Screen content dynamic fuzzy processing method based on user authority verification

Abstract

The present invention relates to the field of electronic data processing technology and discloses a method for dynamically blurring screen content based on user permission verification. The method comprises: collecting the original screen content on the screen sharing end and dividing it into regions, assigning target permission levels based on the regional content; obtaining the user permission level on the viewing end and comparing it with the target permissions of the layers, blurring the layers with insufficient permissions; determining the screen gaze point based on the clarity of the user image and eye tracking, and adjusting the user permissions; and finally, pushing the blurred screen content based on the adjusted permissions. By dynamically adjusting permissions and blurring, this method ensures the protection of sensitive information while improving the user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic data processing, and in particular to a method for dynamic fuzzy processing of screen content based on user authority verification. Background Art

[0002] Traditional permission management methods such as password authentication and role assignment can control user access rights, but it is difficult to effectively protect the security of specific areas or partial data when multiple users share the same screen content.

[0003] Existing screen content protection technologies mostly rely on static blurring or hiding processing, which has the following shortcomings: a unified blurring strategy is adopted for all users and cannot be dynamically adjusted according to user permissions; static blurring processing may make important information completely invisible, affecting normal work processes; and there is a lack of real-time and intelligent permission verification mechanisms. Summary of the Invention

[0004] In view of this, the present invention proposes a screen content dynamic fuzzy processing method based on user authority verification to solve the problems in the prior art.

[0005] The present invention proposes a method for dynamic fuzzy processing of screen content based on user authority verification, comprising:

[0006] Collecting the original screen content of the screen sharing terminal, dividing the original screen content into a number of regions, inputting each region into a different layer, and assigning a target permission level to the layer according to the original screen content in each region;

[0007] Obtaining the user authority level of each user on the viewing end, comparing the user authority level with the target authority level of each layer, and performing fuzzy processing on the layers whose user authority level is lower than the target authority level;

[0008] Overlaying the blurred layer with the remaining unblurred layer to form blurred screen content, and pushing the blurred screen content;

[0009] The user image of the viewing end after receiving the blurred screen content is collected according to the time series, and the screen gaze point of the viewing end user is determined based on the eye tracking technology; the trustworthiness of the screen gaze point is calculated according to the clarity of the user image; the user permission level is adjusted according to the relationship between the screen gaze point and each of the layers to obtain the modified user permission;

[0010] Collecting a user portrait of the screen sharing terminal according to a time series, determining the screen gaze point based on the user image, and adjusting the user authority level based on the screen gaze points of the screen sharing terminal and the viewing terminal to obtain the final user authority;

[0011] Subsequent blurred screen content is pushed based on the end user permissions and the target permission level.

[0012] Furthermore, when the original screen content is recognized and divided into several areas, it includes:

[0013] The window borders in the original screen content are obtained based on Canny edge detection, the area enclosed by all the window borders is calculated, the window border with the smallest area is selected and input into a layer, and the content of the window border is removed from the original screen content. The window border with the smallest area is selected again and input into another layer, and the selection is repeated until the original screen content is completely divided.

[0014] Furthermore, when assigning a target permission level to the layer according to the original screen content in each of the areas, it includes:

[0015] Establish a sensitive word library, which includes sensitive keywords and corresponding viewing permission levels;

[0016] Performing OCR text recognition and application interface feature recognition on the original screen content in each layer to determine whether the layer contains sensitive keywords or a preset specific application program interface;

[0017] When it is determined that the preset feature application interface is included, the maximum target permission level is assigned to the current layer; when it is determined that the sensitive keywords and the preset specific application interface are not included, the minimum target permission level is assigned to the current layer;

[0018] When it is determined that only sensitive keywords are included, the position coordinates of the sensitive keywords in the current layer are obtained, and it is determined whether to re-divide the layer based on the position coordinates.

[0019] Furthermore, when there is no need to redivide the layers or after the layers have been redivided, the viewing permission level corresponding to each sensitive keyword is obtained according to the sensitive word library, the highest viewing permission level is selected, and the highest viewing permission level is used as the target permission level of the current layer.

[0020] Furthermore, when determining whether to re-divide the layers according to the position coordinates, the method includes:

[0021] Calculate the straight-line distance from each position coordinate to the same edge of the current layer and the interval distance from the edge to the opposite edge respectively, and re-divide the layer when max(di)≤D / 3 is satisfied;

[0022] Wherein, di represents the straight-line distance from the position coordinate to edge i in the current layer, di={x1, x2, x3, ..., xn}, i is a positive integer; the maximum value of i is the total number of edges in the current layer, xn represents the straight-line distance from the nth position coordinate to the same edge of the current layer, max(di) represents the maximum value in the set di, and D represents the interval distance.

