Meeting document access control methods, equipment and media

Abstract

This application provides a method, device, and medium for controlling access to meeting documents, relating to the field of paperless conferencing technology. The method includes: decomposing the meeting document into a content fragment tree; establishing access permission profiles for participants, including each participant's basic access level, temporary access, and behavioral characteristics; responding to a preview request for the meeting document; generating a dynamic watermark based on the preview request; and rewriting the meeting document based on the content fragment tree and access permission profiles to generate a rewritten document, the dynamic watermark including information about the participant corresponding to the preview request; and sending the target document, incorporating the dynamic watermark and the rewritten document, to the participant corresponding to the preview request. This application's embodiments enable fine-grained access control for meeting documents, generating corresponding rewritten documents for participants with different access levels, and recording participant information for the preview request through dynamic watermarking, thus meeting the needs for refined access control and improving the security of meeting document usage.

Description

Technical Field

[0001] This application relates to the field of paperless conferencing technology, and more specifically, to a method, device, and medium for controlling access to meeting documents. Background Technology

[0002] Paperless meeting systems replace traditional paper materials with electronic documents. This not only reduces meeting costs but also improves the efficiency of distributing and organizing meeting materials. Therefore, it has been widely used in corporate meetings, government meetings, and other scenarios.

[0003] However, in actual use, meeting documents often contain content with different security levels, and attendees with different identities have different access permissions. Existing access control technologies mostly set overall access permissions for the entire document, which cannot achieve granular access control for different content segments within the document. Once a document is leaked or unauthorized access occurs, it is difficult to trace the source of the leak, and it cannot meet the needs of fine-grained access control. Summary of the Invention

[0004] The purpose of this application is to address the shortcomings of existing technologies by providing a method, device, and medium for controlling access to meeting documents, which can solve the problems of existing technologies being unable to achieve granular access control of document content and being unable to trace the source of leakage.

[0005] To achieve the above objectives, this application adopts the following technical solution: This application provides a method for controlling access to meeting documents, the method comprising: The meeting document is broken down into a content fragment tree, and a permission profile for each participant is established. The permission profile includes each participant's basic permission level, temporary permissions, and behavioral characteristics. In response to the preview request of the meeting document, a dynamic watermark is generated based on the preview request, and the meeting document is rewritten based on the content fragment tree and the permission file to generate a rewritten document. The dynamic watermark includes information about the participants corresponding to the preview request. The target document, which integrates the dynamic watermark and the rewritten document, is sent to the participant corresponding to the preview request.

[0006] In one possible implementation, the step of decomposing the meeting document into a content fragment tree includes: The system receives uploaded meeting documents, calls a document parser to perform structured parsing on the meeting documents, and obtains the content fragment tree. The skeleton of the content fragment tree is formed by the document logical structure of the meeting documents, and the content units of the meeting documents constitute the leaf nodes of the content fragment tree, with each leaf node associated with an access control tag.

[0007] In one possible implementation, establishing the participant permission profile includes: The basic permission level is determined based on the role type of the participant, the temporary permissions granted to the participant during the meeting are recorded using a temporary permission list, and the behavioral characteristics of the participant are recorded during the meeting. The role vector of the participant is generated based on the basic permission registration, the temporary permissions, and the behavioral characteristics.

[0008] In one possible implementation, generating the dynamic watermark based on the preview request includes: Obtain the participant information corresponding to the preview request. The participant information includes at least one of the following: login username, user ID, terminal IP address, request timestamp, and unique identifier of this session. Generate an encoding of the participant information, and map the encoding onto a preset matrix to generate a first matrix; The first matrix is ​​randomly distributed, and a first watermark is generated based on the randomly distributed first matrix.

[0009] In one possible implementation, the generation of the target document includes: Align the first watermark with the rewritten document; The first watermark and the rewritten document are merged based on the color channels.

[0010] In one possible implementation, the generation of the target document further includes: Obtain a fused image that combines the first watermark and the rewritten document, and divide the fused image into multiple image blocks; Modify the mid-frequency coefficients of the DCT coefficient matrix of the image patch, and generate the image of the target document based on the modified DCT coefficient matrix.

[0011] In one possible implementation, rewriting the meeting document based on the content fragment tree and the permission file includes: Based on the permission profile, determine whether the meeting file / document needs to be rewritten. If so, then traverse the leaf nodes of the content fragment tree and rewrite the meeting document based on the comparison result between the permission flag of the leaf node and the permission file.

[0012] In one possible implementation, sending the target document, which integrates the dynamic watermark and the rewritten document, to the participant corresponding to the preview request includes: The target document is assembled and rendered in memory, and a preset response header is set in the target document, which is used to instruct the terminal's browser to open the target document in preview mode; The target document is sent to the participant's terminal, and the target document is bound to the participant's current session; Once the sending is complete, release the target document from memory.

