printed matter" can be understood herein to refer to the physical result of the printing process, for example, a printed version including a graphic reference layout (also referred to as its printed reproduction, its printed sample, or printed output). Various examples of graphic reference layouts (also known as reference graphics) indicate the arrangement (e.g., positional relationships, sizes, etc.) of multiple graphic patterns to be printed. A graphic reference layout for a banknote may include one or more of the following graphic patterns: a graphic background; a graphic representation of the denomination; a graphic security feature; an identification number; and a picture element. The term "picture element" can include all types of elements (e.g., decorative, non-functional, for example) (e.g., pictures or portions thereof) that are not necessarily functional (e.g., for identification, security, naming, etc.). Examples of picture elements may include portraits, landscapes, ornaments, drawings, buildings, etc.

[0017] The term "multi-stage" printing process can be understood herein to refer to any (e.g., simulated) printing process that uses multiple colorants and, for each of the multiple colorants, includes stages (also called printing process stages) configured to transfer the colorant to the surface. In various examples, the configuration of a multi-stage printing process (e.g., the number of stages, specific colorants, etc.) is related to a color model, such as the CMYK color model, which involves the use of colorants in cyan, magenta, yellow, and key (black). For example, multiple colorants may be used to reproduce different colors and / or individually reproduce graphic patterns (e.g., having one or more common colors). In various examples, printing process stages may be applied sequentially, for example, to reproduce multiple layout components layer by layer.

[0018] In relation to a color space (or a base color model), a graphic reference layout can be represented as an overlay of multiple individual components (also called layout components), each associated with one of a plurality of colorants, and reproducible by a printing process configured to transfer the colorant to a surface. Each layout component may include one or more graphic patterns of the graphic reference layout and may be associated with a printing layer through which the layout component is reproduced.

[0019] Similarly, multi-stage printing apparatuses can be configured to perform multi-stage printing processes. For each stage of a multi-stage printing process, the multi-stage printing apparatus may include a printing stage configured to provide the printing stage, for example, by adding a printing layer to a surface. In various examples, the material to be printed (e.g., a sheet) may be continuously supplied to the printing stages, for example, by transferring colorants layer by layer. Multiple colorants may be transferred one over another onto the surface (also known as overprinting), such that their superposition provides a printed reproduction of a graphic reference layout. In various examples, each printing stage may include a colorant carrier (e.g., a plate or roller) that carries colorants according to the printing stage provided by the printing stage. Transferring the colorant carried by the colorant carrier to the surface may include pressing the colorant carrier against the surface.

[0020] In various examples, a multi-stage printing process can be a multi-type printing process, where the stages of the printing process differ from one another in terms of printing type (e.g., printing technology). Examples of different printing types include offset printing, gravure printing, and / or security feature printing.

[0021] For the sake of understanding, the result of printing layout components through printing process stages (e.g., colorant transferred from the surface by printing process stages) is also referred to herein as a printed layer, where reference to it can be similarly applied to non-layered printing processes. For example, each printed layer may include one or more printed graphic patterns based on a graphic reference layout. For the sake of understanding, a reference printed layer in the printed layers (e.g., the bottommost printed layer) is referred to as a "background printed layer," and subsequent printed layers are referred to as "foreground printed layers." Similarly, the corresponding layout components are referred to as "background layout components" or "foreground layout components," which also applies to their graphic patterns.

[0022] In the context of one or more objects (e.g., layout components and / or graphic patterns), the term "position" (or "positional") may be understood herein as an indication of the location and / or orientation of one or more objects. The relative position of two objects may be understood herein as an indication of the position (e.g., distance) and / or orientation (e.g., by using one of the two objects as a reference) of two objects relative to each other, such as the distance between them. The position of a spatially extended object may be a function of the position of a reference point of the object, e.g., expressed in coordinates. The orientation of a spatially extended object may be a function of the positions of two reference points of the object relative to each other.

[0023] Due to variations in the multi-stage printing process (e.g., alignment variations, apparatus variations, etc.), the position on the surface (at which the graphic pattern of the graphic reference (e.g., the entire layout assembly) is reproduced by the printing process) can vary from the printed output of the graphic reference to the printed output (also known as print-dependent positional variation). For example, two printed outputs of the same graphic reference may differ from each other in the position at which the same graphic pattern is reproduced on the surface of the printed output. The same applies to the position at which two different graphic patterns are reproduced on the surface of the same printed output. For example, the distance and / or misalignment of two or more graphic patterns relative to each other can be used as the positional relationship between two or more graphic patterns relative to each other. For example, for a larger number of printed outputs, print-dependent positional variation can be a function of parameters (e.g., variance and / or deviation) of the statistical distribution of the positional relationship. It is understood that any other suitable parameter can be used to represent the positional relationship.

[0024] The term "model" in this document can be understood as a data-based (e.g., digital and / or virtual) representation of an original (e.g., physical or virtual) object (e.g., apparatus) or process (e.g., printing process). Determining a model may include, for example, determining one or more parameters (also called model parameters) of the model representing the original properties by mapping attributes to model parameters. To determine a model, attributes may be abstracted, parameterized, and / or simplified. Examples of attributes presented by model parameters (also called representations of attributes) may include: one or more geometric attributes (e.g., length, distance, volume, and / or position, as examples), weight, volume, composition, color-related attributes (e.g., represented as coordinates in a color space), logical information (links, sequences, couplings, interrelationships, dependencies, etc.), time-related information (e.g., time, total duration, frequency, period duration, etc.), and / or functional information (e.g., current, effect, working point space, force, degrees of freedom, etc.).

[0025] For example, a model of a graphic reference layout can be provided component-by-component, for instance, by implementing a representation for each layout component. The representation of a layout component can represent one or more geometric properties (e.g., length, distance, volume, and / or position, as an example) and / or one or more color-related properties (e.g., represented as coordinates in a color space). The representation of a layout component can be provided as a representation of a graphic reference layout, where one or more graphic patterns that are not part of the layout component are masked (e.g., as transparency or as reference color values). Individual representations of the model can be independent of each other, for example, provided as separate images. Similarly, a model of a printed output (e.g., a banknote) can include representations of one or more geometric properties (e.g., length, distance, volume, and / or position, as an example) and / or one or more color-related properties (e.g., represented as coordinates in a color space) of one or more printed graphic patterns (or the entire printed graphic layout) of the printed output.

[0026] The term "overlay" (also referred to as superposition in the context of graphics and images) for two objects (e.g., image data and / or graphic patterns) can be understood herein to refer to the result of superimposing two objects, such as the overlay of two objects. An overlay (e.g., each pixel thereof) can be a (not necessarily linear) function of each of the two objects (e.g., their pixels), such as in the case where a semi-transparent version of one of the two objects overlays the other (also known as transparent overlay). An exemplary overlay of two pixels can include averaging the two pixels, which can be applied to each pixel of the object. However, other functions can be used to determine the overlay (e.g., if the image contains more than two objects). An overlay (e.g., each pixel thereof) can also be a (not necessarily linear) function of only one of the two objects (e.g., their pixels), such as in the case where an opaque version of one of the two objects overlays the other (also known as opaque overlay).

