Anti-fake printed matter preparation and identification method and system based on halftone frequency compensation

Abstract

This invention discloses a method and system for preparing and identifying anti-counterfeiting printed materials based on halftone frequency compensation. The method includes determining the compensation frequency of the embedded texture to be embedded in the original digital image based on the frequency of the visible moiré patterns induced when the anti-counterfeiting printed material is copied; designing an embedded texture based on halftone frequency compensation according to the compensation frequency to generate a digital image containing the embedded texture; and outputting the digital image containing the embedded texture onto a printing substrate through a printing device to obtain the anti-counterfeiting printed material. This invention aims to utilize the active induction mechanism of moiré patterns to actively induce moiré features in the frequency domain and micro-texture at the moment of imaging of the copy, transforming the identification of copies from searching for weak signals to detecting significant interference, thereby improving the stability of anti-counterfeiting printed material identification in complex media and cross-brand device application scenarios.

Description

Technical Field

[0001] This invention relates to the field of anti-counterfeiting technology for printed materials, specifically to a method and system for preparing and identifying anti-counterfeiting printed materials based on halftone frequency compensation. Background Technology

[0002] In anti-counterfeiting technology for printed materials, embedded texture is a high-security anti-counterfeiting method based on the physical properties of materials. Its core idea is to embed unique, unreplicable microstructure textures, whether natural or artificially manufactured, into the printed material, and then verify authenticity through image acquisition and comparison. However, current technologies use a passive embedding method. While this allows for the creation of visible moiré patterns when the anti-counterfeiting printed material is copied, the moiré signal is weak and the contrast is not obvious when the anti-counterfeiting printed material is used in complex media and cross-brand device applications. This results in insufficient stability of the anti-counterfeiting printed material authentication technology in complex media and cross-brand device applications. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for the preparation and identification of anti-counterfeiting printed materials based on halftone frequency compensation, which addresses the above-mentioned problems of the prior art. The present invention aims to actively induce moiré features in the frequency domain and micro-texture at the moment of imaging of the copy by utilizing the active induction mechanism of moiré patterns, thereby changing the identification of the copy from searching for weak signals to detecting significant interference, and improving the stability of anti-counterfeiting printed material identification in application scenarios with complex media and cross-brand equipment.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0005] A method for preparing anti-counterfeiting printed materials based on halftone frequency compensation includes the following steps: determining the frequency of the visible moiré pattern induced when the anti-counterfeiting printed material is reproduced according to the input requirements. Determine the compensation frequency of the embedded texture to be embedded in the original digitized image. According to the compensation frequency The design incorporates embedded textures based on halftone frequency compensation to generate digital images that conceal these embedded textures. These digital images are then printed onto a substrate using a printing device to produce anti-counterfeiting printed materials.

[0006] Optionally, the compensation frequency for determining the embedded texture to be embedded in the original digitized image is... The function expression is:

[0007] ;

[0008] in, The frequency of the visible moiré pattern induced when counterfeit-proof printed materials are copied. The sampling spatial frequency of the replication device, It represents the harmonic order.

[0009] Optionally, the compensation frequency The design of a halftone frequency-compensated embedded texture to generate a digital image with hidden embedded textures includes: preprocessing the original digital image, including histogram equalization, image denoising, and background region extraction; using the background region as the embedding region of the embedded texture; and adjusting the pattern duty cycle of the halftone frequency-compensated embedded texture to improve the contrast of the generated digital image with hidden embedded textures. By satisfying the given constraints, the final embedded texture is obtained; the final embedded texture and the denoised digital image are superimposed to obtain a digital image with the embedded texture hidden.

[0010] Optionally, the functional expression of the given constraint is:

[0011] ;

[0012] ;

[0013] in, For perception correction coefficients, It is a contrast sensitivity function. To hide the spatial frequencies of a digitized image with embedded texture, the spatial frequencies of a digitized image with embedded texture. Compensation frequency of embedded texture It is obtained by superimposing the spatial frequencies of the original digitized image. These are empirical constant coefficients. It is a natural constant.

