A method and system for preventing screen information leakage

Abstract

This invention provides a method and system for preventing screen information leakage, relating to the field of information security. Based on preset parameters of the hidden camera device and the energy distribution of the original information in the transform domain, a target moiré fringe pattern capable of effectively interfering with the hidden camera device is calculated and generated. The hash value of the original information content and dynamic security parameters are logically operated and input into a hyperchaotic system to generate a chaotic phase mask bound to the content and time. By combining the amplitude and frequency information of the moiré fringe with the mask, a preset imaging optical path model is inversely optimized to obtain a stable and effective anti-leakage perturbation pattern within a preset range. The anti-leakage perturbation pattern is pixel-level superimposed with the original information through a multi-scale fusion algorithm to generate a display frame and output it to the screen. The perturbation pattern is updated with each frame or periodically to resist attacks that use multi-frame averaging to eliminate interference.

Description

Technical Field

[0001] This application belongs to the field of information security, and in particular relates to a method and system for preventing the leakage of screen information. Background Technology

[0002] While cameras, smartphones, tablets, and other electronic devices bring convenience, unauthorized individuals can use them, including pinhole cameras, to remotely photograph or record screens, stealing sensitive data such as trade secrets, financial information, and personal privacy. This electronic spying poses a significant threat to information security. Physical privacy films prevent side viewing by limiting the screen's viewing angle, but they reduce screen brightness and clarity, affecting the normal user experience, and are almost ineffective against frontal spying. Early software-based anti-leakage methods tended to overlay fixed noise patterns or digital watermarks on the screen, but static interference patterns are highly regular, and attackers can easily filter them out using simple image processing techniques such as multi-frame averaging and image filtering to recover the original information, resulting in weak protection. Moiré fringes are low-frequency interference patterns generated when the spatial frequencies of two or more objects with periodic structures, such as the pixel grid of a screen, overlap with the CMOS / CCD sensor array of a spying device. By overlaying a carefully designed perturbation pattern onto the original image, strong moiré fringes that disrupt information readability can be induced during imaging by a covert camera. However, the perturbation patterns generated by moiré fringe-based anti-leakage methods are often designed for fixed, idealized covert shooting parameters. For covert cameras of different models, distances, and orientations, the interference effect is unstable, may be significantly weakened or even disappear, and cannot effectively resist hacking attacks based on multi-frame image processing. There is no inherent security correlation between the perturbation pattern and the displayed content; once the perturbation pattern is cracked, all displayed content will be at risk. How to design an anti-leakage method that can adapt to a wide range of covert shooting conditions, possesses time-varying characteristics, and is tightly bound to the displayed content is a pressing technical problem to be solved in this field. Summary of the Invention

[0003] To address the above problems, in a first aspect of the present invention, a method and system for preventing screen information leakage are proposed, comprising the following steps:

[0004] Based on a preset range of typical hidden camera parameters, and combined with the energy distribution characteristics of the original information to be displayed in a specific transform domain, a target moiré pattern that can generate strong visual interference to potential hidden camera devices within the parameter range is calculated and generated.

[0005] The content hash value of the original information is logically operated on with the dynamic security parameter, and the operation result is used as the initial condition to be input into the multidimensional hyperchaotic system. A chaotic phase mask that is dynamically bound to the display content and time is iteratively generated. The dynamic security parameter is a timestamp or a pseudo-random number.

[0006] By combining the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and by performing reverse diffraction propagation optimization calculations on the preset imaging optical path model, a leak-proof disturbance pattern that can stably generate interference effects within a preset range is obtained.

[0007] The anti-leakage disturbance pattern is superimposed on the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames.

[0008] Preferably, the calculation to generate a target moiré pattern capable of causing strong visual interference to potential covert cameras within the parameter range includes:

[0009] Obtain the parameter range of a typical covert camera device, including the image sensor pixel size, lens focal length, shooting distance, and shooting tilt angle;

[0010] The dominant spatial frequency of the aliasing effect, generated by the interaction of the periodic structure of the screen pixel grid and the device sensor grid, is calculated based on Fourier optics theory within the specified parameter range.

[0011] One or more of the dominant spatial frequencies are selected as target frequencies, and corresponding spatial domain periodic patterns or superimposed patterns are synthesized as the target moiré fringe patterns, wherein the spectral energy of the target moiré fringe patterns is concentrated at the target frequencies.

[0012] Preferably, the step of performing logical operations on the content hash value of the original information and dynamic security parameters, and inputting the operation result as an initial condition into a multidimensional hyperchaotic system to iteratively generate a chaotic phase mask dynamically bound to the displayed content and time, includes:

[0013] A hash algorithm is used to calculate the digest value of the original information content, and logical operations are performed with the dynamic security parameters to obtain a seed sequence of fixed length;

[0014] The seed sequence is divided into several data blocks, and each data block is used as the state variable for initializing the multidimensional hyperchaotic system after numerical transformation.