[0023] Furthermore, when the layers are re-divided, it includes:

[0024] When multiple max(di)s are less than or equal to D / 3, the smallest max(di) is selected and the current layer is split into two sub-layers, one of which contains all sensitive keywords.

[0025] Furthermore, when calculating the trustworthiness of the screen gaze point according to the clarity of the user image, the method includes:

[0026] The user image is converted into a grayscale image, and the clarity satisfies the following relationship:

[0027] ;

[0028] The clarity threshold is set to 100. When the clarity is less than or equal to 100, it is judged as low confidence. When the clarity is greater than 100, the confidence is calculated according to the normalization method. The confidence satisfies the following relationship:

[0029] ;

[0030] in, For clarity, is the image resolution, which represents the product of the image width and height, In the grayscale image The pixel value at is the Laplace value, is the average of all Laplace values; For trust, is the clarity threshold, For maximum clarity.

[0031] Furthermore, the user authority level is adjusted according to the relationship between the screen gaze point and each of the layers to obtain a modified user authority, including:

[0032] Obtain the layer where the screen gaze point is located, and calculate the dwell time on the layer. When a preset time threshold is met, determine whether the layer contains sensitive keywords; if not, replace the modified user authority with the user authority level; if yes, perform the following steps:

[0033] Obtain the screen gaze points of the remaining viewing terminals and record them as reference gaze points. Count the number of the reference gaze points in the layer and calculate the attention degree of the layer, where the attention degree is the ratio of the number of reference gaze points in the layer to the total number of viewing terminals. When the ratio is less than the attention threshold, lower the user authority level of the viewing terminal corresponding to the screen gaze point by one level and record it as the modified user authority. When the ratio is greater than or equal to the attention threshold, use the user authority level as the modified user authority.

[0034] Furthermore, when the user authority level is adjusted according to the screen gaze points of the screen sharing terminal and the viewing terminal to obtain the final user authority, it includes:

[0035] When the screen gaze points of the screen sharing end and the viewing end are in the same layer, the modified user permissions of the viewing end are used as the final user permissions;

[0036] When the screen gaze points of the screen sharing end and the viewing end are not in the same layer, the dwell time is calculated. When the time threshold is met, the modified user is downgraded by one level to obtain the final user authority.

[0037] Furthermore, it also includes:

[0038] When the user authority level changes, the blur level of the layer is adjusted to the highest level. Compared with the prior art, the beneficial effects of the present invention are:

[0039] By dividing screen content into multiple layers and assigning target permission levels to each layer, differentiated permission control is achieved for different content areas. Compared to a unified fuzzy strategy, this approach provides higher levels of protection for sensitive information areas while maintaining visibility in non-sensitive areas, thereby improving the balance between information security and user experience. Using eye tracking technology to capture each viewer's screen gaze in real time, the system dynamically adjusts user permission levels based on layer content and gaze behavior. This effectively identifies anomalous viewing behavior or unauthorized peeping, enhancing security control capabilities in multi-user sharing scenarios such as conferences and remote collaboration. The concept of "attention level" is introduced. By considering group statistics of gaze points and user behavior, it intelligently determines whether a specific area is a sensitive hotspot and dynamically adjusts individual user permissions based on this information. This enhances the system's adaptability to user group behavior and the accuracy of sensitive area protection. By using dynamic fuzzing instead of static hiding, the system prevents unauthorized users from identifying information while avoiding "over-obstruction" of content that could disrupt normal user experience, thereby improving the flexibility and practicality of information protection. The credibility of the user's gaze point is assessed through image clarity indicators to avoid misjudgments caused by abnormal conditions such as image blur and posture occlusion, ensuring the rationality and robustness of permission adjustments. Supports coordinated permission adjustment based on the gaze points of the screen sharing end and the viewing end. Using the gaze point of the user on the screen sharing end as a reference for permission adjustment, a linkage judgment logic is formed between the content creator and the viewer, improving the contextual relevance and rationality of permission control decisions. Based on Canny edge detection and sensitive word position judgment, the system can re-divide the screen layers as needed to ensure that sensitive content is clustered rather than dispersed, which helps improve the logical clarity of the layers and the accuracy of permission control. By building a sensitive word library and combining OCR and interface feature recognition methods, it can flexibly adapt to a variety of application scenarios (such as text, finance, medical care, etc.) and has good scalability and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0041] Figure 1 This is a flowchart of a method for dynamic fuzzy processing of screen content based on user authority verification provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The exemplary embodiments disclosed in the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0043] See Figure 1 As shown, this embodiment provides a method for dynamic blur processing of screen content based on user authority verification, including:

