Loss assessment method, device and equipment for accidental health insurance and storage medium

Abstract

The invention discloses a loss assessment method, device and equipment for accidental health insurance and a storage medium, and belongs to the technical field of artificial intelligence. Firstly, an image of a to-be-processed invoice is acquired, and the invoice type is identified by using a convolutional neural network. And selecting a corresponding text field extraction rule according to the invoice type, and performing text extraction on the invoice image to obtain specific text information in the invoice. And then, associating the extracted invoice text field with the medical knowledge graph, and identifying a medical service item code and a medicine code. Then, through a preset cost standard, in combination with the codes of the medical service items and the medicines, the item cost amount in the invoice is calculated; and finally, according to the project cost amount and the claim settlement amount limit of the accidental health insurance, evaluating and determining a loss assessment risk. The invention also relates to the technical field of block chains, and the invoice image is stored in a block chain network. According to the method, the loss assessment efficiency is greatly improved, the dependence on professional manpower is reduced, and the insurance case processing speed is remarkably improved.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to a loss assessment method, device, equipment and storage medium for accidental health insurance. Background Art

[0002] In the claims process of accidental health insurance (referred to as "accidental health insurance"), damage assessment, that is, the step of accurately calculating the amount of compensation, is a crucial link. The automatic damage assessment process of cases is a key link in the digital transformation of the modern insurance industry, aiming to significantly improve claims efficiency, reduce operating costs, and ultimately improve customer satisfaction. At present, many property and casualty insurance companies are actively deploying intelligent technologies to innovate the handling of accidental health insurance cases through automation, especially for relatively simple cases, such as claims involving only medical expenses or hospitalization allowances.

[0003] In the traditional claims settlement model, claims specialists have to deal with complicated and ever-changing medical invoices. Invoice formats vary from region to region, and there are also differences in damage assessment rules. This process is time-consuming and prone to errors. What's worse, when the invoice information is incomplete, it is necessary to manually enter the medical insurance settlement form or hospitalization expense details, the latter of which can easily number hundreds. Manual operation is not only inefficient, but also greatly consumes the energy of professionals, limits the speed of case processing, and pushes up operating costs, which in turn affects the overall customer experience. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, computer equipment and storage medium for accidental health insurance damage assessment, so as to solve the technical problems that the existing accidental health insurance damage assessment schemes are inefficient, greatly consume the energy of professionals, limit the speed of case processing, increase operating costs, and thus affect the overall customer experience.

[0005] In order to solve the above technical problems, the embodiment of the present application provides a method for assessing damages of accidental health insurance, which adopts the following technical solution:

[0006] A method for assessing loss of accidental health insurance, comprising:

[0007] Obtaining an image of an invoice to be processed for accidental health insurance, and importing the image of the invoice to be processed into a pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, where the invoice type includes an electronic invoice and a paper invoice;

[0008] Determine a matching text field extraction rule according to the invoice type of the invoice to be processed, and obtain a target extraction rule;

[0009] Use the target extraction rule to perform text extraction on the invoice image to be processed to obtain the invoice text field;

[0010] Associating the invoice text fields with the pre-established medical knowledge graph to identify the medical service item codes and drug codes in the invoices to be processed;

[0011] Obtain the preset fee standards and calculate the item fee amounts of the invoices to be processed based on the medical service item codes and drug codes;

[0012] Determine the loss risk of accidental health insurance based on the project cost amount and claim amount limit.

[0013] Furthermore, the text field extraction rules include text field extraction rules for electronic invoices and text field extraction rules for paper invoices. The steps of extracting text from the invoice image to be processed using the target extraction rules to obtain the invoice text field specifically include:

[0014] When the invoice to be processed is an electronic invoice, extracting the text field of the electronic invoice according to a pre-established electronic invoice template to obtain a first text field;

[0015] When the invoice to be processed is a paper invoice, the text field of the paper invoice is extracted by combining optical character recognition and semantic analysis to obtain a second text field;

[0016] The first text field and the second text field are combined to obtain the invoice text field.

[0017] Furthermore, before obtaining the invoice image to be processed of the accidental health insurance and importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, the method further includes:

[0018] Obtain the original invoice image, and use image preprocessing technology to remove noise and enhance the original invoice image;

[0019] For the preprocessed original invoice image, the image quality assessment algorithm is used to calculate the image clarity index and integrity index to obtain the image quality assessment result;

[0020] Compare the image quality assessment result with the preset image quality threshold to determine whether the original invoice image meets the quality requirements;

[0021] If the image quality assessment result is lower than the image quality threshold, a prompt message is output to prompt the user to re-acquire the invoice image;

[0022] If the image quality assessment result is higher than or equal to the image quality threshold, the original invoice image is output as the invoice image to be processed and saved.

[0023] Furthermore, the convolutional neural network is preset with the layout features of the electronic invoice and the paper invoice, and the step of importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed specifically includes:

[0024] Inputting the invoice image to be processed into a convolutional neural network to extract features of the invoice image to be processed, thereby obtaining invoice image features;

[0025] Compare the invoice image features with the layout features of the electronic invoice and the layout features of the paper invoice, and calculate the feature similarity score;

[0026] Based on the feature similarity scores, the invoice type of the invoice to be processed is identified.