[0027] Typically, superimposing image data from multiple (e.g., multiple) printed outputs of the same graphic pattern can reduce granularity and / or reduce printing-related variations (e.g., due to defects and other artifacts). Furthermore, image sharpness and / or image contrast can be functions of the positional relationship between multiple printed outputs relative to each other, where they are superimposed. Image sharpness refers to the level of clarity and / or detail in an image. When an image is sharp, it has clear, well-defined edges and a high level of detail. When an image is not sharp, it may appear blurry or out of focus.

[0028] For example, the sharpness of the superimposed image can increase as positional variation (e.g., through alignment) decreases among multiple print outputs, for example, exceeding the average sharpness of the multiple print outputs. Conversely, the resulting superimposed image sharpness can decrease as positional variation (e.g., through misalignment) among multiple print outputs (e.g., below the average sharpness of the multiple print outputs) decreases (through blurred superposition, for example, similar to motion blur). In some embodiments, the sharpness of the superimposed image can be determined, for example, relative to a reference sharpness (e.g., average sharpness) (e.g., its variation (enhancement or reduction)).

[0029] For example, sharpness can be determined (e.g., parameterized) by determining one or more image parameters (e.g., edge gradient, noise, granularity, and / or contrast, etc.) and / or mapping one or more image parameters to sharpness information (e.g., its numerical form). Additionally or alternatively, sharpness can be determined (e.g., parameterized) using metrics (such as, for example, Laplacian edge detection, Sobel edge detection, or Canney edge detection) and / or mapping metrics to sharpness information (e.g., its numerical form). In some embodiments, sharpness can satisfy a sharpness criterion, for example, represented by a threshold. For example, a sharpness criterion can be satisfied when sharpness (e.g., sharpness information) exceeds a threshold and / or when changes in sharpness remain below a predetermined threshold (e.g., to avoid excessive reduction in sharpness).

[0030] The term "masking" as used herein can be understood to refer to a process of manipulating image data by selecting one or more portions of the image data according to a masking function (also known as masking). A masking function may be expressed mathematically (e.g., as part of an algorithm) and / or equivalently, and where appropriate, by additional image data (also referred to in this context as an image processing mask or simply a mask) that can be used as an overlay. In an exemplary implementation, the mask comprises a grayscale image whose grayscale values ​​indicate the level of selection. The result of masking may include a modified version of the image data (e.g., through a filter) in which one or more portions are marked as selected (also known as masking), which in one or more embodiments may include filtering the image data according to a masking function (e.g., such that only one or more portions are filtered). The masking function can be used, for example, to isolate portions from residues of image data by changing the color of the residue.

[0031] The term "filtering" in this document can be understood to refer to the process of manipulating image data by transforming (or masking) one or more portions of the image data according to a filter function (also known as an image processing filter), for example, in terms of color, transparency, and / or geometry. The filter function can be expressed mathematically, for example (e.g., as part of an algorithm). As an example, filtering may include setting one or more portions to black or another reference color value, or setting them to transparent. For example, a masking function indicates (e.g., marks) one or more portions of the image data to be filtered.

[0032] In some examples, masking may include selecting a portion of image data and hiding or revealing that portion by applying a mask. In some examples, masking may be configured to be reversible, for example, when the original image is still fully included in the result of the masking, but it does not need to be reversible. For example, image data may be manipulated pixel by pixel, such as by transforming each pixel according to its coordinates (e.g., transforming to or through a predetermined color space vector). Less complex implementations of masking may include binary graphical masking, where white areas indicate invariant portions of the image data, and where black areas indicate portions of the image data set to common preset color values ​​(e.g., black, white, or other hues).

[0033] The term "classification" (also known as categorization) can be understood herein as referring to the process of assigning an object to one of a plurality of predetermined categories. Examples of such categories may include categories for condition (e.g., new, used, damaged), categories for contamination (e.g., none, less, moderate, severe), categories for damage (e.g., none, common, severely damaged), and categories for integrity (e.g., complete, incomplete). Some or more categories detailed herein (also known as fitness representation categories) may represent the acceptable state of an object for further use, while other categories represent the unacceptable state of the object. For example, in the case of banknotes, when banknotes are damaged and / or severely contaminated, banknotes may be classified as unacceptable (also known as unsuitable).

[0034] The term “predetermined” herein may be understood to indicate one or more of the following properties: immutable, determined in the past, stored (e.g., in a computer-readable medium), read from a computer-readable medium, a constraint, or a boundary condition. For example, predetermined information used in a process (e.g., a method) may be immutable over multiple iterations of the process, may be determined before the process begins, may be stored in a computer-readable medium, such as in a code segment that includes instructions according to the process, and so on.

[0035] To facilitate understanding, for example, regarding the various states detailed herein, the result of a printing process (e.g., provided as printed output) may be indicated by the prefix "printing", the result of an imaging process (e.g., provided as image data) may be indicated by the prefix "imaging", and the result of a modeling process (e.g., provided as a model) may be indicated by the prefix "modeling".

[0036] In the following detailed description, reference is made to banknotes as an example printed output. It will be understood that reference to this can be similarly applied to other printed outputs, such as other printed cash, stamps, etc.

[0037] Various aspects are based on the finding that stain and graffiti detection algorithms (SGD) can allow for improved detectability of stains and graffiti, for example, especially in the field of banknote printing. Some SGD methods use banknote models, also known as banknote numeric models (BNMs). To improve the performance of SGD, such BNMs can be generated for each individual printed banknote being tested, which allows for consideration of the relative positions (also known as positional relationships) of graphic patterns contained in different printing layers or security features. Illustratively, such positional relationships can vary from banknote to banknote (also known as positional variations). To create an individual BNM, for example, for a banknote grayscale image expected to have (e.g., minimum, average, median, or maximum values), the following can be determined: - Select the corresponding background (BG) pattern as the BG reference; - Foreground (FG) patterns are overlaid one after another on the background (BG) references, one after another, to continuously create the background numeric pattern (BNM). Thus, the individual position of each FG reference on the BNM can be adjusted based on the estimated position of its corresponding printed layer relative to the background printed layer, which can be provided by so-called anchor point detection (APD). Typically, the overlay FG patterns can include opaque (non-transparent) or transparent overlays.

[0038] - Mark BNM areas, which should be completely obscured. These areas may have a fixed position (e.g., banknote border) or a variable position (e.g., security thread or serial number).

[0039] To enable BG and FG references to overlap under variable positional relationships relative to each other, a background reference can be provided for the entire banknote. As an example, it is possible to select the offset layer as the BG reference based on the printing plate used for the offset layer. Unfortunately, for many (if not most) monetary systems, the printing plates are not public, and therefore data representing individual printing layers can be determined (e.g., learned, estimated) based on one or more printed samples of the banknote (also known as a test set).

[0040] According to various embodiments, multiple printed banknotes (BNs) can be provided as a test set, where each BN can be imaged. Based on this, a BG reference can be determined in the following manner: - Transform each imaged BN (also known as a BN image) into a normalized position (which produces a position-aligned BN image) so that the offset layers of all banknotes are aligned. - The desired (e.g., minimum or median) reference is calculated by solving for the expected value for each pixel from the BN image aligned to the set of positions.