[0013] In another aspect, this application provides an electronic device including a processor and a memory, wherein the processor is communicatively connected to the memory, the memory stores program data, and the program data is executed by the processor to implement the method described in any of the above embodiments.

[0014] In another aspect, this application provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the methods described in any of the above embodiments.

[0015] The beneficial effects of the technical solutions provided in this application are: The meeting document access control method provided in this application includes: decomposing the meeting document into a content fragment tree, establishing access permission files for participants, including each participant's basic access level, temporary access, and behavioral characteristics; responding to a preview request for the meeting document, generating a dynamic watermark based on the preview request, and rewriting the meeting document based on the content fragment tree and access permission files to generate a rewritten document, with the dynamic watermark including information about the participant corresponding to the preview request; and sending the target document, which integrates the dynamic watermark and the rewritten document, to the participant corresponding to the preview request. This application's embodiments enable fine-grained access control for meeting documents, generating corresponding rewritten documents for participants with different access levels, only granting access to content fragments within their access scope, and recording the participant information for the preview request through dynamic watermarking. In the event of document leakage, the source of the leakage can be traced by extracting the watermark, meeting the needs of refined access control, improving the security of meeting document usage, and solving the problems of coarse-grained overall document control and difficulty in tracing leaks in existing technologies. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.

[0017] Figure 1 A flowchart illustrating the meeting document access control method provided in this application embodiment; Figure 2 A flowchart illustrating the workflow for meeting document access control provided in this application embodiment; Figure 3A structural diagram of the electronic device provided in the embodiments of this application; Figure 4 This is a structural diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0018] The embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions of the embodiments of this application.

[0019] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the terms “comprising” and “including” as used in embodiments of this application mean that the corresponding feature can be implemented as the presented feature, information, data, step, operation, element, and / or component, but do not exclude implementation as other features, information, data, step, operation, element, component, and / or combinations thereof supported by the art. It should be understood that when we say that an element is “connected” or “coupled” to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element establish a connection relationship through an intermediate element. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The term “and / or” as used herein indicates at least one of the items defined by the term; for example, “A and / or B” indicates implementation as “A,” or implementation as “A,” or implementation as “A and B.”

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0021] The technical solutions of this application and their effects are described below through several exemplary embodiments. It should be noted that the following embodiments can be referenced, borrowed from, or combined with each other. Identical terms, similar features, and similar implementation steps in different embodiments will not be repeated.

[0022] The meeting document access control method, device and medium provided in this application are intended to solve at least one technical problem existing in the prior art.

[0023] Optionally, the meeting document access control method of this application can be used on computers, mobile phones, servers, cloud platforms, and other terminals capable of running intelligent agents to perform business operations.

[0024] Optionally, such as Figure 1 , Figure 2As shown, the meeting document access control method of this application includes: S101: Decompose the meeting document into a content fragment tree and establish a permission profile for each participant.

[0025] Optionally, the meeting document is decomposed into a content fragment tree, including: receiving the uploaded meeting document, calling a document parser to perform structured parsing of the meeting document to obtain the content fragment tree. The skeleton of the content fragment tree is formed by the document logical structure of the meeting document, and the content units of the meeting document constitute the leaf nodes of the content fragment tree, with each leaf node associated with an access permission tag.

[0026] In one embodiment, the object executing the method of this application is a server or a server-side management system for a conference. Let's describe this object as a server. The conference document received by the server can be a document uploaded by the conference host or other participants, or a document retrieved by the server from a storage address according to received instructions. After receiving the conference document, the server calls a document parser to perform structured parsing of the document, breaking it down into a content fragment tree. This content fragment tree uses the document's logical structure (chapter, section, paragraph, table, image) as its skeleton, and indivisible content units (paragraph, table cell, image object) as leaf nodes. Each leaf node is associated with an access control flag, which includes a sensitivity level field (e.g., public, internal, confidential, top secret) and a policy identifier field. The document parser can be an existing document parsing model or other tools capable of parsing documents and obtaining their logical structure.

[0027] Optionally, permission tags can be automatically generated based on a preset sensitive word library and semantic analysis model (the semantic analysis model can match the sensitive word library with the content of the leaf nodes to obtain the matching sensitive words, and set the permission tags according to the permissions corresponding to the sensitive words), or they can be set by the document uploader. The document uploader can manually adjust the permission tags of each leaf node through a visual interface. The permission tags can include the permission level corresponding to different leaf nodes, the personnel who can view them, and other content used for viewing permissions.

