Document identifier identification method and device, computer device, and storage medium

Abstract

The application relates to the technical field of artificial intelligence, can be applied to business system platforms such as medical health and financial technology, and discloses a document identifier identification method and device, computer equipment and a storage medium, wherein a target document is obtained by preprocessing a to-be-identified document; a pre-trained visual large model is used to identify the identifier of the target document multiple times, and a plurality of corresponding identification results are generated; context information of the plurality of identification results is constructed according to the content of the target document, and the context information and a pre-trained language large model are used to sequentially perform semantic analysis on the plurality of identification results, and a plurality of corresponding analysis results are generated; based on the plurality of analysis results, the probability value of the to-be-identified document existing an identifier is calculated, and whether the to-be-identified document exists an identifier is determined according to the probability value and a preset judgment strategy; and the application can effectively improve the efficiency and accuracy of document identifier identification.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and specifically to a document identifier recognition method, apparatus, computer device, and computer-readable storage medium. Background Technology

[0002] Currently, in document processing and information management systems, seals serve as an important identifier, widely used in various official documents to verify their legality and validity. The identification and verification of seals are crucial for document classification, review, and management, especially in the healthcare and fintech sectors.

[0003] In the healthcare field, medical records are among the most important documents, documenting a patient's medical history, diagnoses, and treatment processes. The doctor's or medical institution's seal on the medical record is a key identifier of its legitimacy and validity. Seal recognition technology can automatically verify the authenticity of medical records, ensuring accurate entry and management of medical record documents in electronic medical record systems. By recognizing seals, approved and unapproved medical records can be quickly distinguished, improving the efficiency and security of medical record management.

[0004] In the fintech field, financial contracts are among the most important documents, recording the rights and obligations between financial institutions and their customers. The seal on the contract is a key identifier of its legality and validity. Seal recognition technology can automatically verify the authenticity of contracts, ensuring the accurate entry and management of contract documents in electronic contract management systems. By recognizing seals, signed and unsigned contracts can be quickly distinguished, improving the efficiency and security of contract management.

[0005] However, in both the healthcare and fintech sectors, the complex visual features of seals (such as shape, color, and texture) and the diversity of document backgrounds mean that traditional manual seal recognition methods or single-model seal recognition methods currently suffer from poor recognition efficiency and accuracy. Therefore, providing a document identifier recognition method, apparatus, computer device, and computer-readable storage medium that can effectively improve the efficiency and accuracy of document identifier recognition is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of the shortcomings of the prior art, the purpose of this invention is to provide a document identifier recognition method, apparatus, computer device and computer-readable storage medium, aiming to solve the problem of how to effectively improve the efficiency and accuracy of document identifier recognition.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a document identifier recognition method, comprising: Obtain the document to be identified, and preprocess the document to be identified to obtain the target document; Using a pre-trained large visual model, the identifiers of the target document are identified multiple times, generating multiple corresponding recognition results; Based on the content of the target document, contextual information of the multiple recognition results is constructed, and semantic analysis is performed on the multiple recognition results sequentially using the contextual information and a pre-trained language model to generate corresponding multiple analysis results; Based on the multiple analysis results, the probability value of the document to be identified containing an identifier is calculated, and the existence of an identifier in the document to be identified is determined according to the probability value and a preset judgment strategy.

[0009] Secondly, the present invention provides a document identifier recognition device, comprising: The acquisition module is used to acquire the document to be identified, preprocess the document to be identified, and obtain the target document; The recognition module is used to recognize the identifiers of the target document multiple times using a pre-trained large visual model, and generate multiple corresponding recognition results. The analysis module is used to construct contextual information of the multiple recognition results based on the content of the target document, and use the contextual information and a pre-trained language model to perform semantic analysis on the multiple recognition results in sequence to generate corresponding multiple analysis results; The determination module is used to calculate the probability value of the existence of an identifier in the document to be identified based on the multiple analysis results, and to determine whether the document to be identified contains an identifier according to the probability value and a preset determination strategy.

[0010] Thirdly, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the document identifier recognition method as described above.

[0011] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the document identifier recognition method as described above.