[0015] The multidimensional hyperchaotic system is iterated N times, where N is a preset value. The generated N-dimensional chaotic sequence is transformed into a two-dimensional chaotic matrix with the same screen resolution through interpolation or tiling algorithm. The element values ​​of the two-dimensional chaotic matrix are linearly mapped to the interval [0, 2π] as the chaotic phase mask.

[0016] Preferably, the step of combining the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and performing reverse diffraction propagation optimization calculations on a preset imaging optical path model, yields an anti-leakage disturbance pattern that can stably generate interference effects within a preset range, including:

[0017] An iterative Fourier transform algorithm is used to perform iterative optimization calculations between the initial plane representing the screen and the target plane representing the sensor of the hidden camera device;

[0018] A uniform amplitude constraint is applied to the initial plane, and an amplitude constraint consisting of the spectral amplitude of the target moiré fringe pattern is applied to the target plane, with the chaotic phase mask serving as the initial phase of the target plane.

[0019] After a predetermined number of iterations, the phase distribution obtained on the initial plane will be... Through function Convert the leak prevention perturbation pattern to real values, where Let be a periodic function, A be the disturbance intensity factor, C be the DC bias, p(x,y) be the position (x,y) in the anti-leakage disturbance pattern, and be the value at position p(x,y) in the anti-leakage disturbance pattern.

[0020] Preferably, the step of pixel-level overlaying of the anti-leakage perturbation pattern and the original information based on a multi-scale fusion algorithm to generate a display frame and output it to the screen includes:

[0021] The original information image and the anti-leakage perturbation pattern are decomposed using Laplacian pyramid decomposition to obtain a low-frequency approximate component and a series of bandpass detail components at different scales.

[0022] The low-frequency approximation components of the original information and the low-frequency approximation components of the perturbation pattern are fused by weighted average, with the weight of the original information components being higher than that of the perturbation pattern components.

[0023] For the bandpass detail components at each level, a fusion rule based on the absolute value of the corresponding coefficient or local energy is adopted to select coefficients that can enhance the interference effect for fusion;

[0024] The display frame is reconstructed by performing an inverse Laplacian pyramid transform using the fused low-frequency components and bandpass detail components at each level.

[0025] Preferably, the anti-leakage disturbance pattern is updated with each display frame or a preset time period, triggered by at least one of the following conditions:

[0026] a) The preset time period has expired;

[0027] b) A change in the original information content exceeding a preset threshold is detected;

[0028] After the update is triggered, the steps of generating the chaotic phase mask, calculating the anti-leakage disturbance pattern, and fusing it with the original information are re-executed using the new dynamic security parameters to generate a new display frame.

[0029] In a second aspect of the present invention, a screen information leakage prevention system is provided, comprising the following modules:

[0030] The moiré fringe generation module is used to calculate and generate a target moiré fringe pattern that can produce strong visual interference to potential hidden cameras within a preset range of typical hidden camera parameters, combined with the energy distribution characteristics of the original information to be displayed in a specific transform domain.

[0031] The mask generation module is used to perform logical operations on the content hash value of the original information and the dynamic security parameter, and input the operation result as the initial condition into the multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the display content and time. The dynamic security parameter is a timestamp or a pseudo-random number.

[0032] The perturbation pattern generation module is used to combine the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and obtain a leak-proof perturbation pattern that can stably generate interference effects within a preset range by performing reverse diffraction propagation optimization calculations on a preset imaging optical path model.

[0033] The overlay module is used to overlay the anti-leakage disturbance pattern with the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames.

[0034] This invention uses hash calculations to apply information about the display content itself as one of the initial conditions of a chaotic system, ensuring a unique and secure binding between the generated perturbation pattern and the protected content. The complex and unpredictable patterns reduce the risk of the perturbation patterns being cracked and reused. The mechanism of updating the perturbation pattern with each display frame effectively counters cracking methods that eliminate interference through multi-frame averaging, thus enhancing the strength of the protection. Through backpropagation optimization and multi-scale fusion, it can maintain strong interference against those attempting to take photos while maximizing the clarity of the original information for legitimate users, achieving a good balance between protection effectiveness and user experience. Attached Figure Description

[0035] Figure 1 A schematic diagram of the target moiré stripe pattern;

[0036] Figure 2 This is a schematic diagram of a chaotic phase mask.

