Artificial intelligence-based design file semantic online processing and paperless auditing system

Abstract

This invention discloses an online semantic processing and paperless review system for design documents based on artificial intelligence, belonging to the field of semantic processing technology. The system includes core modules such as template management, parsing, online editing, AI proofreading, and review processes. Based on Adobe Illustrator templates, it completes parsing and business field mapping, supports self-service editing by operations, and synchronously and automatically detects material specifications and font consistency. It also links with the enterprise compliance library to achieve AI-powered intelligent proofreading. A rule-based paperless approval workflow is established, enabling full-link log tracking, version snapshot comparison and rollback. A multi-tenant permission module achieves brand data isolation and fine-grained permission control. After approval, multi-channel specification source files can be exported in batches, and hash digest values ​​are generated for tamper-proof verification. This invention improves the efficiency of the entire design document creation and review process, avoids compliance risks, and achieves process standardization and version traceability.

Description

Technical Field

[0001] This invention relates to the field of semantic processing technology, specifically to an online semantic processing and paperless review system for design documents based on artificial intelligence. Background Technology

[0002] The current advertising poster production process typically involves designers creating Adobe Illustrator source files offline, operations staff submitting modification requests via communication tools, designers repeatedly revising and outputting preview images, and finally, after multiple levels of approval, the materials are output for printing or distribution. This process has the following shortcomings: Low iteration efficiency: Frequent changes in operational copy, pricing, timing, channel specifications, QR code links, etc., require designers to repeatedly modify source files, resulting in high communication costs and uncontrollable scheduling.

[0003] The review process is not standardized: preview images are circulated via chat / email, resulting in scattered opinions and a lack of unified process nodes, timeliness control, and clear boundaries of responsibility.

[0004] Version traceability is difficult: There is a lack of structured records and verifiable evidence regarding "what was changed, who made the changes, when the changes were made, and whether the proofreading and approval were passed".

[0005] The disconnect between online editing and offline production: Traditional web page editors struggle to directly reuse the layer structure, font attributes, element coordinates, color modes, and high-resolution output capabilities of AI project files.

[0006] Material specifications and compliance risks: Issues such as substandard material resolution, incorrect aspect ratio, missing fonts, and omissions of sensitive words / compliance clauses often only become apparent in the later stages.

[0007] Managing multiple brands and teams presents challenges: the risk of mixing templates and materials from different brands / tenants is high, and there is a lack of isolation and access control. Summary of the Invention

[0008] To address the aforementioned issues, an online document semantic processing and paperless review system based on artificial intelligence is needed to enable self-service modification, high-precision preview, AI proofreading, paperless approval, version rollback, batch export, and traceable evidence storage.

[0009] The present invention provides an online semantic processing and paperless review system for design documents based on artificial intelligence, comprising: The template management module is used to upload and store standard Adobe Illustrator format template files, generate template objects, and maintain versions. The template parsing module is used to parse templates and extract canvas, layer, text element attributes, and image placeholder element information. The field configuration module is used to map parsed elements to business fields, form a field mapping table, and configure validation rules. The online editing module is used to display template previews on the terminal and provides interactive editing features such as text replacement, material replacement, element adjustment, and layer control. The material specification verification module is used to verify specifications including resolution, aspect ratio, and color mode when materials are uploaded. The font consistency detection module is used to detect the font consistency and missing character risk in edited text. The preview generation module is used to generate high-precision review preview images with status watermarks; The AI ​​proofreading module is used to automatically proofread the edited content and output proofreading results and risk warnings; The approval process module is used to initiate paperless approvals and supports multi-node approval operations; The version management module is used to generate version snapshots and supports version difference comparison and rollback. The log tracking module is used to record the entire operation log and associate it with version and approval information; The source file export module is used to export the source file after the review is approved; The multi-tenant permissions module is used to isolate data by brand / tenant and configure role-based operation permissions.

[0010] Preferably, the template parsing module is specifically used to parse Adobe Illustrator format files, and the extracted text element information includes content, bounding box coordinates, font, font size, color, and line spacing attributes, while the image element information includes position, size, and cropping boundaries.

[0011] Preferably, the field validation rules configured by the field configuration module include mandatory fields, length, format validation, regular expression validation, and channel field validation.

[0012] Preferably, the online editing module also supports automatically generating QR codes or barcodes from input parameters, automatically replacing the corresponding placeholder elements in the template, and recording the replacement behavior in the version log.

[0013] Preferably, the verification items of the material specification verification module also include DPI, file size, and transparency channel verification, and the submission is blocked or a risk warning is given when the verification fails.