[0027] Furthermore, the step of identifying the invoice type of the invoice to be processed according to the feature similarity score specifically includes:

[0028] Determine the relationship between the feature similarity score and a preset first similarity threshold;

[0029] If the feature similarity score between the invoice image feature and the format feature of the electronic invoice exceeds a preset first similarity threshold, the invoice to be processed is determined to be an electronic invoice;

[0030] If the feature similarity score between the invoice image feature and the format feature of the paper invoice exceeds a preset first similarity threshold, the invoice to be processed is determined to be a paper invoice.

[0031] Furthermore, the steps of associating the invoice text field with the pre-established medical knowledge graph and identifying the medical service item codes and drug codes in the invoice to be processed specifically include:

[0032] By performing natural language processing on the invoice text field, the names of medical service items and medicines in the invoice text field are obtained;

[0033] The names of medical service items are matched with the standard item names in the medical knowledge graph to calculate the item name similarity;

[0034] If the similarity of the project name exceeds a preset second similarity threshold, the code corresponding to the standard project name is obtained as the medical service project code;

[0035] Match the drug names with the standard drug names in the medical knowledge graph to calculate the drug name similarity;

[0036] If the similarity of the drug name exceeds a preset second similarity threshold, the code corresponding to the standard drug name is obtained as the drug code.

[0037] Furthermore, the steps of determining the loss risk of accidental health insurance based on the project cost amount and the claim amount limit specifically include:

[0038] If the project cost amount is greater than the claim amount limit, the accident health insurance loss risk is determined to be high risk, and a high risk warning message is output;

[0039] If the project cost amount is less than or equal to the claim amount limit, the loss risk of accidental health insurance is determined to be low risk.

[0040] In order to solve the above technical problems, the embodiment of the present application also provides a loss assessment device for accidental health insurance, which adopts the following technical solution:

[0041] A loss assessment device for accidental health insurance, comprising:

[0042] An invoice type recognition module is used to obtain an image of an invoice to be processed for accidental health insurance, and import the image of the invoice to be processed into a pre-trained convolutional neural network to recognize the invoice type of the invoice to be processed, where the invoice type includes an electronic invoice and a paper invoice;

[0043] An extraction rule matching module, used for determining a matching text field extraction rule according to the invoice type of the invoice to be processed, and obtaining a target extraction rule;

[0044] A text field extraction module, used for performing text extraction on the invoice image to be processed using a target extraction rule to obtain the invoice text field;

[0045] The knowledge graph association module is used to associate the invoice text field with the pre-established medical knowledge graph to identify the medical service item codes and drug codes in the invoice to be processed;

[0046] The fee amount calculation module is used to obtain the preset fee standard and calculate the item fee amount of the invoice to be processed according to the medical service item code and drug code;

[0047] The loss risk assessment module is used to determine the loss risk of accidental health insurance based on the project cost amount and the claim amount limit.

[0048] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0049] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the loss assessment method for accidental health insurance as described in any one of the above items are implemented.

[0050] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0051] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the method for assessing loss of accidental health insurance as described in any one of the above.

[0052] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0053] The present application discloses a method, device, equipment and storage medium for assessing accidental health insurance, and belongs to the field of artificial intelligence technology. The solution of the present application aims to solve the problems of low efficiency and waste of professional manpower in existing accidental health insurance assessment solutions. Through automated image recognition and text extraction technology, the efficiency of assessment is greatly improved. Convolutional neural networks are used to identify invoice types, match corresponding text extraction rules, automatically extract key data from invoice images, and automatically identify medical service items and drug codes through medical knowledge graphs to realize the intelligent assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 An exemplary system architecture diagram is shown in which the present application can be applied;

[0056] Figure 2 A flow chart showing an embodiment of a method for assessing damages for accidental health insurance according to the present application;

[0057] Figure 3 A schematic structural diagram of an embodiment of a loss assessment device for accidental health insurance according to the present application is shown;

[0058] Figure 4 A schematic structural diagram of an embodiment of a computer device according to the present application is shown. DETAILED DESCRIPTION

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0060] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0061] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0062] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0063] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0064] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0065] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0066] It should be noted that the damage assessment method for accidental health insurance provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the damage assessment device for accidental health insurance is generally set in the server / terminal device.

[0067] It should be understood that Figure 1 The numbers of terminal devices, networks and servers in the embodiment are merely illustrative. The above system may have any number of terminal devices, networks and servers according to implementation requirements.

[0068] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for assessing damages for accidental health insurance according to the present application. The method for assessing damages for accidental health insurance comprises the following steps:

[0069] S201, obtaining an image of an invoice to be processed for accidental health insurance, and importing the image of the invoice to be processed into a pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, wherein the invoice type includes an electronic invoice and a paper invoice.