[0037] Figure 3 Schematic diagram to prevent leakage and disturbance;

[0038] Figure 4 This is a schematic diagram of the original information;

[0039] Figure 5 This is a schematic diagram of the display frame. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with relevant laws, regulations, and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0041] In the first embodiment, this application proposes a method for preventing the leakage of screen information, including the following steps:

[0042] S1, for a preset range of typical hidden camera parameters, combined with the energy distribution characteristics of the original information to be displayed in a specific transform domain, calculate and generate a target moiré pattern that can produce strong visual interference to potential hidden camera devices within the parameter range.

[0043] The distance range of the hidden camera is set to 1 to 5 meters, with horizontal and vertical deflection angles ranging from -60 degrees to +60 degrees, and sensor pixel size ranging from 1.2 micrometers to 2.8 micrometers. The imaging optical path is approximated using a thin lens model. A Fast Fourier Transform (FFT) is performed on the original information frame to be displayed to analyze its main energy concentration areas in the frequency domain. Any image can be viewed as a superposition of sine waves or similar waveforms of different frequencies and directions. Low-frequency components represent large areas of smooth change in the image, such as solid-color backgrounds, the sky, and skin tones; high-frequency components represent rapidly changing, detailed areas, such as text edges, object textures, and fine lines. Through a Fourier transform, the image can be converted from the visible pixel arrangement to the frequency domain. The frequency domain spectrum, or spectrogram, shows which frequencies constitute the image and the energy of each frequency. If the calculated optimal target interference frequency overlaps with or is very close to a major frequency of the original information, direct use will severely damage the image. Further research is needed to find a suboptimal frequency near the target interference frequency that still produces a good interference effect while avoiding the main information frequency, or to synthesize multiple weaker frequencies distributed in the gaps of the information spectrum and combine them to achieve the interference effect. A genetic algorithm is used for optimization, with the goal of maximizing the contrast of the moiré fringes generated after simulation imaging within the above parameter range, to generate a spectral distribution of the target moiré fringe pattern composed of a two-dimensional sinusoidal grating or chirp signal.

[0044] More specifically, transform domain analysis is performed on the original information to obtain the energy spectrum, physical optical analysis is performed on the hidden camera to find the target frequency range that can generate strong interference, and one or a group of frequencies are selected from the target frequency range. These frequencies must simultaneously meet the two conditions of being able to generate strong interference and being located in the low range of the original information energy spectrum. Based on the decision, the target moiré fringe pattern is calculated and generated.

[0045] S2, perform logical operations on the content hash value of the original information and the dynamic security parameter, and input the operation result as the initial condition into the multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the display content and time. The dynamic security parameter is a timestamp or a pseudo-random number.

[0046] The SHA-256 hash algorithm is used to calculate the hash digest of all pixel values ​​in the current original information frame, resulting in a 256-bit binary sequence. The current system's millisecond-level timestamp is obtained and converted into a binary sequence of the same length. These two binary sequences are then XORed bitwise to obtain a new 256-bit sequence. This sequence is divided into four segments, each 64 bits, and after normalization, serves as the four initial state values ​​x0, y0, z0, and w0 of the Chen's four-dimensional hyperchaotic system. Iterative calculations are performed based on the dynamic equations of the Chen's four-dimensional hyperchaotic system, with the number of iterations equal to the total number of pixels on the display screen. The output sequence of one state variable, such as x, is taken and rearranged into a two-dimensional matrix with the same resolution as the screen. The values ​​in the matrix are linearly mapped to the interval from 0 to 2π, forming a chaotic phase mask.

[0047] S3, combining the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and performing reverse diffraction propagation optimization calculations on the preset imaging optical path model, an anti-leakage disturbance pattern that can stably generate interference effects within a preset range is obtained.

[0048] An improved Gerchberg-Saxton algorithm is used for iterative optimization. In the initial step of the algorithm, on the virtual spy camera sensor plane (i.e., the frequency domain plane), the spectrum of the target moiré fringe pattern generated in the first step is used as the amplitude, and the chaotic phase mask generated in the second step is used as the phase, forming a complex amplitude distribution. An inverse Fraunhofer diffraction transform is performed on this complex amplitude distribution to simulate light propagation from the sensor back to the computer screen plane, obtaining the complex amplitude distribution on the screen plane. The amplitude component is retained as a preliminary estimate of the perturbation pattern, while the phase component is discarded. Another Fraunhofer diffraction transform is performed on this complex amplitude distribution to simulate light propagation from the screen to the sensor plane. On the sensor plane, the currently calculated amplitude is replaced with the target moiré fringe spectrum amplitude obtained in the first step, but the currently calculated phase is retained. This iterative process of reverse and forward propagation is repeated, for example, 100 times, until the complex amplitude distribution calculated on the screen plane converges. The converged complex amplitude distribution map is the anti-leakage perturbation pattern.