[0041] This approach may be suitable for median references and / or specific BN designs, such as the back side of a EURBN consisting of a single BG printed layer. However, for calculating the (e.g., minimum) BG reference of the front side of the EUR BN, this method, similar to statistical note calculation, results in unsatisfactory BNM. This is due to the movement of the FG pattern relative to the BG pattern, resulting in altered occlusion of the BG pixels from BN to BN to reduce occlusion. Exemplary implementations may include the following: - Identify those FG printed layers and their graphic patterns that interfere with the BG reference estimation; - Create an APD adapter for estimating the location of the BG and related FG printed layers. - Create a binary FG mask for each FG printing layer to identify the position of the contained pattern in a normalized coordinate system. -BN for each image in the test set: a) Transform that BN image to a normalized position so that the offset layer is aligned with its normalized reference frame (e.g., based on the APD). b. Convert each binary FG mask to a normalized BG reference system using APD, taking into account the offset between FG and BG. c. Overlay a binary FG mask onto the transformed BN image, and mark the pixels in the transformed BN image that belong to FG as invalid; - Determine the desired BG reference from the transformed and masked BN image, while ignoring BN-specific invalid pixels in this calculation. - Calculate estimates for those pixels in the BG reference where no statistics can be solved because those pixels are marked as invalid for all transformations of the BN image of the test set, for example, by classical extrapolation.

[0042] Various more general embodiments are described in detail below.

[0043] Figure 1A schematic process diagram 100 illustrates a printing process (e.g., multi-stage) according to various aspects. A graphic reference layout 102 may include various graphic patterns, exemplarily including a background pattern 102a, a first foreground pattern 102b, and a second foreground pattern 102c. The graphic reference layout 102 may be provided as an overlay of a plurality of layout components 104a, exemplarily including a first layout component 104a including the background 102a, a second layout component 104b including the first foreground pattern 102b, and a third layout component 104c including the second foreground pattern 102c. It should be understood that, for example, in a preferred embodiment, the background 102a fills the entire banknote and one or more foreground patterns 102b cover the background 102a, such that the blank areas shown for the first layout component 104a do not necessarily have to exist.

[0044] A graphic reference layout 102 (e.g., each layout component thereof) can be supplied to a multi-stage printing apparatus 110, which includes multiple printing stages configured to print the graphic reference layout 102 (e.g., each layout component thereof) on each of a plurality of sheets 112. Thus, the printing apparatus 110 can output a plurality of print outputs 114, each print output 114 comprising a sheet and a reproduction of the graphic reference layout 102 printed on the sheet 112.

[0045] In the example of a banknote as printed output 114, the printing process may include various printing techniques, such as offset printing, intaglio printing, and / or security feature printing, as examples. Offset printing may include transferring an inked image from a plate (also called a printing plate) to a rubber blanket that serves as a colorant carrier, and then onto the banknote, which can be used to print most of the banknote (e.g., its background printing layer, including one or more background images, text, and other elements). Intaglio printing may include pressing an inked plate, which serves as a colorant carrier, onto the banknote to create a raised surface that can be used to produce detailed images, such as portraits, and one or more security features (e.g., watermarks and / or lines). Security feature printing may be used to add security features to the banknote, which may include holograms, UV-sensitive inks, and / or microprinting, as examples.

[0046] In some examples, the printing process may include overprinting. Overprinting may include, for example, printing a graphic pattern (e.g., including images and / or text) on top of a (pre-existing) graphic pattern using different printing plates. The inks used for overprinting are typically different colors from those already applied to the printed material, which helps to make the overprinted text or image stand out.

[0047] Due to defects in the printing process (e.g., misalignment), at least two printed outputs 114 may differ from each other in the position of the printed graphic pattern (then also referred to as a position-variable pattern), for example, in its position and / or orientation. Such defects may be based, for example, on the deliberate alteration of the security thread.

[0048] Therefore, the positional relationship between the two printed graphic patterns of print output 114 relative to each other can vary, as shown in Figure 100a as a distribution n as a function of distance d as an example positional relationship. For the sake of understanding, reference will be made to individual position-variable patterns, where each pattern is part of or forms part of a position-variable layout component. If a position-variable layout component comprises two or more graphic patterns (whose positions are fixed relative to each other), then a reference to one of the graphic patterns as a position-variable pattern can be similarly applied to the entire layout component.

[0049] Figure 2 Several methods according to various embodiments in schematic diagram 200 are illustrated, which are linked to a model 250 of a graphical reference layout (also referred to as a modeling layout or layout model), as detailed herein. The layout model 250 may implement multiple representations 250a, 250b, 250c, wherein each representation represents a layout component of the graphical reference layout (e.g., formed by one or more graphical patterns).

[0050] It is understandable that although three representations 250a, 250b, and 250c are shown as examples, the number of multiple representations 250a, 250b, and 250c can be greater than or less than three. Therefore, as an example, this document will also refer to two representations 250a, 250b, and 250c, where reference to them can be similarly applied to more than two representations 250a, 250b, and 250c.

[0051] The first method 300 is configured to determine a layout model 250 based on multiple print outputs 114 (also referred to as print output set 114 or test set) of a graphical reference layout. The second method 700 is configured to assign one or more print outputs 114 to one of multiple predetermined categories 280 (also referred to as classification). The one or more print outputs 114 supplied to method 300 may, but need not, be as follows: Figure 1 The result of the printing process shown.

[0052] Figure 3The method 300 according to various embodiments is illustrated schematically, wherein the printed output 114 includes a plurality of printed graphic patterns, which are exemplary in the form of a star. The method 300 includes, in 201 (also referred to as pattern determination), determining (e.g., identifying) two or more graphic patterns (also referred to as position-variable patterns) of a graphic reference layout 102, which satisfy a criterion of printing-related positional variation relative to each other, such that a reference to the position-variable patterns of the layout components can be similarly applied to the remaining graphic patterns of the layout components.

[0053] In some aspects, pattern determination 201 may be based on a comparison 203 of multiple print outputs 114 (also referred to as print output comparison 203), for example, based on the results of imaging 202 of multiple print outputs 114. The results of imaging 202 of multiple print outputs 114 may include image data 202a, 202b, 202c of print outputs 114 for each of the multiple print outputs 114 (also referred to as imaged print outputs 202a, 202b, 202c), a working example of which is presented below.

[0054] In some embodiments, image data for each of the plurality of printed outputs 114 can be sensed by the same imaging device 204. For example, the imaging device 204 may be part of a bank terminal or at least of the same type as the imaging device of a bank terminal. This increases the reliability of the layout model due to the increased data correlation.

[0055] Print output comparison 203 may include a comparison of the imaged print outputs 202a, 202b, 202c with each other. As an illustrative example, a stack of imaged print outputs 202a, 202b, 202c is depicted, wherein the imaged print outputs 202a, 202b, 202c are positioned relative to each other such that the upper right four-pointed star 203a (as an exemplary reference pattern) of each imaged print output 202a, 202b, 202c is arranged in a common position (also referred to as aligned). As can be seen from the stack, the position of the upper right four-pointed star 203a varies within the print output relative to the surrounding edges of the print output and relative to the six-pointed star 203b, but remains constant relative to the remaining four-pointed stars. Based on this, the six-pointed star 203b and the four-pointed star group can be determined by print output comparison 203 as meeting the criteria for satisfying the print-related positional variations relative to each other.

