Visual intelligent document polishing editing system, method and equipment and storage medium

Abstract

The invention relates to the technical field of document processing, in particular to a visual intelligent document polishing editing system, method and equipment and a storage medium. The invention provides a visual intelligent document polishing editing system, method and device and a storage medium, and the system comprises a text analysis module which analyzes an original document uploaded by a user, recognizes an obvious problem, and outputs an analysis result; the retouching and optimizing module is used for automatically retouching and optimizing the memorability of the original document to generate a modified file based on the analysis result; the visual editing interface module is used for providing an interface, allowing a user to check the modified file and the original file, and editing and adjusting suggestions provided by the system in real time; and the user feedback and learning module is used for collecting feedback of the user on the modified file, and learning and improving the retouching optimization module based on the feedback. The invention aims to solve the problems before and after modification through visual comparison and facilitate manual checking of customers.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of document processing, and in particular to a visual intelligent document polishing editing system, method, device and storage medium. BACKGROUND

[0002] Natural Language Processing Model (NLP Model) is a branch of artificial intelligence (AI) that focuses on enabling computers to understand, generate and process human language (such as Chinese, English, etc.). Simply put, it is a tool or algorithm used by computers to "understand" and "write" human language.

[0003] Chinese invention patent application No. CN119129567A disclosed a writing polishing method, system, device and storage medium on December 13, 2024, by obtaining the target user's to-be-polished text, historical reading data and historical writing data in a specified time period; determining the writing style of the target user according to the to-be-polished text, historical reading data and historical writing data of the target user; generating a personalized writing template based on the writing style and the to-be-polished text; inputting the personalized writing template and the to-be-polished text into the trained writing model, and the trained writing model outputs the target polishing text corresponding to the to-be-polished text with the personalized writing template as the prompt word. Although this document generates polishing text through personalized template, it does not provide an interface for users to compare before and after document polishing, and users cannot know the specific modification place, resulting in time-consuming and laborious checking process.

[0004] Therefore, it is necessary to propose a visual intelligent document polishing editing system, method, device and storage medium, which can intuitively compare before and after modification and facilitate customers to manually check. SUMMARY

[0005] The present application aims to overcome the shortcomings of the prior art and provide a visual intelligent document polishing editing system, method, device and storage medium, which can intuitively compare before and after modification and facilitate customers to manually check.

[0006] To achieve the above-mentioned purpose, in the first aspect, the present application provides a visual intelligent document polishing editing system, characterized in that it comprises: a text analysis module that analyzes the original document uploaded by the user and outputs an analysis result after identifying obvious problems; a polishing optimization module that automatically polishes and optimizes the original document based on the analysis result to generate a modified file; The visual editing interface module provides an interface that allows users to view the modified file and the original file, edit and adjust the suggestions provided by the system in real time; The user feedback and learning module collects user feedback on the modified file, learns and improves the polishing optimization module based on the feedback.

[0007] As a preferred embodiment, the visual editing interface module includes: The list module lists all the modifications in chronological order; The zoom-in module selects the corresponding part of the original file and the modified file in one of the list module to display together; The box selection module selects the modified content in the zoom-in module; The elimination module eliminates the content in the box selection module with one click.

[0008] As a preferred embodiment, the text analysis module differentiates the identified obvious problems, including syntax errors, inconsistent styles, and inconsistent semantics.

[0009] As a preferred embodiment, the polishing optimization module includes: The syntax analysis module performs fine-grained syntax analysis on the original document to identify potential errors and automatically correct them; The style adjustment module automatically replaces vocabulary or sentence patterns based on the user-provided style to form a text style that better meets the user's needs; The semantic consistency checking module checks the semantic coherence across sentences / paragraphs in the original document to ensure smooth overall logic and no conflicts between the preceding and following text; The tone and style consistency preservation module prioritizes preserving the user's expression method and core meaning in key sentences when performing syntax and style optimization.

[0010] As a preferred embodiment, the elimination module includes: The cloud module stores the content of the zoom-in module in the cloud, and the box selection module displays the selected content locally or entirely; The local and full deletion switching module allows one-click full deletion of the content in the box selection module, or selects appropriate words for deletion based on user needs; The local locking module selects part of the content in the box selection module to lock, and retains the locked content when deleting.