[0014] Optionally, the embedded texture is composed of subpixel-level dots or stripes, where subpixel-level means that the physical size of the dots is smaller than the smallest physical dot of the printing equipment or the size corresponding to a single photosensitive unit of the scanning equipment.

[0015] This invention also provides a method for identifying anti-counterfeiting printed materials based on halftone frequency compensation, comprising the following steps: extracting texture noise from the non-image and text areas of the printed material image to obtain a denoised printed material image; correcting dot distortion in the denoised printed material image; extracting multi-dimensional features embedded with texture from the dot distortion corrected printed material image; using a pre-trained classifier to predict whether the printed material to be identified is an anti-counterfeiting printed material prepared by the method based on halftone frequency compensation, wherein the classifier is a machine learning model and has been trained to establish a mapping relationship between the multi-dimensional features embedded with texture and the judgment result of whether the printed material to be identified is an anti-counterfeiting printed material.

[0016] Optionally, the step of extracting texture noise from the non-textual area and denoising the printed image to obtain a denoised printed image includes: extracting high-frequency texture components from the non-textual area of ​​the printed image of the printed material to be identified using a high-pass filter, calculating the mean and variance of the high-frequency texture components, constructing a Gaussian white noise model using the mean and variance of the high-frequency texture components as texture noise, and subtracting the texture noise from the printed image of the printed material to be identified to obtain a denoised printed image.

[0017] Optionally, the dot distortion correction of the denoised printed image includes: selecting a local image patch of the region of interest in the denoised printed image; extracting the light intensity response curve of the local image patch using the step response function (ESF); numerically differentiating the light intensity response curve of the local image patch using the line spread function (LSF) to obtain the lateral light intensity diffusion distribution of the local image patch; generating the spatial light intensity response of the local image patch by two-dimensional deconvolution of the lateral light intensity diffusion distribution; and applying a deconvolution operation to the spatial light intensity response of the local image patch to obtain the dot distortion corrected printed image.

[0018] Optionally, when extracting multi-dimensional features embedded with texture from the printed image after dot distortion correction, the multi-dimensional features include some or all of the following: high-frequency energy, axial imbalance, high-frequency energy concentration, total wavelet energy, wavelet detail energy ratio, noise granularity, texture contrast, texture uniformity, texture entropy, energy peak E(u,v) of preset frequency domain coordinates (u,v), local binary pattern code (LBP), gray-level co-occurrence matrix (GLCM), attribute energy, and image contrast.

[0019] The present invention also provides an anti-counterfeiting printed matter processing system, including a microprocessor and a memory interconnected thereto, the microprocessor being programmed or configured to execute the anti-counterfeiting printed matter preparation method based on halftone frequency compensation or the anti-counterfeiting printed matter identification method based on halftone frequency compensation.

[0020] Compared with existing technologies, the present invention mainly achieves the following beneficial effects: The present invention includes the frequency of visible moiré patterns induced by the input requirements when the anti-counterfeiting printed material is copied. Determine the compensation frequency of the embedded texture to be embedded in the original digitized image. According to the compensation frequency This invention provides a method for preparing anti-counterfeiting printed materials based on embedded textures using halftone frequency compensation. This method transforms the process from passively searching for residual features to actively embedding induced features. During the original printing plate-making stage, the technology pre-generates hidden embedded textures within the halftone dot structure. When the copier scans, a pre-defined physical interference is generated. This induced mechanism allows the copier to excite structured moiré patterns in the frequency domain and microstructure at the moment of imaging, thus changing the copier's discrimination from "finding weak signals" to "detecting significant interference." This solution not only effectively counteracts the smoothing algorithm built into the copier, ensuring high contrast of the discrimination features under various repair and compensation conditions, but also has low resolution requirements for the acquisition device. It can effectively identify and maintain system stability even in complex media and cross-brand applications. Attached Figure Description

[0021] Figure 1 This is a schematic flowchart of the method for preparing anti-counterfeiting printed materials in an embodiment of the present invention.