[0070] Specifically, first obtain the invoice images to be processed provided by the customer during the insurance claim process. The images may come from electronic invoice images uploaded by the customer or scanned copies of paper invoices. These invoice images are pre-processed, such as image clarity optimization and noise removal, to ensure that the image quality meets the subsequent recognition requirements. Next, use the pre-trained convolutional neural network (CNN) to analyze the invoice images. CNN is a deep learning model that can automatically extract features from images and identify patterns and structures in images through multiple convolutional layers. In this step, the convolutional neural network identifies the type of invoice to be processed based on the layout, format and other features of the invoice, that is, distinguishes whether the invoice is an electronic invoice or a paper invoice. Through automated invoice type recognition, not only is the processing efficiency improved, but it also provides an accurate basis for subsequent text extraction.

[0071] S202, determining a matching text field extraction rule according to the invoice type of the invoice to be processed, and obtaining a target extraction rule.

[0072] Specifically, based on the invoice type (electronic invoice or paper invoice) identified in the previous step, the corresponding text field extraction rules are determined. Due to the differences in format, layout and information arrangement between electronic invoices and paper invoices, different extraction strategies need to be used. For example, electronic invoices may have a more standardized structure and clear numeric fields, which is convenient for directly extracting text using a pre-established electronic invoice template; while the quality of paper invoices may be limited by the resolution and image clarity of the scanning device, and its layout is usually not as consistent as that of electronic invoices, requiring specific preprocessing algorithms to enhance image quality and adjust extraction rules. In addition, the position and format of key fields of different types of invoices may also be different. For example, the arrangement order and identifiers of date, project name, amount, etc. may vary depending on the type of invoice. Therefore, according to the type of invoice, select the corresponding text field extraction rules to ensure that the required invoice text fields can be accurately extracted in the next step.

[0073] S203, using the target extraction rule to perform text extraction on the invoice image to be processed, to obtain the invoice text field.

[0074] Specifically, the system uses the target extraction rules determined in the previous step to perform text extraction on the invoice image to be processed. According to different target extraction rules, the text information in different types of invoice images (electronic invoices or paper invoices) can be converted into editable text data. For example, for electronic invoices, since their format is relatively standard, the extraction accuracy is usually high, and the system directly uses the processing method of the electronic invoice template. For paper invoices, especially paper invoices with low clarity, the system uses a combination of OCR optical character recognition and semantic analysis to extract the text fields of paper invoices to ensure the accuracy of text extraction. In this process, the system will focus on identifying key information fields in the invoice, such as invoice number, invoice date, medical service items, drug name, amount, hospital name, etc. In addition, the system will also handle some common text recognition errors, such as misrecognition of similar characters (such as "0" and "O", "1" and "I", etc.), to ensure that the extracted text is as accurate as possible. After this step is completed, the system will generate structured text data.

[0075] S204, associating the invoice text field with the pre-established medical knowledge graph to identify the medical service item codes and drug codes in the invoice to be processed.

[0076] Specifically, by matching the invoice text fields with the medical knowledge graph, the standard codes of medical service items and drugs in the invoice are automatically identified. The medical knowledge graph is a structured data representation method that contains a large amount of knowledge base about medical services, drugs, diagnosis and treatment items and their related codes. At this stage, the system compares and matches the extracted invoice text fields with the information in the knowledge graph. For example, the specific medical service items mentioned in the invoice may use non-standardized names or abbreviations, and the medical knowledge graph can associate these non-standardized project names with the standard medical coding system through semantic matching or fuzzy matching technology, and finally identify the corresponding standard medical service item codes (such as CPT codes or ICD codes) and drug codes (such as ATC codes or generic drug codes). This automated association and identification method not only improves the standardization and consistency of the data, but also provides a more accurate basis for subsequent cost calculations.

[0077] S205, obtaining a preset fee standard, and calculating the item fee amount of the invoice to be processed according to the medical service item code and the drug code.

[0078] Specifically, the system first extracts the cost data related to the medical service item code and drug code from the preset cost standard library. The cost standard library is usually pre-established by insurance companies or medical institutions and contains cost reference values ​​for different medical items and drugs. These cost standards may vary according to factors such as region, hospital grade, and service type, so the system needs to select the corresponding cost standard based on the service provider or regional information in the invoice. Then, the system matches these standard costs with the medical service item code and drug code in the invoice to calculate the actual cost amount of each medical service or drug. If the cost in the invoice differs from the standard cost, the system may also conduct further analysis to determine whether the difference is reasonable. This automated cost calculation step greatly reduces the workload of manual accounting, and ensures the transparency and accuracy of cost calculation through standardized cost matching.

[0079] S206, determining the loss risk of accidental health insurance based on the project cost amount and the claim amount limit.

[0080] Specifically, in the last step, the system determines the loss risk of the claim based on the project cost amount calculated in the previous step and the claim limit in the insurance contract. First, the system compares the actual project cost amount in the invoice with the claim limit in the customer's insurance plan to determine whether it exceeds the insurance compensation standard. For example, some accidental health insurance products may have an upper limit on the compensation amount for specific medical items or drugs, and the system will automatically adjust the compensation amount based on these limits. If the actual cost exceeds the claim limit, the system will mark the excess as non-compensation, thereby reducing unnecessary disputes. In addition, the system may also assess the claim risk of the case based on the insurance company's risk control model. For example, for some invoices with high medical expenses, the system may trigger further review processes to prevent fraud risks. Through this automated loss risk assessment, not only the claim processing speed is accelerated, but also the fairness and accuracy of the loss assessment results are ensured, effectively improving the insurance company's operational efficiency and customer satisfaction.