[0049] S4, the anti-leakage disturbance pattern is superimposed on the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames.

[0050] A Laplacian pyramid fusion algorithm is employed. A five-layer Laplacian pyramid is constructed using both the original information image and the leak-proof perturbation pattern generated in the previous step. For each layer of the pyramid, the original information layer and the perturbation pattern layer are weighted and fused. For example, the top layer containing high-frequency details is given higher weight to the original information, while the middle and bottom layers containing mid-to-low frequencies are given higher weight to the perturbation pattern. All fused pyramid layers are then reverse-engineered to generate the display frame image. For dynamic content such as video playback, all the above steps are repeated for each frame; for static documents or image displays, the chaotic phase mask and perturbation pattern are regenerated and the display frame is refreshed at a fixed frequency of 20 times per second (i.e., every 50 milliseconds) using a new timestamp.

[0051] In an optional embodiment, the calculation to generate a target moiré pattern capable of causing strong visual interference to potential spying devices within the parameter range includes:

[0052] Obtain the parameter range of a typical covert camera device, including the image sensor pixel size, lens focal length, shooting distance, and shooting tilt angle;

[0053] The dominant spatial frequency of the aliasing effect, generated by the interaction of the periodic structure of the screen pixel grid and the device sensor grid, is calculated based on Fourier optics theory within the specified parameter range.

[0054] One or more of the dominant spatial frequencies are selected as target frequencies, and corresponding spatial domain periodic patterns or superimposed patterns are synthesized as the target moiré fringe patterns, wherein the spectral energy of the target moiré fringe patterns is concentrated at the target frequencies.

[0055] To cover mainstream smartphones and miniature cameras, in another embodiment, the parameters of a typical covert surveillance device are set as follows: image sensor pixel size between 1.0 micrometers and 2.5 micrometers, lens equivalent focal length between 24 millimeters and 70 millimeters, shooting distance between 0.5 meters and 3 meters, and shooting tilt angle between -60 degrees and +60 degrees in both the horizontal and vertical directions. This range ensures the universality of this method for the vast majority of potential covert surveillance scenarios. These parameters collectively determine the period and orientation of the projection grid of the covert surveillance device's sensor grid onto the display screen plane, which forms the basis for subsequent calculations.

[0056] This embodiment utilizes Fourier optics theory to analyze the interference effect between the two two-dimensional periodic grids of the screen and the sensor. For example, a display screen with a resolution of 3840×2160 has pixel grids with specific spatial frequencies fs. When a mobile phone with a sensor pixel size of 1.4 micrometers and a lens focal length of 28 mm takes a picture from 1 meter away, its sensor grid will form a projection frequency fc on the screen. Due to the sampling rate mismatch, aliasing will inevitably occur, which manifests in the frequency domain as the occurrence of the dominant spatial frequency fm, which is approximately equal to the vector difference between fs and fc. By scanning and calculating the entire parameter range, a set of the most likely and visually most interfering fm values ​​can be obtained; these are the dominant spatial frequencies.

[0057] Suppose two dominant spatial frequencies are obtained through calculation, one corresponding to horizontal fringe interference and the other to fringe interference tilted at 45 degrees. This method selects these two frequencies as target frequencies. A corresponding periodic pattern is generated in the spatial domain, for example, by superimposing two cosine functions to construct a two-dimensional grid-like or rhomboid pattern. This pattern is designed so that after Fourier transform, the energy spectrum exhibits sharp energy peaks only at the two selected target frequency positions and their conjugate positions, with extremely low energy in other frequency regions. This synthesized pattern is the target moiré fringe pattern, such as... Figure 1 As shown, it defines the interference patterns presented in candid images.

[0058] In an optional embodiment, the step of performing logical operations on the hash value of the original information content and dynamic security parameters, and inputting the operation result as an initial condition into a multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the displayed content and time, includes:

[0059] A hash algorithm is used to calculate the digest value of the original information content, and logical operations are performed with the dynamic security parameters to obtain a seed sequence of fixed length;

[0060] The seed sequence is divided into several data blocks, and each data block is used as the state variable for initializing the multidimensional hyperchaotic system after numerical transformation.

[0061] The multidimensional hyperchaotic system is iterated N times, where N is a preset value. The generated N-dimensional chaotic sequence is transformed into a two-dimensional chaotic matrix with the same screen resolution through interpolation or tiling algorithm. The element values ​​of the two-dimensional chaotic matrix are linearly mapped to the interval [0, 2π] as the chaotic phase mask.