[0011] In a second aspect, the present application proposes a visual intelligent document polishing editing method applied to the visual intelligent document polishing editing system, The method analyzes the user-uploaded original document, identifies obvious problems, and outputs the analysis results; Based on the analysis result, the original document is automatically polished and the modified file is generated; An interface is provided to allow users to view the modified file and the original file, and to edit and adjust the suggestions provided by the system in real time; Collect user feedback on the modified file, and learn and improve the system's polishing and optimization of the original document based on the feedback.

[0012] As a preferred, after generating the modified file, a list is automatically generated, which is listed in order before and after modification, and is outside the document; Select the modified content to be viewed in the list, enlarge the content to be viewed, and display it in the annular frame; By changing the width and height of the enlarged annular frame, you can select part of the content to be viewed; After adjusting the annular frame, select one-key clear to eliminate unnecessary content.

[0013] As a preferred, part of the content to be viewed that cannot be completely displayed in the annular frame will be automatically uploaded to the cloud; After adjusting the annular frame to display the area, generate a picture format of the area to be deleted before deleting it.

[0014] In a third aspect, the present application provides a bag device, wherein the processor is used to run a program, the electronic device comprises a processor and a memory; the memory stores a computer readable program that can be executed by the processor; the processor executes the computer readable program to realize the writing polishing method.

[0015] In a fourth aspect, the present application provides a processor and a storage medium, wherein the computer readable storage medium stores one or more computer readable programs, the one or more computer readable programs can be executed by one or more processors, and the computer readable program is executed by the processor to realize the writing polishing method.

[0016] Compared with the prior art, the visual intelligent document polishing and editing system, method, device and storage medium provided by the present application have the following advantages: 1. An intuitive and easy-to-use interface is provided, which allows users to view the comparison before and after document polishing, edit and adjust the suggestions provided by the system in real time, and through the visual editing interface, users can easily understand the modification suggestions of the system and provide immediate feedback, improving the convenience of operation and the transparency of polishing.

[0017] 2. By adding an elimination module in the visual editing interface module function, one-key clear is realized, which facilitates editing text.

[0018] The features and advantages of the present application will be described in detail in conjunction with the embodiments and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 Workflow diagram of the method in the embodiment of the present application.

[0020] Figure 2 Internal structure block diagram of step S2 in the embodiment of the present application.

[0021] Figure 3 Internal structure block diagram of the syntax evaluation model in the embodiment of the present application.

[0022] Figure 4 Internal structure block diagram of the visual intelligent document polishing editing system in the embodiment of the present application.

[0023] Figure 5 Internal structure block diagram of the polishing optimization module in the embodiment of the present application.

[0024] Figure 6 Internal structure block diagram of the visual editing interface module in the embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to make the objectives, technical solutions and advantages of the present application clearer and more comprehensible, the present application will be further described in detail below with the aid of the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the scope of the present application. In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present application.

[0026] Despite the significant progress made in the field of document generation with artificial intelligence (AI), which has high efficiency and innovation in generating text, one of the main challenges faced by AI generation technology is the randomness and unpredictability of its content. These characteristics often lead to the generated document being difficult to fully meet the specific needs and expectations of users, especially in terms of accuracy and professional standards. Current AI document generation technologies, such as systems based on GPT models, are capable of automatically producing text with a certain logic and coherence. These systems learn from a large amount of text data to simulate different styles and structures of writing. Despite the maturity of the technology, AI-generated text often lacks sensitivity to specific uses and details, and the randomness of the generated content is greater, which makes the generated document not precise enough in details or not in line with specific professional requirements in expression.

[0027] The present application proposes a visual intelligent document polishing editing system based on natural language processing. The system is specially designed to solve AI-generated documents are polished and personalized to ensure that the generated documents are not only technically efficient, but also meet the specific requirements and professional standards of the user, significantly improving the quality of the final document and user satisfaction. The system uses advanced natural language processing technology and visual editing tools to allow users to see the polishing suggestions and make real-time adjustments, ensuring that the final output of the document meets the efficient generation of AI and meets the personalized and detailed needs.

[0028] Referring to Figure 1 , the embodiment of the present application provides a visual intelligent document polishing editing method, which comprises: S1: Analyze the user uploaded original document, identify the obvious problems and output the analysis results.