[0081] In the above embodiment, the solution of the present application aims at the problems of low efficiency and waste of professional manpower in existing accidental health insurance loss assessment solutions. Through automated image recognition and text extraction technology, the loss assessment efficiency is greatly improved. A convolutional neural network is used to identify the invoice type, match the corresponding text extraction rules, automatically extract key data from the invoice image, and automatically identify medical service items and drug codes through the medical knowledge graph to realize the intelligent loss assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of loss assessment results.

[0082] Furthermore, the text field extraction rules include text field extraction rules for electronic invoices and text field extraction rules for paper invoices. The steps of extracting text from the invoice image to be processed using the target extraction rules to obtain the invoice text field specifically include:

[0083] When the invoice to be processed is an electronic invoice, extracting the text field of the electronic invoice according to a pre-established electronic invoice template to obtain a first text field;

[0084] When the invoice to be processed is a paper invoice, the text field of the paper invoice is extracted by combining optical character recognition and semantic analysis to obtain a second text field;

[0085] The first text field and the second text field are combined to obtain the invoice text field.

[0086] In this embodiment, the solution distinguishes between different processing methods for electronic invoices and paper invoices to improve the accuracy and efficiency of text extraction. For electronic invoices, due to the high degree of format standardization, the system can pre-establish a template library for electronic invoices, which stores the layout structure, field location and other key information of different types of electronic invoices. After obtaining the electronic invoice, the system will quickly identify and extract the text fields in the invoice according to the template to generate the first text field. The standardized template of the electronic invoice can ensure the efficient and accurate extraction process and minimize the recognition errors caused by differences in invoice formats.

[0087] On the other hand, for paper invoices, since their format is not as standard as electronic invoices, the system uses optical character recognition (OCR) technology for preliminary text extraction, and combines semantic analysis technology to semantically understand and correct the extracted text to generate a second text field. Semantic analysis can help identify common OCR recognition errors and correct incorrect characters or ambiguous content through contextual information. Subsequently, the system combines the first text field extracted from the electronic invoice with the second text field extracted from the paper invoice to form a complete invoice text field.

[0088] Through the above steps, efficient text extraction for different types of invoices is achieved. The accuracy and consistency of the extraction process are guaranteed by combining template matching of electronic invoices and OCR and semantic analysis of paper invoices.

[0089] Furthermore, before obtaining the invoice image to be processed of the accidental health insurance and importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, the method further includes:

[0090] Obtain the original invoice image, and use image preprocessing technology to remove noise and enhance the original invoice image;

[0091] For the preprocessed original invoice image, the image quality assessment algorithm is used to calculate the image clarity index and integrity index to obtain the image quality assessment result;

[0092] Compare the image quality assessment result with the preset image quality threshold to determine whether the original invoice image meets the quality requirements;

[0093] If the image quality assessment result is lower than the image quality threshold, a prompt message is output to prompt the user to re-acquire the invoice image;

[0094] If the image quality assessment result is higher than or equal to the image quality threshold, the original invoice image is output as the invoice image to be processed and saved.

[0095] In this embodiment, the original invoice image uploaded by the user is first obtained, and the image is optimized using image preprocessing technology, including noise removal and image enhancement. Noise removal can effectively eliminate impurities in the image, such as noise and spots generated when scanning equipment or mobile phone shooting, so as to improve the overall clarity of the image; image enhancement technology makes the text, table and lines in the invoice more prominent by improving the contrast, brightness and other parameters of the image. After the image preprocessing is completed, the system further evaluates the image quality of the processed image, and calculates the clarity and completeness index of the image through a special image quality evaluation algorithm. The clarity index measures whether the image is clear enough to ensure the legibility of characters and details, while the completeness index evaluates whether the image is comprehensive and whether there are problems such as cropping, folding or occlusion. Then, the image quality evaluation result is compared with the preset quality threshold to determine whether the image meets the recognition requirements. If the evaluation result is lower than the set threshold, the system will output a prompt message to remind the user to re-upload a clear invoice image. On the contrary, if the image quality meets the standard, the system will mark the image as qualified and save it as an invoice image to be processed.

[0096] Through the above steps, the overall quality of image processing is effectively improved. Through noise removal, image enhancement and image quality assessment, the invoice images entering the subsequent recognition and processing stages are ensured to be clear and complete. This not only improves the accuracy of image recognition, but also reduces the recognition error rate caused by poor image quality, thereby optimizing the overall damage assessment and claims process.

[0097] Furthermore, the convolutional neural network is preset with the layout features of the electronic invoice and the paper invoice, and the step of importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed specifically includes:

[0098] Inputting the invoice image to be processed into a convolutional neural network to extract features of the invoice image to be processed, thereby obtaining invoice image features;

[0099] Compare the invoice image features with the layout features of the electronic invoice and the layout features of the paper invoice, and calculate the feature similarity score;

[0100] Based on the feature similarity scores, the invoice type of the invoice to be processed is identified.