[0012] Compared to existing technologies, this invention provides a document identifier recognition method, apparatus, computer device, and computer-readable storage medium. The method involves: acquiring a document to be identified; preprocessing the document to obtain a target document; using a pre-trained visual large-scale model to repeatedly identify identifiers in the target document, generating multiple corresponding recognition results; constructing contextual information for the multiple recognition results based on the content of the target document; and sequentially performing semantic analysis on the multiple recognition results using the contextual information and a pre-trained language large-scale model, generating multiple corresponding analysis results; calculating the probability value of the existence of identifiers in the document to be identified based on the multiple analysis results; and determining whether identifiers exist in the document to be identified based on the probability value and a preset judgment strategy. Therefore, this invention's document identifier recognition method effectively improves the efficiency and accuracy of document identifier recognition by combining a multi-step processing flow of a visual large-scale model and a language large-scale model. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram illustrating the application environment of a document identifier recognition method provided in an embodiment of the present invention.

[0015] Figure 2 This is a flowchart illustrating a document identifier recognition method according to an embodiment of the present invention.

[0016] Figure 3 This is a schematic diagram of the program module of a document identifier recognition device provided in an embodiment of the present invention.

[0017] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention.

[0018] Figure 5 This is another structural schematic diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0021] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0022] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once [the described condition or event] is detected," or "in response to detection of [the described condition or event]."

[0023] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0025] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0026] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0027] An embodiment of the present invention provides a document identifier recognition method, which can be applied to, for example... Figure 1 In the application environment shown, the client and server communicate via a network. The client includes, but is not limited to, handheld computers, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud computing devices, and personal digital assistants (PDAs). The server can be a standalone server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0028] Please see Figure 2 An embodiment of the present invention provides a document identifier recognition method, wherein the method includes the following steps: S100: Obtain the document to be identified, and preprocess the document to be identified to obtain the target document; S200. Using a pre-trained large visual model, the identifiers of the target document are identified multiple times, generating multiple corresponding identification results. S300. Based on the content of the target document, construct contextual information for the multiple recognition results, and use the contextual information and a pre-trained language model to sequentially perform semantic analysis on the multiple recognition results to generate corresponding multiple analysis results; S400. Based on the multiple analysis results, calculate the probability value of the document to be identified containing an identifier, and determine whether the document to be identified contains an identifier according to the probability value and a preset judgment strategy.

[0029] In practical implementation, the document identifier recognition method of this embodiment significantly improves the efficiency and accuracy of document identifier (such as seals, symbols, etc.) recognition by combining a multi-step processing flow of a visual large model and a language large model. First, in the preprocessing stage (S100), operations such as format verification, noise removal, and content standardization effectively improve document quality, providing clearer and more consistent target documents for subsequent recognition. Second, in the multiple recognition stage (S200), the pre-trained visual large model can be used to perform multiple identifier recognitions on the target document, generating diverse recognition results. This not only increases the number of recognition samples but also reduces misjudgments caused by single recognition errors through multiple recognitions, improving the robustness of recognition. Next, in the result analysis stage (S300), the constructed contextual information and the pre-trained language large model can be used to perform semantic analysis on multiple recognition results. The contextual information is used to verify the rationality of the identifiers, further enhancing the accuracy of the recognition results. Finally, in the determination stage (S400), the probability value of the identifier's existence is calculated based on multiple analysis results, and a comprehensive determination is made in conjunction with a preset determination strategy. This comprehensive determination method based on probability values ​​and determination strategies not only improves the flexibility of determination but also adapts to different types of documents through mechanisms such as dynamic threshold adjustment, further enhancing the accuracy and reliability of recognition. Overall, the document identifier recognition method in this embodiment effectively solves the limitations of traditional methods in complex documents and noisy environments through multi-model fusion and multi-step optimization, achieving high-efficiency and high-accuracy document identifier recognition.

[0030] Understandably, the document identifier recognition method provided in this embodiment of the invention can be applied to document identifier recognition scenarios in the medical and health field. The following is a specific example: Scenario: A hospital's electronic medical record system needs to digitize paper medical records and ensure their authenticity. The doctor's signature and the hospital's seal on the medical record are crucial for verifying its authenticity.

[0031] Application: Using the document identifier recognition method of this invention, the uploaded medical record document is first preprocessed to remove scanning noise and standardize the document format. Next, a pre-trained visual model is used to recognize the seal in the medical record multiple times, generating multiple recognition results. Then, using the constructed context information and a pre-trained language model, semantic analysis is performed on the multiple recognition results sequentially, generating multiple analysis results. Finally, based on these analysis results, the probability value of the seal's existence is calculated, and the authenticity of the seal is determined according to a preset judgment strategy. If the seal is determined to be authentic, the medical record will be marked as reviewed and stored in the electronic medical record system; if the seal is determined to be forged or missing, the system will prompt for manual review.