[0028] Optionally, the permission profile includes each participant's basic permission level, temporary permissions, and behavioral characteristics. The behavioral characteristics represent the participant's actions during the meeting. The basic permission level, temporary permissions, and behavioral characteristics can be aggregated into a unified multi-dimensional vector, which is used to determine the participant's permissions. Specifically, the structure of this multi-dimensional vector is defined as follows: R = { "base_level": 2, / / Basic permission level, value 0-4, derived from step (1) "temp_permissions": [ / / List of temporary permissions, derived from step (2)] { "fragment_id": "frag_023", / / Accessible content fragment ID "expire_time": "14:30" / / Expiration time; the time after which it will automatically expire. }, { "fragment_id": "frag_089", "expire_time": "15:00" } ], "behavior_flags": { / / Behavioral characteristic flags, derived from step (3) "has_annotated_secret": true, / / Whether annotations have been added to confidential content. "high_focus_chapter": "ch_03", / / High-focus chapter "request_count": 3 / / Total number of requests } } Optionally, an access permission profile is established for each participant, including: determining the basic access level based on the participant's role type; using a temporary access permission list to record the temporary access permissions granted to participants during the meeting; and recording the participants' behavioral characteristics during the meeting; and generating a role vector for each participant based on the basic access permission registration, temporary access permissions, and behavioral characteristics.

[0029] In one embodiment, a basic permission level can be assigned to each participant based on their role type. This basic permission level can be represented by an integer, with higher levels granting greater permissions. The mapping relationship between the basic permission level and role type is shown in Table 1. Table 1

[0030] Optionally, participants can synchronously obtain their basic permission level from an object storing their role type (such as an enterprise directory service (e.g., LDAP)) when they log into the meeting system. This basic permission level is generally not changed automatically during the meeting (the host can manually adjust it).

[0031] Optionally, temporary permissions for participants can be acquired during the meeting and recorded in a temporary permission list. This temporary permission list can store the authorization information acquired by participants during the meeting, which can be stored in the form of fields. Specifically, the fields in the temporary permission list are shown in Table 2: Table 2

[0032] In one embodiment, the application scenario for temporary permissions can be: When a department head enters the stage of an issue they are responsible for, they are automatically granted temporary access to view the financial data for that issue, which is automatically revoked after the issue ends. The moderator temporarily sets a regular participant as an "issue presenter". After the setting is completed, the participant is automatically granted document annotation permissions. Once the participant is no longer an "issue presenter", the document annotation permissions are revoked.

[0033] Optionally, during the meeting, participants' actions can be recorded. These actions are used to determine whether automatic privilege escalation is necessary. The recorded actions are shown in Table 3. Table 3

[0034] In one embodiment, corresponding triggering rules can be set based on the participants' actions. Specifically, when a user's (participant's) actions on a specific content segment exceed a threshold, it is determined that the user has an "actual information need" for the content, and a suggestion to "grant temporary permissions" is automatically pushed to the host, instead of directly granting permissions automatically. This achieves both intelligence and preserves the host's final decision-making power.

[0035] S102: In response to a preview request for the meeting document, generate a dynamic watermark based on the preview request, and rewrite the meeting document based on the content fragment tree and permission profile to generate a rewritten document.

[0036] Optionally, the dynamic watermark includes information about the attendees corresponding to the preview request. To further prevent leakage, a one-time access token can be generated based on the preview request, and this token can be bound to the user session, document identifier, and generation time. Furthermore, it is sent along with the preview URL of the target document when the target document is distributed.

[0037] Alternatively, the token may be invalidated under any of the following conditions: Close the preview page or browser tab corresponding to the target document; Session timeout (default 15 minutes of inactivity); Attendees actively clicked the "End Preview" button; The token is marked as used immediately after being used once (to prevent the URL from being forwarded).

[0038] After the token expires, when an attendee attempts to refresh the page or reopen the link, the server returns a "401 Expired" or other message indicating that the token has expired, and the attendee needs to reinitiate an authorization request or a preview request.

[0039] Optionally, generating a dynamic watermark according to the preview request includes: obtaining attendee information corresponding to the preview request, where the attendee information includes at least one of a login user name, a user ID, a terminal IP address, a request timestamp, and a unique identifier of the current session; generating an encoding of the attendee information, mapping the encoding to a preset matrix to generate a first matrix; performing random scattering processing on the first matrix, and generating a first watermark according to the randomly scattered first matrix. The generation of the dynamic watermark can be implemented by invoking an image processing extension (GD2 or Imagick).

[0040] Optionally, the first watermark may be a translucent watermark image. When generating the encoding of the attendee information, the collected attendee information may be spliced into a string format, for example: "user:Zhang San|id:10086|ip:192.168.1.5|time:14:30:25|session:sess_abc123", then performing Base64 encoding or binary conversion on the string to obtain an original watermark bit sequence, which may be a one-dimensional bit sequence.