[0101] In this embodiment, a convolutional neural network (CNN) is used to extract features from invoice images, and the invoice type is automatically identified by comparing the image features with a preset layout feature library. First, the system inputs the invoice image to be processed into a pre-trained convolutional neural network. CNN extracts features at different levels from the image through multi-layer convolution, pooling and activation function operations. These features may include edges, textures, image structures and higher-level patterns. The extracted features represent the unique structural information of the image. Next, the system compares these invoice image features with the pre-established layout feature library of electronic invoices and paper invoices. The layout feature library stores a large amount of structured feature information of electronic invoices and paper invoices, including the layout, text distribution, key field positions, etc. of the invoices. By calculating the similarity score between the invoice image to be processed and the two layout features, the system can determine which type of invoice the invoice to be processed is more similar to. Generally, there are significant differences in the layout features of electronic invoices and paper invoices. For example, the structure of electronic invoices is more standardized, and the position of text fields is relatively fixed, while paper invoices may have different layouts and irregular structures due to the scanning and shooting methods. Based on the feature similarity score, the system will eventually automatically identify the type of invoice. If the electronic invoice has a higher similarity score, the system will identify it as an electronic invoice, otherwise it will be a paper invoice.

[0102] Through the above steps, the powerful feature extraction capability and precise image comparison mechanism of convolutional neural networks can be used to quickly and accurately identify invoice types. By calculating feature similarity, the system can effectively distinguish between electronic invoices and paper invoices, avoiding the inefficiency and errors of manual classification.

[0103] Furthermore, the step of identifying the invoice type of the invoice to be processed according to the feature similarity score specifically includes:

[0104] Determine the relationship between the feature similarity score and a preset first similarity threshold;

[0105] If the feature similarity score between the invoice image feature and the format feature of the electronic invoice exceeds a preset first similarity threshold, the invoice to be processed is determined to be an electronic invoice;

[0106] If the feature similarity score between the invoice image feature and the format feature of the paper invoice exceeds a preset first similarity threshold, the invoice to be processed is determined to be a paper invoice.

[0107] In this embodiment, the invoice type identification process is further refined by comparing the similarity scores between the invoice image features and the layout features of the electronic invoice and the paper invoice. First, after extracting the invoice image features and calculating the similarity scores between them and the layout features of the electronic invoice and the paper invoice, the system compares these similarity scores with a preset first similarity threshold. This first similarity threshold is defined in advance and is used to determine whether the similarity between the image features and the features of a certain type of invoice is high enough to meet the standard for reliable identification. If the similarity score between the invoice image features and the layout features of the electronic invoice exceeds the first similarity threshold, the system will determine that the invoice is an electronic invoice, that is, the overall layout and key field distribution of the invoice are highly consistent with the template features of the electronic invoice, meeting the identification standard of the electronic invoice. Similarly, if the similarity score between the invoice image features and the layout features of the paper invoice is higher than the first similarity threshold, the system determines that the invoice to be processed is a paper invoice.

[0108] In some cases, the features of the invoice image may have a certain degree of similarity with both layout features, but the system will make the final classification based on which score is the highest and exceeds the first similarity threshold. For example, the feature similarity score Q1 between the invoice image feature and the layout feature of the electronic invoice is 0.8, and the feature similarity score Q2 between the invoice image feature and the layout feature of the paper invoice is 0.9. The first similarity threshold is set to 0.5, and both Q1 and Q2 are greater than 0.5. At this time, the system will compare the size of Q1 and Q2 to determine the type of the invoice to be processed, and will eventually determine the type of the invoice to be processed as a paper invoice (Q2 = 0.9 > Q1 = 0.8). In this way, the system can make accurate judgments in complex environments, reduce misjudgments and ambiguities, and ensure the accuracy of classification.

[0109] Through the above steps, the invoice type is judged by presetting the similarity threshold, making the classification process more accurate. By comparing the similarity of the layout features of electronic invoices and paper invoices, the system can quickly make an accurate judgment on the invoice type.

[0110] Furthermore, the steps of associating the invoice text field with the pre-established medical knowledge graph and identifying the medical service item codes and drug codes in the invoice to be processed specifically include:

[0111] By performing natural language processing on the invoice text field, the names of medical service items and medicines in the invoice text field are obtained;

[0112] The names of medical service items are matched with the standard item names in the medical knowledge graph to calculate the item name similarity;

[0113] If the similarity of the project name exceeds a preset second similarity threshold, the code corresponding to the standard project name is obtained as the medical service project code;

[0114] Match the drug names with the standard drug names in the medical knowledge graph to calculate the drug name similarity;

[0115] If the similarity of the drug name exceeds a preset second similarity threshold, the code corresponding to the standard drug name is obtained as the drug code.

[0116] In this embodiment, natural language processing (NLP) technology is used to automatically identify the names of medical service items and drug names in the invoice text fields, and match them with the standard names in the medical knowledge graph to achieve accurate coding of medical items and drugs.