[0044] S1: Collect the original screen content of the screen sharing terminal, divide the original screen content into several areas, input each area into a different layer, and assign the target permission level to the layer according to the original screen content in each area;

[0045] S2: Obtain the user permission level of each user on the viewing end, compare the user permission level with the target permission level of each layer, and perform fuzzy processing on the layers whose user permission level is lower than the target permission level;

[0046] S3: Overlaying the blurred layer with the remaining unblurred layers to form blurred screen content, and pushing the blurred screen content;

[0047] S4: Collect user images of the viewing end after receiving the blurred screen content based on the time series, and determine the screen gaze point of the viewing end user based on eye tracking technology; calculate the trustworthiness of the screen gaze point based on the clarity of the user image; adjust the user permission level based on the relationship between the screen gaze point and each layer to obtain the corrected user permission;

[0048] S5: Collect user portraits of the screen sharing end based on the time series, determine the screen gaze point based on the user image, and adjust the user permission level based on the screen gaze points of the screen sharing end and the viewing end to obtain the final user permission;

[0049] S6: Pushing subsequent blurred screen content based on the end user's permissions and the target permission level.

[0050] Specifically, after obtaining the user image, any means of processing the image so that the image can participate in subsequent operations belongs to the existing technology. In order to more clearly introduce the technical solution of this embodiment, the above-mentioned "any means" can be: improving the resolution and frame rate of the viewing end camera, adjusting the light source conditions when obtaining the image; image sharpening processing; and image super-resolution processing, etc.

[0051] It's important to note that dynamic blurring technology effectively protects sensitive information by adjusting the visibility of screen content based on user privilege levels. For example, when a user's privileges are insufficient, the system automatically blurs areas that don't match their permissions, preventing them from viewing or accessing inappropriate content. This approach not only ensures data confidentiality but also prevents human error or leaks, improving overall information security.

[0052] The system dynamically adjusts permissions based on real-time user behavior (such as screen gaze point and image clarity), making permission management more flexible and adaptive. For example, if the area a user is gazing at on the screen is blurred, and the clarity of their gaze point is sufficiently high, the system will adjust their permissions based on the trustworthiness of the gaze point. This dynamic adjustment mechanism based on user behavior enables the system to respond to user needs and behaviors in real time, ensuring accurate and personalized information display.

[0053] Dynamic blur processing allows content to be displayed based on the permissions of different users, providing a personalized screen content experience. For example, high-privilege users can see the entire content, while low-privilege users can only see the blurred portion. This approach not only meets security requirements but also flexibly adjusts content display based on user permissions and behavior, avoiding interference from irrelevant content and thus improving the user experience.

[0054] Through eye tracking technology, the system can accurately detect the user's gaze point, enabling more precise permission control based on user behavior. If the user's gaze point approaches a sensitive area, the system can calculate the trust level based on the clarity of the user's image and adjust the user's permissions in real time, improving the accuracy and flexibility of permission control and ensuring that only users who meet the permission requirements can access high-privilege content.

[0055] Dynamic fuzzification technology not only performs preliminary screening based on preset permissions but also further calibrates and adjusts permissions based on time series, user behavior, and eye tracking information. This mechanism avoids information leaks caused by initial permission allocation errors, ensuring that the system always makes the most accurate permission adjustments based on real-time data, effectively preventing the risk of leaks.

[0056] This method supports real-time interaction and collaboration among multiple users and is particularly suitable for scenarios requiring screen sharing and collaborative work. In a multi-user environment, the system dynamically adjusts the visible area based on each user's permissions and behavior. For viewing users, the system uses eye tracking technology and image clarity calculations to determine permissions changes in real time, ensuring that each user only has access to appropriate content during the collaborative process, enhancing the smoothness and security of the workflow.

[0057] The system automatically adjusts permissions based on real-time user behavior, eliminating the need for manual intervention or review. This automated process not only improves system efficiency but also reduces bias and errors caused by manual intervention, reducing labor costs and potential operational errors.

[0058] In multi-user permission management scenarios, user permissions and behaviors are subject to change. The system can flexibly adjust and modify permissions based on real-time data, making the entire process highly adaptable and fault-tolerant. Even when user behavior or permission settings change, the system can still effectively respond, ensuring a balance between information security and user experience.

