Resume document recognition method and apparatus, computing device, and storage medium

Abstract

The application provides a resume document identification method and device, a computing device and a storage medium. The resume document identification method comprises: obtaining a target document to be identified, and extracting document information from the target document; performing resume document identification processing on the target document based on the document information to obtain an identification result of whether the target document is a resume document, wherein the resume document identification processing comprises rule verification and classification identification based on a neural network. Through the above method, the target document is identified by combining the verification rule with the resume identification model, effectively improving the accuracy of the resume document identification. At the same time, since the rule verification is performed in the resume document identification process, the calculation amount of the resume identification model is reduced, thereby improving the efficiency of the resume document identification.

Description

Technical Field

[0001] This application relates to the field of document recognition technology, and in particular to a method for recognizing resume documents. This application also relates to a resume document recognition device, a computing device, and a computer-readable storage medium. Background Technology

[0002] With the development of internet technology, intelligent resume management has been widely used in recruitment, greatly improving the processing efficiency for both job seekers and recruiters. Intelligent resume management is a collective term encompassing resume parsing, job matching, automatic integration of resume resources, intelligent resume deduplication, and resume screening. Resume document recognition is a crucial prerequisite for intelligent resume management; therefore, improving the accuracy of resume document recognition is key to achieving effective intelligent resume management. Summary of the Invention

[0003] In view of this, embodiments of this application provide a resume document recognition method to address the technical deficiencies existing in the prior art. Embodiments of this application also provide a resume document recognition device, a computing device, and a computer-readable storage medium.

[0004] According to a first aspect of the embodiments of this application, a resume document recognition method is provided, including:

[0005] Obtain the target document to be identified;

[0006] Extract document information from the target document;

[0007] Based on document information, the target document is processed for resume document recognition to obtain the recognition result of whether the target document is a resume document. The resume document recognition process includes rule verification and neural network-based classification recognition.

[0008] Optionally, the steps for extracting document information from the target document include:

[0009] Identify the document format of the target document;

[0010] If the target document is in text format, then extract the text content of type string from the target document;

[0011] The steps for performing resume document recognition processing on the target document based on document information to obtain the recognition result of whether the target document is a resume document include:

[0012] The text content is validated using preset resume text validation rules.

[0013] If the rule verification of the text content is successful, the text content is input into the first resume recognition model to obtain the recognition result of whether the target document is a resume document. The first resume recognition model is a text classification model that is pre-trained on a preset neural network using multiple sample resume texts.

[0014] Optionally, after the step of extracting string-type text content from the target document, the above method further includes:

[0015] If extraction fails, the target document's document format will be converted to an image format.

[0016] Optionally, after the step of converting the target document's document format to an image format, the above method further includes:

[0017] Extract image information for each image from the target document;

[0018] The steps for performing resume document recognition processing on the target document based on document information to obtain the recognition result of whether the target document is a resume document include:

[0019] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0020] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0021] Optionally, after performing rule validation on the text content using preset resume text validation rules, the above method further includes:

[0022] If the text content rule validation fails, the target document is determined to be a non-resume document.

[0023] Optionally, the steps of performing rule validation on the text content using preset resume text validation rules include:

[0024] Extract keywords from the text content;

[0025] Each keyword is matched against the core keywords in the preset core keyword library;

[0026] If the number of keywords that successfully match the core keywords reaches the first preset threshold, then the rule verification of the text content is considered successful.

[0027] Optionally, after matching each keyword with core keywords in a preset core keyword library, the method further includes:

[0028] If the number of keywords that successfully match the core keyword does not reach the first preset threshold, then each keyword will be matched with the secondary keywords in the preset secondary keyword library.

[0029] If the number of keywords that successfully match the secondary keyword reaches the second preset threshold, then the rule verification for determining the text content is successful.

[0030] If the number of keywords that successfully match the secondary keyword does not reach the second preset threshold, the rule verification for determining the text content will fail.

[0031] Optionally, the steps for extracting document information from the target document include:

[0032] Identify the document format of the target document;

[0033] If the target document is in image format, then extract the image information of each image from the target document;

[0034] The steps for performing resume document recognition processing on the target document based on document information to obtain the recognition result of whether the target document is a resume document include:

[0035] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0036] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0037] Optionally, the step of identifying the document format of the target document includes:

[0038] Read the filename extension of the target document;

[0039] Determine the document format of the target document based on the document name extension.

[0040] Optionally, after the step of performing resume document recognition processing on the target document based on document information to obtain the recognition result of whether the target document is a resume document, the above method further includes:

[0041] The recognition results are sent to the client for display.

[0042] According to a second aspect of the embodiments of this application, a resume document recognition device is provided, comprising:

[0043] The acquisition module is configured to acquire the target document to be identified.

[0044] The extraction module is configured to extract document information from the target document.

[0045] The resume recognition module is configured to perform resume document recognition processing on the target document based on document information, and obtain the recognition result of whether the target document is a resume document. The resume document recognition processing includes rule verification and neural network-based classification recognition.

[0046] Optionally, the extraction module is further configured as follows:

[0047] Identify the document format of the target document;

[0048] If the target document is in text format, then extract the text content of type string from the target document;

[0049] The resume recognition module is further configured as follows:

[0050] The text content is validated using preset resume text validation rules.

[0051] If the rule verification of the text content is successful, the text content is input into the first resume recognition model to obtain the recognition result of whether the target document is a resume document. The first resume recognition model is a text classification model that is pre-trained on a preset neural network using multiple sample resume texts.

[0052] Optionally, the extraction module is further configured as follows:

[0053] If extraction fails, the target document's document format will be converted to an image format.

[0054] Optionally, the extraction module is also configured to extract image information for each image from the target document;

[0055] The resume recognition module is further configured as follows:

[0056] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0057] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0058] Optionally, the resume recognition module is further configured as follows:

[0059] If the text content rule validation fails, the target document is determined to be a non-resume document.