[0056] Illustratively, the six-pointed star 203b and the group of four-pointed star 203b can be determined by comparing the printed output 203, and can be determined as the result of different printing process stages. For example, each of the four-pointed star in the group can be determined as the result of the first printing process stage of a multi-stage printing process, and the six-pointed star can be determined as the result of the second printing process stage of a multi-stage printing process.

[0057] The depicted stacking of imaging print outputs 202a, 202b, and 202c can be understood as an example of superimposing imaging print outputs 202a, 202b, and 202c to perform print output comparison 203. Other examples of superposition can include any multi-input algorithm configured to map imaging print outputs 202a, 202b, and 202c to their single superposition. Such a multi-input algorithm can include sub-algorithms such as alignment algorithms, transparency algorithms, pattern recognition algorithms, one or more image post-processing algorithms (e.g., for adding transparency, normalizing contrast, etc.). The superposition of imaging print outputs 202a, 202b, and 202c can also be understood as an example component of print output comparison 203, which facilitates print output comparison 203. Additional or alternative implementations of the print output comparison 203 can be configured to determine, for each imaged print output 202a, 202b, 202c, the positional relationship of each pattern relative to another pattern identifiable in the imaged print outputs 202a, 202b, 202c (e.g., by a pattern recognition algorithm). Additional or alternative implementations of the print output comparison 203 can be based on one or more trained algorithms (e.g., including artificial neural networks) and / or on another type of machine learning algorithm. It is understood that artificial neural networks are an example of machine learning algorithms.

[0058] Method 300 includes determining a layout model 250 for a graphic reference layout in 205 (also known as layout model determination 205). The layout model determination is based on the results of pattern determination 201, for example, based on the determined graphic pattern, which meets the criteria for print-related positional variations relative to each other, and / or based on its determined position.

[0059] For each graphic pattern, the printing model 250 implements representations 250a, 250b (e.g., modeling layout components) of a graphic reference layout (also referred to as a printing layer representation), wherein each graphic reference layout is associated with an individual printing process stage and therefore lacks at least one graphic pattern. For example, the printing model 250 implements a first printing layer representation 250a for printing a group of four-pointed stars lacking a six-pointed star and a second printing layer representation 250b for printing a group of six-pointed stars lacking a four-pointed star. As an example, each printing layer representation 250a, 250b may include image data of the graphic pattern based on the imaging print outputs 202a, 202b, 202c (e.g., based on their overlay). For example, each printing layer representation may include an extraction of the overlay.

[0060] In an exemplary implementation where print output 114 includes a background print layer and one or more foreground print layers, pattern determination 201 may include determining (e.g., identifying) one or more foreground print layers and / or one or more printed graphic patterns based on multiple print outputs 114. For example, pattern determination 201 may include determining only those print components (e.g., one or more foreground print layers and / or one or more graphic patterns) that are represented by interference from print outputs 114.

[0061] Optionally, determining each printed layer representation 250a, 250b may include, for example, masking one or more graphic patterns 350 by means of one or more masks. For example, method 300 may include: for each positional variable graphic pattern as generated from pattern determination 201, determining a mask that indicates the positional variable graphic pattern to be masked. The determination of the first printed layer representation 250a may be based on a first mask indicating the 6-pointed star 203b to be masked and one or more of the imaging printed outputs 202a, 202b, 202c, wherein the 6-pointed star 203b, as an exemplary first positional variable graphic pattern, is masked. This may be applied to one or more second printed layer representations 250b, 250c. In its exemplary implementation, one or more imaging printed outputs 202a, 202b, 202c (e.g., their superposition) may be superimposed by a first mask to mask the first positional variable graphic pattern and / or reveal one or more second positional variable graphic patterns (here, the right 4-star 203a of this group is an example).

[0062] It is understandable that the positional variation of the foreground graphic pattern (here, the six-pointed star 203b as an example) allows for the determination of a larger area of ​​the printed layer (e.g., the background printed layer) beneath the foreground graphic pattern. In this case, only a portion of the foreground graphic pattern is masked or at least marked, such that the total area of ​​the masked or at least marked area is smaller than the total area of ​​the foreground graphic pattern. This allows for the determination of a larger portion of the underlying printed layer, which increases the reliability of the layout model 250. For example, due to the positional variation of the security threat, a portion of the background printed layer hidden by the security threat of the first banknote may be visible to the second banknote.

[0063] This application illustrates image data 202 of multiple banknotes as an example of a printed output 114 according to various embodiments, indicating printing-related positional variations of a printed foreground pattern within the banknote, for example, relative to a background printing layer serving as a positional reference. For this exemplary banknote, one or more foreground printing layers are printed using optically variable ink (OVI) via intaglio printing, and include a foil / security thread, while the background printing layer is printed using offset printing. Moiré patterns are visible artifacts in the imaging process and can be suppressed by using sensors with higher pixel density.

[0064] This application illustrates multiple masks (also known as image processing masks) used to determine (also known as background representation determination) the representation of background layout components (also known as background typographic layer representation), wherein the graphic patterns of the remaining foreground layout components are masked.

[0065] Background representation determination may include determining a position reference (e.g., anchor point) for each printed layer of printed output 114 based on printed output 114 (e.g., represented as coordinates), for example, based on its image data. For example, the position of the position reference of the printed layer on printed output 114 may be determined as the printing position of the printed layer. It is understood that determining the printing position may be an example of determining a printing-related positional change, which may additionally or alternatively use another printed layer as a positional reference to determine the positional relationship.

[0066] This helps determine the positions of the background printed layer and one or more foreground printed layers. Background representation determination may also include determining, for each foreground printed layer, a (monochrome, e.g., binary) mask (then also called a foreground mask) and / or its position based on a position reference of the foreground printed layer. Each foreground mask may indicate the position of one or more (e.g., each) graphic patterns of the foreground printed layer, for example, represented as coordinates and / or in a normalized coordinate system.

[0067] Background representation determination may further include: for each printed sample of a banknote (e.g., based on its image data), transforming the image data of the printed banknote (e.g., a banknote image) into a normalized position such that the background printing layer (e.g., is an offset layer) (e.g., its position reference) is aligned with a normalized coordinate system (e.g., defined by the position reference); transforming each (e.g., binary) foreground mask into the normalized coordinate system based on the position reference (e.g., anchor point), which may take into account printing-related positional variations between the background printing layer and one or more foreground printing layers; and masking the result of transforming the image data (also known as the position-normalized banknote image) based on the result of transforming each (e.g., binary) foreground mask (also known as the position-normalized foreground mask).

[0068] For example, a position-normalized banknote image may include, for example, overlaying (e.g., covering) the position-normalized banknote image onto the foreground mask for each foreground mask, and, based on the result, marking pixels of the foreground printing layer revealed by the foreground mask as invalid in the position-normalized banknote image (e.g., by removing the corresponding pixels from the position-normalized banknote image).

[0069] Background representation determination may include: determining the representation of background layout components based on the results of a banknote image with masked location normalization (also known as a banknote image with location normalization and masking), while ignoring the masking components (e.g., pixels) of the banknote image with location normalization and masking for each banknote image with location normalization and masking.