[0059] This method is applicable to various scenarios where screen content display must be tailored to user permissions, such as remote conferencing, online education, enterprise file sharing, and the viewing of sensitive information. In scenarios with high information security requirements, dynamically adjusting permissions and obfuscating screen content can significantly improve the efficiency and accuracy of data protection and access control.

[0060] In some embodiments of the present application, when performing content recognition on the original screen content and dividing it into several regions, the process includes:

[0061] The original screen content is detected based on Canny edge detection to obtain the window borders in the original screen content, and the area enclosed by all window borders is calculated. The window border with the smallest area is selected and input into a layer, and the content of the window border is removed from the original screen content. The window border with the smallest area is selected again and input into another layer. The selection is repeated until the original screen content is completely divided.

[0062] It's important to note that Canny edge detection is a classic edge detection algorithm that effectively identifies edges within screen content, particularly for extracting window borders. This method allows the system to accurately identify the boundaries of different regions (such as windows) and segment content based on those boundaries. This segmentation method divides screen content into multiple independent regions, facilitating subsequent blurring, permission verification, and content display.

[0063] The area enclosed by each window border is divided according to its size, allowing the system to prioritize smaller window areas. This segmentation method effectively avoids confusion when processing complex screen layouts, allowing each area to be processed separately according to specific logic and priority. By prioritizing the smallest windows, the system can ensure that the smallest areas are processed first when blurring, gradually covering the entire screen, and avoiding omissions and duplications.

[0064] By removing the already processed areas (window border content) and continuing the segmentation based on the remaining areas, the integrity of the original screen content is preserved. By selecting the smallest window and removing its content each time, we can effectively avoid missing or duplicating areas during the segmentation process, ensuring that all content is properly and completely divided into different layers, providing complete basic data for subsequent fuzzy processing and permission control.

[0065] Using Canny edge detection to extract window borders enables highly accurate content recognition by clearly extracting edge information. For complex screen layouts (such as multiple windows or application interfaces), this method effectively distinguishes the boundaries of each window, resulting in more accurate content segmentation results, avoiding errors that can occur during manual segmentation, and enhancing the stability of the segmentation process.

[0066] The size, shape, and number of windows may vary across different screen layouts. By gradually processing the smallest windows, the system can adapt to different screen layouts. Regardless of window size, the system automatically prioritizes processing from smallest to largest, ensuring that each area is prioritized. This approach offers flexibility and can adapt to screen layouts of varying types and complexities.

[0067] By placing each segmented area into a separate layer for permission assignment and fuzzy processing, each area can be processed independently, avoiding interference between areas. This segmentation method allows the system to more precisely control the permission level of each area in subsequent steps and dynamically adjust the display content for users with different permissions. Through this refined regional processing, the system's processing efficiency is greatly improved, especially when dealing with complex screen content.

[0068] Assigning each region to a different layer and applying blurring based on each region's permission level makes this method ideal for scenarios requiring dynamic permission control. For example, when certain users have lower permissions, the system can blur the areas they need to view, while users with higher permissions can see clear content. This precise region division allows the system to better target blurring to different areas, improving security and personalizing the experience.

[0069] The edge detection-based segmentation method ensures that the system maintains good segmentation across a variety of screen layouts and content structures. This is especially true in complex screen sharing scenarios (such as multiple windows and multiple applications interacting), where the system effectively identifies each window and correctly segments it, enhancing the system's adaptability and fault tolerance.

[0070] In multi-user scenarios (such as video conferencing and online collaboration), each user may have different permission levels. Through refined segmentation and permission control, different levels of obfuscation can be presented to each user based on their permission level. This flexible approach supports the permission management requirements of multi-client collaboration scenarios, ensuring that each user can only access information consistent with their permissions.

[0071] In some embodiments of the present application, when assigning target permission levels to layers based on the original screen content within each area, the following steps are included:

[0072] Establish a sensitive word library, which includes sensitive keywords and corresponding viewing permission levels;

[0073] Perform OCR text recognition and application interface feature recognition on the original screen content in each layer to determine whether the layer contains sensitive keywords or preset specific application interfaces;

[0074] Specifically, the preset specific applications include but are not limited to WeChat, QQ and Alipay, etc. Since these applications are logged in through private accounts, the privacy protection of the above applications needs to be carried out.

[0075] When it is determined that the preset feature application interface is included, the maximum target permission level is assigned to the current layer; when it is determined that the sensitive keywords and the preset specific application interface are not included, the minimum target permission level is assigned to the current layer;

[0076] When it is determined that only sensitive keywords are included, the position coordinates of the sensitive keywords in the current layer are obtained, and it is determined whether to re-divide the layer based on the position coordinates.