[0029] Responsible for comprehensive analysis of the original document uploaded by the user, identify the syntax errors, style inconsistencies, semantic inconsistencies and possible logical flaws in the text. Mark the identified problems and enter the next stage. Use natural language processing techniques such as syntax analysis and semantic analysis algorithms to automatically detect and mark problem areas in the document. This process includes but is not limited to part-of-speech tagging, syntax dependency analysis and entity recognition to ensure the accuracy of problem identification.

[0030] S2: Based on the analysis results, automatically polish and optimize the original document and generate a modified file.

[0031] Based on the analysis results, automatically polish and optimize the content of the document, including correcting grammatical errors, improving expression style and enhancing the logical coherence of the content. Use machine learning models, especially models based on the Transformer architecture such as BERT, to provide context-aware text polishing suggestions. In addition, style conversion algorithms will be applied to adjust the writing style of the document to meet specific application requirements or personal preferences.

[0032] Specific adjustments are as follows (see Figure 2 ): Grammar correction: Through fine-grained syntax analysis and scoring of the document text, potential spelling errors, part-of-speech errors, tense errors, punctuation errors, etc. are found and corrected automatically.

[0033] Style adjustment: According to the user-provided writing tone (formal, semi-formal, friendly, academic, etc.) and style preferences (such as whether to be more lively or more rigorous), automatically replace some vocabulary or sentence patterns to form a text style that better meets the user's expectations.

[0034] Semantic consistency check: Check the semantic coherence across sentences / paragraphs in the document to ensure smooth overall logic and no conflicts between the preceding and following text User's original tone and style are maintained: Through instruction engineering and context analysis, the system prioritizes preserving the user's expression method and core meaning in key sentences during grammar and style optimization, rather than blindly rewriting.

[0035] Reference Figure 3 The method for grammar evaluation uses a grammar evaluation model, which specifically includes text representation construction, grammar feature extraction, and grammar scoring network.

[0036] The method for text representation construction is to split the document text by character or word level. For Chinese, BPE (Byte Pair Encoding) or [CLS] / [SEP] tokenization can be used, and for English, WordPiece or SentencePiece can be combined. Through pre-trained language models (such as BERT, RoBERTa, etc.) or lightweight embedding models, each word / subword is vectorized to obtain the context representation of the entire concatenated text.

[0037] The method for grammar feature extraction is to input the vector sequence of the concatenated text into the backbone network of the grammar evaluation model (which can be Transformer, bidirectional LSTM, or multi-layer perceptron MLP) to obtain grammar feature vectors rich in grammar information. As described in the example, the feature vectors can be further processed to obtain intermediate representations for grammar scoring.

[0038] The grammar scoring network uses a "second multi-layer perceptron + second normalization layer" structure to map the above grammar features to a predicted grammar score value (between 0 and 1). A higher score indicates that the text is more in line with grammar specifications; a lower score indicates that the text has grammar or spelling errors and needs to be corrected.

[0039] The grammar correction process includes instruction engineering-based correction prompts, error detection and local replacement, and grammar and context linkage.

[0040] The instruction engineering correction prompt provides clear instructions to the model during correction, such as: Instruction: Please make minor grammar modifications to the following text while maintaining the original meaning, correcting spelling or tense errors, and without changing the core meaning or style. This Prompt Engineering ensures that large models or specialized grammar correction models do not significantly alter the text logic when rewriting.

[0041] Error detection and local replacement involves comparing detected low-score sentences or words, using edit distance, common error lists, and tense rules to make refined suggestions. The final output is a new candidate text, and the grammar scores before and after correction are compared. If there is a significant improvement and no semantic loss is found, the correction result is adopted.

[0042] Grammar and context linkage is to prevent rewriting from causing a break in context continuity. After correction, the system will check the sentence before and after again to ensure consistency in tense, pronoun, and reference.

[0043] Instructional grammar correction can fully utilize the context understanding ability of large models, ensuring that the core content is not altered during correction; the secondary scoring mechanism is used to verify the rationality of the modification, avoiding excessive or incorrect correction.

[0044] Style adjustment and style unification include style preference acquisition and modeling, and instruction engineering driven style rewriting.