[0032] Results: This method allows hospitals to significantly improve the efficiency and accuracy of medical record management, reduce the workload of manual review, and ensure the authenticity and completeness of medical records.

[0033] Understandably, the document identifier recognition method provided in this embodiment of the invention can also be applied to document identifier recognition scenarios related to the financial technology field. The following is a specific example: Scenario: Financial institutions need to manage a large number of financial contracts, and ensuring the authenticity of these contracts is a crucial aspect of contract management. The financial institution's seal on the contract is key to verifying its authenticity.

[0034] Application: Using the document identifier recognition method of this invention, the uploaded contract document is first preprocessed to remove scanning noise and standardize the document format. Next, a pre-trained visual model is used to recognize the seal in the contract multiple times, generating multiple recognition results. Then, using the constructed contextual information and a pre-trained language model, semantic analysis is performed on the multiple recognition results sequentially, generating multiple analysis results. Finally, based on these analysis results, the probability value of the seal's presence is calculated, and the authenticity of the seal is determined according to a preset judgment strategy. If the seal is determined to be authentic, the contract will be marked as signed and stored in the electronic contract management system; if the seal is determined to be forged or missing, the system will prompt for manual review.

[0035] Results: This method allows financial institutions to significantly improve the efficiency and accuracy of contract management, reduce the workload of manual review, and ensure the authenticity and completeness of contracts.

[0036] The document identifier recognition method of this invention has broad application prospects in the fields of healthcare and fintech. By combining a multi-step processing flow of large visual and large language models, this method not only improves the efficiency of document identifier recognition but also significantly enhances its accuracy. In the healthcare field, this method can effectively manage medical records, ensuring their authenticity and integrity; in the fintech field, it can effectively manage financial contracts, prevent financial fraud, and improve the automation and security of financial contract management.

[0037] Furthermore, in one embodiment, the document identifier recognition method, wherein obtaining the document to be recognized and preprocessing the document to be recognized to obtain the target document specifically includes the following steps: Receive the document to be identified uploaded by the client and perform format verification on the document to be identified; If the format verification passes, noise removal and content standardization are performed on the document to be identified to generate the target document.

[0038] In practice, the specific implementation process of the steps in this embodiment is roughly as follows: (1) Receive the document to be identified 1. Document Upload: Receive user-uploaded documents for recognition via a client (such as a web interface, mobile application, or file transfer interface). Supports multiple file formats, including PDF, Word documents, and image files (such as JPEG and PNG).

[0039] 2. Document Storage: Uploaded documents are temporarily stored in a specified directory on the server for subsequent processing.

[0040] (2) Format validation 1. File Type Check: Verifies whether the uploaded document conforms to the supported file format. Checks the file extension (such as .pdf, .docx, .jpg, etc.) to quickly determine the file type.

[0041] 2. File Size Check: Ensure the document size is within the preset limit (e.g., maximum file size is 10MB). If the file size exceeds the limit, return an error message and prompt the user to re-upload.

[0042] 3. Format Compliance Check: For documents in specific formats (such as PDF or Word), further check whether the internal structure of the file is complete. For example, check whether a PDF file contains valid page information and whether a Word document contains readable text content.

[0043] (3) Noise Removal 1. Image Document Processing: If the document is an image file (such as a scanned document), use image processing algorithms (such as median filtering and Gaussian filtering) to remove noise. These algorithms can effectively remove random noise generated during the scanning process and improve image quality.

[0044] 2. Document background processing: For documents with complex backgrounds, use background segmentation algorithms (such as threshold-based segmentation or deep learning models) to separate the document content from the background, thereby enhancing the clarity of the document content.

[0045] (4) Content standardization 1. Unified text encoding: For documents containing text content (such as PDF or Word documents), unify the text encoding to UTF-8 format to ensure that there will be no encoding problems during subsequent processing.

[0046] 2. Formatting: Adjust the document's formatting to conform to preset standards. For example, align all text lines, adjust font size and style, and ensure consistency in document content.

[0047] 3. Image resizing: For image files, adjust the image size and resolution to meet the input requirements of the pre-trained model. For example, uniformly adjust the image resolution to 300 dpi and the size to A4 paper size (210mm × 297mm).