[0117] First, the system processes the text fields in the invoice and extracts the names of medical service items and drug names. Since the names of items and drugs in the invoice may contain abbreviations, spelling errors or different expressions, it is difficult to ensure accuracy by direct comparison. Therefore, the system uses natural language processing technology, including word segmentation, semantic analysis, etc., to standardize these names so that they can better match the standard names in the knowledge graph. Then, the system uses a similarity algorithm to compare the extracted project names with the standard project names in the knowledge graph and calculate the similarity of the names. If the similarity exceeds the preset second similarity threshold, it is determined that the project name is highly similar to the standard name in the knowledge graph, and the system will use the code corresponding to the standard name as the code of the medical service item. This process also applies to the matching of drug names, that is, the system compares the drug name extracted from the invoice with the standard drug name in the knowledge graph through similarity calculation. If the similarity is high enough (the similarity exceeds the preset second similarity threshold), the standard drug code will be used as the corresponding code of the drug.

[0118] Through the above steps, natural language processing and similarity matching technology are used to effectively improve the coding accuracy of medical items and drugs. Automated similarity matching reduces manual intervention, avoids mismatching due to differences in expression, and ensures the standardization of item and drug information in invoices.

[0119] Furthermore, the steps of determining the loss risk of accidental health insurance based on the project cost amount and the claim amount limit specifically include:

[0120] If the project cost amount is greater than the claim amount limit, the accident health insurance loss risk is determined to be high risk, and a high risk warning message is output;

[0121] If the project cost amount is less than or equal to the claim amount limit, the loss risk of accidental health insurance is determined to be low risk.

[0122] In this embodiment, the loss risk of accidental health insurance is evaluated by comparing the project cost amount with the claim amount limit. First, the system obtains the project cost amount of the insured and the claim amount limit agreed in its insurance contract. Then, the system compares the project cost amount with the claim amount limit: if the project cost amount is greater than the claim amount limit, the system will determine that the case is high risk and send a high-risk warning message to the relevant department or personnel, indicating that further review or processing is required; if the project cost amount is less than or equal to the claim amount limit, the system will determine that the case is low risk and handle it according to the normal process.

[0123] Through the above steps, the automated assessment of accidental health insurance loss risk is achieved, ensuring that high-risk cases can be identified and handled in a timely manner, reducing the insurance company's compensation risk and maintaining the order of the insurance market.

[0124] In the above embodiments, the present application discloses a method for assessing damages for accidental health insurance, which belongs to the field of artificial intelligence technology. The solution of the present application aims to address the problems of low efficiency and waste of professional manpower in existing accidental health insurance damage assessment solutions. Through automated image recognition and text extraction technology, the efficiency of damage assessment is greatly improved. A convolutional neural network is used to identify the invoice type, match the corresponding text extraction rules, automatically extract key data from the invoice image, and automatically identify medical service items and drug codes through a medical knowledge graph to realize the intelligentization of the damage assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of damage assessment results.

[0125] In this embodiment, the method for assessing the loss of accidental health insurance is executed on an electronic device (eg Figure 1 The server shown in the figure) can receive instructions or obtain data through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0126] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned invoice image information, the above-mentioned invoice image information can also be stored in a blockchain node.

[0127] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0128] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0129] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0130] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0131] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0132] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for assessing damages of accidental health insurance. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0133] like Figure 3 As shown, the damage assessment device 300 for accidental health insurance described in this embodiment includes:

[0134] An invoice type recognition module 301 is used to obtain an image of an invoice to be processed for accidental health insurance, and import the image of the invoice to be processed into a pre-trained convolutional neural network to recognize the invoice type of the invoice to be processed, wherein the invoice type includes an electronic invoice and a paper invoice;

[0135] An extraction rule matching module 302 is used to determine a matching text field extraction rule according to the invoice type of the invoice to be processed, and obtain a target extraction rule;

[0136] A text field extraction module 303, used to extract text from the invoice image to be processed using a target extraction rule to obtain an invoice text field;

[0137] The knowledge graph association module 304 is used to associate the invoice text field with the pre-established medical knowledge graph to identify the medical service item code and drug code in the invoice to be processed;

[0138] The fee amount calculation module 305 is used to obtain the preset fee standard and calculate the item fee amount of the invoice to be processed according to the medical service item code and the drug code;

[0139] The damage assessment risk determination module 306 is used to determine the damage assessment risk of accidental health insurance based on the project cost amount and the claim amount limit.

[0140] Furthermore, the text field extraction rules include text field extraction rules for electronic invoices and text field extraction rules for paper invoices. The text field extraction module 303 is specifically used for:

[0141] When the invoice to be processed is an electronic invoice, the text field of the electronic invoice is extracted according to the pre-established electronic invoice template to obtain the first text field; when the invoice to be processed is a paper invoice, the text field of the paper invoice is extracted by combining optical character recognition and semantic analysis to obtain the second text field; the first text field and the second text field are combined to obtain the invoice text field.