[0045] Among them, the acquisition and modeling of style preferences include user-initiated selection and automatic detection and analysis. User-initiated selection is for users to explicitly inform the desired writing style of "formal", "friendly", "academic" and the like during interaction, and can specify whether to "emphasize personal advantages" or "maintain modesty" and other characteristics. Automatic detection and analysis is that if the user provides some demonstration text or past documents, the system can automatically induce the user's common style and preferred vocabulary through sentiment analysis, word frequency statistics, and sentence pattern detection.

[0046] Instruction engineering driven style rewriting includes multi-level Prompt design, mood words, sentence pattern library, and cross-section unification.

[0047] Among them, multi-level Prompt design is a step-by-step instruction for large models to adjust the style. The specific steps are as follows: first, detect the differences between the current text and the user's desired style (such as overly colloquial language, and not concise enough sentence patterns), and second, provide rewriting solutions while maintaining the core meaning.

[0048] Using mood words and sentence pattern libraries, a "style corpus" can be introduced, containing common expressions in different scenarios (e.g., "Sincerely, thank you", "My heartfelt thanks", "My research direction", etc.). The model can call this library for replacement when rewriting.

[0049] Cross-section unification maintains the consistency of the entire document style by aligning sentence patterns and word usage across multiple sections of the text (such as detecting synonymous replacements and using similar adjectives).

[0050] By combining user-initiated selection and automatic detection, the system can accurately capture the user's unique style. Through Prompt Engineering, the model can also adhere to the principle of "not changing the core content" when rewriting style, only adjusting the mood, word choice, and sentence length.

[0051] Semantic consistency checking includes semantic relevance and conflict detection, and Chain-of-Thought (CoT) self-checking.

[0052] Semantic consistency and conflict detection includes sentence vectors and chapter vectors, and topic consistency and conflict points. Sentence vectors and chapter vectors use the Transformer model (such as Sentence-BERT) to generate sentence vectors for each sentence and chapter vectors for the entire document to analyze the closeness, similarity, and logical coherence between contexts. If a paragraph mentions "apply for major A", but the latter mentions "I am very interested in major B" and the two are contradictory, the model can identify the semantic conflict point and prompt the editor or automatically adjust the expression.

[0053] Chain-of-Thought (CoT) self-checking includes thought chain checking and instructional fine-tuning. Thought chain checking is a consistency check that allows the model to first output its thought chain: "Paragraph 1 states that the target school is X, and Paragraph 2 states that the research interest is Y. Are they in conflict?" If potential contradictions are detected, the model will explain the reason in the thought chain and give correction suggestions. Instructional fine-tuning uses instructions to the model, "unify the expression of the target major / project while maintaining the core facts, and ensure consistency in the context." In the process of semantic consistency checking, not only vectorization technology is used for static semantic comparison, but also the thought chain of the large model is used for "self-checking" reasoning, which can more subtly find the logical inconsistency or ambiguity within the document. Allow users to intervene in the automatic revision or manual confirmation of "contradictory paragraphs" to ensure that the user has the final control over the "content" and "facts" of the document.

[0054] Among them, maintaining the user's original tone and style includes context preservation strategy and multi-round comparison and user intervention. The context preservation strategy explicitly states in the Prompt that the core expression of the target text is preserved, and only the parts that are obviously inappropriate or do not conform to the target style are minimally modified. By measuring the difference in sentence structure (Levenshtein distance, etc.), if the difference is too large, the system will prompt again or perform a lighter rewriting plan. Multi-round comparison and user intervention generate two versions after completing a grammar and style processing: Version A: difference comparison between the original text and the modified version; Version B: final merged modified version. Users can view all modifications in the visual interface and choose to keep or roll back to maximize the preservation of the user's "personalized" language.

[0055] Maintaining the user's original tone and style uses the "difference comparison + user intervention" mode, which is different from full-automatic rewriting. This mode can better preserve the user's personalized expression and reduce overwriting. Through instruction engineering, the large model performs "minimal modification" while ensuring grammatical correctness, style uniformity, and logical fluency.