[0070] Background representation determination may include: for each location-normalized banknote and masked banknote image, determining estimates for those pixels in the representation of the background layout components for which statistics cannot be solved because those pixels are masked (e.g., marked as invalid). The resulting estimates may be determined, for example, by classical extrapolation based on neighboring valid estimates and / or by deep learning-based algorithms.

[0071] It is understandable that the aspects detailed above regarding the background printing layer can be similarly applied to one or more foreground printing layers.

[0072] This application illustrates two exemplary results of masked location normalized banknote images (e.g., along with statistical calculations per pixel), one of which (or a superposition of the two results) can be determined as a representation of the background printing layer.

[0073] Comparing the results of properly masking the obscured portion of the banknote with those of a blurred portion reveals: - The modeled background printing layer has no artifacts caused by foreground objects; that is, only those areas where the foreground would be clearly located in a real banknote are not part of the background reference. - It can even completely eliminate some foreground graphic patterns, such as foil / security lines in the intermediate reference.

[0074] Figure 4 A method 700 according to various embodiments is illustrated schematically, wherein one or more printed outputs 114 include a plurality of printed graphic patterns, which exemplarily take the form of a star. Method 700 includes determining a model 450 (also referred to as a reproduction model 450) of the printed outputs 114 in 701 (also referred to as reproduction modeling). Reproduction modeling 701 may be based on the result of imaging 202 of the printed outputs 114, for example, including an imaged printed output 202a. Preferably, the imaged printed output 202a may be sensed by an imaging device 204 of the construction type on which the layout model 250 is based.

[0075] The reproduction model 450 is based on the layout model 250, which implements multiple printed layer representations 250a, 250b (e.g., including a background printed layer representation and one or more foreground printed layer representations), wherein each printed layer representation represents at least one position-variable pattern and / or differs from each other in at least one position-variable pattern as thus represented.

[0076] For each of the multiple printed layer representations 250a, 250b, the reproduction model 450 may be based on a position on the print output 114 at which a graphic pattern represented by the printed layer representation (e.g., its anchor point) is printed. For example, reproduction modeling 701 may include: for each printed graphic pattern on the print output 114, determining the position of the printed graphic pattern as a printing position, for example, expressed in coordinates and / or as a positional relationship with a positional reference on the print output 114 (e.g., an anchor point or another printed graphic pattern). Additionally or alternatively, modeling reproduction 701 may include locating each printed layer representation 250a, 250b based on the print output 114 (e.g., based on image data 202 and / or printing position of the print output 114).

[0077] Method 700 includes, in 703 (also known as classification), classifying the print output 114 based on a reproduction model 450 (e.g., based on a comparison of the print output 202a with the reproduction model 450, e.g., an imaged print output 202a). For example, the reproduction model 450 can be used as a reference for classification, to which the print output 114 can be compared.

[0078] For example, classification 703 can be based on a comparison between the printed output 114 (e.g., its image data 202a) and the reproduction model 450. For example, classification 703 can be determined based on one or more deviations between the printed output 114 (e.g., its image data 202a) and the reproduction model 450, for example, based on their comparison.

[0079] In an exemplary implementation, the reproduction model 450 can represent the reference state of the printed output 114 as a result of a multi-stage printing process, taking into account printing-related positional changes during the multi-stage printing process (e.g., in the case where the printed output is produced by a multi-stage printing process). For example, the reference state may deviate from the actual state of the printed output 114 due to its use. One or more deviations may indicate changes from the reference state to the actual state, such as indicating contamination, damage, visible signs of use, and / or missing portions of the printed output 114.

[0080] As an example, classification 703 can be implemented using a stain and graffiti detection algorithm (SGD), which is configured to detect stains and graffiti based on reproduction model 450, particularly in areas of printed banknotes. In this exemplary implementation, the result of classification 703 can indicate whether the printed output 114 is acceptable, for example, for further use.

[0081] In the case of banknote printing output 114, layout model 250 can select from a plurality of layout models that differ from each other in the type of banknote they represent, for example, in terms of banknote denomination, issuing institution, and monetary system.

[0082] In the following, additional working examples according to various embodiments are described in detail based on a printed banknote as an example of printed output 114. In the context of banknotes, reproduction model 450 may also be referred to as banknote model 450 (BNM), and graphic reference layout may also be referred to as banknote reference layout. The graphic pattern may include a banknote background pattern; a graphic representation of the banknote denomination (e.g., a number) (e.g., as a banknote foreground pattern); one or more graphic security features of the banknote (e.g., as a banknote foreground pattern); the banknote identification number (e.g., as a banknote foreground pattern); and / or image elements (e.g., as a banknote foreground pattern), such as portraits, landscapes, etc.

[0083] This application illustrates multiple banknote models 450 according to various embodiments, which are determined based on layout model 250 and differ from each other in the base printed samples of the banknotes, such as median models representing three different banknote images. For example, banknote models 450 used for stain and graffiti detection algorithms can be generated for each individual printed sample of the banknote, for example, because the positional relationships of one or more printing layers and / or one or more security features can vary from banknote to banknote.

[0084] In an exemplary working example, for each printed sample of a banknote, determining the banknote model 450 (e.g., for the minimum, average, median, or maximum expected banknote grayscale image 202a) may include: selecting a printed layer representation 250a (e.g., in which one or more foreground patterns are marked) as a background reference; overlaying one or more printed layer representations 250b, 250c (e.g., in which one or more background patterns are marked) as corresponding foreground references one after another on the background reference to continuously construct the banknote model 450 of the printed sample; wherein the overlay may optionally include, for example, the positioning of each foreground reference relative to the background reference based on the printed sample (e.g., positional variations determined therefrom); optionally determining (e.g., marking) one or more disturbances of the banknote model 450 as (e.g., completely) masked.

[0085] For example, the position of each individual foreground reference in banknote model 450 can be determined, for instance, by adjusting for the estimated position of their corresponding printed layer relative to the background printed layer (or relative to another position reference). Examples of interference can have positions that are invariant with respect to the printing process (e.g., such as printed output edges) or positions that are functions of the printing process (e.g., such as security threads). Printed output edges can be the physical edges of the printed output (e.g., part of a circumference).

[0086] For some or most currencies, the details of the banknote's graphic reference layout and / or multi-stage printing process are not publicly known. For example, the layout of the printing plate used for offset printing may be unknown, which complicates the determination of the offset printing layer used as a background reference. In such cases, aspects as detailed herein allow for the determination of a more accurate layout model 250 (e.g., and its printed layer representations 250a, 250b), and thus allow for the determination of a more accurate BNM 450. To enable the background and foreground references to be overlaid in varying positions, the background reference for the entire banknote is determined. In such cases, individual layout components can be determined (e.g., through reconstruction, learning, and / or estimation) based on available printed samples of the banknote.

[0087] To facilitate understanding, this application illustrates comparative examples of overlays according to various embodiments, depicting median overlay and minimum overlay of a CNY100d background (offset printing, 100dpi, G channel) against a GNY100d background (offset printing, 100dpi, G channel), where print-related positional variations are set to zero. Multiple banknote printing outputs 114 may include examples of public denominations and public currency systems, each sensed to provide multiple banknote images as imaging printing outputs 202a, 202b, 202c.