[0022] Figure 2 This is a flowchart illustrating the method for identifying anti-counterfeit printed materials in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram of the test results of the SVM classifier in an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram illustrating the contribution analysis results of multi-dimensional indicators in an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described in detail below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] like Figure 1 As shown, the method for preparing anti-counterfeiting printed materials based on halftone frequency compensation in this embodiment includes the following steps: according to the input requirements, the frequency of the visible moiré pattern induced when the anti-counterfeiting printed material is copied. Determine the compensation frequency of the embedded texture to be embedded in the original digitized image. According to the compensation frequency The design incorporates embedded textures based on halftone frequency compensation to generate a digital image containing these embedded textures. This digital image is then printed onto a substrate using a printing device to obtain anti-counterfeiting printed materials. This embodiment of the method actively pre-places sub-pixel-level induced compensation embedded textures within the halftone dot structure during the original printing plate-making stage, constructing microscopic frequency traps. Utilizing the frequency aliasing effect generated during the scanning and resampling process of the photocopying equipment, the copied copy is forced to generate macroscopic moiré patterns with high contrast. During detection, the response intensity of the induced stripes can be captured by performing frequency domain energy distribution analysis or spatial texture topology recognition on the acquired image. This embodiment of the method overcomes the limitations of traditional passive recognition techniques, which are susceptible to interference from photocopying machine smoothing algorithms, achieving a technological leap from "passively searching for residual features" to "actively embedding induced features."

[0027] In order to achieve the active induction of visually perceptible frequencies using the beat difference principle, In this embodiment, the compensation frequency for the embedded texture to be embedded in the original digitized image is determined. The function expression is:

[0028] ;

[0029] in, The frequency of the visible moiré pattern induced when counterfeit-proof printed materials are copied. The sampling spatial frequency of the copying device (copying / scanning device). As a harmonic order, this carefully designed embedded texture is completely hidden under normal printing observation conditions. The embedded texture is in the low-frequency interference range that the copying equipment cannot perfectly reproduce. Only when it is sampled a second time by the copying equipment will it produce significant physical interference fringes (moiré patterns) due to the frequency interference effect, thus exposing the characteristic traces of the copying operation.

[0030] In this embodiment, based on the compensation frequency The design of a halftone frequency-compensated embedded texture to generate a digital image with hidden embedded textures includes: preprocessing the original digital image (e.g., a trademark image), including histogram equalization, image denoising, and background region extraction; using the background region as the embedding region of the embedded texture; and adjusting the pattern duty cycle of the halftone frequency-compensated embedded texture (which controls the droplet distribution density) to improve the contrast of the generated digital image with hidden embedded textures. By satisfying the given constraints, the final embedded texture is obtained; the final embedded texture and the denoised digital image are superimposed to obtain a digital image with the embedded texture hidden.

[0031] In this embodiment, the functional expression for the given constraint is:

[0032] ;

[0033] ;

[0034] in, For perception correction coefficients, It is a contrast sensitivity function. To hide the spatial frequencies of a digitized image with embedded texture, the spatial frequencies of a digitized image with embedded texture. Compensation frequency of embedded texture It is obtained by superimposing the spatial frequencies of the original digitized image. These are empirical constant coefficients. is a natural constant. By using the contrast sensitivity function CSF to optimize the contrast of the embedded texture, the embedded texture can be placed in the blind spot of the human eye, thereby achieving the purpose of hiding the embedded texture in the digital image.

[0035] In this embodiment, the embedded texture is composed of subpixel-level dots or stripes. Subpixel-level means that the physical size of the dots is smaller than the smallest physical dot of the printing equipment or the size corresponding to a single photosensitive unit of the scanning equipment. Alternatively, other types of embedded textures can be used as needed.

[0036] Finally, the digital image with embedded textures is output onto the printing substrate using printing equipment to obtain anti-counterfeiting printed materials. Through calibrated printing equipment, the digital image with embedded textures is converted into a high-quality physical print. During the printing process, by strictly controlling dot gain and ink diffusion, it is ensured that the pre-embedded textures are presented on the printing substrate with sub-pixel precision, achieving a lossless conversion from digital domain trap design to physical domain feature embedding, ultimately forming anti-counterfeiting printed materials with self-authentication capabilities.