[0062] The SHA-256 hash algorithm is applied to the original information currently displayed on the screen, such as image data of a confidential document, to generate a 256-bit digest value. A dynamic security parameter, also 256 bits, is obtained, for example, generated from the current system timestamp and a random number. A 256-bit seed sequence is generated by performing a bitwise XOR operation on these two 256-bit values. This seed sequence is uniquely related to the displayed content and is time-varying, ensuring that each generated phase mask is unpredictable and unreproducible.

[0063] The obtained 256-bit seed sequence is divided into four 64-bit data blocks. Preferably, a four-dimensional hyperchaotic system, such as a Chern hyperchaotic system, is used, whose state is determined by four variables: x, y, z, and w. These four 64-bit data blocks are converted into double-precision floating-point numbers to set the initial state values ​​x0, y0, z0, and w0 of the Chern hyperchaotic system. This initialization process injects the entire entropy of the seed into the chaotic system as the starting point for its evolution. The chaotic system undergoes, for example, N equals 500,000 iterative calculations. In one embodiment, N is not equal to the total number of pixels on the display screen or is not an integer multiple of the total number of pixels on the display screen. The value of N is deliberately chosen to avoid integer multiples of the total number of screen pixels, such as 1920 × 1080, which is approximately two million, to prevent potential periodic artifacts.

[0064] After 500,000 iterations, a one-dimensional real-valued chaotic sequence of length 500,000 is obtained from a state variable, such as x, of the hyperchaotic system. To apply this to a 1920×1080 resolution screen, this one-dimensional sequence needs to be expanded into a two-dimensional matrix. This method uses bicubic interpolation to smoothly expand this shorter sequence into a 1920×1080 two-dimensional data matrix. All element values ​​in the two-dimensional data matrix are normalized, linearly mapping them from the original chaotic attraction domain to the interval between 0 and 2π radians. The resulting two-dimensional matrix is ​​the chaotic phase mask, whose numerical distribution appears to be random noise, such as... Figure 2 As shown, it is actually a deterministic pseudo-random phase distribution with a complex internal structure, which is completely determined by the initial seed.

[0065] In an optional embodiment, the step of combining the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and performing reverse diffraction propagation optimization calculations on a preset imaging optical path model, yields an anti-leakage disturbance pattern that can stably generate interference effects within a preset range, including:

[0066] An iterative Fourier transform algorithm is used to perform iterative optimization calculations between the initial plane representing the screen and the target plane representing the sensor of the hidden camera device;

[0067] A uniform amplitude constraint is applied to the initial plane, and an amplitude constraint consisting of the spectral amplitude of the target moiré fringe pattern is applied to the target plane, with the chaotic phase mask serving as the initial phase of the target plane.

[0068] After a predetermined number of iterations, the phase distribution obtained on the initial plane will be... Through function Convert the leak prevention perturbation pattern to real values, where Let be a periodic function, A be the disturbance intensity factor, C be the DC bias, p(x,y) be the position (x,y) in the anti-leakage disturbance pattern, and be the value at position p(x,y) in the anti-leakage disturbance pattern.

[0069] The numerical model of optical propagation constructed in this embodiment includes two core planes: the initial plane represents the display screen, and the target plane represents the sensor of the hidden camera. The core is to repeatedly perform forward and inverse Fourier transforms between these two planes to simulate the diffraction and propagation of the light field. The goal of the entire process is to solve for a pure phase distribution loaded on the screen, which, after diffraction propagation, forms a preset interference image with a specific energy spectrum on the camera sensor. The Gerchberg-Saxton algorithm is preferably used to achieve the phase retrieval task.

[0070] The iterative calculation begins on the target plane, i.e., the camera sensor plane. Two constraints are applied to this plane: amplitude constraints and phase constraints. The amplitude constraint is directly taken from the Fourier spectrum amplitude of the target moiré fringe pattern generated in the previous steps; for example, a matrix that has values ​​only at specific spatial frequency points. The phase constraint uses a chaotic phase mask generated in the previous step, consistent with the screen resolution. This complex amplitude field is then inversely Fourier transformed and propagated to the initial plane, i.e., the screen plane. Upon reaching the screen plane, constraints are applied to the initial plane: its amplitude is forced to a uniform value of 1, as a pure phase perturbation is desired, preserving its calculated phase distribution. The new complex amplitude field is then propagated back to the target plane via a forward Fourier transform, completing one iteration. This process is repeated, for example, 50 times, until the system converges.

[0071] After 50 iterations of optimization, a stable phase distribution matrix is ​​obtained on the initial plane. This is a two-dimensional array containing radian values ​​from negative π to positive π, which cannot be directly displayed. To convert it into a visible image, a mapping function is used. Preferably, a cosine function is used as the periodic function f, i.e. The perturbation intensity factor A is set to a small value, such as 0.1, to control the visibility of the perturbation pattern on the screen and ensure that its impact on normal viewing is minimized; the DC bias C is set to 0.5 to shift the output range of the cosine function from -1 to 1 to between 0 and 1, making it a valid grayscale image intensity value, such as... Figure 3 As shown. In an alternative embodiment, to prevent the periodic function from simplifying the phase distribution, Real-valued images generated for non-periodic functions such as activation functions. These are anti-leakage perturbation patterns that can be superimposed on the original information.