[0088] In the comparative example, the overlay includes: transforming each banknote image 202a into a position-normalized banknote image having a common offset layer position (e.g., by aligning the offset layers with each other); for each pixel of the position-normalized banknote image, solving for an expected value from the position-normalized banknote image to compute a minimum, average, median, or maximum reference to the expected value. This uniform position normalization can result in acceptable results for median references and / or less complex banknote reference layouts, such as the reverse side of an EUR banknote, which consists of a single printed layer serving as the banknote background. In the case of more complex banknote reference layouts, such as the front or back of an EUR or CNY banknote, uniform position normalization can produce a banknote model 450 with blurred areas due to printing-related positional variations. For example, the position of each foreground graphic pattern printed on the printed output relative to the background can be a function of the printed output. For example, printing-related positional variations can cause alterations in the occlusion of background pixel fragments, which vary per sample of the banknote.

[0089] Note that the higher the printing-related positional variation, the more information can be determined about the graphic (e.g., background) pattern based on multiple print outputs 114. Illustratively, a portion of a graphic (e.g., background) pattern overprinted on one print output (e.g., a printed sample of a banknote) may be visible on another print output (e.g., a printed sample of a banknote). This information can be included in the layout model 250, for example, at least in the print layer representation, where the graphic (e.g., background) pattern is unmasked. As an exemplary result, the total masking area of ​​the layout model 250 may be less than 100%, and / or at least one graphic pattern may be partially (incompletely) masked by the print layer representation, where the graphic background pattern is unmasked.

[0090] Figure 5 A schematic diagram illustrates a bank terminal 1000 and its components according to various embodiments, which are communicatively coupled to each other (e.g., via a network) as indicated by arrows (e.g., via the data interface of the respective components).

[0091] The bank terminal 1000 may include a human-machine interface 1004, such as a cash transfer device 1004a (e.g., configured to receive cash from and / or distribute cash to a person), a keypad 1004b (e.g., for receiving instructions from a person), and a display 1004c (e.g., for presenting information to a person). The bank terminal 1000 may also include a cash storage device 1008 (e.g., including a safe and / or one or more cash storage boxes) for storing cash (e.g., banknotes) received and / or distributed by the bank terminal 1000, for example, via the human-machine interface 1004. The bank terminal 1000 may also include a transport device 1012 for transporting cash between the cash storage device 1008 and the human-machine interface 1004 (e.g., to and / or from the cash storage device 1008).

[0092] The bank terminal 1000 may also include an imaging device 1006 configured to sense cash (e.g., one or more banknotes) received by the bank terminal 1000, for example, via a human-machine interface 1004. For example, the imaging device 1006 may be configured to sense cash transported by a transport device 1012.

[0093] The bank terminal 1000 also includes a control device 1002 and one or more of its components, which are not necessarily physically housed within the housing 1010 of the bank terminal 1000. For example, the control device 1002 or one or more of its components may be provided, at least partially via cloud computing, as a shared computing system for multiple bank terminals 1000, or as a remote control device. The control device 1002 may be configured to control one or more of the following components (e.g., via the data interface of the respective component): imaging device 1006, cash storage device 1008, human-machine interface 1004, and / or transport device 1012.

[0094] The control device 1002 may include at least one or more processors 1002a and / or at least one memory 1002b, for example storing code segments (e.g., provided as bit strings) including instructions comprising one or more methods (e.g., method 300 and / or method 700) as detailed herein. The instructions may be configured to, when executed by at least one processor 1002a, direct at least one processor 1002a to perform one or more methods detailed herein, such as method 300 and / or method 700. Additionally or alternatively, one or more of the following may be stored by at least one memory 1002b: sharpness criteria, criteria for print-related positional variations, one or more layout models 250, image data, and one or more predefined categories.

[0095] For example, control device 1002 may be configured to receive image data of banknotes from imaging device 1006 (e.g., via cash transfer device 1004a) and determine reproduction model 450 based thereon. Furthermore, control device 1002 may be configured to sort banknotes and, based on the sorting results, instruct the rejection of banknotes and / or the storage of banknotes in cash storage device 1008 (e.g., in its respective cash box).

[0096] Additionally or alternatively, control device 1002 may be configured to receive image data of one or more banknotes from imaging device 1006 (e.g., via cash transfer device 1004a) and determine (e.g., update) layout model 250 based on the image data. This reduces the need for maintenance. For example, control device 1002 may be configured to determine changes in one or more optical properties of imaging device 1006 based on the image data and update layout model 250 based on the changes. This allows for at least partial compensation for the changes and thus increases the lifespan of bank terminal 1000. Additionally or alternatively, control device 1002 may be configured to determine a new layout model 250 from scratches based on image data. This helps adapt bank terminal 1000 to changes in banknote design and / or monetary system (e.g., during normal operation) or allows operators to build new bank terminals 1000 without manufacturer assistance.

[0097] As an example, updating layout model 250 may be beneficial in situations where layout model 250 was initially determined based on a set of banknotes (e.g., 100) and subsequently determined to be insufficient, for instance, when modeling 700 or more additional banknotes (e.g., 10,000) reveals that the number of banknotes in that set is less to obtain an appropriate layout model 250. In such cases, layout model 250 can be updated based on image data of the additional banknotes, as determined during the modeling of the additional banknotes 700.

[0098] As another example, updating the layout model 250 may be beneficial in cases where the printing process has been altered, for example, by changing the colorant. In such cases, during the modeling of the results of the altered printing process 700, the layout model 250 can be updated based on image data of additional results from the altered printing process.

[0099] This application illustrates multiple banknote models 450 according to various embodiments, wherein one or more additional (e.g., individual banknotes) graphic patterns (e.g., security features) are masked. Examples of individual banknote patterns may include: identification numbers, serial numbers, and / or security threats. Masking portions of the banknote models allows for maximizing the reliability of classification, for example, when variations in individual banknotes are minimized. As an example, the location of the masked individual banknote may optionally be a function of the printing-related positional variations of the security threat location, as is visible to the security threat. This allows for further maximization of the reliability of classification.

[0100] It is understood that such a mask (e.g., its geometry and / or location) can be determined in a manner similar to that described above, for example, based on printing-related positional variations. Additionally or alternatively, the representation of one or more printed layers can be manually modified, for example when the available reproducible variations are too large, such as in the case of optically variable inks (OVI), where masking the area during detection may be better.

[0101] Various examples are provided below with reference to the above aspects.

[0102] Example 1 is a method comprising: determining two or more graphic patterns of a graphic reference layout based on multiple printed reproductions (e.g., comparisons) of the graphic reference layout, the two or more graphic patterns satisfying a criterion of print-related positional variations relative to each other; and determining (e.g., generating or updating) a model of the graphic reference layout based on the result of determining the two or more graphic patterns, wherein, for each of the two or more graphic patterns, the model implements a (e.g., a print result) (e.g., individual and / or partial) representation of the graphic reference layout, according to which the graphic pattern (e.g., completely or only partially) is masked (e.g., by preset color values) or at least marked (e.g., by preset color values).

[0103] Example 2 is the method of Example 1, wherein the result of determining two or more graphic patterns includes positional variations of the two or more graphic patterns as reproduced by printing reproduction.

[0104] Example 3 is a method as described in Example 1 or 2, wherein determining the two or more graphic patterns includes determining that the two or more graphic patterns meet the criteria and / or determining the positional variations of the two or more graphic patterns as reproduced by the printing reproduction.