[0056] S3: Provide an interface that allows users to view the modified file and the original file, edit and adjust the suggestions provided by the system in real time; After the original file is polished, a polished file is output, and the modified part is displayed by bold or without color. A new interface is displayed on the side or top of the polished file. After generating the modified file, a list is automatically generated, which is listed in order before and after modification and is outside the document. Specifically, one side of the interface displays a list. The list can be set in the form of 3X3, the first column displays the serial number and the stable position, which can intuitively reflect how many modified parts there are, and the serial numbers are numbered from top to bottom according to the order of the text. The second column displays the text before modification, which corresponds to each serial number in the first column. The third column displays the text after modification, which corresponds to each serial number in the first column.

[0057] Select the modified content in the list that needs to be viewed, and enlarge the content to be viewed and display it in the ring frame. Specifically, the list is generally set in a small area, and the cells in the list can be selected according to actual needs. The content of the cell will be enlarged and displayed. After enlargement, if the text is too much, the overall size will still be small, and observation will not change. The content to be enlarged can be selected by the ring frame. The selected content will occupy the entire interface, and the unselected content will be automatically uploaded to the cloud.

[0058] The width and height of the enlarged ring frame can be changed to select part of the content to be viewed. Since the content displayed in the text may be text, formulas, or other content, different formats have different display needs. For example, formulas or special symbols.

[0059] The part of the content to be viewed that cannot be completely displayed in the ring frame will be automatically uploaded to the cloud, so that deletion will not affect other content; the area to be deleted is first generated in picture format before the area to be deleted is generated in picture format, and then the area to be deleted is generated in picture format. The picture can be uniformly set to a recycle bin, and if it is deleted by mistake, it can be retrieved. The area of the ring frame can also be edited, and the text can be added or deleted according to needs. When adding, the display area and the number of displayed text content or symbols remain unchanged, and some originally displayed characters will be squeezed out of the display area and stored in the cloud. The stored content will be synchronized to the modified text.

[0060] The ring frame and the deletion function can more efficiently process the text scheme, and the processing process is more convenient.

[0061] S4: Collect user feedback on the modified file, and improve the system's polishing of the original document based on the feedback.

[0062] The feedback of the user on the modified file is converted into data features, and the parameters or rules of the polishing model are adjusted in real time or near real time using the feedback data; the converted data is analyzed for weight, and data with a small corresponding weight is discarded; and the screened data features are output to the polishing model.

[0063] During the modification process, each modification is marked, such as setting an agreed modification to 1 and a disagreed modification to 0. Instead of adopting data every time, some data is subjected to noise reduction processing, such as incorrect judgments and fatigue caused by long-time modification. By analyzing the weight of this type of data, if the weight is too small, the rationality of this data is crossed, and feedback should not be collected to avoid affecting the accuracy of the original model.

[0064] In an optional embodiment, referring to Figure 4 , the present application provides the above-mentioned visual intelligent document polishing editing method, and further provides a visual intelligent document polishing editing system, which can realize all functions of the visual intelligent document polishing editing method. The system includes a text analysis module, a polishing optimization module, a visual editing interface module, and a user feedback and learning module.

[0065] The text analysis module analyzes the uploaded original document of the user, identifies obvious problems, and outputs analysis results. In the text analysis module, the identified obvious problems include grammatical errors, inconsistent styles, and inconsistent semantics, which are marked differently.

[0066] The polishing optimization module automatically polishes and optimizes the original document based on the analysis results to generate a modified file. Referring to Figure 5 , the polishing optimization module includes: a grammar analysis module that performs fine-grained grammar analysis on the original document to find potential errors and automatically correct them; a style adjustment module that automatically replaces vocabulary or sentence patterns according to the user-provided style to form a text style that better meets the user's needs; a semantic consistency checking module that checks the semantic coherence across sentences / paragraphs in the original document to ensure smooth overall logic and no conflicts between preceding and following texts; and a tone and style consistency maintaining module that, when performing grammar and style optimization, prioritizes preserving the user's expression method and core meaning in key sentences.

[0067] The visual editing interface module provides an interface that allows the user to view the modified file and the original file, and to edit and adjust the suggestions provided by the system in real time.

[0068] Referring to Figure 6The visual editing interface module comprises: a list module, which lists all the modifications in chronological order; an enlargement module, which enlarges and displays the corresponding parts of the original file and the modified file in the list module; a box selection module, which selects the modified content in the enlargement module through a ring-shaped box; and an elimination module, which eliminates the content in the box selection module with one key.