[0037] like Figure 2As shown, the anti-counterfeiting printed material identification method based on halftone frequency compensation in this embodiment includes the following steps: for the printed material image to be identified, extract the texture noise in the non-image and text areas and denoise the printed material image to obtain a denoised printed material image; perform dot distortion correction on the denoised printed material image; extract multi-dimensional features embedded with texture from the dot distortion corrected printed material image, and use a pre-trained classifier to predict whether the printed material to be identified is an anti-counterfeiting printed material obtained by the anti-counterfeiting printed material preparation method based on halftone frequency compensation. The classifier is a machine learning model and has been trained to establish a mapping relationship between the multi-dimensional features embedded with texture and the judgment result of whether the printed material to be identified is an anti-counterfeiting printed material.

[0038] In this embodiment, extracting texture noise from areas without images and text and denoising the printed image to obtain a denoised printed image includes: extracting high-frequency texture components from areas without images and text in the printed image of the printed product to be identified using a high-pass filter; calculating the mean and variance of the high-frequency texture components; using the mean and variance of the high-frequency texture components to construct a Gaussian white noise model as texture noise; and subtracting the texture noise from the printed image of the printed product to be identified to obtain a denoised printed image, thereby improving the accuracy of anti-counterfeiting printed product identification on old or low-quality paper.

[0039] In this embodiment, the dot distortion correction of the denoised printed image includes: selecting a local image block of the region of interest in the denoised printed image; extracting the light intensity response curve of the local image block using the step response function (ESF); numerically differentiating the light intensity response curve of the local image block using the line spread function (LSF) to obtain the lateral light intensity diffusion distribution of the local image block; generating the spatial light intensity response of the local image block by two-dimensional deconvolution of the lateral light intensity diffusion distribution; and applying a deconvolution operation to the spatial light intensity response of the local image block to obtain the printed image after dot distortion correction.

[0040] In this embodiment, when extracting multi-dimensional features embedded with texture from the printed image after dot distortion correction, a high-resolution image acquisition device is used to acquire local and overall microscopic morphological images of the trademark to be tested. Image enhancement and preprocessing techniques are then applied to improve the discernibility of details. This process focuses on capturing various frequency and spatial features, such as macroscopic moiré effects caused by multiple copying and resampling operations, grayscale step changes in halftone dot edges, and morphological distortions and distribution anomalies in dot topology during the copying process. By constructing multi-dimensional features and inputting them into a deep learning model pre-trained with a large number of samples, the system can effectively suppress interference factors such as the original paper texture and scanning noise, thereby accurately identifying essential feature differences caused by copying behavior and achieving highly accurate copy traceability and authenticity determination. As an optional implementation, the multi-dimensional features in this embodiment include some or all of the following: high-frequency energy, axial imbalance, high-frequency energy concentration, total wavelet energy, wavelet detail energy ratio, noise granularity, texture contrast, texture uniformity, texture entropy, energy peak E(u,v) of preset frequency domain coordinates (u,v), local binary pattern code (LBP), gray-level co-occurrence matrix (GLCM), attribute energy, and image contrast.

[0041] The expression for the calculation function of high-frequency energy is:

[0042] ;

[0043] in, High-frequency energy, The Fourier transform spectrum of the image. This is the filtered image obtained by passing the image through a Butterworth high-pass filter.

[0044] The expression for the function of calculating axial unevenness is:

[0045] ;

[0046] in, For axial unevenness, The horizontal component of the Fourier transform spectrum. This represents the axial component of the Fourier transform spectrum.

[0047] The function expression for calculating high-frequency energy concentration is:

[0048] ;

[0049] in, For high-frequency energy concentration, The peak value of the Fourier transform spectrum.

[0050] The expression for the function to calculate the total energy of the wavelet is:

[0051] ;

[0052] in, For the total energy of the wavelet, The wavelet decomposition of the image yields the first... Wavelet detail coefficients at each wavelet decomposition level.

[0053] The expression for the wavelet detail energy ratio is:

[0054] ;

[0055] in, For wavelet detail energy ratio, The high-frequency detail components obtained from wavelet decomposition of the image. The low-frequency approximation component is obtained from the wavelet decomposition of the image.