[0014] Preferably, the preview generation module also generates a text-free preview image for comparison and checking the compliance of the base image and background; and automatically adds a pending or reviewed watermark according to the instance status to prevent unreviewed manuscripts from being misused.

[0015] Preferably, the AI ​​proofreading module's proofreading scope includes typos, grammar, date and price formats, compliance with sensitive words, and brand specifications. It also links with the enterprise's sensitive word library and compliance template library to output risk levels and suggested replacement text.

[0016] Preferably, the review process module automatically matches the approval flow based on brand, channel, activity level, or amount threshold, and supports approval operations such as signing, countersigning, transfer, rejection and return, and automatic reminders for timeout.

[0017] Preferably, the version snapshot of the version management module includes field values, material references, preview images, and a summary of proofreading results, and supports field difference comparison, preview image comparison, and one-click rollback to a specified version.

[0018] Preferably, the source file export module also supports batch exporting finished files of different sizes and DPIs according to different channel specifications and packaging them for archiving; and generates a hash digest value for the final exported file, which is bound to the approval record for storage to achieve tamper-proof verification.

[0019] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention supports operations personnel to complete content editing independently based on standard AI templates, reducing designers' repeated revisions, lowering cross-functional communication costs and scheduling management difficulties; the online editing effect is highly consistent with the final exported AI source file, and supports batch export of materials according to channel specifications, adapting to the production needs of multiple scenarios, and improving the iteration and implementation efficiency of the entire process of advertising poster production.

[0020] 2. This invention uses AI proofreading to link sensitive words and compliance databases of enterprises, automatically verifies copywriting issues and provides rectification suggestions. Combined with material specifications and font consistency detection, it avoids production risks such as substandard material specifications, missing fonts, and compliance violations from the source. With the addition of preview watermarks and finished product hash archiving, it reduces the probability of file misuse and tampering, and ensures the production quality and circulation security of design materials.

[0021] 3. This invention constructs a paperless approval closed loop, which can automatically match the approval flow according to rules, clarify the approval nodes, time limits and responsibility boundaries, and support diverse approval operations such as transfer and countersigning; at the same time, it realizes full-link operation log recording and version management, supports version difference comparison and rollback, forms a verifiable chain of evidence, and can also realize multi-brand / tenant isolation and fine-grained permission control, solving the problems of non-standard approval, difficulty in tracing versions, and chaotic cross-team management. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall system structure of the present invention. Detailed Implementation

[0023] refer to Figure 1 This invention provides a system architecture that enables operators to modify operational content and image materials online based on standard AI templates. The system automatically generates high-precision preview images for paperless approval. After approval, it exports workable Adobe AI source files and performs versioned log tracking and anti-tampering archiving of the entire process of editing, proofreading, review, and exporting. It also automatically checks for material specifications, font consistency, sensitive words, and compliance requirements, thereby improving production efficiency, reducing communication costs, and minimizing content risks. This invention's AI-based online semantic processing and paperless review system for design documents includes: Template Management Module: The module is used for uploading, storing, and managing standard template files; the standard template files include at least Adobe Illustrator format files; the template management module is used to generate template objects and maintain template versions.

[0024] Template parsing module: Connected to the template management module, it is used to parse the Adobe Illustrator format file and extract at least the following: artboard information, layer hierarchy, text element information (content, bounding box coordinates, font, font size, color, line spacing, and other attributes), and image element or image placeholder element information (position, size, cropping boundaries, etc.).

[0025] Field-based configuration module (can be a sub-module of the template management module): This is used to establish a mapping relationship between the parsed text / image elements and business fields, forming an editable field mapping table; the field mapping table supports field validation rules, including required fields, length, format (date / amount / discount), regular expression validation, channel field validation, etc.

[0026] Online editing module: Used to display template previews and provide editing interactions on the terminal side; the editing interactions include at least: Text content replacement and style constraint hints; Image material replacement, cropping, and adaptation; Adjusting element position / size; Layer visibility control; QR codes / barcodes are automatically generated and the specified placeholder elements are replaced.

[0027] Material specification verification module: This is used to verify materials during upload or replacement. Verification items include at least resolution, aspect ratio, DPI, file size, color mode (RGB / CMYK), and alpha channel. If the specifications are not met, submission will be blocked or a risk warning will be given.

[0028] Font consistency and missing character detection module: Used to identify the font set required for the template, and to perform font consistency checks and provide risk warnings for missing / alternative fonts on the text after online editing.