[0048] (5) Generate the target document 1. Document Integration: Integrate the documents that have undergone noise removal and content standardization into a target document. For image files, save them as high-quality image formats (such as PNG); for text files, save them as documents in a uniform format (such as PDF).

[0049] 2. Document storage and tagging: Store the target document in the processing directory on the server and tag it as "preprocessed" for subsequent identifier recognition steps.

[0050] Through the above steps, this embodiment can efficiently receive, verify, and preprocess the document to be identified, ensuring the quality and consistency of the target document and providing a good foundation for subsequent identifier recognition.

[0051] Furthermore, in one embodiment, the document identifier recognition method, wherein the step of using a pre-trained large visual model to recognize the identifier of the target document multiple times and generate corresponding multiple recognition results specifically includes the following steps: Preload pre-trained large visual models; Based on the input requirements of the aforementioned visual large model, data augmentation processing is performed on the target document; Using the aforementioned large visual model, the target document after data augmentation is subjected to multiple identifier recognitions, generating multiple corresponding recognition results.

[0052] In practice, the specific implementation process of the steps in this embodiment is roughly as follows: (1) Preload pre-trained large visual model 1. Model Selection: Based on the type of the target document and the characteristics of the identifier, select a suitable pre-trained large-scale visual model. For example, for seal recognition, a convolutional neural network (CNN) model trained on a large number of seal images can be selected.

[0053] 2. Model Loading: Load the pre-trained large-scale vision model in the computing environment (such as a server-side or local computing environment). Ensure that the model's parameters and structure are loaded correctly for subsequent identifier recognition tasks.

[0054] 3. Model Configuration: Adjust the model parameters (such as input size, color channels, etc.) according to the characteristics of the target document to optimize the model's recognition performance.

[0055] (2) Perform data augmentation processing on the target document according to the input requirements of the visual big data model. 1. Input Requirements Analysis: Analyze the input requirements of the pre-trained large-scale visual model, including the size, color channels, and data format of the input image.

[0056] 2. Data Augmentation Strategy Selection: Based on the model's input requirements and the characteristics of the target document, select an appropriate data augmentation strategy. Common data augmentation methods include random cropping, rotation, scaling, and color adjustment.

[0057] 3. Data Augmentation: Perform data augmentation on the target document to generate multiple variants. For example, randomly crop and rotate image documents to simulate different recognition scenarios and improve the robustness of the model.

[0058] (3) Multiple identifier recognition through large visual models 1. Identification Task Initialization: Initialize the identifier identification task, set the number of identification attempts (e.g., 3 or 5), and prepare multiple document variants after data augmentation.

[0059] 2. Multiple Recognition Executions: Utilizing a pre-trained large-scale visual model, the data-augmented target document undergoes multiple identifier recognition operations. Each recognition can employ different data augmentation variants to generate diverse recognition results.

[0060] 3. Result Collection: Collect the results of each recognition, including the location, content, and confidence level of the identifier. Store these results as multiple recognition results for subsequent analysis and fusion.

[0061] Through the steps described above, this embodiment can efficiently utilize a pre-trained large visual model to perform multiple identifier recognitions on the target document, generating a variety of recognition results. This method not only improves the robustness and accuracy of recognition but also provides a rich data foundation for subsequent analysis and verification.

[0062] Further, in one embodiment, the document identifier recognition method, wherein the step of constructing contextual information of the multiple recognition results based on the content of the target document, and using the contextual information and a pre-trained language model to sequentially perform semantic analysis on the multiple recognition results to generate corresponding multiple analysis results, specifically includes the following steps: Preload pre-trained large language models; Extract text content related to the identifier from the target document, and generate context information corresponding to each recognition result based on the extracted text content; Using the contextual information and the language model, semantic analysis is performed on each recognition result in turn to generate multiple corresponding analysis results.

[0063] In practice, the specific implementation process of the steps in this embodiment is roughly as follows: (1) Preload pre-trained large language model 1. Model Selection: Based on the type of the target document and the characteristics of the identifiers, select a suitable pre-trained large language model. For example, for documents containing text content, a Transformer model trained on a large amount of text data (such as BERT, GPT, etc.) can be selected.

[0064] 2. Model Loading: Load the pre-trained large language model in the computing environment (such as a server-side or local computing environment). Ensure that the model's parameters and structure are loaded correctly for subsequent result analysis tasks.