[0060] Optionally, the resume recognition module is further configured as follows:

[0061] Extract keywords from the text content;

[0062] Each keyword is matched against the core keywords in the preset core keyword library;

[0063] If the number of keywords that successfully match the core keywords reaches the first preset threshold, then the rule verification of the text content is considered successful.

[0064] Optionally, the resume recognition module is further configured as follows:

[0065] If the number of keywords that successfully match the core keyword does not reach the first preset threshold, then each keyword will be matched with the secondary keywords in the preset secondary keyword library.

[0066] If the number of keywords that successfully match the secondary keyword reaches the second preset threshold, then the rule verification for determining the text content is successful.

[0067] If the number of keywords that successfully match the secondary keyword does not reach the second preset threshold, the rule verification for determining the text content will fail.

[0068] Optionally, the extraction module is further configured as follows:

[0069] Identify the document format of the target document;

[0070] If the target document is in image format, then extract the image information of each image from the target document;

[0071] The resume recognition module is further configured as follows:

[0072] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0073] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0074] Optionally, the extraction module is further configured as follows:

[0075] Read the filename extension of the target document;

[0076] Determine the document format of the target document based on the document name extension.

[0077] Optionally, the device further includes:

[0078] The sending module is configured to send the recognition results to the client for display.

[0079] According to a third aspect of the embodiments of this application, a computing device is provided, comprising:

[0080] Memory and processor;

[0081] The memory is used to store computer-executable instructions, and the processor executes the computer-executable instructions to implement the steps of any of the above resume document recognition methods.

[0082] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided that stores computer-executable instructions, which, when executed by a processor, implement the steps of any of the above-described resume document recognition methods.

[0083] The resume document recognition method provided in this application obtains the target document to be recognized, extracts document information from the target document, and performs resume document recognition processing on the target document based on the document information to obtain the recognition result of whether the target document is a resume document. By combining verification rules with a resume recognition model to recognize the target document, the accuracy of resume document recognition is effectively improved. At the same time, since rule verification is performed during the resume document recognition process, the computational load of the resume recognition model is reduced, thereby improving the efficiency of resume document recognition. Attached Figure Description

[0084] Figure 1 This is a schematic diagram of the structure of a resume document recognition system provided in one embodiment of this application;

[0085] Figure 2 This is a flowchart of a resume document recognition method provided in one embodiment of this application;

[0086] Figure 3 This is a flowchart of another resume document recognition method provided in one embodiment of this application;

[0087] Figure 4 This is a flowchart of another resume document recognition method provided in an embodiment of this application;

[0088] Figure 5 This is a flowchart of a rule-based validation method for a target document provided in an embodiment of this application;

[0089] Figure 6This is a flowchart of another method for rule validation of a target document provided in an embodiment of this application;

[0090] Figure 7 This is a flowchart illustrating how to determine the document format of a target document according to an embodiment of this application;

[0091] Figure 8 This is a flowchart of another resume document recognition method provided in an embodiment of this application;

[0092] Figure 9 This is a schematic diagram of the structure of a resume document recognition device provided in one embodiment of this application;

[0093] Figure 10 This is a structural block diagram of a computing device provided in one embodiment of this application. Detailed Implementation

[0094] Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific implementations disclosed below.

[0095] The terminology used in one or more embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the scope of one or more embodiments of this application. The singular forms “a,” “the,” and “the” used in one or more embodiments of this application and in the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” used in one or more embodiments of this application refers to and includes any or all possible combinations of one or more associated listed items.

[0096] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this application, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this application, and similarly, second may also be referred to as first.

[0097] First, the terminology used in one or more embodiments of the present invention will be explained.

[0098] Open Neural Network Exchange (ONNX) format: A standard for representing deep learning models, it is an open file format designed for machine learning. It allows different artificial intelligence frameworks to store model data in the same format and interact with each other.

[0099] This application provides a method for recognizing resume documents. This application also relates to a resume document recognition device, a computing device, and a computer-readable storage medium, which will be described in detail in the following embodiments.

[0100] Figure 1 The diagram shows a structural schematic of a resume document recognition system provided in one embodiment of this application.

[0101] The resume document recognition system of this application includes a training end and an application end. The training end is used to train the model, and the application end is used to provide samples to the training end and receive the trained model from the training end. The model is used to process the target document to be recognized to obtain the recognition result. The application end and the training end can be different applications in the same terminal device or applications in different terminal devices.

[0102] In this embodiment, the resume recognition model can be trained by the following method: the application selects resume document data as positive samples, and then selects document data related to the resume document (such as personal information form, resume description, job application information, etc.) as negative samples. The positive and negative samples are sent to the training end as the training sample set of the model. The training end inputs the positive and negative samples in the sample set into the neural network model. The neural network model that has been trained through multiple iterations and finally reaches the preset threshold is used as the trained model.

[0103] The application can identify the target document using the following method: First, acquire the target document to be identified, extract document information from it, and then perform resume document recognition processing based on this information to determine whether the target document is indeed a resume document. This resume document recognition processing includes rule validation and neural network-based classification (i.e., the trained model mentioned above).

[0104] By combining the verification rules with the resume recognition model, the accuracy of resume document recognition is effectively improved. At the same time, since rule verification is performed during the resume document recognition process, the computational load of the resume recognition model is reduced, thereby improving the efficiency of resume document recognition.

[0105] Figure 2 The flowchart illustrates a resume document recognition method according to an embodiment of this application, which specifically includes the following steps:

[0106] S202. Obtain the target document to be identified.

[0107] The target document to be identified refers to the document for which resume recognition is performed. Its document format can be text, portable document format (PDF), or image format, etc. When there is a need to recognize a resume from a target document, the user can select / input the target document to be recognized on the client side. The executing entity can retrieve the target document, which can be a server or a client.