[0072] In an optional embodiment, the step of pixel-level overlaying of the anti-leakage perturbation pattern and the original information based on a multi-scale fusion algorithm to generate a display frame and output it to the screen includes:

[0073] The original information image and the anti-leakage perturbation pattern are decomposed using Laplacian pyramid decomposition to obtain a low-frequency approximate component and a series of bandpass detail components at different scales.

[0074] The low-frequency approximation components of the original information and the low-frequency approximation components of the perturbation pattern are fused by weighted average, with the weight of the original information components being higher than that of the perturbation pattern components.

[0075] For the bandpass detail components at each level, a fusion rule based on the absolute value of the corresponding coefficient or local energy is adopted to select coefficients that can enhance the interference effect for fusion;

[0076] The display frame is reconstructed by performing an inverse Laplacian pyramid transform using the fused low-frequency components and bandpass detail components at each level.

[0077] For the original information image, such as Figure 4 As shown, for example, a page containing text and charts, and the leak prevention perturbation pattern generated in the previous step, are each subjected to a 4-level Laplacian pyramid decomposition. After decomposition, each image obtains a minimum-sized low-frequency approximation component, representing the overall outline and basic tone of the image, as well as four bandpass detail components at different scales, which capture high-frequency information such as edges and textures at different levels of detail. This decomposition allows for the implementation of different fusion strategies at different information levels.

[0078] The fusion process is performed in layers. For the low-frequency approximation components, i.e., the top layer of the pyramid, a weighted average method is used. For example, the fused low-frequency components consist of 95% original information low-frequency components and 5% perturbation pattern low-frequency components. This high weight of original information can preserve the main content and readability of the original image to the greatest extent. For the bandpass detail components of the four layers, the rule of selecting the coefficient with the largest absolute value is adopted. That is, at each pixel position of each layer, the absolute values ​​of the original information detail coefficient and the perturbation pattern detail coefficient are compared, and the coefficient with the larger absolute value is used as the fused coefficient value. The rule of the largest absolute value coefficient tends to preserve more significant edges and textures, and can effectively and seamlessly inject high-frequency interference details in the perturbation pattern into the details of the original image. Figure 5 This is the final display frame generated.

[0079] After all components at all levels have been fused according to their respective rules, a new fused Laplacian pyramid is obtained, containing one fused low-frequency component and four fused bandpass detail components. The inverse transform reconstruction process of the Laplacian pyramid is then performed. Starting from the finest detail layer, it is upsampled layer by layer and added to the image of the next layer, until the bottom of the pyramid. This inverse process recombines all the fused multi-scale information to generate a display frame with the same resolution as the original image. The display frame clearly presents the original information content while embedding anti-leakage perturbations at different scales, achieving a balance between information protection and visual experience.

[0080] In an optional embodiment, the anti-leakage disturbance pattern is updated with each display frame or a preset time period, triggered by at least one of the following conditions:

[0081] a) The preset time period has expired;

[0082] b) A change in the original information content exceeding a preset threshold is detected;

[0083] After the update is triggered, the steps of generating the chaotic phase mask, calculating the anti-leakage disturbance pattern, and fusing it with the original information are re-executed using the new dynamic security parameters to generate a new display frame.

[0084] One update triggering mechanism is based on a fixed time period. An internal timer is set, for example, with a period of 200 milliseconds. Every 200-millisecond interval, a complete perturbation pattern update process is automatically triggered. This means that even if the content displayed on the screen remains completely unchanged, the anti-leakage perturbation pattern will refresh at a frequency of 5 times per second. To the human eye, such rapid and subtle pattern changes are almost imperceptible, but for cameras with exposure times typically in the tens of seconds, each capture is a different perturbation field, making attempts to eliminate perturbation through multi-frame averaging extremely difficult.

[0085] Another update triggering mechanism is based on content changes. Specifically, it monitors data in the display buffer in real time and calculates the difference between the current frame and the previous frame. For example, it calculates the percentage of pixels whose pixel values ​​have changed out of the total number of pixels. When this rate of change exceeds a preset threshold, such as 8%, a significant content update is determined, such as when the user scrolls the page or switches applications. At this point, a perturbation pattern update is immediately triggered. Because the content has changed, its hash value also changes. Combined with new dynamic security parameters, a completely new perturbation pattern closely associated with the new content is generated, ensuring the targeting and effectiveness of the protection and keeping the protection always synchronized with the currently displayed sensitive information.