[0041] Optionally, the preset matrix may be a two-dimensional matrix, and the original watermark bit sequence is filled into the two-dimensional matrix in a row-major order. The size of the two-dimensional matrix may be dynamically determined according to the page resolution of the meeting document (for example, if the document page is 1920×1080 pixels, the two-dimensional matrix may be set to 240×135, with each matrix unit corresponding to an 8×8 pixel block). Each element in the two-dimensional matrix takes a value of 0 or 1, indicating whether there is a watermark point at that position.

[0042] Optionally, a random distribution manner is adopted to avoid the watermark from showing obvious patterns (such as a grid pattern) so as to reduce the probability of the watermark being removed. A random perturbation algorithm may be used to offset the positions corresponding to the watermark points in the first matrix. The specific method may be: adding random offsets (Δx, Δy) to the center coordinates (x, y) of each matrix unit in the preset matrix, where Δx and Δy may follow a uniform distribution U(-3, 3) pixels, so that the final positions of the watermark points are irregularly scattered.

[0043] Optionally, when generating the first watermark, each watermark dot can be rendered as a semi-transparent dot or short line segment (raster) or stripes arranged in a certain direction (striped). The size and transparency of the dots are adjusted using gradient parameters, such as 0.3 for the center and 0.6 for the edge, to further enhance the integration with the background content.

[0044] Optionally, rewriting the meeting document based on the content fragment tree and permission profile includes: detecting whether the meeting document needs to be rewritten based on the permission profile; if so, traversing the leaf nodes of the content fragment tree and rewriting the meeting document based on the comparison result between the permission tags of the leaf nodes and the permission profile; if not, no rewriting is performed. When it is determined that rewriting is needed based on the permission profile, rewriting can be performed directly without traversing the leaf nodes, or rewriting can be performed after traversing the leaf nodes and based on the traversal results. The specific rewriting method can be determined according to actual needs.

[0045] Optionally, when determining whether rewriting is necessary, a three-tiered progressive logic can be adopted: first, consider temporary permissions (highest priority), then basic permissions (policy matrix), and finally, decide whether to recommend dynamic adjustment (rewriting) based on behavioral characteristics.

[0046] Optionally, when performing temporary permission checks, the decision engine can be used to check the temp_permissions (temporary permission list) in the multidimensional role vector R to determine whether the content fragment ID accessed by the current preview request is in the list and within its validity period: If it exists and has not expired, it will be directly judged as "allowed" without rewriting the meeting document and skipping the subsequent basic permission check; If the rule does not exist or has expired, proceed to the next rule matching layer to determine the basic permission level and behavioral characteristics.

[0047] In the rule matching layer, a comparison and judgment of basic permission levels and sensitivity levels (levels of meeting documents) can be performed based on a predefined policy matrix. An example of this policy matrix is ​​shown in Table 4: Table 4

[0048] The output judgment result is based on the strategy matrix, and this judgment result can be any one of allow, deny, or anonymize. Specifically, if the judgment result is allow, no modification is made, and the meeting document is sent to participants after adding a watermark. If the judgment result is deny, no modification is made, but a message indicating insufficient permissions can be sent to participants. If the judgment result is anonymize, the document can be modified according to the participants' permission profiles.

[0049] Optionally, if the judgment result indicates desensitization, the semantic understanding module can be invoked to rewrite the original content. This module is based on a generative language model for rewriting. The input to this module is a multi-dimensional vector corresponding to the original meeting document and the user's permission profile, and the output is a semantically equivalent but less sensitive representation.

[0050] In one embodiment, the attendees are department heads, and an example of rewriting the content of the meeting document is shown in Table 5: Table 5

[0051] Optionally, after rewriting and sending the meeting document to participants, permissions can be adjusted based on participants' behavioral characteristics during the meeting. A new meeting document can then be generated and sent to the participants based on the adjusted results. Specifically, behavioral characteristics can be represented as a behavioral feature vector. When a participant's behavioral feature vector r_dynamic exceeds a preset threshold (e.g., annotating the same confidential section ≥ 2 times consecutively, or staying for ≥ 2 minutes), permissions are not automatically changed; instead, permissions are adjusted based on the meeting host's actions. A relevant example could be: Send a suggested message to the host: "User Zhang San has shown high interest in the [Confidential Data] section. Should temporary viewing permission be granted?" Once the host confirms, a permission adjustment record is automatically added to temp_permissions (including the segment ID and validity period corresponding to the adjusted permission). Users can see the meeting document content corresponding to the new permission without refreshing.

[0052] Optionally, after determining that de-identification processing is required based on the permission file, the content fragment tree can be traversed, and de-identification processing can be performed based on the traversal results. When traversing the content fragment tree, a depth-first traversal algorithm can be used, visiting each leaf node of the content fragment tree layer by layer starting from the root node. For each node visited, the permission determination engine (the three-layer judgment logic of the aforementioned permission file) is invoked to obtain the judgment result for that node.