[0069] The elimination module comprises: a cloud module, in which the content of the enlargement module is stored, and the box selection module displays the content in the cloud module partially or entirely; a partial and entire deletion switching module, which can delete the content in the box selection module with one key or delete the appropriate words according to the needs; and a partial locking module, which locks part of the content in the box selection module and retains the locked content when deleting.

[0070] Specifically, the original file is polished to output a polished file, and the modifications are displayed through bolding or without color. A new interface is displayed on the side or top of the polished file. After generating the modified file, a list is automatically generated, and the modifications before and after are listed in order and outside the document. Specifically, one side of the interface displays a list. The list can be set in the form of 3X3, the first column displays the serial number and the stable position, which can intuitively reflect how many modifications, and is numbered from top to bottom according to the chronological order of the text. The second column displays the text before the modification, which corresponds to each serial number in the first column. The third column displays the text after the modification, which corresponds to each serial number in the first column.

[0071] The modified content in the list that needs to be viewed is enlarged and displayed in a ring-shaped box. Specifically, the list is generally set in a small area, and the cells in the list can be selected according to actual needs, and the content of the cells will be enlarged and displayed. After enlargement, if the text is too much, the whole will still be relatively small even if it is enlarged, and observation will not change. The content to be enlarged can be selected through a ring-shaped box, and the selected content will occupy the entire interface, and the unselected content will be automatically uploaded to the cloud.

[0072] By changing the width and height of the enlarged ring-shaped box, part of the content in the content to be viewed can be selected. Since the content displayed in the text can be text, formula or other content, different formats have different display needs. For example, it is a formula or a special symbol.

[0073] For the need to view the content can not be fully displayed in the ring box part will be automatically uploaded to the cloud, so delete the time will not affect other content; ring box adjustment display area, the area to be deleted first generated picture format before one key clear, clear the picture can be set a recycle bin, if wrong delete can be found back. Also can edit the area of the ring box, can be according to the demand, to increase or reduce the text, increase the display area, the number of displayed text content or symbol does not change, some of the original display character will be squeezed out of the display area into the cloud storage.

[0074] Through the ring box and delete function can more efficient processing of text scheme, the process is more convenient.

[0075] User feedback and learning module, collect user feedback on the modified file, based on feedback and improve the polishing optimization module.

[0076] In an optional embodiment, the present application provides an electronic device, the processor is used to run the program, wherein, the electronic device includes processor and memory; the memory is stored with computer readable program which can be executed by the processor; the processor executes the computer readable program to realize the writing polishing method.

[0077] In an optional embodiment, the present application provides a storage medium, characterized in that, the computer readable storage medium is stored with one or more computer readable programs, the one or more computer readable programs can be executed by one or more processors, the computer readable program is executed by the processor, realizes the writing polishing method.

[0078] In the above embodiment, the description of each embodiment has its own emphasis, and the part not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0079] It can be understood that the related features in the above method and system can be mutually referred. In addition, "first", "second" and the like in the above embodiments are used to distinguish each embodiment, and do not represent the advantages and disadvantages of each embodiment.

[0080] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above described system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

[0081] The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems can be used with programs in accordance with the teachings herein, or it can prove convenient to construct more specialized apparatus to perform the required method steps. The required structure for a variety of these systems will be apparent from the description above. In addition, the present application is not described with reference to any particular programming language. It will be appreciated that a variety of programming languages can be used to implement the teachings of the application as described herein, and any particular language should not be construed as limiting the scope of the application.

[0082] Moreover, the memory can include non-transitory memory in a computer-readable medium, a Random Access Memory (RAM), and / or a non-volatile memory such as a Read Only Memory (ROM) or a flash memory, for example, and the memory includes at least one memory chip.

[0083] As will be appreciated by one of skill in the art, embodiments of the present application can be comprised of a method, a system, or a computer program product. Accordingly, the present application can be embodied in a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer program code thereon for use by or in connection with an instruction execution system.

[0084] The present application is described in reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0085] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0086] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 Figure 1 The flowchart blocks or blocks in the multiple flows and / or blocks

[0087] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0088] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the computer stores information such as computer program instructions. Memory is an example of computer readable media.