[0029] Preview the generated module: This module is used to generate online preview images for review based on editing instructions generated online; the preview generation module supports outputting high-precision preview images at a preset DPI, and supports outputting: Full preview image; Remove the text preview image (or the sensitive layer preview image) for comparison and inspection; Preview images with the "Pending / Reviewed" watermark are used for circulation and to prevent misuse.

[0030] AI proofreading module: This tool is used to automatically proofread operational content and output proofreading results and modification suggestions. The proofreading scope includes at least: typos / grammar, date and price formats, sensitive words / compliance prompts, and brand specification prompts. It also supports linkage with the enterprise's sensitive word library and compliance template library to output risk levels and suggested replacement text.

[0031] Review process module: Used to initiate a paperless approval process for preview images and related operational information; the approval process at least supports: submission, approval, rejection, transfer, countersigning or signing, timeout policy, and automatic matching of approval flow according to rules such as channel / brand / amount threshold (approval policy capability).

[0032] Version management and comparison rollback module: Used to generate version numbers and store version snapshots for each save / commit, supporting version difference comparison (field differences, preview differences) and rollback to a specified version.

[0033] Modify the log tracing and audit evidence storage module: This is used to record the operation logs of template instances in the stages of editing, proofreading, submission, approval, rejection, and export, and to associate them with version number, preview image, proofreading result, approval task, and approval comments to form a full-link timeline; preferably, a summary value (hash) is generated for the final approved result and archived for anti-tampering verification.

[0034] Source file export module: This tool is used to output usable Adobe Illustrator source files based on the final editing results after approval; and supports batch export of finished images or packaged files of different sizes and DPIs according to different channel specifications.

[0035] Multi-tenant / multi-brand isolation and access control module: This tool is used to isolate templates, materials, examples, approval workflows, logs, and exported files by brand / tenant, and to set visibility and operation permissions for different roles (designer, operations, legal, manager, administrator).

[0036] Storage module: It is used to store template files, material files, preview images, exported source files, version snapshots, log records, proofreading results, and approval records.

[0037] The AI-based online semantic processing and paperless review system for design documents achieves online semantic processing and paperless review of design documents through the following methods: Template import, parsing, and field configuration: Designers upload standard AI template files to the template management module. The template parsing module parses the template, extracting structural information such as canvases, layers, text elements, and image placeholders. The field configuration module maps business fields such as "event title, event time, price, promotion rules, store address, disclaimer, and QR code link" to corresponding elements and sets validation rules (required, length, format, and channel differences).

[0038] Online editing (text / images / positions / layers) and material specification verification: The operators create poster examples and enter the online editing module.

[0039] When operations staff replace image materials, the material specification verification module verifies the image resolution, aspect ratio, DPI, color mode (RGB / CMYK), alpha channel, etc. If the image does not meet the specifications, a prompt or submission will be blocked.

[0040] When the operations team modifies the text, the system will prompt that the length exceeds the limit or the format is incorrect based on the field rules, and may also prompt brand guidelines (such as "certain words are not allowed" or "font size range").

[0041] Font consistency and missing character detection: The system identifies the font set required for the template and detects whether there are any missing characters, font substitutions, or font inconsistencies in the text of the current instance; the risk items are written into the proofreading results or quality inspection report as mandatory inspection items before submission for approval.

[0042] Automatic generation and replacement of QR codes / barcodes: For QR code placeholder areas in the template, the system supports inputting a URL or activity parameters to automatically generate a QR code or barcode image and replace the placeholder element; at the same time, the generation parameters and replacement behavior are recorded in the version log.

[0043] High-precision preview generation, text removal comparison, and watermark transfer: The preview generation module renders and outputs a preview image at a preset DPI. Preferably, the system also generates a text-free preview image for comparison and inspection (e.g., to observe whether the base image is obscured or whether the background is compliant).

[0044] When an instance is pending review, the system automatically adds a "pending review" watermark to the preview image; once approved, it is replaced with a "reviewed" watermark to prevent misuse of unreviewed manuscripts.

[0045] AI proofreading works in conjunction with sensitive word / compliance databases: Before submitting for approval, the operations team triggers the AI ​​proofreading module. This module performs typos / grammar checks and date / price format verification on the operational copy, and integrates with the company's sensitive word library and compliance clause library to output risk levels and suggested replacement text. The proofreading results and revision history are stored in a version snapshot.