[0053] Optionally, the judgment result can include any of the following: allow, reject, or desensitize. Different judgment results will result in different operations. The specific operations are shown in Table 6: Table 6

[0054] Optionally, for table-type content segments in meeting documents, more granular control than that for paragraphs can be applied. The relevant control methods include: Row-level authorization: Dynamically decide whether to show or hide certain rows based on the role type of the participants. For example, in a budget table, the "Total Revenue" row is visible to all employees, while the "Executive Salaries" row is only visible to decision-makers.

[0055] Column-level authorization: Dynamically decide whether to show or hide certain columns based on the role type of the participants. For example, in the same financial table, the "Account Name" column is visible to all participants, while the "Budget Amount" column is only visible to senior executives.

[0056] Specifically, when traversing table nodes, the permission determination engine is called to determine permissions for each row and column, dynamically constructing a "filtered" table structure.

[0057] Optionally, after the meeting document is rewritten, the processed content fragment tree can be reassembled into a standard document format, supporting the following output formats: PDF: Suitable for formal preview and printing, with the best watermark blending effect; HTML: Suitable for real-time rendering by browsers, supporting dynamic refresh and partial updates; OFD: Compliant with domestic electronic document standards.

[0058] The restructured meeting document view has been cropped and anonymized according to the participants' permissions, and can be directly accessed for watermark fusion and output.

[0059] S103: Send the target document with integrated dynamic watermark and rewritten document to the participant corresponding to the preview request.

[0060] Optionally, the generation of the target document includes: aligning the first watermark and rewriting the document; and merging the first watermark and the rewritten document according to the color channels. Specifically, when the meeting document does not require rewriting, it is directly designated as the rewritten document, and the document is aligned with the first watermark.

[0061] Optionally, this process can be performed after the rewritten document is generated but before the target document is rendered into the final output format (PDF / HTML / OFD), ensuring that the watermark becomes an integral part of the document content, rather than a separate layer added later. Each page of the document independently incorporates its corresponding watermark image. The user identification information (username, ID, IP, timestamp, session identifier) ​​in the watermark image is dynamically generated with each request. Different users and different times requesting the same document will see different watermark information, achieving "one code per person, one code per time".

[0062] Optionally, the image of the first watermark can be merged with the image of the rewritten document (which can be a bitmap rendered from each page of the meeting document in PDF format or a directly uploaded image) to achieve pixel-level overlay and ensure that each output document carries the unique identifier of the current user.

[0063] Optionally, the images of the first watermark and the rewritten document can be aligned to ensure that the images of the first watermark and the rewritten document have the same size and resolution (if they are not the same, the images of the first watermark are scaled and centered first).

[0064] Optionally, when fusing based on color channels, the following fusion algorithm is performed on the R (red), G (green), and B (blue) color channels for each pixel position in the image: P_out(channel)=α×P_watermark(channel)+(1-α)×P_original(channel) In the formula, P_out(channel) is the pixel value (0-255 integer) of the fused color channel, P_watermark(channel) is the pixel value of the watermark image in that channel, P_original(channel) is the pixel value of the corresponding channel in the rewritten document image, and α is the pixel-level fusion coefficient, which can range from 0 < α ≤ 0.3. Through this method, the information of the first watermark is simultaneously integrated into the R, G, and B color channels, rather than being superimposed on a separate "watermark layer." Traditional PDF watermarks often exist as an "attached layer," which malicious users can directly delete using PDF editing tools (such as Adobe Acrobat). This application uses independent mixing of the RGB channels, making the watermark information an organic component of the image content itself, eliminating the need for a separate, removable layer. Furthermore, the watermark covers the entire page background in a randomly distributed dot matrix pattern, rather than being concentrated in the header, footer, or edges. Even if a malicious user crops off 20% of the document's perimeter, the remaining portion still contains a sufficient number of watermark dots (evenly distributed, with approximately 80% retained). Similarly, if a user obscures an area with a black rectangle, the watermark dots in the unobscured areas can still be used to extract source information. The merged document image visually maintains the original content's readability, while the watermark cannot be removed using conventional methods such as layer deletion, edge cropping, or partial obscuration.

[0065] Optionally, the selection strategy for the pixel fusion coefficient can be: the larger the α value, the more obvious the watermark, but the more severely it obscures the original content; the smaller the α value, the more concealed the watermark, but it may be ignored in extreme cases; in this application, α can be 0.15, which achieves the best balance between the existence of the watermark that can be recognized by the naked eye and not affecting the normal reading of the document.