[0105] Example 4 is a method of one of Examples 1 to 3, in which two or more graphic patterns are determined based on criteria.

[0106] Example 5 is a method of one of Examples 1 to 4, wherein the criteria for print-related positional variations are stored (e.g., on a storage medium) and / or remain unchanged for each reproduction.

[0107] Example 6 is a method of one of Examples 1 to 5, wherein determining (e.g., a model of a graphic reference layout and / or two graphic patterns) and / or comparing based on multiple print reproductions (e.g., their image data) (e.g., image-based overlay) and / or based on image data of multiple print reproductions (e.g., each) (e.g., overlay), wherein preferably, the image data is caused by optically sensing multiple print reproductions.

[0108] Example 7 is a method of one of Examples 1 to 6, wherein the representation is based on (e.g., each) multiple print reproductions (e.g., their image data) and / or on (e.g., each) image data of multiple print reproductions (e.g., superimposed), wherein preferably, the image data is caused by optically sensing multiple print reproductions.

[0109] Example 8 is a method of one of Examples 1 to 7, wherein the comparison is based on first image data of a first print reproduction of the plurality of print reproductions and second image data of a second print reproduction of the plurality of print reproductions; wherein, preferably, the comparison is based on at least one superposition (e.g., multiple superpositions) of the first image data and the second image data, wherein each superposition is a function of positional variation.

[0110] Example 9 is the method of Example 8, wherein at least one superposition of the first image data and the second image data includes a first superposition and a second superposition, the first superposition and the second superposition being different from each other in the positional relationship of the first image data and / or the second image data to which they are superimposed.

[0111] Example 10 is a method of one of Examples 1 to 9, wherein the comparison is based on the (e.g., optical) superposition of multiple print reproductions (e.g., their image data); and / or wherein a criterion of print-related positional variation is met when the superposition of multiple print reproductions of their image data (e.g., first image data and second image data) alters the sharpness of the print graphic pattern (e.g., relative to the image data).

[0112] Example 11 is a method of one of Examples 1 to 10, wherein a criterion for positional variation is met when the positional variation in multiple printed reproductions exceeds a threshold.

[0113] Example 12 is a method of one of Examples 1 to 11, wherein the result of determining two or more graphic patterns includes positional variations of the two or more graphic patterns, such as in multiple reproductions and / or relative to a positional reference (e.g., a graphic reference layout), wherein the positional reference can be determined by anchor point detection.

[0114] Example 13 is a method (e.g., the method of any one of Examples 1 to 12) comprising: determining a model of a printed reproduction of a graphic reference layout, wherein the graphic reference layout comprises two or more graphic patterns, wherein the model of the printed reproduction is based on a model of the graphic reference layout, the model of the graphic reference layout comprising, for each of the two or more graphic patterns, a (e.g., a printed result) (e.g., individual and / or partial) representation of the graphic reference layout, wherein, according to the representation (e.g., by means of the representation), the graphic reference layout (e.g., completely or only partially) is occluded (e.g., by a preset color value) or at least marked (e.g., by a preset color value); and classifying the printed reproduction based on the model of the printed reproduction and preferably based on the printed reproduction (e.g., its image data), for example, based on a comparison between the model of the printed reproduction and the printed reproduction.

[0115] Example 14 is the method as described in Example 13, wherein the model of the print reproduction is determined based on the image data of the print reproduction, preferably optically and / or by an imaging device of the construction type on which the model of the graphic reference layout is based, sensing the result of the print reproduction.

[0116] Example 15 is a method of one of Examples 13 or 14, further comprising: determining a change in one or more optical properties of the imaging device; and updating the layout model based on the change, preferably at least partially compensating for the change.

[0117] Example 16 is a method of one of Examples 13 to 15, and further includes, for example, selecting a model of the print reproduction from a plurality of predetermined models of the print reproduction based on the print reproduction (or based on its image data).

[0118] Example 17 is a method of one of Examples 13 to 16, wherein the model of the print reproduction is based on a location where each of at least two or more graphic patterns is reproduced on the print reproduction of the graphic reference layout.

[0119] Example 18 is a method of one of Examples 13 to 17, wherein the model of the graphic reference layout includes multiples of representation (e.g., stacking), wherein determining the model of print reproduction includes determining the positional relationship of the multiples of representation (e.g., stacking) relative to each other based on the position of each of at least two or more graphic patterns reproduced on the print reproduction of the graphic reference layout (e.g., where they are combined into the model of the graphic reference layout, for example, by overlay).

[0120] Example 19 is a method from Examples 13 through 18, wherein the classification of print reproductions can be determined based on one or more deviations between the print reproduction and a model of the print reproduction, for example, based on their comparison.

[0121] Example 20 is a method of one of Examples 13 to 19, wherein classifying the reproduction of a print includes assigning the reproduction of a print to a conditional state representing a category (e.g., suitable for further use).

[0122] Example 21 is a method from Examples 13 through 20, in which stain and graffiti detection algorithms are used to classify print reproductions.

[0123] Example 22 is a method of one of Examples 13 to 21, wherein classifying print reproductions includes assigning print reproductions to one of a set of categories that are distinct from each other in the conditional state represented therefrom.

[0124] Example 23 is a method of one of Examples 1 to 22, wherein two or more graphic patterns include at least a graphic background pattern and / or at least one graphic foreground pattern.

[0125] Example 24 is a method of one of Examples 1 to 23, wherein the model of the graphic reference layout includes multiple (e.g., stacks) representations, which are preferably separated from each other (e.g., independent and / or far apart) with respect to their positions, and / or preferably different from each other in the graphic pattern in which they are masked.

[0126] Example 25 is a method of Example 24, wherein the plurality of representations of the graphic reference layout includes a first representation and a second representation, wherein at least one of two or more graphic patterns (e.g., a foreground graphic pattern) is masked, wherein at least one of two or more graphic patterns (e.g., a background graphic pattern) is masked, wherein preferably, at least one first graphic pattern and at least one second graphic pattern are different from each other (e.g., at least in the graphic patterns and / or positions where they are masked).

[0127] Example 26 is a method of one of Examples 1 to 25, wherein the model of the graphic reference layout represents a multi-stage printing process, preferably used to reproduce the graphic reference layout, for example for printing the graphic reference layout or each reproduction.

[0128] Example 27 is a method of one of Examples 1 to 26, wherein a graphic pattern is masked by a portion of a representation of the graphic pattern's location and / or set to a preset color value (e.g., its pixels).

[0129] Example 28 is a method of one of Examples 1 through 27, wherein the model of the graphical reference allows (e.g., implementation) changes in the position of the representation.

[0130] Example 29 is a method of one of Examples 1 to 28, wherein the or each reproduction is provided as (e.g., physically) a print and / or optically sensed.

[0131] Example 30 is a method of one of Examples 1 to 29, wherein each of the graphic reference layouts in which the graphic pattern is masked indicates the location where the graphic pattern is masked, wherein the location is preferably based on print-related positional variations and / or on the reproduction of each print.

[0132] Example 31 is a method of one of Examples 1 to 30, wherein the or each reproduction is the result of multiple printing processes, preferably wherein at least one first printing process includes offset printing, and wherein at least one second printing process includes gravure printing.