[0046] Approval strategies and paperless processes: The review process module automatically matches the approval flow based on rules such as brand, channel, activity level, or amount threshold (e.g., "e-commerce main image requires final review by legal department + person in charge," "high-risk keyword hits require additional compliance checkpoints"). It supports signature / countersignature, transfer, rejection / return, and automatic timeout reminders. Approval opinions and checkpoint actions are written to the process timeline and logs.

[0047] Version management, difference comparison and rollback: Each save or commit generates a version snapshot (including field values, source material references, preview images, and a summary of proofreading results). System supports: Field difference comparison (values ​​before and after change); preview image comparison; one-click rollback to a specified version; and write the comparison and rollback behavior to the log.

[0048] After approval, export the AI ​​source files and batch export multiple specifications: Once approved, the source file export module writes the final editing results back to the corresponding elements in the template AI file, generating a workable Adobe Illustrator source file for printing or distribution. Preferably, finished images of different sizes / DPIs can be exported in batches according to channel specifications and packaged into ZIP archives.

[0049] Meanwhile, to prevent tampering with the evidence, the system generates a hash value for the final exported file and binds it to the approval record for safekeeping.

[0050] Multi-tenant / multi-brand isolation and access control: The system isolates templates, materials, instances, approvals, and logs by tenant / brand. The access control module restricts access: operations staff can only edit template instances for their own tenant; legal staff can only review instances for their respective brands; administrators can configure approval policies and sensitive word lists, thereby reducing the risk of misuse across teams.

[0051] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0052] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An online document semantic processing and paperless review system based on artificial intelligence, characterized in that: include: The template management module is used to upload and store standard Adobe Illustrator format template files, generate template objects, and maintain versions. The template parsing module is used to parse templates and extract canvas, layer, text element attributes, and image placeholder element information. The field configuration module is used to map parsed elements to business fields, form a field mapping table, and configure validation rules. The online editing module is used to display template previews on the terminal and provides interactive editing features such as text replacement, material replacement, element adjustment, and layer control. The material specification verification module is used to verify specifications including resolution, aspect ratio, and color mode when materials are uploaded. The font consistency detection module is used to detect the font consistency and missing character risk in edited text. The preview generation module is used to generate high-precision review preview images with status watermarks; The AI ​​proofreading module is used to automatically proofread the edited content and output proofreading results and risk warnings; The approval process module is used to initiate paperless approvals and supports multi-node approval operations; The version management module is used to generate version snapshots and supports version difference comparison and rollback. The log tracking module is used to record the entire operation log and associate it with version and approval information; The source file export module is used to export the source file after the review is approved; The multi-tenant permissions module is used to isolate data by brand / tenant and configure role-based operation permissions.

2. The system according to claim 1, characterized in that, The template parsing module is specifically used to parse Adobe Illustrator format files. The extracted text element information includes content, bounding box coordinates, font, font size, color, and line spacing attributes, while the image element information includes position, size, and cropping boundaries.

3. The system according to claim 1, characterized in that, The field validation rules configured by the field configuration module include mandatory fields, length, format validation, regular expression validation, and channel field validation.

4. The system according to claim 1, characterized in that, The online editing module also supports automatically generating QR codes or barcodes from input parameters and automatically replacing the corresponding placeholder elements in the template, while recording this replacement behavior in the version log.

5. The system according to claim 1, characterized in that, The material specification verification module also includes verification items such as DPI, file size, and transparency channel. If the verification fails, it will prevent submission or give a risk warning.

6. The system according to claim 1, characterized in that, The preview generation module also generates a text-free preview image for comparison and checking the compliance of the base image and background; and automatically adds a pending or reviewed watermark according to the instance status to prevent unreviewed manuscripts from being misused.

7. The system according to claim 1, characterized in that, The AI ​​proofreading module's scope includes typos, grammar, date and price formats, compliance with sensitive words, and brand specifications. It also links with the enterprise's sensitive word library and compliance template library to output risk levels and suggested replacement text.

8. The system according to claim 1, characterized in that, The review process module automatically matches the approval flow based on brand, channel, activity level, or amount threshold, and supports approval operations such as signing, countersigning, transfer, rejection and return, and automatic reminders for timeout.

9. The system according to claim 1, characterized in that, The version management module's version snapshot includes field values, material references, preview images, and a summary of proofreading results. It supports field difference comparison, preview image comparison, and one-click rollback to a specified version.

10. The system according to claim 1, characterized in that, The source file export module also supports batch exporting finished files of different sizes and DPIs according to different channel specifications and packaging them for archiving; and generates a hash digest value for the final exported file, which is bound to the approval record for storage to achieve tamper-proof verification.

Images