[0108] S204. Extract document information from the target document.

[0109] After obtaining the target document to be identified, the executing entity extracts the corresponding document information from the target document. The document information refers to the data information set in the target document, such as text set, character set, pixel set, etc.

[0110] There are various ways to extract document information from a target document. For example, neural network models (such as Natural Language Processing (NLP) and Convolutional Neural Network (CNN)) can be used to extract document information from the target document, or unsupervised algorithms (such as Term Frequency-Inverse Document Frequency (TF-IDF)) can be used to extract document information from the target document. This application does not limit the method of extracting document information from the target document.

[0111] In this embodiment, the reason for extracting document information from the target document is that the target document has multiple formats. Different methods are used to identify the resume of the target document in different formats, thereby improving the accuracy of resume document recognition.

[0112] S206. Based on document information, perform resume document recognition processing on the target document to obtain the recognition result of whether the target document is a resume document. The resume document recognition processing includes rule verification and neural network-based classification recognition.

[0113] Resume document recognition and processing includes rule verification and neural network-based classification. Rule verification refers to checking whether the target document conforms to preset verification rules. When the target document does not conform to the preset verification rules, it can be determined that the target document is not a resume document. There are various rules for verifying target documents. For example, similarity comparison can be used to compare the similarity of the target document with preset resume samples, and the similarity comparison results can be used to verify whether the target document conforms to the verification rules. Alternatively, keyword matching can be used to match the target document with a preset keyword list, and the matching results can be used to verify whether the target document conforms to the verification rules.

[0114] Neural network-based classification and recognition refers to using neural network models to classify and recognize target documents. For example, NLP neural network models, CNN neural network models, etc., can be used to classify and recognize target documents, thereby identifying whether the target document is a resume document.

[0115] By combining rule-based validation with neural networks to identify target documents, the accuracy of resume document recognition can be effectively improved.

[0116] By combining the verification rules with the resume recognition model, the accuracy of resume document recognition is effectively improved. At the same time, since rule verification is performed during the resume document recognition process, the computational load of the resume recognition model is reduced, thereby improving the efficiency of resume document recognition.

[0117] In one implementation of this application embodiment, after S206, the resume document recognition method may further include the following step: sending the recognition result to the client for display.

[0118] Sending the recognition results to the client for display means sending the result of whether the target document is a resume document to the client for display.

[0119] There are several ways to send the recognition results to the client for display. One approach is to set up a pop-up window on the client side; when the executing entity receives the recognition results, it sends them to the client, and the client, upon receiving the results, displays a window asking whether the target document is a resume. Alternatively, a display area can be pre-defined on the client side; when the executing entity receives the recognition results, it sends them to the client, and the client, upon receiving the results, directly displays them in the corresponding display area. This application does not impose any restrictions on the display method.

[0120] In this embodiment, the recognition results are sent to the client for display, allowing users to quickly see whether the target document is a resume document, thus improving their work efficiency.

[0121] Figure 3 The flowchart illustrates another resume document recognition method according to an embodiment of this application, which specifically includes the following steps:

[0122] S302. Identify the document format of the target document. If the document format of the target document is text, extract the text content of string type from the target document.

[0123] In this embodiment, the document format of the target document is identified. Since the document format can be text, portable document format (PDF), or image format, different document formats require different processing methods. If the target document is identified as text, string-type text content needs to be extracted from it. Specifically, string extraction methods can be used to extract the text content, that is, to identify the strings in the target document and extract them according to sentences, paragraphs, etc. This extraction method can convert natural language text into string-type text, which is more conducive to machine recognition and improves the recognition efficiency of resume documents.

[0124] In one implementation of this application embodiment, after S302, the resume document recognition method may further include the following steps:

[0125] If extracting text content of type string from the target document fails, the document format of the target document will be converted to an image format.

[0126] Extraction failure refers to the inability to extract string-type text content from a text-formatted target document, or the extracted string content containing garbled characters or missing information.

[0127] In this embodiment, after the step of extracting string-type text content from the target document, due to differences in element styles or the influence of character encoding, sometimes it may be impossible to extract string-type text content from the target document, or the extracted string content may contain garbled characters or be missing. When text content extraction fails, the document format of the target document is converted to an image format.

[0128] After being converted to an image format, Optical Character Recognition (OCR) technology can be used to extract text from the target document. For example, for a PDF document, OCR technology can be used to identify text regions, then the text regions can be segmented into rectangles to break them down into different characters, and then the characters can be classified to identify the text content. Alternatively, resume document recognition can be performed directly based on the image format target document.

[0129] In one implementation of this application, after converting the document format of the target document into an image format, image information of each image can be extracted from the target document. Accordingly, the step of performing resume document recognition processing on the target document based on the document information to obtain a recognition result of whether the target document is a resume document can be specifically implemented through the following steps:

[0130] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0131] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0132] The second resume recognition model is an image classification model obtained by training a preset neural network using multiple sample resume images. The image classification model is saved in ONNX format.

[0133] In one possible implementation of this application, the neural network model can be deployed on a client using the ONNX format, such as on a PC or a mobile device, like a smartphone or tablet. First, the image classification model is trained. For example, resume document data is selected as positive samples, and document data related to the resume document (such as personal information forms, resume descriptions, job application information, etc.) is selected as negative samples. These positive and negative samples form the training sample set for the image recognition model. The sample set is then input into the neural network model. The neural network model that has undergone multiple iterations of training and finally reaches a preset threshold is used as the trained image classification model, and this image classification model is saved in ONNX format.

[0134] The image information of each extracted image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image.

[0135] After obtaining the recognition results of whether each image is a resume image, based on the recognition results of whether each image is a resume image, the preset resume image verification rules are used to obtain the recognition result of whether the target document is a resume document.