[0142] Furthermore, the accidental health insurance damage assessment device 300 further includes an image quality assessment module, which is specifically used for:

[0143] The original invoice image is obtained, and the image preprocessing technology is used to remove noise and enhance the image of the original invoice image; for the preprocessed original invoice image, the image clarity index and integrity index are calculated by the image quality assessment algorithm to obtain the image quality assessment result; the image quality assessment result is compared with the preset image quality threshold to determine whether the original invoice image meets the quality requirements; if the image quality assessment result is lower than the image quality threshold, a prompt message is output to prompt the user to re-acquire the invoice image; if the image quality assessment result is higher than or equal to the image quality threshold, the original invoice image is output and saved as the invoice image to be processed.

[0144] Furthermore, the convolutional neural network is preset with the layout features of electronic invoices and paper invoices, and the invoice type recognition module 301 is specifically used for:

[0145] The invoice image to be processed is input into a convolutional neural network to extract the features of the invoice image to be processed and obtain the invoice image features; the invoice image features are compared with the layout features of the electronic invoice and the layout features of the paper invoice, and the feature similarity score is calculated; according to the feature similarity score, the invoice type of the invoice to be processed is identified.

[0146] Furthermore, the invoice type identification module 301 is also used for:

[0147] Determine the relationship between the feature similarity score and the preset first similarity threshold; if the feature similarity score between the invoice image feature and the layout feature of the electronic invoice exceeds the preset first similarity threshold, determine that the invoice to be processed is an electronic invoice; if the feature similarity score between the invoice image feature and the layout feature of the paper invoice exceeds the preset first similarity threshold, determine that the invoice to be processed is a paper invoice.

[0148] Furthermore, the knowledge graph association module 304 is specifically used for:

[0149] By performing natural language processing on the invoice text field, the medical service item name and drug name in the invoice text field are obtained; the medical service item name is matched with the standard item name in the medical knowledge graph for similarity, and the item name similarity is calculated; if the item name similarity exceeds a preset second similarity threshold, the code corresponding to the standard item name is obtained as the medical service item code; the drug name is matched with the standard drug name in the medical knowledge graph for similarity, and the drug name similarity is calculated; if the drug name similarity exceeds the preset second similarity threshold, the code corresponding to the standard drug name is obtained as the drug code.

[0150] Furthermore, the damage risk determination module 306 is specifically used to:

[0151] If the project cost amount is greater than the claim amount limit, the loss assessment risk of the accidental health insurance is determined to be high risk, and a high-risk warning message is output; if the project cost amount is less than or equal to the claim amount limit, the loss assessment risk of the accidental health insurance is determined to be low risk.

[0152] In the above embodiment, the present application discloses a loss assessment device for accidental health insurance, which belongs to the field of artificial intelligence technology. The solution of the present application aims at the problems of low efficiency and consumption of professional manpower in existing accidental health insurance loss assessment solutions. Through automated image recognition and text extraction technology, the loss assessment efficiency is greatly improved. A convolutional neural network is used to identify the invoice type, match the corresponding text extraction rules, automatically extract key data from the invoice image, and automatically identify medical service items and drug codes through the medical knowledge graph to realize the intelligent loss assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of loss assessment results.

[0153] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0154] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0155] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0156] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the loss assessment method of accidental health insurance, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0157] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions stored in the memory 41 or process data, such as computer-readable instructions for running the loss assessment method of the accidental health insurance.

[0158] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0159] In the above embodiment, the present application discloses a computer device, which belongs to the field of artificial intelligence technology. The solution of the present application is aimed at the problems of low efficiency and consumption of professional manpower in the existing accidental health insurance loss assessment scheme. Through automated image recognition and text extraction technology, the loss assessment efficiency is greatly improved. The convolutional neural network is used to identify the invoice type, match the corresponding text extraction rules, automatically extract key data from the invoice image, and automatically identify medical service items and drug codes through the medical knowledge graph to realize the intelligent loss assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves the case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of loss assessment results.

[0160] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the loss assessment method for accidental health insurance as described above.

[0161] In the above embodiments, the present application discloses a computer-readable storage medium, which belongs to the field of artificial intelligence technology. The solution of the present application is aimed at the problems of low efficiency and consumption of professional manpower in existing accidental health insurance loss assessment solutions. Through automated image recognition and text extraction technology, the loss assessment efficiency is greatly improved. Convolutional neural networks are used to identify invoice types, match corresponding text extraction rules, automatically extract key data from invoice images, and automatically identify medical service items and drug codes through medical knowledge graphs to realize the intelligent loss assessment process. Compared with traditional manual processing methods, this solution reduces dependence on professional manpower, significantly improves case processing speed, reduces operating costs, and improves customer experience. In addition, the application of automation technology also reduces the risk of human error and improves the accuracy of loss assessment results.

[0162] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0163] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0164] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for assessing loss of accidental health insurance, characterized in that: include: Acquire an image of an invoice to be processed for accidental health insurance, and import the image of the invoice to be processed into a pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, wherein the invoice type includes an electronic invoice and a paper invoice; Determine a matching text field extraction rule according to the invoice type of the invoice to be processed to obtain a target extraction rule; Using the target extraction rule to perform text extraction on the invoice image to be processed, to obtain an invoice text field; Associating the invoice text field with a pre-established medical knowledge graph to identify the medical service item code and the drug code in the invoice to be processed; Obtaining a preset fee standard, and calculating the item fee amount of the invoice to be processed according to the medical service item code and the drug code; The loss risk of the accidental health insurance is determined based on the project cost amount and the claim amount limit.