[0065] 3. Model configuration: Adjust the model parameters (such as context window size, attention mechanism weights, etc.) according to the characteristics of the target document to optimize the model's analysis performance.

[0066] (2) Construct contextual information for multiple recognition results based on the content of the target document. 1. Content Extraction: Extracting key text content from the target document, including text information, document structure (such as headings, paragraphs, tables, etc.), and the contextual location of identifiers. For example, extracting the text information of the paragraph containing the identifier, as well as the logical relationship between the identifier and surrounding text.

[0067] 2. Context Information Construction: Based on the extracted text content, context information is constructed for each recognition result. Context information can include the text fragments before and after the identifier, the topic of the paragraph, the document category, etc. For example, if the identifier is a stamp, the context information can include the title of the page containing the stamp and the paragraph content.

[0068] 3. Information Formatting: Format the constructed contextual information into an input format that the language model can process. For example, combine the contextual information and recognition results into a structured text sequence as input to the model.

[0069] (3) Using contextual information and a large language model, perform result analysis on multiple recognition results in sequence. 1. Analysis Task Initialization: Initialize the result analysis task, set the analysis order (e.g., from high to low confidence of the recognition results) and analysis parameters (e.g., context window size).

[0070] 2. Successive Analysis: Using a pre-trained language model, each recognition result is analyzed sequentially. During each analysis, the semantic reasonableness and logical consistency of the recognition result are evaluated in conjunction with contextual information. For example, it checks whether the stamp appears in a reasonable document location (such as a footer or signature area) and whether the stamp is related to the surrounding text content.

[0071] 3. Analysis Result Generation: Based on the model's output, analysis results are generated for each recognition result. These results may include semantic evaluation of the identifier, confidence adjustment, and contextual relevance scoring. For example, if the model determines that the seal is highly relevant to the context, its confidence level is increased; if it determines that the seal does not match the context, its confidence level is decreased.

[0072] Through the steps described above, this embodiment can efficiently utilize a pre-trained large language model to perform detailed analysis on multiple recognition results, generating a variety of diverse analysis outcomes. This method not only improves the accuracy and reliability of the analysis but also provides a rich data foundation for subsequent judgments.

[0073] Further, in one embodiment, the document identifier recognition method, wherein calculating the probability value of the existence of an identifier in the document to be identified based on the plurality of analysis results, and determining whether an identifier exists in the document to be identified according to the probability value and a preset judgment strategy, specifically includes the following steps: A strategy for determining whether an identifier exists in the document to be identified is pre-constructed; Based on the multiple analysis results, calculate the number of times C1 is found that the document to be identified contains an identifier, and calculate the total number of times C2 is performed on the document to be identified to identify the identifier. Then the probability value of the document to be identified containing an identifier is C1 / C2. Based on the probability value and the judgment strategy, it is determined whether an identifier exists in the document to be identified.

[0074] Furthermore, in the document identifier recognition method, the judgment strategy includes a preset threshold, and the step of determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy specifically includes the following steps: When the probability value is greater than the preset threshold, it is determined that an identifier exists in the document to be identified; When the probability value is less than or equal to the preset threshold, it is determined that there is no identifier in the document to be identified.

[0075] Furthermore, the document identifier recognition method, after determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy, further includes the following steps: The generated identifier determination results are formatted. According to the preset judgment result display strategy, the formatted identifier judgment result is sent to the client for display.

[0076] In practice, the specific implementation process of the steps in this embodiment is roughly as follows: (1) Pre-construct judgment strategy 1. Strategy Definition: Based on the application scenario and user needs, predefine the judgment strategy. The judgment strategy includes preset thresholds (such as 0.8 for high confidence, 0.5 for medium confidence, etc.) and possible contextual relevance evaluation rules (such as whether the identifier appears in a specific location or is associated with specific text).

[0077] 2. Threshold Setting: Set an appropriate threshold based on the document type and identifier characteristics. For example, for medical records in the healthcare field, a higher threshold (such as 0.9) can be set to ensure high accuracy of the recognition results; for invoices in the fintech field, a medium threshold (such as 0.7) can be set.

[0078] 3. Strategy storage: Store the decision-making strategy in the system so that it can be called in subsequent steps.

[0079] (2) Calculate the probability value 1. Statistical Recognition Results: Based on multiple analysis results, calculate the number of times the identifier was successfully recognized (C1) and the total number of identifier recognitions (C2). For example, if 5 identifier recognition attempts were performed, and the identifier was successfully recognized 3 times, then C1=3 and C2=5.