[0056] The expression for the calculation function of noise granularity is:

[0057] ;

[0058] in, For noise granularity, and For the height and width of the image, For the original image, The image is the denoised image after the original image has been processed using a Gaussian white noise model.

[0059] The expression for calculating texture contrast is:

[0060] ;

[0061] in, For texture contrast, The position in the gray-level co-occurrence matrix of the image The probability density.

[0062] The expression for the function to calculate texture uniformity is:

[0063] ;

[0064] in, For texture uniformity.

[0065] The expression for calculating texture entropy is:

[0066] ;

[0067] in, This represents the texture entropy.

[0068] The calculation of the energy peak E(u,v) of the preset frequency domain coordinates (u,v) includes: performing a fast Fourier transform on the printed image after dot distortion correction, and extracting the energy peak E(u,v) of the preset frequency domain coordinates (u,v) from the spectrum obtained by the fast Fourier transform.

[0069] The calculation of the Local Binary Pattern Code (LBP) includes: traversing the local dot region of the printed image after dot distortion correction, taking the center pixel value of the local dot region as a threshold, comparing it with its 8 neighboring pixels, recording values ​​greater than the threshold as 1, and values ​​less than or equal to the threshold as 0, thereby arranging the resulting 8-bit binary code clockwise and converting it to decimal as the Local Binary Pattern Code (LBP) for the local dot region.

[0070] The calculation of the gray-level co-occurrence matrix (GLCM) includes: calculating the probability of two pixels with gray levels i and j appearing simultaneously at a given distance and direction in the printed image after halftone distortion correction.

[0071] The attribute energy is the sum of the energies of all dots in the printed image after dot distortion correction, and its function expression is:

[0072] ;

[0073] in, As attribute energy, For the first The weighting coefficient of each network point Let be the energy of the k-th dot in the printed image after dot distortion correction.

[0074] In this embodiment, when using a pre-trained classifier to predict whether a printed item is an anti-counterfeiting printed item prepared by the halftone frequency compensation method, the multi-dimensional features embedded in the texture are used to predict the anti-counterfeiting printed item. The classifier is a machine learning model that has been trained to establish a mapping relationship between the multi-dimensional features embedded in the texture and the judgment result of whether the printed item is an anti-counterfeiting printed item. The machine learning model can adopt the required classifier model as needed. For example, in this embodiment, a support vector machine (SVM) is selected to map the extracted multi-dimensional features embedded in the texture, such as frequency domain energy, spatial texture, and topological distortion, to the feature space to explore the optimal classification hyperplane to distinguish between the original and the copy. The reasoning process focuses on analyzing the statistical distribution anomalies caused by the "active inducement" mechanism, such as the high-frequency energy HFE peak shift and significant improvement in texture contrast caused by resampling interference in the copy. These features have strong linear separability in the SVM space, thereby reducing the model complexity while demonstrating the efficiency and reliability of the classification results through the analysis of physical principles. Based on the probability scores and classification labels output by the machine learning model, and combined with a preset threshold strategy, the final decision is made: when the feature matching degree conforms to the statistical law of the induced interference model, the system determines the sample to be a photocopy; when the feature distribution matches the original design parameters and no resampling distortion features appear, it is determined to be the original printed material. The entire judgment process adopts a multi-feature weighted fusion strategy to ensure the accuracy and reliability of the identification results.

[0075] To verify the effectiveness of the anti-counterfeiting printing method based on halftone frequency compensation in this embodiment, multiple sets of original printed samples and corresponding quantities of photocopies from different brands of photocopying equipment were selected as datasets. Principal component analysis (PCA) was used to reduce the dimensionality of the extracted multi-dimensional features (Principal Component 1 and Principal Component 2), and the data were then input into an SVM classifier for training and testing. The final results are shown below. Figure 3 As shown. Figure 3 As shown, the original sample (blue dot) and the copy sample (red dot) exhibit obvious clustering characteristics in the feature space after dimensionality reduction, and the SVM decision boundary can effectively separate the two types of samples. This proves that the interference features generated by the 'active induction' mechanism in the anti-counterfeiting printing preparation method based on halftone frequency compensation in this embodiment have extremely strong linear separability in statistics and can reliably identify copying behavior.