[0066] Optionally, to further enhance the anti-attack capability of the target document, the generation of the target document also includes: obtaining a fused image of the document with the first watermark and rewriting it; dividing the fused image into multiple image blocks; modifying the intermediate frequency coefficients of the DCT coefficient matrix of the image blocks; and generating the image of the target document based on the modified DCT coefficient matrix. This method of embedding the frequency domain watermark forms a time-domain + frequency-domain dual watermarking mechanism, enhancing the robustness of the watermark (resisting screenshot compression, format conversion, smearing attacks, etc.).

[0067] In one embodiment, the fused image can be divided into non-overlapping image blocks of 8×8 pixels (or other sizes); a discrete cosine transform (DCT) is performed on each image block to convert the pixel domain (spatial domain) signal to the frequency domain, resulting in the corresponding DCT coefficient matrix (also 8×8).

[0068] In each 8×8 DCT coefficient matrix, mid-frequency coefficients are selected as embedding locations (e.g., the mid-frequency coefficients corresponding to coordinates (4,5), (4,6), (5,4), and (5,5) in the matrix). The reason for choosing mid-frequency coefficients is that low-frequency coefficients carry the main visual information of the image, and modifying them can easily lead to image distortion; high-frequency coefficients are easily lost during JPEG compression, scaling, and other operations; mid-frequency coefficients achieve the best balance between visual sensitivity and robustness. After determining the embedding locations, the mid-frequency coefficients are modified, and the image is generated based on the modified DCT coefficient matrix to obtain the image of the target document.

[0069] Optionally, the code corresponding to the first watermark (i.e., the watermark bit sequence) can be obtained, and the watermark bit sequence can be embedded into the selected intermediate frequency coefficients in a 1-quantized index modulation manner to modify the DCT coefficient matrix.

[0070] In one embodiment, the embedding rule can be as follows: Let the selected intermediate frequency coefficient be C, and the quantization step size be Δ (default Δ=12); if the bit to be embedded is 1, then C is adjusted to the nearest odd multiple Δ / 2; if the bit to be embedded is 0, then C is adjusted to the nearest even multiple Δ / 2. The modified intermediate frequency coefficient is denoted as C', and C' is used to replace the original intermediate frequency coefficient to obtain a new DCT coefficient matrix.

[0071] Optionally, after modifying the DCT coefficient matrix of all image blocks, an inverse discrete cosine transform (IDCT) is performed on the DCT coefficient matrix to restore the frequency domain signal back to the pixel domain, obtaining the image of the target document. When it is necessary to extract source information from leaked images, even if the temporal watermark has been destroyed (e.g., excessive compression, erasure, cropping), the frequency domain watermark can still be extracted according to the following process: Perform block-based DCT transformation on the leaked image; Read the coefficient values ​​at the same mid-frequency coefficient location; The coefficient is determined to be an odd multiple or an even multiple Δ / 2 based on the quantization step size Δ, and then decoded as 0 or 1 bit. By reconstructing and decoding the bit sequence, traceability information such as the attendees' identities and timestamps can be restored.

[0072] Using the methods described above, time-domain watermarking provides intuitive warnings that are easily identifiable to the naked eye, while frequency-domain watermarking provides covert backups that resist attacks. The two complement each other, and even after the image has been processed by screenshotting, compression, smearing, format conversion, etc., accurate traceability can still be achieved.

[0073] Optionally, the target document with integrated dynamic watermark and rewritten document is sent to the participant corresponding to the preview request, including: assembling and rendering the target document in memory and setting a preset response header in the target document, the preset response header being used to instruct the terminal's browser to open the target document in preview mode; sending the target document to the participant's terminal and binding the target document to the participant's current session; confirming that the sending is complete and releasing the target document in memory.

[0074] In one embodiment, the merged target document is not written to disk; assembly and rendering are completed only in memory. Furthermore, the content of the target document is directly output to the participants' front-end browsers via HTTP response streams in a chunked manner. The document data in memory is released immediately after the target document is sent, without retaining any copies, thus achieving diskless storage of meeting documents (zero data residue).

[0075] Optionally, the default response header can be an HTTP response header, which can be represented as: Content-Disposition: inline; filename="document_preview.pdf" Content-Type: application / pdf The `nline` option instructs the browser to open the PDF in preview mode (with an embedded PDF viewer) instead of displaying a download and save dialog box. While mainstream browsers (Chrome, Edge, Firefox) do not support one-click disabling of the "Save As" button in their PDF preview plugins, disabling the right-click menu and shortcut (Ctrl+S) using front-end JavaScript can significantly reduce the likelihood of attendees saving copies.

[0076] Optionally, the document stream of the target document can be strongly bound to the participant's current user session, as follows: An encrypted signature is embedded in the document content (such as metadata fields in a PDF or hidden fields in an HTML document). The signature includes at least one of the following: user ID, session ID, unique document identifier, and generation timestamp. The signature can be encrypted using HMAC-SHA256 with the server's private key to prevent user tampering. The participant's front-end preview page periodically (e.g., every 30 seconds) sends a heartbeat request to the server to verify that the session is still valid.