2. The method for assessing loss of accidental health insurance according to claim 1, characterized in that: The text field extraction rules include text field extraction rules for electronic invoices and text field extraction rules for paper invoices. The step of using the target extraction rules to extract text from the invoice image to be processed to obtain the invoice text field specifically includes: When the invoice to be processed is an electronic invoice, extracting the text field of the electronic invoice according to a pre-established electronic invoice template to obtain a first text field; When the invoice to be processed is a paper invoice, extracting the text field of the paper invoice by combining optical character recognition and semantic analysis to obtain a second text field; The first text field and the second text field are combined to obtain the invoice text field.

3. The method for assessing loss of accidental health insurance according to claim 1, characterized in that: Before the step of obtaining the invoice image to be processed of the accidental health insurance and importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, the method further includes: Acquire an original invoice image, and use image preprocessing technology to remove noise and enhance the image of the original invoice image; For the pre-processed original invoice image, an image quality assessment algorithm is used to calculate the clarity index and the integrity index of the image to obtain an image quality assessment result; Compare the image quality assessment result with a preset image quality threshold to determine whether the original invoice image meets the quality requirements; If the image quality evaluation result is lower than the image quality threshold, a prompt message is output to prompt the user to re-acquire the invoice image; If the image quality evaluation result is higher than or equal to the image quality threshold, the original invoice image is output and saved as the invoice image to be processed.

4. The method for assessing loss of accidental health insurance according to claim 1, characterized in that: The convolutional neural network is preset with the layout features of the electronic invoice and the paper invoice. The step of importing the invoice image to be processed into the pre-trained convolutional neural network to identify the invoice type of the invoice to be processed specifically includes: Inputting the invoice image to be processed into the convolutional neural network to extract features of the invoice image to be processed and obtain invoice image features; Comparing the invoice image features with the layout features of the electronic invoice and the layout features of the paper invoice, respectively, and calculating feature similarity scores; The invoice type of the invoice to be processed is identified according to the feature similarity score.

5. The method for assessing loss of accidental health insurance according to claim 4, characterized in that: The step of identifying the invoice type of the invoice to be processed according to the feature similarity score specifically includes: Determine the magnitude relationship between the feature similarity score and a preset first similarity threshold; If the feature similarity score between the invoice image feature and the format feature of the electronic invoice exceeds a preset first similarity threshold, then the invoice to be processed is determined to be an electronic invoice; If the feature similarity score between the invoice image feature and the format feature of the paper invoice exceeds a preset first similarity threshold, the invoice to be processed is determined to be a paper invoice.

6. The method for assessing loss of accidental health insurance according to claim 1, characterized in that: The step of associating the invoice text field with the pre-established medical knowledge graph to identify the medical service item code and the drug code in the invoice to be processed specifically includes: Obtaining the medical service item name and the drug name in the invoice text field by performing natural language processing on the invoice text field; Perform similarity matching between the medical service item name and the standard item name in the medical knowledge graph, and calculate the item name similarity; If the similarity of the item name exceeds a preset second similarity threshold, obtaining the code corresponding to the standard item name as the medical service item code; Perform similarity matching between the drug name and the standard drug name in the medical knowledge graph, and calculate the drug name similarity; If the similarity of the drug name exceeds a preset second similarity threshold, the code corresponding to the standard drug name is obtained as the drug code.

7. The method for assessing loss of accidental health insurance according to claim 1, characterized in that: The step of determining the loss risk of the accidental health insurance based on the project cost amount and the claim amount limit specifically includes: If the project fee amount is greater than the claim amount limit, the loss assessment risk of the accidental health insurance is determined to be high risk, and a high risk warning message is output; If the project cost amount is less than or equal to the claim amount limit, the loss risk of the accidental health insurance is determined to be low risk.

8. A device for assessing damages for accidental health insurance, characterized in that: include: An invoice type recognition module, used to obtain an image of an invoice to be processed for accidental health insurance, and import the image of the invoice to be processed into a pre-trained convolutional neural network to identify the invoice type of the invoice to be processed, wherein the invoice type includes an electronic invoice and a paper invoice; An extraction rule matching module, used to determine a matching text field extraction rule according to the invoice type of the invoice to be processed, and obtain a target extraction rule; A text field extraction module, used to extract text from the invoice image to be processed using the target extraction rule to obtain the invoice text field; A knowledge graph association module, used to associate the invoice text field with a pre-established medical knowledge graph, and identify the medical service item code and drug code in the invoice to be processed; A fee amount calculation module is used to obtain a preset fee standard and calculate the item fee amount of the invoice to be processed according to the medical service item code and the drug code; The damage assessment risk determination module is used to determine the damage assessment risk of the accidental health insurance based on the project cost amount and the claim amount limit.

9. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, it implements the steps of the loss assessment method for accidental health insurance as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the method for assessing damages for accidental health insurance as described in any one of claims 1 to 7.

Images