[0080] 2. Probability Calculation: Calculate the probability of the identifier existing, P = C1 / C2. In the example above, the probability P = 3 / 5 = 0.6, indicating that the identifier has a 60% chance of existing.

[0081] 3. Record probability values: Record the calculated probability values ​​in the system for subsequent judgment and analysis.

[0082] (3) Make a judgment based on the probability value and the judgment strategy. 1. Threshold Comparison: The calculated probability value P is compared with a preset threshold. If P is greater than the preset threshold, the identifier is determined to exist; if P is less than or equal to the preset threshold, the identifier is determined to not exist.

[0083] 2. Contextual Relevance Verification (Optional): For high-confidence probability values, further examine the contextual relevance of the identifier. Utilize a large language model to analyze the semantic position and logical relationships of the identifier in the document, ensuring logical consistency between the identifier and the document content.

[0084] 3. Record the judgment results: Record the judgment results in detail, including probability values, judgment conclusions (identifier exists or does not exist), contextual relevance assessment results, etc., for subsequent review and analysis.

[0085] (4) Format the generated identifier determination results. 1. Results Compilation: Compile the judgment results into a unified format, including information such as the location and content of the identifier, confidence level, and contextual relevance score.

[0086] 2. Formatting: Convert the processed results into a format suitable for client display, such as JSON, XML, or HTML.

[0087] (5) According to the preset judgment result display strategy, send the formatted identifier judgment result to the client for display. 1. Display Strategy Configuration: Configure the display strategy for the judgment results based on user needs and application scenarios. For example, decide where to display the results on the client side, and whether to highlight the identifier position, etc.

[0088] 2. Result Transmission: The formatted identifier determination result is sent to the client over the network. The client can be a web interface, mobile application, or desktop software.

[0089] 3. Result Display: The identifier determination results are displayed on the client side according to a preset result display strategy. For example, in the web interface, the location of the identifier in the document can be highlighted, and detailed information about the identifier (such as location, content, confidence level, etc.) can be displayed.

[0090] Through the above steps, this embodiment can not only accurately calculate and determine the probability value of the existence of the identifier, but also display the identifier determination result to the user in a user-friendly way, thereby improving the system's usability and user experience.

[0091] As can be seen from the above method embodiments, the document identifier recognition method provided by the present invention includes: acquiring a document to be identified; preprocessing the document to be identified to obtain a target document; using a pre-trained visual large model to identify the identifiers of the target document multiple times, generating multiple corresponding recognition results; constructing context information of the multiple recognition results based on the content of the target document, and using the context information and a pre-trained language large model to sequentially perform semantic analysis on the multiple recognition results, generating multiple corresponding analysis results; calculating the probability value of the existence of identifiers in the document to be identified based on the multiple analysis results, and determining whether identifiers exist in the document to be identified based on the probability value and a preset judgment strategy. Thus, the method of the present invention can effectively improve the efficiency and accuracy of document identifier recognition.

[0092] It should be understood that although this application provides the method operation steps as described in the embodiments or flowcharts, conventional or non-inventive labor may include more or fewer operation steps, and these operation steps are not necessarily executed sequentially according to the order of the embodiments or flowcharts. The order of steps listed in the embodiments or flowcharts is merely one way of executing many steps and does not represent the only execution order. It should be noted that there is no necessary sequential order between the above steps. Those skilled in the art can understand from the description of the embodiments of the present invention that the above steps may have different execution orders in different embodiments, that is, they may be executed in parallel or in exchange, etc. Moreover, at least some steps in the embodiments or flowcharts may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed in turn, alternately, or synchronously with other steps or at least a part of the sub-steps or stages of other steps.

[0093] Based on the above method embodiments, please refer to Figure 3 Another embodiment of the present invention also provides a document identifier recognition device, wherein the device includes: The acquisition module 11 is used to acquire the document to be identified, preprocess the document to be identified, and obtain the target document; The recognition module 12 is used to recognize the identifiers of the target document multiple times using a pre-trained large visual model, and generate multiple corresponding recognition results. Analysis module 13 is used to construct context information of the multiple recognition results based on the content of the target document, and use the context information and a pre-trained language model to perform semantic analysis on the multiple recognition results in sequence to generate corresponding multiple analysis results; The determination module 14 is used to calculate the probability value of the existence of an identifier in the document to be identified based on the multiple analysis results, and to determine whether the identifier exists in the document to be identified according to the probability value and a preset determination strategy.