[0136] For example, after acquiring the target document to be identified, it is determined that the document format is text. String text content is extracted from the target document. If extraction fails, the document format is converted from text to image format, and image information for each image is extracted. Then, the image information for each image is input into a trained image classification model. If the image classification model outputs "1", it indicates that the image is a resume image; if the image classification model outputs "0", it indicates that the image is not a resume image. After determining whether each image in the target document is a resume image, based on the identification result of each image, a preset resume image verification rule is used to obtain the identification result of whether the target document is a resume document.

[0137] There can be multiple preset rules for verifying resume images. For example, when a target document contains two images, if the first image is a resume image, the target document can be determined to be a resume document; if the first image is not a resume image, the target document can be determined to be a non-resume document. Alternatively, when a target document contains multiple images, a preset threshold can be set. When the number of images identified as resume images in the target document reaches a preset threshold, or when the proportion of images identified as resume images in the target document reaches a preset proportion threshold, the target document is determined to be a resume document.

[0138] In the above implementation, after failing to extract string-type text content from a text-formatted target document, the document format is converted from text to image format. Then, image information for each image is extracted from the target document. This image information is then input into a second resume recognition model to obtain a result indicating whether each image is a resume image. Based on this result, a preset resume image verification rule is used to determine whether the target document is indeed a resume document. This method not only identifies text-formatted target documents as resume documents but also reprocesses text-formatted target documents after text content extraction failures to further identify image-formatted target documents as resume documents. This satisfies the diverse format requirements for resume document recognition and solves the problem of target document recognition failure caused by string content errors during text content extraction. Furthermore, combining verification rules with the resume recognition model improves the accuracy of resume document recognition results. Furthermore, since neural network models saved in ONNX format can run on different platforms and devices and have the advantages of being small and lightweight, saving image classification models in ONNX format makes it easier to deploy them on PCs or mobile devices.

[0139] S304. Verify the text content using preset resume verification rules.

[0140] Resume verification rules refer to the rules used to verify whether a target document is a resume document. These verification rules can be keywords, rule lookup tables, similarity matching algorithms, etc., pre-stored in a database.

[0141] In this embodiment of the application, the resume verification rule is a keyword preset in the database. When the text content in the target document is extracted, the text content is matched with the preset keywords in the database. If the number of keywords in the text content that successfully match the preset keywords in the database reaches a preset threshold, the text content verification is determined to be successful; otherwise, the text content verification fails, thereby realizing the verification of the text content.

[0142] By using preset resume verification rules to verify text content, text content that does not meet the verification rules can be filtered out, thereby reducing the computational load of the subsequent resume recognition model and improving the efficiency of resume document recognition.

[0143] S306. If the text content verification is successful, input the text content into the first resume recognition model to obtain the recognition result of whether the target document is a resume document.

[0144] In this embodiment, the text content is verified using preset resume text verification rules. If verification is successful, the text content is input into a first resume recognition model to obtain a result indicating whether the target document is a resume document. The first resume recognition model is a text classification model pre-trained on a preset neural network model using multiple sample resume texts. The preset neural network model can be a Long Short-Term Memory (LSTM) model, a Deep Convolutional Network (DCNN) model, a Recurrent Neural Network (RNN) model, or a Convolutional Recurrent Neural Network (CRNN) model, etc. When the text content verification is successful, the text content is input into the neural network model (i.e., the trained text classification model). The neural network model classifies and recognizes the input text content, thereby determining whether the target document is a resume document.

[0145] In one possible implementation of this application, the ONNX format can be used to deploy the neural network model on a client, such as on a PC, or on a mobile device, such as a mobile phone or tablet. The ONNX format is an open neural network exchange format, a standard for representing deep learning models. It is an open file format designed for machine learning, which allows different artificial intelligence frameworks to use the same format to store model data and interact with each other. This enables neural network models saved in the ONNX format to run on different platforms and devices, while also being small and lightweight.

[0146] In one implementation of this application embodiment, after S304, the resume document recognition method may further include the following steps:

[0147] If the text content rule validation fails, the target document is determined to be a non-resume document.

[0148] In the above implementation, if the rule verification of the text content fails, it indicates that the target document does not meet the preset resume verification rules, and the target document is directly determined to be a non-resume document.

[0149] The methods described above make extracting string-type text content from target documents more conducive to machine recognition, thus improving the efficiency of resume document recognition. Converting the target document's format to an image format avoids errors in resume document recognition caused by text extraction failures, thereby improving recognition efficiency. Furthermore, combining validation rules with the resume recognition model to recognize target documents effectively improves accuracy while reducing computational load and efficiency by filtering out text content that does not conform to validation rules. Additionally, since neural network models saved in ONNX format can run on different platforms and devices and have the advantages of being small and lightweight, saving image classification models in ONNX format makes them easier to deploy on PCs or mobile devices.

[0150] Figure 4 The flowchart illustrates another resume document recognition method according to an embodiment of this application, which specifically includes the following steps:

[0151] S402. Identify the document format of the target document. If the document format of the target document is an image format, extract the image information of each image from the target document.

[0152] If the target document is identified as being in image format, then the image information of each image in the target document needs to be extracted. Specifically, a pixel extraction method can be used to extract the image information, that is, to identify each pixel in the target document and extract the pixel set one image at a time.

[0153] S404. Input the image information of each image into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is pre-trained on a preset neural network using multiple sample resume images.

[0154] The second resume recognition model is an image classification model obtained by training a preset neural network using multiple sample resume images. The image classification model is saved in ONNX format.