[0076] In this embodiment, the contribution analysis of multi-dimensional features in the halftone frequency compensation-based anti-counterfeiting printing method includes texture entropy, texture uniformity, texture contrast, noise granularity, wavelet detail energy ratio, total wavelet energy, high-frequency energy concentration, axial imbalance, and high-frequency energy. Figure 4The contribution analysis results show that texture uniformity, high-frequency energy concentration, and axial unevenness contribute significantly more than other indicators. Experimental results indicate that due to the frequency aliasing effect during the copying process, the copy will exhibit significant interference distortion in specific frequency domain coordinates and spatial texture distribution. These key characteristics are core elements for maintaining the stability of the system in cross-brand equipment application scenarios.

[0077] In summary, this embodiment constructs a closed-loop identification system based on halftone frequency compensation for anti-counterfeiting printed materials, employing an "active induction-intelligent judgment" approach. Its core advantages lie in: actively embedding sub-pixel-level latent frequency compensation dots or stripes during the original document preprocessing stage, transforming uncontrollable random noise during copying into strong interference moiré signals with definite physical topological laws, fundamentally solving the problem of smoothing and masking degraded features in modern copying algorithms; and in the back-end detection stage, using machine learning algorithms to perform multi-dimensional joint extraction and attribute determination on salient features forcibly activated by the "physical aliasing effect," effectively achieving a technological leap from single macroscopic stripe detection to high-dimensional microscopic texture classification. Compared with traditional solutions, the innovation and advantages of this solution are: 1. The identification shifts from randomness to determinism. Unlike traditional solutions that passively rely on weak traces left by copying, this solution actively embeds sub-pixel-level frequency compensation dots during the trademark preprocessing stage, forcibly promoting feature manifestation through the frequency aliasing effect during physical sampling. This actively generated moiré pattern is a hardware-level physical interference result, effectively resisting the image restoration algorithms built into modern copiers, such as denoising and smoothing filters, ensuring that the discrimination features still possess extremely high determinism and indelibility in various high-fidelity copying scenarios. 2. A machine learning model is used to perform multi-dimensional nonlinear feature fusion on the induced moiré pattern features, dot topological distortion, and edge step patterns. After model training, the system can accurately learn the essential differences between the original and the copy at the microscale, effectively filtering out interference from paper fiber texture and random environmental noise, significantly improving the recognition robustness and classification accuracy under different paper media and complex imaging environments. 3. In the preprocessing stage, by performing visual threshold matching on the energy distribution and spatial frequency of the compensated dot patterns, the anti-counterfeiting information is deeply embedded within the halftone structure of the trademark. This design achieves a high degree of concealment of the anti-counterfeiting features under human visual observation, without compromising the design aesthetics and visual quality of the original trademark.

[0078] Furthermore, this embodiment also provides an anti-counterfeiting printed matter processing system, including a microprocessor and a memory interconnected thereto, wherein the microprocessor is programmed or configured to execute the anti-counterfeiting printed matter preparation method based on halftone frequency compensation or the anti-counterfeiting printed matter identification method based on halftone frequency compensation.