[0133] Example 32 is a method of one of Examples 1 to 31, wherein determining the model of print reproduction includes: for each of two or more graphic patterns, determining a positional transformation of a representation of a graphic reference layout based on position, according to the positional transformation (e.g., by means of the positional transformation) at least a portion of the graphic pattern is masked (e.g., removed) or at least marked.

[0134] Example 33 is a method of one of Examples 1 to 32, wherein each of the multiple print reproductions of the graphic reference layout is the result of a multi-stage printing process, preferably represented by a model of the graphic reference layout.

[0135] Example 34 is a method of one of Examples 1 to 33, wherein the model of the graphic reference layout includes multiples of representations, wherein each representation is based on the superposition of multiple print reproductions, preferably based on the superposition of image data of multiple print reproductions (e.g., each print reproduction).

[0136] Example 35 is a method of one of Examples 1 to 34, wherein an image processing mask is used to mask a graphic pattern, wherein the mask is based on multiple print reproductions, preferably on image data of multiple print reproductions (e.g., each print reproduction), and more preferably on the superposition of image data.

[0137] Example 36 is a method of one of Examples 1 to 35, wherein the two or more graphic patterns include one or more of the following graphic patterns: graphic background; graphic representation of denomination; graphic security feature; identification number; picture element.

[0138] Example 37 is a method of one of Examples 1 through 36, wherein the graphic reference layout is a banknote reference layout.

[0139] Example 38 is a method of one of Examples 1 to 37, wherein the representation is a superposition of first image data of a first print reproduction based on the plurality of print reproductions and second image data of a second print reproduction based on the plurality of print reproductions.

[0140] Example 39 is a method of one of Examples 1 to 38, wherein, for each of two or more graphic patterns, a model of the graphic reference layout implements a graphic representation of the graphic pattern (preferably determined based on multiple printed reproduction image data) that is unmasked.

[0141] Example 40 is one or more non-transitory computer-readable media on which instructions are stored, which, when executed by at least one processor, instruct the at least one processor to perform the methods of Examples 1 to 39.

[0142] Example 41 is a control device (e.g., implemented by one or more processors) configured to perform the methods of Examples 1 to 39, preferably including one or more non-transitory computer-readable media of Example 40.

[0143] Example 42 is a bank terminal that includes the control device of Example 41.

[0144] Example 43 is a bank terminal of Example 42, and further includes: an imaging device, wherein the model (e.g., of a print reproduction and / or a graphic reference layout) is based on image data of one or more print reproductions of the graphic reference layout sensed by the imaging device.

[0145] Example 44 is a model for determining the printed reproduction of a graphic reference layout using a model that includes two or more graphic patterns (e.g., a banknote reference layout), wherein for each of the two or more graphic patterns, the model implements a representation of the graphic reference layout according to which at least a portion of the graphic pattern is masked (e.g., removed) or at least marked.

[0146] Example 45 is to determine a model of a graphic reference layout using multiple printed reproductions (e.g., image data thereof) of a graphic reference layout comprising two or more graphic patterns, wherein, for each of the two or more graphic patterns, the model of the graphic reference layout implements a representation of the graphic reference layout according to which at least a portion of the graphic pattern is masked (e.g., removed) or at least marked.

[0147] While this disclosure has been specifically shown and described with reference to particular aspects, those skilled in the art will understand that various changes in form and detail may be made without departing from the spirit and scope of this disclosure as defined by the appended claims. Therefore, the scope of this disclosure is indicated by the appended claims, and all changes falling within the meaning and scope of the claims equivalents are thus intended to be included.

Claims

1. A method comprising: • Based on multiple printed reproductions (114) of the graphic reference layout (102), determine (201) two or more graphic patterns of the graphic reference layout (102), the two or more graphic patterns satisfying the criteria for positional changes related to printing with respect to each other; The graphic pattern mentioned above is a position-variable pattern; The graphic reference layout (102) mentioned therein is a printed layer representation; • Based on the result of determining (201) the two or more graphic patterns, determine (205) a layout model (250) of the graphic reference layout (102), wherein the layout model (250) implements modeling layout components (250a, 250b, 250c) of the graphic reference layout (102) for each of the two or more graphic patterns, according to the modeling layout components (250a, 250b, 250c), at least a portion of the graphic pattern is masked or at least marked, wherein each modeling layout component (250a, 250b, 250c) is associated with an individual printing process stage and therefore lacks at least one graphic pattern; • Determine a reproduction model (450) of the printed reproduction (114) of the graphic reference layout (102), wherein the graphic reference layout (102) includes the two or more graphic patterns, wherein the reproduction model (450) of the printed reproduction (114) is based on the layout model (250) of the graphic reference layout (102), the layout model (250) including the modeling layout components (250a, 250b, 250c) for each of the two or more graphic patterns. • The printing reproduction (114) is classified (280) based on the printing reproduction (114) and the reproduction model (450) based on the printing reproduction (114).

2. The method according to claim 1, wherein, The determination (201) of the two or more graphic patterns is based on the comparison (203) of the plurality of print reproductions (114) with each other and / or on the image data of the plurality of print reproductions (114).

3. The method according to claim 1 or 2, wherein, The graphic pattern is masked using an image processing mask (802, 804), wherein the mask (802, 804) is based on the image data of each of the plurality of print reproductions (114).

4. The method according to any one of claims 1 to 3, wherein, The result of determining (201) the two or more graphic patterns includes the positional changes of the two or more graphic patterns, preferably relative to the positional reference of the graphic reference layout.

5. The method according to any one of claims 1 to 4, wherein, The layout model (250) of the graphic reference layout (102) includes multiples of the modeling layout components (250a, 250b, 250c), each representation of which is based on the superposition of the plurality of print reproductions, preferably based on the superposition of image data of each print reproduction in the plurality of print reproductions (114).

6. The method according to claim 1, wherein, The reproduction model (450) for determining the print reproduction (114) is based on the image data of the print reproduction (114).

7. The method according to any one of claims 1 to 6, wherein, The reproduction model (450) of the print reproduction (114) is based on a position, at which each of the at least two or more graphic patterns is reproduced by the print reproduction (114) of the graphic reference layout (102).

8. The method according to claim 7, wherein, The determination of the reproduction model (450) of the print reproduction (114) includes: for each of two or more graphic patterns, determining a positional transformation of the representation of the graphic reference layout (102) based on the position of the graphic pattern reproduced by the print reproduction (114) of the graphic reference layout (102), according to the positional transformation, at least a portion of the graphic pattern is masked or at least marked.

9. The method according to any one of claims 1 to 8, wherein, The graphic reference layout (102) is a banknote reference layout.

10. The method according to any one of claims 1 to 9, wherein, The two or more graphic patterns include one or more of the following graphic patterns: Graphic background; Graphical representation of denominations; Graphical security features; Identification number; and / or Image elements.

11. One or more non-transitory computer-readable media having instructions stored thereon, the instructions, when executed by at least one processor, instructing the at least one processor to perform the method according to claims 1 to 10.

12. A control device (1002) configured to perform the method according to claims 1 to 10, preferably comprising one or more non-transitory computer-readable media according to claim 11.

13. A bank terminal (1000) comprising the control device according to claim 12.

Images