[0077] Optionally, during the meeting, three types of trigger events can be continuously monitored. Upon triggering of an event, permission adjustments and view refreshes can be completed without interrupting the meeting or reloading the entire document. Specifically, the types of the three trigger events, their related triggering conditions, and processing actions are shown in Table 7: Table 7

[0078] After updating participant permissions based on the aforementioned events, the temporary permissions list for affected users can be updated, or the policy matrix can be switched. Permission determination can then be re-executed, and the target document updated based on the determination result. Specifically, the old and new views of the target document can be compared, and only the changed content fragments (JSON format, typically a few KB) can be packaged and sent to the participants (via WebSocket push to their terminals) to achieve incremental updates.

[0079] Optionally, to achieve behavior auditing and end-to-end traceability, every document preview operation can be fully recorded, forming a dual traceability mechanism of audit logs and pixel-level watermarks. Each time a user requests to preview a document, the audit log records the following information from Table 8: Table 8

[0080] When tracing back through audit logs (post-event query), if a user actively leaks information (e.g., forwarding screenshots), the audit logs can be used to query "who viewed which content at what time". Furthermore, audit logs are linked to meetings and support retrieval by time, attendee, and document.

[0081] Optionally, when tracing back through watermark information (such as in a photo leak scenario), watermark information can be extracted from the leaked image (even after being photographed, cropped, or compressed, the watermark can still be partially identified). The extracted watermark includes: username, timestamp, and session ID → accurately locating the leaker and the time of the leak.

[0082] Compared with the prior art, this application has the following beneficial effects: 1. Pixel-level watermark fusion, cannot be removed. The watermark information is deeply integrated with the document content at the pixel level, with each RGB channel independently mixed, making it impossible to remove using conventional methods such as cropping, masking, or layer deletion. Frequency domain watermarking further enhances its resistance to attacks; even if the temporal domain watermark is damaged, traceability information can still be extracted.

[0083] 2. Precise traceability, one code per person The watermark information includes personalized information such as the current logged-in username, IP address, and timestamp, achieving "one code per person, one code per time." After a photo leak, the leaker and the time of the leak can be accurately located by extracting the watermark information.

[0084] 3. Atomic-level permission granularity By breaking down documents into paragraphs and cells using a content fragment tree, a differentiated view of "one document, a thousand faces" can be achieved, eliminating the need to create multiple copies of the document and significantly reducing pre-meeting preparation costs and version consistency risks.

[0085] 4. Intelligent desensitization based on semantic awareness The desensitization and rewriting technology based on generative language models can automatically convert sensitive information into equivalent low-sensitivity expressions according to the role, achieving a balance between the "need to know" principle and full information sharing.

[0086] 5. Dynamic strategy with real-time response A multi-dimensional role model and an event-driven strategy adjustment mechanism support flexible changes in permissions during meetings. User behavior and permission policies form a closed-loop feedback loop, improving meeting collaboration efficiency.

[0087] 6. Zero data residue Documents are generated and output in real time, without being stored on disk; a one-time access token mechanism ensures that documents cannot be accessed again after the preview is closed. No file copies are left on the terminal, fundamentally eliminating the risk of data residue.

[0088] 7. Complete auditability Behavioral auditing and watermark tracing provide dual protection, meeting compliance requirements such as Cybersecurity Classified Protection 2.0 and GDPR, and providing complete technical support for post-event tracing and liability determination.

[0089] Based on the same inventive concept, this application also proposes an electronic device, please refer to [link to relevant documentation]. Figure 3 , Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. (In conjunction with...) Figure 3 The electronic device described in this application is described in detail.

[0090] In this embodiment, the smart terminal includes a processor and a memory. The processor and the memory are communicatively connected. The memory stores program data, which is used to execute the method described in the above embodiment.

[0091] In this embodiment, the processor is the control center of the smart terminal. It connects various parts of the smart terminal via various interfaces and lines, and performs various functions and processes data by running or executing software programs and / or modules stored in the memory, and by calling data stored in the memory, thereby providing overall monitoring of the smart terminal. Optionally, the processor may include one or more processing units; preferably, the processor may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into the processor.

[0092] The smart terminal also includes a power supply (such as a battery) to power various components. Preferably, the power supply can be connected to the processor logic through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system.

[0093] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing of the smart terminal by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area can store data created according to the use of the smart terminal (such as audio data, phone book, etc.). In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0094] Based on the same inventive concept, this application also proposes a computer-readable storage medium, please refer to [link to relevant documentation]. Figure 4 , Figure 4 This is a structural diagram of an embodiment of the computer-readable storage medium of this application, in conjunction with... Figure 4 The computer-readable storage medium of this application is described.