[0079] Those skilled in the art will understand that the technical solutions provided by this invention may take the form of a method, system, or computer program product. Therefore, this invention may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this invention may take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, produce an implementation of the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for preparing anti-counterfeiting printed materials based on halftone frequency compensation, characterized in that, The process includes the following steps: determining the frequency of visible moiré patterns induced when the anti-counterfeiting printed material is copied, based on the input requirements. Determine the compensation frequency of the embedded texture to be embedded in the original digitized image. According to the compensation frequency Design an embedded texture based on halftone frequency compensation to generate a digital image with hidden embedded texture; A digital image with embedded textures is printed onto a substrate using a printing device to obtain anti-counterfeiting printed materials; the compensation frequency for the embedded textures to be embedded in the original digital image is determined. The function expression is: ； in, The frequency of the visible moiré pattern induced when counterfeit-proof printed materials are copied. The sampling spatial frequency of the replication device, The harmonic order; the compensation frequency The design of a halftone frequency-compensated embedded texture to generate a digital image with hidden embedded textures includes: preprocessing the original digital image, including histogram equalization, image denoising, and background region extraction; using the background region as the embedding region of the embedded texture; and adjusting the pattern duty cycle of the halftone frequency-compensated embedded texture to improve the contrast of the generated digital image with hidden embedded textures. Satisfying the given constraints, the final embedded texture is obtained; the final embedded texture and the denoised digital image are superimposed to obtain a digital image with the embedded texture hidden; the functional expression of the given constraints is: ； ； in, For perception correction coefficients, It is a contrast sensitivity function. To hide the spatial frequencies of a digitized image with embedded texture, the spatial frequencies of a digitized image with embedded texture. Compensation frequency of embedded texture It is obtained by superimposing the spatial frequencies of the original digitized image. These are empirical constant coefficients. It is a natural constant.

2. The method for preparing anti-counterfeiting printed materials based on halftone frequency compensation according to claim 1, characterized in that, The embedded texture is composed of subpixel-level dots or stripes. Subpixel-level means that the physical size of the dots is smaller than the smallest physical dot of the printing equipment or the size corresponding to a single photosensitive unit of the scanning equipment.

3. A method for identifying a security print based on halftone frequency compensation, characterized in that, The process includes the following steps: for the printed image of the printed matter to be identified, extract the texture noise in the non-image and text areas and denoise the printed image to obtain a denoised printed image; then, perform dot distortion correction on the denoised printed image. Multi-dimensional features embedded with texture are extracted from the printed image after dot distortion correction. The multi-dimensional features embedded with texture are then used by a pre-trained classifier to predict whether the printed product to be identified is an anti-counterfeiting printed product obtained by the halftone frequency compensation method described in claim 1 or 2. The classifier is a machine learning model that has been trained to establish a mapping relationship between the multi-dimensional features embedded with texture and the judgment result of whether the printed product to be identified is an anti-counterfeiting printed product.

4. The halftone frequency compensation-based security print discrimination method according to claim 3, characterized in that, The step of extracting texture noise from the non-textual areas and denoising the printed image to obtain a denoised printed image includes: extracting high-frequency texture components from the non-textual areas of the printed image of the printed material to be identified using a high-pass filter; calculating the mean and variance of the high-frequency texture components; constructing a Gaussian white noise model using the mean and variance of the high-frequency texture components as texture noise; and subtracting the texture noise from the printed image of the printed material to be identified to obtain a denoised printed image.

5. The method for identifying anti-counterfeiting printed materials based on halftone frequency compensation according to claim 3, characterized in that, The dot distortion correction of the denoised printed image includes: selecting a local image block of the region of interest in the denoised printed image; extracting the light intensity response curve of the local image block using the step response function (ESF); obtaining the lateral light intensity diffusion distribution of the local image block by numerical differentiation of the light intensity response curve using the line spread function (LSF); generating the spatial light intensity response of the local image block by two-dimensional deconvolution of the lateral light intensity diffusion distribution; and applying a deconvolution operation to the spatial light intensity response of the local image block to obtain the printed image after dot distortion correction.

6. The halftone frequency compensation based security print discrimination method according to claim 3, characterized in that, When extracting multi-dimensional features embedded with texture from a printed image after dot distortion correction, the multi-dimensional features include high-frequency energy, axial imbalance, high-frequency energy concentration, total wavelet energy, wavelet detail energy ratio, noise granularity, texture contrast, texture uniformity, texture entropy, energy peak E(u,v) of preset frequency domain coordinates (u,v), local binary pattern code (LBP), gray-level co-occurrence matrix (GLCM), attribute energy, and some or all of the following: image contrast.

7. A security document processing system comprising a microprocessor and a memory interconnected, characterised in that, The microprocessor is programmed or configured to execute the method for preparing anti-counterfeiting printed materials based on halftone frequency compensation as described in claim 1 or 2, or the method for identifying anti-counterfeiting printed materials based on halftone frequency compensation as described in any one of claims 3 to 6.

Images