[0086] In a second embodiment, this application provides a screen information leakage prevention system, including the following modules:

[0087] The moiré fringe generation module is used to calculate and generate a target moiré fringe pattern that can produce strong visual interference to potential hidden cameras within a preset range of typical hidden camera parameters, combined with the energy distribution characteristics of the original information to be displayed in a specific transform domain.

[0088] The mask generation module is used to perform logical operations on the content hash value of the original information and the dynamic security parameter, and input the operation result as the initial condition into the multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the display content and time. The dynamic security parameter is a timestamp or a pseudo-random number.

[0089] The perturbation pattern generation module is used to combine the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and obtain a leak-proof perturbation pattern that can stably generate interference effects within a preset range by performing reverse diffraction propagation optimization calculations on a preset imaging optical path model.

[0090] The overlay module is used to overlay the anti-leakage disturbance pattern with the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames.

[0091] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, compact disc read-only memory (CD-ROM), optical storage, etc.) containing computer-usable program code.

[0092] 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, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the technical solution of this application.

Claims

1. A method for preventing the leakage of screen information, characterized in that, Includes the following steps: Based on the preset range of typical hidden camera parameters, and combined with the energy distribution characteristics of the original information to be displayed in the fast Fourier transform domain, a target moiré fringe pattern that can generate strong visual interference to potential hidden camera devices within the parameter range is calculated and generated. The content hash value of the original information is logically operated on with the dynamic security parameter, and the operation result is used as the initial condition to be input into the multidimensional hyperchaotic system. A chaotic phase mask that is dynamically bound to the display content and time is iteratively generated. The dynamic security parameter is a timestamp or a pseudo-random number. By combining the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and by performing reverse diffraction propagation optimization calculations on the preset imaging optical path model, a leak-proof disturbance pattern that can stably generate interference effects within a preset range is obtained. The anti-leakage disturbance pattern is superimposed on the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames. The method combines the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and through inverse diffraction propagation optimization calculation on a preset imaging optical path model, obtains an anti-leakage disturbance pattern that can stably generate interference effects within a preset range, including: An iterative Fourier transform algorithm is used to perform iterative optimization calculations between the initial plane representing the screen and the target plane representing the sensor of the hidden camera device; A uniform amplitude constraint is applied to the initial plane, and an amplitude constraint consisting of the spectral amplitude of the target moiré fringe pattern is applied to the target plane, with the chaotic phase mask serving as the initial phase of the target plane. After a predetermined number of iterations, the phase distribution obtained on the initial plane will be... Through function The leakage prevention disturbance pattern is converted to a real value, where f(·) is a periodic function, A is the disturbance intensity factor, and C is the DC bias; p(x,y) is the position (x,y) in the leakage prevention disturbance pattern. To prevent leaks and disturbances, the value at position p(x,y) in the pattern is determined.

2. The method according to claim 1, characterized in that, The calculation generates a target moiré pattern that can strongly visually interfere with potential surreptitious devices within the parameter range, including: Obtain the parameter range of a typical covert camera device, including the image sensor pixel size, lens focal length, shooting distance, and shooting tilt angle; The dominant spatial frequency of the aliasing effect, generated by the interaction of the periodic structure of the screen pixel grid and the device sensor grid, is calculated based on Fourier optics theory within the specified parameter range. One or more of the dominant spatial frequencies are selected as target frequencies, and corresponding spatial domain periodic patterns or superimposed patterns are synthesized as the target moiré fringe patterns, wherein the spectral energy of the target moiré fringe patterns is concentrated at the target frequencies.

3. The method according to claim 1, characterized in that, The step of performing logical operations on the content hash value of the original information and dynamic security parameters, and using the operation result as an initial condition input into a multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the displayed content and time, includes: A hash algorithm is used to calculate the digest value of the original information content, and logical operations are performed with the dynamic security parameters to obtain a seed sequence of fixed length; The seed sequence is divided into several data blocks, and each data block is used as the state variable for initializing the multidimensional hyperchaotic system after numerical transformation. The multidimensional hyperchaotic system is iterated N times, where N is a preset value. The generated N-dimensional chaotic sequence is transformed into a two-dimensional chaotic matrix with the same screen resolution through interpolation or tiling algorithm. The element values ​​of the two-dimensional chaotic matrix are linearly mapped to the interval [0, 2π] as the chaotic phase mask.