[0095] In this embodiment, a computer-readable storage medium stores a computer program that is used to perform the methods described in the above embodiments.

[0096] The computer-readable storage medium may include, but is not limited to, floppy disks, optical disks, CD-ROMs (compact disc-read-only memory), magneto-optical disks, ROMs (read-only memory), RAMs (random access memory), EPROMs (erasable programmable read-only memory), EEPROMs (electrically erasable programmable read-only memory), magnetic cards or optical cards, flash memory, or other types of media / machine-readable media suitable for storing machine-executable natural language instructions. The computer-readable storage medium may be a product not connected to a computer device or a component used in a computer device.

[0097] The terms "first," "second," "third," "fourth," "1," "2," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that shown in the illustrations or text descriptions.

[0098] It should be understood that although arrows indicate various operation steps in the flowcharts of this application's embodiments, the order in which these steps are implemented is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of this application's embodiments, the implementation steps in each flowchart can be executed in other orders as required. Furthermore, some or all steps in each flowchart, based on the actual implementation scenario, may include multiple sub-steps or multiple stages. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage can also be executed at different times. In scenarios where execution times differ, the execution order of these sub-steps or stages can be flexibly configured according to requirements, and this application's embodiments do not limit this.

[0099] The above description is only an optional implementation method for some implementation scenarios of this application. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this application without departing from the technical concept of this application also fall within the protection scope of the embodiments of this application.

Claims

1. A method for controlling access to meeting documents, characterized in that, The method includes: The meeting document is broken down into a content fragment tree, and a permission profile for each participant is established. The permission profile includes each participant's basic permission level, temporary permissions, and behavioral characteristics. In response to the preview request of the meeting document, a dynamic watermark is generated based on the preview request, and the meeting document is rewritten based on the content fragment tree and the permission file to generate a rewritten document. The dynamic watermark includes information about the participants corresponding to the preview request. The target document, which integrates the dynamic watermark and the rewritten document, is sent to the participant corresponding to the preview request.

2. The meeting document access control method according to claim 1, characterized in that, The process of breaking down meeting documents into content fragment trees includes: The system receives uploaded meeting documents, calls a document parser to perform structured parsing on the meeting documents, and obtains the content fragment tree. The skeleton of the content fragment tree is formed by the document logical structure of the meeting documents, and the content units of the meeting documents constitute the leaf nodes of the content fragment tree, with each leaf node associated with an access control tag.

3. The meeting document access control method according to claim 1, characterized in that, The process of creating participant permission profiles includes: The basic permission level is determined based on the role type of the participant, the temporary permissions granted to the participant during the meeting are recorded using a temporary permission list, and the behavioral characteristics of the participant are recorded during the meeting. The role vector of the participant is generated based on the basic permission registration, the temporary permissions, and the behavioral characteristics.

4. The meeting document access control method according to claim 1, characterized in that, The step of generating a dynamic watermark based on the preview request includes: Obtain the participant information corresponding to the preview request. The participant information includes at least one of the following: login username, user ID, terminal IP address, request timestamp, and unique identifier of this session. Generate an encoding of the participant information, and map the encoding onto a preset matrix to generate a first matrix; The first matrix is ​​randomly distributed, and a first watermark is generated based on the randomly distributed first matrix.

5. The meeting document access control method according to claim 4, characterized in that, The generation of the target document includes: Align the first watermark with the rewritten document; The first watermark and the rewritten document are merged based on the color channels.

6. The meeting document access control method according to claim 5, characterized in that, The generation of the target document also includes: Obtain a fused image that combines the first watermark and the rewritten document, and divide the fused image into multiple image blocks; Modify the mid-frequency coefficients of the DCT coefficient matrix of the image patch, and generate the image of the target document based on the modified DCT coefficient matrix.

7. The meeting document access control method according to claim 2, wherein rewriting the meeting document based on the content fragment tree and the access file includes: Based on the permission profile, determine whether the meeting file / document needs to be rewritten. If so, then traverse the leaf nodes of the content fragment tree and rewrite the meeting document based on the comparison result between the permission flag of the leaf node and the permission file.

8. The meeting document access control method according to claim 1, wherein sending the target document, which integrates the dynamic watermark and the rewritten document, to the participant corresponding to the preview request includes: The target document is assembled and rendered in memory, and a preset response header is set in the target document, which is used to instruct the terminal's browser to open the target document in preview mode; The target document is sent to the participant's terminal, and the target document is bound to the participant's current session; Once the sending is complete, release the target document from memory.

9. An electronic device, characterized in that, The method includes a processor and a memory, the processor being communicatively connected to the memory, the memory storing program data, and the program data being used to execute the method as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is used to perform the method as described in any one of claims 1-8.

Images