[0089] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to computing device. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0090] It should also be noted that the terms "comprising", "including", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, an element defined by the statement "comprising a" does not exclude the existence of additional identical elements in the process, method, article or apparatus including the element.

[0091] ​Those skilled in the art should understand that the computing model provided in the embodiments of the present application is not limited to MiniCPM, Whisper and DeepSeek, and the effect of the present application can also be achieved by using other computing models in the same field. The above three models are only a preferred embodiment of the present application.

[0092] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of being implemented on one or more computer usable storage media containing computer usable program codes (including but not limited to disk storage, CD-ROM, optical storage, etc.).

[0093] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement or improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A visual intelligent document redaction editing system, characterized in that, Comprise: Text analysis module, analyze the original documents uploaded by users, identify obvious problems and output analysis results; Polish optimization module, based on the analysis results, automatically polish and optimize the original documents to generate modified files; Visual editing interface module, provide an interface that allows users to view modified files and original files, and real-time edit and adjust the suggestions provided by the system; User feedback and learning module, collect user feedback on modified files, learn based on feedback and improve the polish optimization module.

2. The visual intelligent document redaction editing system of claim 1, wherein, The visual editing interface module comprises: List module, list all modifications in order; Zoom-in module, select the corresponding part of the original file and the modified file in the list module to display together; Box selection module, select the modified content in the zoom-in module through a ring-shaped box; Elimination module, one-key elimination of the content in the box selection module.

3. The visual intelligent document redaction editing system of claim 1, wherein, In the text analysis module, the identified obvious problems include syntax errors, inconsistent styles, and semantic inconsistencies.

4. The visual intelligent document redaction editing system of claim 1, wherein, The polish optimization module comprises: Syntax analysis module, conduct fine-grained syntax analysis on the original documents to find potential errors and automatically correct them; Style adjustment module, automatically replace vocabulary or sentence patterns based on the user-provided style to form a text style that better meets the user's needs; Semantic consistency checking module, check the semantic coherence across sentences / paragraphs in the original document to ensure smooth overall logic and no conflicts between preceding and following texts; Tone and style consistency maintenance module, when performing syntax and style optimization, prioritize preserving the user's expression method and core meaning in key sentences.

5. The visual intelligent document redaction editing system of claim 2, wherein, The elimination module comprises: Cloud module, store the content of the zoom-in module in the cloud, and display the selected content locally or entirely in the remote module; Local and full deletion switching module, the content in the box selection module can be deleted entirely or selectively; Local locking module, select part of the content in the box selection module to lock, and keep the locked content when deleting.

6. A visual intelligent document polishing editing method applied to the visual intelligent document polishing editing system of any one of claims 1-5, characterized by: Analyzing the original documents uploaded by users, identifying obvious problems, and outputting analysis results; Based on the analysis results, automatically polish and optimize the original documents to generate modified files; Provide an interface that allows users to view modified files and original files, and real-time edit and adjust the suggestions provided by the system; Collect user feedback on modified files, learn based on feedback and improve the system's polish optimization of original documents.

7. A visual intelligent document redaction editing method as claimed in claim 5, wherein, After generating the modified file, automatically generate a list, list the modifications in order before and after the modification, and place them outside the document; Select the modified content to be viewed in the list, enlarge the content to be viewed and display it in a ring-shaped box; By changing the width and height of the zoom-in ring-shaped box, you can select part of the content to be viewed; After adjusting the ring-shaped box, select one-key clear to eliminate unnecessary content.

8. The visual intelligent document redaction editing method of claim 6, wherein, For parts of the content to be viewed that cannot be fully displayed in the ring-shaped box, they will be automatically uploaded to the cloud; The annular frame adjusts the displayed area, and the area to be deleted is first generated in a picture format and then deleted.

9. An electronic device, comprising: The processor is configured to run a program, wherein the electronic device comprises a processor and a memory; the memory stores a computer readable program executable by the processor; and the processor implements the writing polishing method according to any one of claims 1-5 when executing the computer readable program.

10. A storage medium, characterized by The computer readable storage medium stores one or more computer readable programs executable by one or more processors, and the computer readable program is executed by the processor to implement the writing polishing method according to any one of claims 1-5.

Images