[0077] It's important to note that by building a sensitive word library and combining it with OCR technology to identify sensitive keywords within layer content, sensitive information on the screen can be dynamically identified. This provides precise protection for privacy, especially when handling private information. Combined with feature recognition for specific applications (such as WeChat, QQ, and Alipay), the system can identify and protect the interfaces of these highly privacy- and security-critical application programs, preventing unauthorized users from accessing sensitive information.

[0078] When a layer is identified as containing sensitive information or an application interface, the system can assign the maximum permission level to that layer, ensuring that authorized users can access the full content while other users cannot view it. Through this permission control mechanism, the system can effectively manage access rights in a multi-user environment, preventing users from accessing sensitive or private data due to insufficient permissions, thereby enhancing overall security.

[0079] After matching the layer content with sensitive word libraries and identifying the application interface, more flexible permission control can be achieved. When the layer does not contain sensitive keywords and specific application interfaces, it can be assigned a minimum permission level to ensure that unnecessary information is not exposed to unauthorized users. When the layer contains sensitive information, the system will automatically adjust the permission level based on the location and type of sensitive words and the characteristics of the application, thereby achieving more refined permission division and preventing the disclosure of user privacy.

[0080] OCR text recognition and application-specific feature recognition automatically determine whether screen content contains sensitive information or private applications without manual intervention. This automated processing improves efficiency and responds in real time to user behavior and changes in screen content. This allows the system to adjust the screen display in real time based on the specific situation when facing different shared content, ensuring intelligent and efficient privacy protection.

[0081] When a layer is identified as containing sensitive keywords, the system not only assigns permissions but also determines whether the layer needs to be re-divided by obtaining the coordinates of the sensitive keywords within the layer. This approach accurately distinguishes sensitive content from non-sensitive content and allocates and adjusts layers appropriately based on factors such as the importance, location, and sensitivity of the content, thereby preventing unnecessary information leaks.

[0082] In multi-user screen sharing or collaboration scenarios, each user may require different access permissions. For example, some users may need to view the entire screen content, while others may only be able to view blurred or partial content. By combining sensitive word recognition, application interface recognition, and location determination, the system can provide personalized permission control based on each user's permission level and the sensitivity of the screen content. This fine-grained permission management not only meets security requirements but also enhances the user experience.

[0083] By identifying the interface features of specific applications (such as WeChat, QQ, and Alipay), the system can effectively identify and protect application interfaces containing sensitive content such as private accounts and transaction information. This cross-application permission management ensures security across diverse application scenarios, ensuring appropriate permission protection for chat apps, payment platforms, and other privacy-sensitive applications.

[0084] By integrating sensitive word libraries, application identification, and coordinate location determination, the system achieves multi-level permission control. Content in different areas can be dynamically adjusted, taking into account not only the sensitivity of the content itself but also its location, specific user permission level, and the environment in which it occurs. This significantly enhances the system's flexibility and adaptability, enabling it to address a wide range of complex scenarios and security requirements.

[0085] In a multi-user environment, this approach, taking into account behavioral differences and potential false triggering, can prevent information leaks caused by misoperation or improper permission settings through precise layer segmentation, sensitive information identification, and dynamic permission adjustment, thereby improving the system's fault tolerance and stability. Furthermore, pre-defined application feature recognition helps avoid over-obfuscation, preventing users from accessing shared content due to insufficient permissions.

[0086] In some embodiments of the present application, when there is no need to redivide the layers or after the layers have been redivided, the viewing permission level corresponding to each sensitive keyword is obtained according to the sensitive vocabulary library, the highest viewing permission level is selected, and the highest viewing permission level is used as the target permission level of the current layer.

[0087] It's important to note that by identifying sensitive terms for each layer and assigning access permissions based on a database of sensitive terms, the system ensures that any portion of a layer containing sensitive content is appropriately protected. If a layer contains sensitive terms, the system automatically increases the layer's permissions, ensuring that only users with the appropriate permissions can access sensitive content, thus preventing unauthorized access to sensitive information.

[0088] When multiple sensitive words appear on the same layer, the highest permission level is selected as the target permission level for the current layer. This avoids permission conflicts between different sensitive information and allows for unified management and allocation of access rights. This approach simplifies the permission management process, ensures uniformity and rationality of content access rights, and improves system management efficiency.

[0089] In a multi-user environment, some layers may contain multiple sensitive keywords. Protecting these layers with the highest permission level effectively improves system security by preventing low-privilege users from viewing high-risk content. This allows the system to promptly lock down and protect highly sensitive areas, preventing inappropriate access and information leaks, and mitigating potential risks.