[0155] In one possible implementation of this application, the neural network model can be deployed on a client using the ONNX format, such as on a PC or a mobile device, like a smartphone or tablet. First, the image classification model is trained. For example, resume document data is selected as positive samples, and document data related to the resume document (such as personal information forms, resume descriptions, job application information, etc.) is selected as negative samples. These positive and negative samples form the training sample set for the image recognition model. The sample set is then input into the neural network model. The neural network model that has undergone multiple iterations of training and finally reaches a preset threshold is used as the trained image classification model, and this image classification model is saved in ONNX format.

[0156] The image information of each extracted image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image.

[0157] S406. Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using the preset resume image verification rules.

[0158] After obtaining the recognition results of whether each image is a resume image, based on the recognition results of whether each image is a resume image, the preset resume image verification rules are used to obtain the recognition result of whether the target document is a resume document.

[0159] For example, after acquiring the target document to be identified, if the document format is determined to be image format, the image information of each image in the target document is extracted. Then, the image information of each image is input into a trained image classification model. If the image classification model outputs "1", it indicates that the image is a resume image; if the image classification model outputs "0", it indicates that the image is not a resume image. After determining whether each image in the target document is a resume image, based on the identification result of each image, a preset resume image verification rule is used to obtain the identification result of whether the target document is a resume document.

[0160] There can be multiple preset rules for verifying resume images. For example, when a target document contains two images, if the first image is a resume image, the target document can be determined to be a resume document; if the first image is not a resume image, the target document can be determined to be a non-resume document. Alternatively, when a target document contains multiple images, a preset threshold can be set. When the number of images identified as resume images in the target document reaches a preset threshold, or when the proportion of images identified as resume images in the target document reaches a preset proportion threshold, the target document is determined to be a resume document.

[0161] In the above implementation, the document format of the target document is identified. If the target document is in image format, the image information of each image is extracted from the target document. Then, the image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. Then, based on the recognition result of whether each image is a resume image, a preset resume image verification rule is used to obtain the recognition result of whether the target document is a resume document. Through the above method, not only can text-type target documents be identified as resume documents, but also image-type target documents can be identified as resume documents, meeting the needs of resume document recognition with diverse formats. Furthermore, the combination of verification rules and the resume recognition model improves the accuracy of resume document recognition results. Moreover, since the neural network model saved in ONNX format can run on different platforms and devices and has the advantages of being small and lightweight, saving the image classification model in ONNX format makes it easier to deploy the image classification model on PCs or mobile devices.

[0162] Figure 5 The flowchart illustrates a method for rule validation of a target document according to an embodiment of this application, which specifically includes the following steps:

[0163] S502. Extract keywords from the text content.

[0164] S504. Match each keyword with the core keywords in the preset core keyword library.

[0165] S506 If the number of keywords that successfully match the core keywords reaches the first preset threshold, then the rule verification of the text content is considered successful.

[0166] In this embodiment, the resume verification rules are keywords preset in the database. First, a core keyword library is preset. After extracting the text content from the target document, a keyword recognition model or the TF-IDF method can be used to extract each keyword from the text content. Then, each extracted keyword is matched with the core keywords in the preset core keyword library. If the number of keywords in the text content that successfully match the core keywords in the preset core keyword library reaches a first preset threshold, the text content verification is determined to be successful; otherwise, the text content verification fails, thereby realizing the verification of the text content.

[0167] In addition to extracting keywords from the text content and matching them with core keywords in the preset core keyword library, the entire text content can also be directly matched with core keywords in the preset core keyword library to determine the number of keywords that successfully match the core keywords.

[0168] Table 1 uses resume verification as an example to illustrate the preset core keywords for verifying text content in resume verification.

[0169] Table 1. Core Keyword Database for Resume Verification

[0170]

[0171] Table 1 shows a pre-built core keyword library, which can include preset core keywords such as "basic information, basic information, personal information, personal profile, personal introduction, educational background, educational experience, work background, work experience, professional experience, project experience, internship experience, school experience, club activities, student work, and educational practice".

[0172] In one possible implementation of this application embodiment, the core keyword library corresponds to a first preset threshold, which is preset to "2". Keywords are extracted from the text content of the target document. For example, the extracted keywords are "basic information, internship experience, work experience, hobbies, personality traits". The keywords are then matched with the core keywords in the core keyword library. If the matched keywords are "basic information, internship experience, work experience", then it is determined that the number of keywords in the target document that are matched with the core keyword library is 3, which reaches the first preset threshold. Therefore, it is determined that the text content of the target document has been successfully verified.

[0173] In the above implementation, keywords are extracted from the text content and matched against core keywords in a preset core keyword library. If the number of keywords successfully matched against the core keywords reaches a first preset threshold, the text content is considered successfully verified. This method utilizes preset resume verification rules to filter out text content that does not conform to the verification rules during resume document recognition, thereby reducing the computational load of the resume recognition model and improving the efficiency of resume document recognition.

[0174] Figure 6 The flowchart illustrates another method for rule validation of a target document according to an embodiment of this application, which specifically includes the following steps:

[0175] S602. Match each keyword with the core keywords in the preset core keyword library.

[0176] S604. If the number of keywords that successfully match the core keyword does not reach the first preset threshold, then each keyword will be matched with the secondary keywords in the preset secondary keyword library.

[0177] S606. If the number of keywords that successfully match the secondary keyword reaches the second preset threshold, then the rule verification of the text content is successful.

[0178] S608. If the number of keywords that successfully match the secondary keyword does not reach the second preset threshold, the rule verification for determining the text content will fail.

[0179] In this embodiment, the resume verification rules are keywords preset in the database. First, a secondary keyword library is preset. After extracting the text content from the target document, a keyword recognition model or the TF-IDF method can be used to extract keywords from the text content. Then, each keyword is matched with the core keywords in the preset core keyword library. If the number of keywords successfully matched with the core keywords in the preset core keyword library does not reach a first preset threshold, then each keyword is matched with the secondary keywords in the preset secondary keyword library. If the number of keywords successfully matched with the secondary keywords reaches a second preset threshold, then the rule verification of the text content is determined to be successful. If the number of keywords successfully matched with the secondary keywords does not reach the second preset threshold, then the rule verification of the text content is determined to be unsuccessful. Table 2 schematically lists the preset resume verification secondary keywords used to verify the text content.