[0094] Furthermore, in one embodiment, the document identifier recognition device, wherein acquiring the document to be recognized and preprocessing the document to be recognized to obtain the target document specifically includes: Receive the document to be identified uploaded by the client and perform format verification on the document to be identified; If the format verification passes, noise removal and content standardization are performed on the document to be identified to generate the target document.

[0095] Furthermore, in one embodiment, the document identifier recognition device, wherein the step of using a pre-trained large visual model to repeatedly recognize the identifier of the target document and generate multiple corresponding recognition results specifically includes: Preload pre-trained large visual models; Based on the input requirements of the aforementioned visual large model, data augmentation processing is performed on the target document; Using the aforementioned large visual model, the target document after data augmentation is subjected to multiple identifier recognitions, generating multiple corresponding recognition results.

[0096] Further, in one embodiment, the document identifier recognition device, wherein constructing context information of the plurality of recognition results based on the content of the target document, and using the context information and a pre-trained language model to sequentially perform semantic analysis on the plurality of recognition results to generate corresponding plurality of analysis results, specifically includes: Preload pre-trained large language models; Extract text content related to the identifier from the target document, and generate context information corresponding to each recognition result based on the extracted text content; Using the contextual information and the language model, semantic analysis is performed on each recognition result in turn to generate multiple corresponding analysis results.

[0097] Further, in one embodiment, the document identifier recognition device, wherein calculating the probability value of the existence of an identifier in the document to be identified based on the plurality of analysis results, and determining whether an identifier exists in the document to be identified according to the probability value and a preset judgment strategy, specifically includes: A strategy for determining whether an identifier exists in the document to be identified is pre-constructed; Based on the multiple analysis results, calculate the number of times C1 is found that the document to be identified contains an identifier, and calculate the total number of times C2 is performed on the document to be identified to identify the identifier. Then the probability value of the document to be identified containing an identifier is C1 / C2. Based on the probability value and the judgment strategy, it is determined whether an identifier exists in the document to be identified.

[0098] Furthermore, in the document identifier recognition device, the judgment strategy includes a preset threshold, and the step of determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy specifically includes: When the probability value is greater than the preset threshold, it is determined that an identifier exists in the document to be identified; When the probability value is less than or equal to the preset threshold, it is determined that there is no identifier in the document to be identified.

[0099] Furthermore, in the document identifier recognition device, after determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy, the method further includes: The generated identifier determination results are formatted. According to the preset judgment result display strategy, the formatted identifier judgment result is sent to the client for display.

[0100] It should be noted that, in the device embodiments of the present invention, the information interaction and execution process between the above modules are based on the same concept as in the method embodiments of the present invention. For details on their specific functions and the resulting technical effects, please refer to the aforementioned method embodiments section, which will not be repeated here.

[0101] Based on the above method embodiments, another embodiment of the present invention also provides a computer device, which can be a server, and its internal structure diagram can be as follows. Figure 4 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the document identifier recognition method server-side as described in any of the above method embodiments.

[0102] Based on the above method embodiments, another embodiment of the present invention also provides a computer device, which can be a client, and its internal structure diagram can be as follows. Figure 5 As shown, the computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program implements the functions or steps of the document identifier recognition method on the client side as described in any of the above method embodiments.

[0103] Those skilled in the art will understand that Figure 4 and Figure 5The structural schematic diagram shown is only a schematic diagram of a part of the structure related to the present invention and does not constitute a limitation on the computer device on which the present invention is applied. The specific computer device may include more components than shown in the figure, or combine certain components, or have different component arrangements.

[0104] The processor referred to herein can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0105] The memory includes readable storage media, internal memory, etc., where internal memory can be the RAM of a computer device. Internal memory provides an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of the computer device, or in other embodiments, it can be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory can include both internal storage units and external storage devices of the computer device. The memory is used to store the operating system, applications, bootloader, data, and other programs, such as program code for computer programs. The memory can also be used to temporarily store data that has been output or will be output.

[0106] Based on the above method embodiments, another embodiment of the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the document identifier recognition method as described in any of the above method embodiments. The computer-readable storage medium may be non-volatile or volatile.

[0107] It should be noted that the functions or steps that can be achieved by the computer-readable storage medium or computer device, and the technical effects brought about by the functions / steps, can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0108] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc. The disclosed memory components or memories of the operating environment described herein are intended to include one or more of these and / or any other suitable types of memory.