[0090] By automatically adjusting permission levels based on a sensitive vocabulary, the system intelligently identifies the sensitivity of layer content and increases permissions as needed. This automated permission adjustment not only improves security but also ensures a smooth user experience. Users can access the appropriate content on demand without losing access to critical information due to misuse or incorrect permissions.

[0091] The use of a sensitive word library provides a dynamic and flexible mechanism for privacy protection. The system can update and identify new sensitive information in real time and adjust the permissions for displayed content accordingly. This mechanism allows the system to better adapt to changing privacy protection needs, ensuring that sensitive data is protected without affecting the normal user experience.

[0092] By selecting the highest permission level in a layer as the target permission level, rather than assigning permissions individually for each keyword, you can effectively avoid inconsistent permissions across multiple keywords and reduce the risk of false triggering or incorrect permission assignments. For example, some sensitive information may be described by multiple words. Using the highest permission level ensures comprehensive protection for the layer and avoids missing sensitive information.

[0093] In some embodiments of the present application, when determining whether to re-divide the layers based on the position coordinates, the process includes:

[0094] Calculate the straight-line distance from each position coordinate to the same edge of the current layer and the distance between the edge and the opposite edge. When max(di)≤D / 3 is satisfied, the layer is re-divided.

[0095] Where di represents the straight-line distance from the position coordinate to edge i in the current layer, di={x1, x2, x3, ..., xn}, i is a positive integer; the maximum value of i is the total number of edges in the current layer, xn represents the straight-line distance from the nth position coordinate to the same edge of the current layer, max(di) represents the maximum value in the set di, and D represents the interval distance.

[0096] It's important to note that by calculating the distance from each location coordinate to the layer edge and comparing it to the edge separation distance, we can intelligently determine whether the layer boundary needs to be adjusted. This method ensures more accurate layer division, especially when there are subtle differences in content or critical locations.

[0097] This method dynamically adjusts the layer structure based on the distance between the position coordinates and the layer edge, allowing the system to adapt to changing content distribution. This enhances the system's adaptability to complex interfaces and avoids the improper layer division caused by fixed boundaries.

[0098] When the location coordinates are close to the edge of the layer, redividing the layers can prevent content that should have been assigned to other layers from being mistakenly included in the current layer, thereby ensuring that the layer content division is more in line with actual needs and avoiding information omission or incorrect display.

[0099] By setting the distance threshold D / 3, the system can flexibly adjust the sensitivity of layer division. When the distance is small and meets the conditions, it will re-divide, avoiding unnecessary divisions and effectively improving the accuracy of layer division. This allows the system to process content more efficiently, thereby enhancing the overall system's responsiveness and effectiveness.

[0100] By accurately categorizing layers, we can more accurately assign appropriate permission levels to each layer, ensuring that the content in each area is protected according to its actual sensitivity. This precise categorization also helps optimize the dynamic adjustment of user permissions and improves the system's adaptability to different user permissions.

[0101] By re-dividing the layers and ensuring that each area has appropriate permissions, we avoid confusion and leakage during the display process. After each area is properly divided, different permissions can be managed according to its sensitivity, ensuring the security and consistency of information.

[0102] It should be noted that by selecting the smallest maximum value among multiple edge distances, sensitive area division is achieved at the minimum cost, ensuring that sensitive keywords are concentrated in the newly generated sub-layer, facilitating subsequent permission and fuzzy processing. Prioritizing the direction with the lowest division cost for cutting avoids unnecessary complex layer reconstruction and improves the system's real-time processing performance. Through refined redivision methods, sensitive information is confined to the minimum necessary area, effectively retaining more original clear content and improving the viewing user experience. This method can automatically determine the optimal segmentation boundary based on the distribution of sensitive words, enhance the intelligence and content adaptability of layer division, and is suitable for a variety of complex sharing scenarios.

[0103] In some embodiments of the present application, when calculating the trustworthiness of the screen gaze point based on the clarity of the user image, the following steps are included:

[0104] The user image is converted into a grayscale image, and the clarity satisfies the following relationship:

[0105] ;

[0106] The clarity threshold is set to 100. When the clarity is less than or equal to 100, it is judged as low confidence. When the clarity is greater than 100, the confidence is calculated according to the normalization method. The confidence satisfies the following relationship:

[0107] ;

[0108] in, For clarity, is the image resolution, which represents the product of the image width and height, In the grayscale image The pixel value at is the Laplace value, is the average of all Laplace values; For trust, is the clarity threshold, For maximum clarity.