[0180] Table 2. Secondary Keyword Database for Resume Verification

[0181]

[0182] Table 2 shows a pre-built secondary keyword library, which contains pre-set secondary keywords such as "entrepreneurial experience, scientific research papers, research direction, academic achievements, awards, professional skills, major courses, on-campus positions, practical research, job intentions, self-evaluation, honors received, skills, and special talents".

[0183] The secondary keyword library corresponds to a second preset threshold, which is a variable threshold. It can be preset as a first variable threshold and a second variable threshold, where the first variable threshold is smaller than the second variable threshold, such as "3" and "5" respectively. The setting of the second preset threshold is related to the number of keywords successfully matched by the core keywords in the preset core keyword library. If the number of successfully matched keywords is large, the second preset threshold can be set smaller; if the number of successfully matched keywords is small, the second preset threshold can be set larger. Specifically, a correspondence between the second preset threshold and the number of successfully matched keywords can be preset. After determining the number of successfully matched keywords, the corresponding second preset threshold can be selected based on this correspondence. In one embodiment, if the number of keywords in the text content that successfully match the core keywords is greater than zero and less than the first preset threshold, and the number that successfully matches the secondary keywords is greater than the first variable threshold, or if the number of keywords in the text content that successfully match the core keywords is zero, and the number that successfully matches the secondary keywords is greater than the second variable threshold, then the rule verification of the text content is determined to be successful. By setting the second preset threshold as a variable threshold, the accuracy of resume recognition can be improved.

[0184] In the first possible implementation of this application embodiment, after extracting the text content of the target document in text format, keywords are extracted from the text content. For example, the extracted keywords are "basic information, extracurricular activities, research direction, academic achievements, major courses". The keywords are then matched with the core keywords in the core keyword library. If the matched keyword is "basic information", it is determined that the number of keywords in the target document that are matched with the core keyword library is 1, which does not reach the first preset threshold. If the matching result does not reach the first preset threshold, the keywords in the target document are matched with the secondary keywords in the secondary keyword library. If the matched keywords are "research direction, academic achievements, major courses", it is determined that the number of keywords in the target document that are matched with the secondary keyword library is 3, which reaches the second preset threshold. Thus, it is determined that the text content of the target document has been successfully verified.

[0185] In a second possible implementation of this application embodiment, after extracting the text content of the target document in text format, keywords are extracted from the text content. For example, if the extracted keywords are "on-campus position, practical research, job intention, self-evaluation, honors received", then it is determined that the number of keywords in the target document that successfully match the core keyword library is 0, which does not reach the first preset threshold. When it is determined that the matching result does not reach the first preset threshold, then each keyword in the target document is matched with the secondary keywords in the secondary keyword library. If the successfully matched keywords are "on-campus position, practical research, job intention, self-evaluation, honors received", then it is determined that the number of keywords in the target document that successfully match the secondary keyword library is 5, which reaches the second preset threshold, and thus it is determined that the text content of the target document has been successfully verified.

[0186] In a third possible implementation of this application embodiment, after extracting the text content of the target document in text format, each keyword is extracted from the text content. For example, if the extracted keywords are "entrepreneurial experience, work display, self-evaluation, communication practice, and language expression", then it is determined that the number of keywords in the target document that successfully match the core keyword library is 0, which does not reach the first preset threshold. When it is determined that the matching result does not reach the first preset threshold, then each keyword in the target document is matched with the secondary keywords in the secondary keyword library. If the successfully matched keyword is "self-evaluation", then it is determined that the number of keywords in the target document that successfully match the secondary keyword library is 1, which does not reach the second preset threshold, and thus it is determined that the text content verification of the target document has failed.

[0187] In the fourth possible implementation of this application embodiment, after extracting the text content of the target document in text format, each keyword is extracted from the text content. For example, if the extracted keywords are "entrepreneurial experience, work display, communication practice, language expression, and logical analysis", then it is determined that the number of keywords in the target document that successfully match the core keyword library is 0, which does not reach the first preset threshold. When it is determined that the matching result does not reach the first preset threshold, then each keyword in the target document is matched with the secondary keywords in the secondary keyword library. If the number of successfully matched keywords is 0, then it is determined that the number of keywords in the target document that successfully match the secondary keyword library does not reach the second preset threshold, and thus it is determined that the text content verification of the target document has failed.

[0188] In the above implementation, keywords are extracted from the text content and matched against core keywords in a preset core keyword library. If the number of keywords successfully matched with core keywords does not reach a first preset threshold, then each keyword is matched against secondary keywords in a preset secondary keyword library. If the number of keywords successfully matched with secondary keywords reaches a second preset threshold, the text content verification is considered successful; otherwise, the text content verification is considered unsuccessful. This method utilizes preset resume verification rules to filter out text content that does not conform to the verification rules during resume document recognition, thereby reducing the computational load of the resume recognition model and improving the efficiency of resume document recognition.

[0189] Figure 7 The flowchart illustrating a method for determining the document format of a target document according to an embodiment of this application is shown, specifically including the following steps:

[0190] S702, Read the file extension of the target document;

[0191] S704. Determine the document format of the target document based on the document name suffix.

[0192] In this embodiment, after the target document to be identified is obtained, the document format of the target document is determined based on the document suffix of the obtained target document.