[0109] Those skilled in the art will understand that, for the sake of convenience and brevity, the embodiments of the device of the present invention are only illustrated by the division of the above-mentioned functional units and modules. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of the present invention. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0110] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0111] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0112] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0113] It should be noted that if any software tools or components not belonging to this company appear in the embodiments of this application, they are merely illustrative examples and do not represent actual use. The above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A document identifier recognition method, characterized in that, include: Obtain the document to be identified, and preprocess the document to be identified to obtain the target document; Using a pre-trained large visual model, the identifiers of the target document are identified multiple times, generating multiple corresponding recognition results; Based on the content of the target document, contextual information of the multiple recognition results is constructed, and semantic analysis is performed on the multiple recognition results sequentially using the contextual information and a pre-trained language model to generate corresponding multiple analysis results; Based on the multiple analysis results, the probability value of the document to be identified containing an identifier is calculated, and the existence of an identifier in the document to be identified is determined according to the probability value and a preset judgment strategy.

2. The document identifier recognition method according to claim 1, characterized in that, The step of obtaining the document to be identified and preprocessing the document to be identified to obtain the target document includes: Receive the document to be identified uploaded by the client and perform format verification on the document to be identified; If the format verification passes, noise removal and content standardization are performed on the document to be identified to generate the target document.

3. The document identifier recognition method according to claim 1, characterized in that, The method utilizes a pre-trained large visual model to repeatedly identify the identifiers of the target document, generating multiple corresponding recognition results, including: Preload pre-trained large visual models; Based on the input requirements of the aforementioned visual large model, data augmentation processing is performed on the target document; Using the aforementioned large visual model, the target document after data augmentation is subjected to multiple identifier recognitions, generating multiple corresponding recognition results.

4. The document identifier recognition method according to claim 1, characterized in that, The step involves constructing contextual information for the multiple recognition results based on the content of the target document, and then using the contextual information and a pre-trained language model to sequentially perform semantic analysis on the multiple recognition results, generating corresponding multiple analysis results, including: Preload pre-trained large language models; Extract text content related to the identifier from the target document, and generate context information corresponding to each recognition result based on the extracted text content; Using the contextual information and the language model, semantic analysis is performed on each recognition result in turn to generate multiple corresponding analysis results.

5. The document identifier recognition method according to claim 1, characterized in that, Based on the multiple analysis results, the probability value of the document to be identified containing an identifier is calculated, and the existence of the identifier in the document to be identified is determined according to the probability value and a preset judgment strategy, including: A strategy for determining whether an identifier exists in the document to be identified is pre-constructed; Based on the multiple analysis results, calculate the number of times C1 is found that the document to be identified contains an identifier, and calculate the total number of times C2 is performed on the document to be identified to identify the identifier. Then the probability value of the document to be identified containing an identifier is C1 / C2. Based on the probability value and the judgment strategy, it is determined whether an identifier exists in the document to be identified.

6. The document identifier recognition method according to claim 5, characterized in that, The judgment strategy includes a preset threshold. The step of determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy includes: When the probability value is greater than the preset threshold, it is determined that an identifier exists in the document to be identified; When the probability value is less than or equal to the preset threshold, it is determined that there is no identifier in the document to be identified.

7. The document identifier recognition method according to claim 5, characterized in that, After determining whether an identifier exists in the document to be identified based on the probability value and the judgment strategy, the method further includes: The generated identifier determination results are formatted. According to the preset judgment result display strategy, the formatted identifier judgment result is sent to the client for display.

8. A document identifier recognition device, characterized in that, include: The acquisition module is used to acquire the document to be identified, preprocess the document to be identified, and obtain the target document; The recognition module is used to recognize the identifiers of the target document multiple times using a pre-trained large visual model, and generate multiple corresponding recognition results. The analysis module is used to construct contextual information of the multiple recognition results based on the content of the target document, and use the contextual information and a pre-trained language model to perform semantic analysis on the multiple recognition results in sequence to generate corresponding multiple analysis results; The determination module is used to calculate the probability value of the existence of an identifier in the document to be identified based on the multiple analysis results, and to determine whether the document to be identified contains an identifier according to the probability value and a preset determination strategy.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the document identifier recognition method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the document identifier recognition method as described in any one of claims 1-7.

Images