4. The method according to claim 1, characterized in that, The step of performing pixel-level overlay of the anti-leakage perturbation pattern and the original information based on a multi-scale fusion algorithm to generate a display frame and output it to the screen includes: The original information image and the anti-leakage perturbation pattern are decomposed using Laplacian pyramid decomposition to obtain a low-frequency approximate component and a series of bandpass detail components at different scales. The low-frequency approximation components of the original information and the low-frequency approximation components of the perturbation pattern are fused by weighted average, with the weight of the original information components being higher than that of the perturbation pattern components. For the bandpass detail components at each level, a fusion rule based on the absolute value of the corresponding coefficient or local energy is adopted to select coefficients that can enhance the interference effect for fusion; The display frame is reconstructed by performing an inverse Laplacian pyramid transform using the fused low-frequency components and bandpass detail components at each level.

5. The method according to claim 1, characterized in that, The anti-leakage disturbance pattern is updated with each display frame or a preset time period, and is triggered by at least one of the following conditions: a) The preset time period has expired; b) A change in the original information content exceeding a preset threshold is detected; After the update is triggered, the steps of generating the chaotic phase mask, calculating the anti-leakage disturbance pattern, and fusing it with the original information are re-executed using the new dynamic security parameters to generate a new display frame.

6. A screen information leakage prevention system, characterized in that, Includes the following modules: The moiré fringe generation module is used to calculate and generate a target moiré fringe pattern that can produce strong visual interference to potential hidden cameras within the preset range of typical hidden camera parameters, combined with the energy distribution characteristics of the original information to be displayed in the fast Fourier transform domain. The mask generation module is used to perform logical operations on the content hash value of the original information and the dynamic security parameter, and input the operation result as the initial condition into the multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the display content and time. The dynamic security parameter is a timestamp or a pseudo-random number. The perturbation pattern generation module is used to combine the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and obtain a leak-proof perturbation pattern that can stably generate interference effects within a preset range by performing reverse diffraction propagation optimization calculations on a preset imaging optical path model. The overlay module is used to overlay the anti-leakage disturbance pattern with the original information at the pixel level based on a multi-scale fusion algorithm to generate a display frame and output it to the screen. The anti-leakage disturbance pattern is updated with the display frame or a preset time period to counteract the leakage method of eliminating interference by averaging multiple frames. The method combines the amplitude-frequency information of the target moiré fringe pattern with the chaotic phase mask, and through inverse diffraction propagation optimization calculation on a preset imaging optical path model, obtains an anti-leakage disturbance pattern that can stably generate interference effects within a preset range, including: An iterative Fourier transform algorithm is used to perform iterative optimization calculations between the initial plane representing the screen and the target plane representing the sensor of the hidden camera device; A uniform amplitude constraint is applied to the initial plane, and an amplitude constraint consisting of the spectral amplitude of the target moiré fringe pattern is applied to the target plane, with the chaotic phase mask serving as the initial phase of the target plane. After a predetermined number of iterations, the phase distribution obtained on the initial plane will be... Through function The leakage prevention disturbance pattern is converted to a real value, where f(·) is a periodic function, A is the disturbance intensity factor, and C is the DC bias; p(x,y) is the position (x,y) in the leakage prevention disturbance pattern. To prevent leaks and disturbances, the value at position p(x,y) in the pattern is determined.

7. The system according to claim 6, characterized in that, The calculation generates a target moiré pattern that can strongly visually interfere with potential surreptitious devices within the parameter range, including: Obtain the parameter range of a typical covert camera device, including the image sensor pixel size, lens focal length, shooting distance, and shooting tilt angle; The dominant spatial frequency of the aliasing effect, generated by the interaction of the periodic structure of the screen pixel grid and the device sensor grid, is calculated based on Fourier optics theory within the specified parameter range. One or more of the dominant spatial frequencies are selected as target frequencies, and corresponding spatial domain periodic patterns or superimposed patterns are synthesized as the target moiré fringe patterns, wherein the spectral energy of the target moiré fringe patterns is concentrated at the target frequencies.

8. The system according to claim 6, characterized in that, The step of performing logical operations on the content hash value of the original information and dynamic security parameters, and using the operation result as an initial condition input into a multidimensional hyperchaotic system to iteratively generate a chaotic phase mask that is dynamically bound to the displayed content and time, includes: A hash algorithm is used to calculate the digest value of the original information content, and logical operations are performed with the dynamic security parameters to obtain a seed sequence of fixed length; The seed sequence is divided into several data blocks, and each data block is used as the state variable for initializing the multidimensional hyperchaotic system after numerical transformation. The multidimensional hyperchaotic system is iterated N times, where N is a preset value. The generated N-dimensional chaotic sequence is transformed into a two-dimensional chaotic matrix with the same screen resolution through interpolation or tiling algorithm. The element values ​​of the two-dimensional chaotic matrix are linearly mapped to the interval [0, 2π] as the chaotic phase mask.

Images