[0109] It should be noted that calculating image clarity using a grayscale image and Laplacian operator effectively identifies image blur and indirectly reflects whether the user's gaze is truly focused on the screen, helping to filter out false gaze data caused by head tilt, blinking, or camera occlusion. The introduction of a clarity threshold and normalized scoring mechanism shifts from a simple binary judgment of gaze validity to assigning trust based on the continuity of image quality, making subsequent permission adjustments smoother and more resilient. This method relies solely on image grayscale and edge sharpness, with low requirements for device conditions such as camera resolution. It is highly adaptable and operates stably across different devices and network environments. Linking user image clarity to gaze point trust provides a quantitative basis for permission adjustments, improving the rationality and security of the system's decisions regarding "fuzzy control" and "sensitive area access."

[0110] In some embodiments of the present application, the user permission level is adjusted based on the relationship between the screen gaze point and each layer to obtain modified user permissions, including:

[0111] Get the layer where the screen gaze point is located and calculate the dwell time on the layer. When the preset time threshold is met, determine whether the layer contains sensitive keywords. If not, replace the user permission level to modify the user permission. If it does, perform the following steps:

[0112] Obtain the screen gaze points of the remaining viewing terminals and record them as reference gaze points. Count the number of reference gaze points in the layer in which they are located and calculate the attention degree of the layer in which they are located. The attention degree is the ratio of the number of reference gaze points in the layer in which they are located to the total number of viewing terminals. When the ratio is less than the attention threshold, lower the user authority level of the viewing terminal corresponding to the screen gaze point by one level and record it as the revised user authority. When the ratio is greater than or equal to the attention threshold, use the user authority level as the revised user authority.

[0113] It should be noted that the optimal range for the attention threshold can be determined through repeated experimentation. By monitoring the duration of a user's gaze on a specific area (layer), permission checks are triggered only when the duration meets a predetermined threshold. This avoids misjudgments due to momentary, accidental gazes and ensures real-time and stable permission adjustments. When a layer contains sensitive keywords, the system further obtains gaze data from other viewing users, counts reference gaze points, and calculates the attention level. This mechanism ensures that a user's permissions are only downgraded by one level when attention is unusually concentrated in a particular area and low, thereby achieving stricter control over sensitive content. By obtaining reference gaze points from other viewing users and using the ratio of reference gaze points to the total number of viewers as an attention level indicator, the system can leverage group behavior to determine whether individual users are engaging in unusual attempts to access highly sensitive information and dynamically adjust their permissions. This approach takes into account both individual actions and the overall security strategy. When a layer does not contain sensitive keywords, the user's permissions are directly replaced with modified permissions, avoiding excessive obfuscation of irrelevant content and improving the overall readability and user experience of screen content sharing. This method forms a closed-loop feedback mechanism by collecting and analyzing user gaze behavior data in real time: the user's gaze behavior affects the permission adjustment, and the permission adjustment further affects the blurred presentation of screen content, ensuring that the system can adapt to changing user behavior and scenario requirements, and provide continuous and effective security protection.

[0114] In some embodiments of the present application, when the user authority level is adjusted according to the screen gaze point of the screen sharing terminal and the viewing terminal to obtain the final user authority, it includes:

[0115] When the screen gaze points of the screen sharing end and the viewing end are in the same layer, the modified user permissions of the viewing end are used as the final user permissions;

[0116] When the screen gaze points of the screen sharing end and the viewing end are not in the same layer, the dwell time is calculated. When the time threshold is met, the modified user is downgraded by one level to obtain the final user authority.

[0117] In addition, when the user permission level changes, the blur level of the layer is adjusted to the highest.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for processing screen content dynamic fuzzy based on user authority verification, characterized in that: include: Collecting the original screen content of the screen sharing terminal, dividing the original screen content into a plurality of regions, inputting each of the regions into a different layer, and assigning a target permission level to the layer according to the original screen content in each region; Obtaining the user authority level of each user on the viewing end, comparing the user authority level with the target authority level of each layer, and performing fuzzy processing on the layers whose user authority level is lower than the target authority level; Overlaying the blurred layer with the remaining unblurred layer to form blurred screen content, and pushing the blurred screen content; Collecting user images of the viewing end after receiving the blurred screen content in a time series, determining the screen gaze point of the viewing end user based on eye tracking technology; calculating the trustworthiness of the screen gaze point based on the clarity of the user image; Adjusting the user authority level according to the relationship between the screen gaze point and each of the layers to obtain a modified user authority; Collecting a user portrait of the screen sharing terminal according to a time series, determining the screen gaze point based on the user image, and adjusting the user authority level based on the screen gaze points of the screen sharing terminal and the viewing terminal to obtain the final user authority; Pushing subsequent blurred screen content based on the end user's permissions and the target permission level; When calculating the trustworthiness of the screen gaze point according to the clarity of the user image, the method includes: The user image is converted into a grayscale image, and the clarity satisfies the following relationship: ； The clarity threshold is set to 100. When the clarity is less than or equal to 100, it is judged as low confidence. When the clarity is greater than 100, the confidence is calculated according to the normalization method. The confidence satisfies the following relationship: ； in, For clarity, is the image resolution, which represents the product of the image width and height, In the grayscale image The pixel value at is the Laplace value, is the average of all Laplace values; For trust, is the clarity threshold, For maximum clarity.

2. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 1, characterized in that: When the original screen content is recognized and divided into several areas, including: The window borders in the original screen content are obtained based on Canny edge detection, the area enclosed by all the window borders is calculated, the window border with the smallest area is selected and input into a layer, and the content of the window border is removed from the original screen content. The window border with the smallest area is selected again and input into another layer, and the selection is repeated until the original screen content is completely divided.

3. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 2, characterized in that: When assigning a target permission level to the layer according to the original screen content within each of the regions, including: Establish a sensitive word library, which includes sensitive keywords and corresponding viewing permission levels; Performing OCR text recognition and application interface feature recognition on the original screen content in each layer to determine whether the layer contains sensitive keywords or a preset specific application program interface; When it is determined that the preset feature application interface is included, the maximum target permission level is assigned to the current layer; when it is determined that the sensitive keywords and the preset specific application interface are not included, the minimum target permission level is assigned to the current layer; When it is determined that only sensitive keywords are included, the position coordinates of the sensitive keywords in the current layer are obtained, and it is determined whether to re-divide the layer based on the position coordinates.

4. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 3, characterized in that: When there is no need to redivide the layers or the layers have been redivided, the viewing permission level corresponding to each sensitive keyword is obtained according to the sensitive word library, the highest viewing permission level is selected, and the highest viewing permission level is used as the target permission level of the current layer.

5. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 4, characterized in that: When determining whether to re-divide the layers according to the position coordinates, the method includes: Calculate the straight-line distance from each position coordinate to the same edge of the current layer and the interval distance from the edge to the opposite edge respectively, and re-divide the layer when max(di)≤D / 3 is satisfied; Wherein, di represents the straight-line distance from the position coordinate to edge i in the current layer, di={x1, x2, x3, ..., xn}, i is a positive integer; the maximum value of i is the total number of edges in the current layer, xn represents the straight-line distance from the nth position coordinate to the same edge of the current layer, max(di) represents the maximum value in the set di, and D represents the interval distance.

6. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 5, characterized in that: When the layers are re-divided, it includes: When multiple max(di)s are less than or equal to D / 3, the smallest max(di) is selected and the current layer is split into two sub-layers, one of which contains all sensitive keywords.

7. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 1, characterized in that: Adjusting the user authority level according to the relationship between the screen gaze point and each of the layers to obtain a modified user authority includes: Obtain the layer where the screen gaze point is located, and calculate the dwell time on the layer. When a preset time threshold is met, determine whether the layer contains sensitive keywords; if not, replace the modified user authority with the user authority level; if yes, perform the following steps: Obtain the screen gaze points of the remaining viewing terminals and record them as reference gaze points. Count the number of the reference gaze points in the layer and calculate the attention degree of the layer, where the attention degree is the ratio of the number of reference gaze points in the layer to the total number of viewing terminals. When the ratio is less than the attention threshold, lower the user authority level of the viewing terminal corresponding to the screen gaze point by one level and record it as the modified user authority. When the ratio is greater than or equal to the attention threshold, use the user authority level as the modified user authority.

8. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 7, characterized in that: When the user authority level is adjusted according to the screen gaze points of the screen sharing terminal and the viewing terminal to obtain the final user authority, it includes: When the screen gaze points of the screen sharing end and the viewing end are in the same layer, the modified user permissions of the viewing end are used as the final user permissions; When the screen gaze points of the screen sharing end and the viewing end are not in the same layer, the dwell time is calculated. When the time threshold is met, the modified user is downgraded by one level to obtain the final user authority.

9. The method for dynamic fuzzy processing of screen content based on user authority verification according to claim 8, characterized in that: Also includes: When the user permission level changes, the blur level of the layer is adjusted to the highest level.

Images