[0193] In one possible implementation of this application embodiment, the target document to be identified can be in text format, portable document format, or image format, etc. After obtaining the target document to be identified, the document name suffix of the target document is read. If the suffix of the target document is ".txt", ".doc", or ".docx", the document format of the target document is determined to be text format; if the suffix of the target document is ".pdf", the document format of the target document is determined to be PDF format; if the suffix of the target document is ".jpg" or ".png", the document format of the target document is determined to be image format.

[0194] By reading the document extension of the target document, the document format can be determined quickly and accurately, thus improving the efficiency of resume document recognition.

[0195] Figure 8 A flowchart of another resume document recognition method provided according to an embodiment of this application is shown.

[0196] Input the target document to be identified. The document format of the target document is determined based on the document name extension. If the document name extension is ".txt", ".doc", or ".docx", then the target document is determined to be in text format, and string text content is extracted from the text format target document. If the document name extension is ".pdf", then the target document is determined to be in PDF format, and OCR technology is then used to recognize the text content of the target document and extract the string text content.

[0197] After extracting the text content, the text content is validated using preset resume validation rules. If the validation fails, the result is output indicating that the target document is not a resume document. If the validation passes, the text content is input into the trained text classification model for resume document recognition. The text classification model outputs the result indicating whether the target document is a resume document or not.

[0198] If the target document to be identified is in text format but text content extraction fails, or if the document name of the target document to be identified ends in "jpg" or ".png", then the target document to be identified is determined to be in image format. Alternatively, if the target document to be identified is identified as an image format, the image format target document to be identified is directly input into the image classification model to obtain the recognition result of whether each image is a resume image. Then, based on the recognition result of whether each image is a resume image, the image rule verification is performed using the preset resume image verification rules to obtain the recognition result of whether the target document is a resume document.

[0199] By combining the verification rules with the resume recognition model, the accuracy of resume document recognition is effectively improved. At the same time, since target documents that do not conform to the verification rules are filtered out during the resume document recognition process, the computational load of the resume recognition model is reduced, thereby improving the efficiency of resume document recognition.

[0200] Corresponding to the above method embodiments, this application also provides an embodiment of a resume document recognition device. Figure 9 A schematic diagram of a resume document recognition device according to an embodiment of this application is shown. Figure 9 As shown, the device includes:

[0201] Module 902 is configured to acquire the target document to be identified.

[0202] Extraction module 904 is configured to extract document information from the target document;

[0203] The resume recognition module 906 is configured to perform resume document recognition processing on the target document based on document information, and obtain the recognition result of whether the target document is a resume document. The resume document recognition processing includes rule verification and neural network-based classification recognition.

[0204] Optionally, the extraction module 904 is further configured as follows:

[0205] Identify the document format of the target document;

[0206] If the target document is in text format, then extract the text content of type string from the target document;

[0207] The resume recognition module 906 is further configured as follows:

[0208] The text content is validated using preset resume text validation rules.

[0209] If the rule verification of the text content is successful, the text content is input into the first resume recognition model to obtain the recognition result of whether the target document is a resume document. The first resume recognition model is a text classification model that is pre-trained on a preset neural network using multiple sample resume texts.

[0210] Optionally, the extraction module 904 is further configured as follows:

[0211] If extraction fails, the target document's document format will be converted to an image format.

[0212] Optionally, the extraction module 904 is also configured to extract image information for each image from the target document;

[0213] The resume recognition module 906 is further configured as follows:

[0214] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0215] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0216] Optionally, the resume recognition module 906 is further configured as follows:

[0217] If the text content rule validation fails, the target document is determined to be a non-resume document.

[0218] Optionally, the resume recognition module 906 is further configured as follows:

[0219] Extract keywords from the text content;

[0220] Each keyword is matched against the core keywords in the preset core keyword library;

[0221] If the number of keywords that successfully match the core keywords reaches the first preset threshold, then the rule verification of the text content is considered successful.

[0222] Optionally, the resume recognition module 906 is further configured as follows:

[0223] If the number of keywords that successfully match the core keyword does not reach the first preset threshold, then each keyword will be matched with the secondary keywords in the preset secondary keyword library.

[0224] If the number of keywords that successfully match the secondary keyword reaches the second preset threshold, then the rule verification for determining the text content is successful.

[0225] If the number of keywords that successfully match the secondary keyword does not reach the second preset threshold, the rule verification for determining the text content will fail.

[0226] Optionally, the extraction module 904 is further configured as follows:

[0227] Identify the document format of the target document;

[0228] If the target document is in image format, then extract the image information of each image from the target document;

[0229] The resume recognition module 906 is further configured as follows:

[0230] The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images.

[0231] Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

[0232] Optionally, the extraction module 904 is further configured as follows:

[0233] Read the filename extension of the target document;

[0234] Determine the document format of the target document based on the document name extension.

[0235] Optionally, the device further includes:

[0236] The sending module is configured to send the recognition results to the client for display.

[0237] By combining validation rules with a resume recognition model, the accuracy of resume document recognition is effectively improved. At the same time, since target documents that do not conform to the validation rules are filtered out during the resume document recognition process, the computational load of the resume recognition model is reduced, thereby improving the efficiency of resume document recognition.

[0238] The above is an illustrative scheme of a resume document recognition device according to this embodiment. It should be noted that the technical solution of this resume document recognition device and the technical solution of the above-described resume document recognition method belong to the same concept. Details not described in detail in the technical solution of the resume document recognition device can be found in the description of the technical solution of the above-described resume document recognition method. Furthermore, the components in the device embodiment should be understood as functional modules necessary to implement each step of the program flow or each step of the method; these functional modules are not actual functional divisions or separations. The device claims defined by such a set of functional modules should be understood as a functional module architecture that primarily implements the solution through the computer program described in the specification, and not as a physical device that primarily implements the solution through hardware.

[0239] Figure 10 A structural block diagram of a computing device according to an embodiment of this application is shown. The components of the computing device 1000 include, but are not limited to, a memory 1010 and a processor 1020. The processor 1020 is connected to the memory 1010 via a bus 1030, and a database 1050 is used to store data.

[0240] The computing device 1000 also includes an access device 1040, which enables the computing device 1000 to communicate via one or more networks 1060. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 1040 may include one or more of any type of wired or wireless network interface (e.g., Network Interface Card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Wi-MAX (World Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0241] In one embodiment of this application, the aforementioned components of the computing device 1000 and Figure 10 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 10 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this application. Those skilled in the art can add or replace other components as needed.

[0242] The computing device 1000 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or PCs. The computing device 1000 can also be a mobile or stationary server.

[0243] The processor 1020 is used to execute computer-executable instructions for the resume document recognition method.

[0244] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the resume document recognition method described above belong to the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the resume document recognition method described above.

[0245] An embodiment of this application also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, are used for a resume document recognition method.

[0246] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium belongs to the same concept as the technical solution of the resume document recognition method described above. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the resume document recognition method described above.

[0247] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0248] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0249] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0250] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0251] The preferred embodiments disclosed above are merely illustrative of this application. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this application. These embodiments are selected and specifically described in this application to better explain the principles and practical applications of this application, thereby enabling those skilled in the art to better understand and utilize this application. This application is limited only by the claims and their full scope and equivalents.

Claims

1. A method for recognizing resume documents, characterized in that, include: Obtain the target document to be identified; Extracting document information from the target document, the step of extracting document information from the target document includes: identifying the document format of the target document; if the document format of the target document is text format, then extracting string-type text content from the target document; if extraction fails, then converting the document format of the target document to image format, extracting image information from the target document, and inputting the image information into a second resume recognition model to obtain a recognition result of whether the target document is a resume document, wherein the second resume recognition model is an image classification model obtained by pre-training a preset neural network using sample resume images; Based on the document information, the target document is subjected to resume document recognition processing to obtain the recognition result of whether the target document is a resume document. The resume document recognition processing includes verifying the target document through rule verification and classifying and recognizing the target document that successfully passes the rule verification based on neural network.

2. The method according to claim 1, characterized in that, The step of performing resume document recognition processing on the target document based on the document information to obtain a recognition result of whether the target document is a resume document includes: The text content is validated using preset resume text validation rules. If the rule verification of the text content is successful, the text content is input into the first resume recognition model to obtain the recognition result of whether the target document is a resume document. The first resume recognition model is a text classification model that is pre-trained on a preset neural network using multiple sample resume texts.

3. The method according to claim 1, characterized in that, After the step of converting the document format of the target document to an image format, the method further includes: Extract image information for each image from the target document; The step of performing resume document recognition processing on the target document based on the document information to obtain a recognition result of whether the target document is a resume document includes: The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images. Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

4. The method according to claim 2, characterized in that, After the step of performing rule validation on the text content using preset resume text validation rules, the method further includes: If the rule validation of the text content fails, the target document is determined to be a non-resume document.

5. The method according to claim 2 or 4, characterized in that, The step of performing rule validation on the text content using preset resume text validation rules includes: Extract keywords from the text content; Each keyword is matched with a core keyword in a preset core keyword library; If the number of keywords that successfully match the core keyword reaches a first preset threshold, then the rule verification of the text content is determined to be successful.

6. The method according to claim 5, characterized in that, After the step of matching each keyword with core keywords in a preset core keyword library, the method further includes: If the number of keywords that successfully match the core keyword does not reach the first preset threshold, then each keyword will be matched with the secondary keywords in the preset secondary keyword library. If the number of keywords that successfully match the secondary keyword reaches the second preset threshold, then the rule verification of the text content is determined to be successful. If the number of keywords that successfully match the secondary keyword does not reach the second preset threshold, then the rule verification of the text content is determined to have failed.

7. The method according to claim 1, characterized in that, The step of extracting document information from the target document includes: Identify the document format of the target document; If the target document is in image format, then extract the image information of each image from the target document; The step of performing resume document recognition processing on the target document based on the document information to obtain a recognition result of whether the target document is a resume document includes: The image information of each image is input into the second resume recognition model to obtain the recognition result of whether each image is a resume image. The second resume recognition model is an image classification model that is trained on a preset neural network using multiple sample resume images. Based on the recognition results of whether each image is a resume image, the recognition results of whether the target document is a resume document are obtained using preset resume image verification rules.

8. The method according to claim 1 or 7, characterized in that, The step of identifying the document format of the target document includes: Read the filename extension of the target document; The document format of the target document is determined based on the document name suffix.

9. The method according to claim 1, characterized in that, After the step of performing resume document recognition processing on the target document based on the document information to obtain the recognition result of whether the target document is a resume document, the method further includes: The recognition results are sent to the client for display.

10. A resume document recognition device, comprising: The acquisition module is configured to acquire the target document to be identified. An extraction module is configured to extract document information from the target document. The step of extracting document information from the target document includes: identifying the document format of the target document; if the document format of the target document is text format, extracting string-type text content from the target document; if extraction fails, converting the document format of the target document to image format, extracting image information from the target document, and inputting the image information into a second resume recognition model to obtain a recognition result of whether the target document is a resume document. The second resume recognition model is an image classification model obtained by pre-training a preset neural network using sample resume images. The resume recognition module is configured to perform resume document recognition processing on the target document based on the document information, and obtain a recognition result of whether the target document is a resume document. The resume document recognition processing includes verifying the target document through rule verification and identifying the target document that has successfully passed the rule verification through neural network-based classification.

11. A computing device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the steps of the resume document recognition method according to any one of claims 1 to 9.

12. A computer-readable storage medium storing computer instructions, characterized in that, When executed by the processor, this instruction implements the steps of the resume document recognition method according to any one of claims 1 to 9.

Images