Business insurance auditing method, system, terminal and storage medium

Abstract

The application provides a commercial insurance auditing method and system, a terminal and a storage medium. The method comprises the following steps: obtaining a commercial insurance reporting document, performing character recognition on the commercial insurance reporting document to obtain document characters, performing data cleaning on the document characters, performing data conversion on the data cleaned document characters to obtain document structured information, performing auditing and proofreading on the document structured information, performing data reduction on the document structured information after the auditing and proofreading, and generating a commercial insurance auditing result according to the document structured information after the data reduction. According to the embodiment of the application, the commercial insurance reporting document can be automatically audited and proofread, manual auditing is not needed, the trademark auditing efficiency is improved, and the user waiting time and service cost are reduced.

Description

Technical Field

[0001] This invention relates to the field of insurance review technology, and in particular to a commercial insurance review method, system, terminal and storage medium. Background Technology

[0002] Currently, with increasing user awareness of security, users are gradually realizing the importance of insurance, and more and more users are joining the ranks of those purchasing commercial insurance. During the commercial insurance application process, insurance companies will enter the user's policy information, such as policyholder information, insured information, insurance period, type of insurance, insured amount, and premium. When a commercial insurance claim occurs, the insurance company needs to review the claim documents first. After the review is completed, the insurance company will process the claim.

[0003] The current commercial insurance review process generally uses manual methods to review commercial insurance claims, resulting in low efficiency in commercial insurance review. Summary of the Invention

[0004] The purpose of this invention is to provide a commercial insurance review method, system, terminal, and storage medium, aiming to solve the problem of low efficiency in existing commercial insurance review processes.

[0005] This invention is implemented as follows: a commercial insurance review method, the method comprising:

[0006] Obtain the commercial insurance claim form and perform text recognition on the commercial insurance claim form to obtain the text of the form;

[0007] The document text is cleaned, and the cleaned document text is converted to obtain the document structure information.

[0008] The structured information of the document is reviewed and verified, and data is reduced from the reviewed and verified structured information of the document.

[0009] The commercial insurance audit result is generated based on the structured information of the documents after data reduction.

[0010] Preferably, the text recognition of the commercial insurance claim form includes:

[0011] The commercial insurance claim form is binarized to obtain a binarized image, and the connected regions in the binarized image are obtained.

[0012] Calculate the average pixel value of each connected region, and perform denoising processing on each connected region based on the average pixel value to obtain a denoised image;

[0013] The denoised image is dilated to obtain a dilated image, and the pixel values ​​of the dilated image are detected.

[0014] The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line.

[0015] The denoised image is corrected according to the rotation angle to obtain a corrected image, and the corrected image is subjected to contour detection to obtain the document text.

[0016] Preferably, the data cleaning of the document text includes:

[0017] Obtain the names of commercial insurance items from the text of the document, and determine the content of the commercial insurance items based on the position of each commercial insurance item name;

[0018] Obtain the number of items in each item column of each commercial insurance item, and determine whether the number of items in each item column is equal to the preset number;

[0019] If the number of items in any of the project columns is less than the preset number, the commercial insurance project items corresponding to the number of items in the project column will be marked as abnormal content.

[0020] Obtain the text type of the text in the item column of each commercial insurance item, and compare the text type of each item column text with the text type of the adjacent item column text;

[0021] If the text type of any item column text fails to match the text type of the adjacent item column text, the commercial insurance item content corresponding to the item column text will be marked as the abnormal content.

[0022] If a deletion instruction is received for the abnormal content, then the abnormal content is deleted.

[0023] Preferably, the data conversion of the document text after data cleaning includes:

[0024] Entity recognition is performed on the text in the item column of each commercial insurance project to obtain the text entity type, and the entity mapping type of the item column in each commercial insurance project is queried.

[0025] The text entity types are matched with the entity mapping types of each item column, and the text in each item column is mapped according to the type matching results to obtain the document structure information.

[0026] Preferably, the step of reviewing and verifying the structured information of the document includes:

[0027] Query the medical insurance deduction ratio for each commercial insurance item name in the structured information of the document, and calculate the target amount based on the medical insurance deduction ratio and the pricing amount in the structured information of the document;

[0028] For each commercial insurance item name, determine whether the amount payable in the structured information of the document is equal to the target amount;

[0029] If the amount due is not equal to the target amount, then the name of the commercial insurance project corresponding to the amount due will be marked as incorrect in the document structure information.

[0030] Preferably, the step of reducing the data in the structured information of the document after review and verification includes:

[0031] The field values ​​of each row in the document structure information are obtained respectively, and the field validation of the document structure information is performed based on the field values ​​to obtain the field validation results;

[0032] The document values ​​of each number in the document structure information are obtained respectively, and the document values ​​are verified according to the preset value range to obtain the value verification results.

[0033] The text of each line in the structured information of the document is obtained respectively, and the text of each line is validated according to a preset dictionary to obtain the text validation result.

[0034] The data in the document's structured information is reduced based on the field validation results, the numerical validation results, and the text validation results.

[0035] Another objective of this invention is to provide a commercial insurance review system, the system comprising:

[0036] The text recognition module is used to obtain commercial insurance claim documents and perform text recognition on the commercial insurance claim documents to obtain the document text;

[0037] The data conversion module is used to clean the text of the document and convert the cleaned text to obtain structured information of the document.

[0038] The review and verification module is used to review and verify the structured information of the document, and to reduce the data in the reviewed and verified structured information of the document.

[0039] The result generation module is used to generate commercial insurance audit results based on the structured information of the document after data reduction.

[0040] Preferably, the text recognition module is further used for:

[0041] The commercial insurance claim form is binarized to obtain a binarized image, and the connected regions in the binarized image are obtained.

[0042] Calculate the average pixel value of each connected region, and perform denoising processing on each connected region based on the average pixel value to obtain a denoised image;

[0043] The denoised image is dilated to obtain a dilated image, and the pixel values ​​of the dilated image are detected.

[0044] The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line.

[0045] The denoised image is corrected according to the rotation angle to obtain a corrected image, and the corrected image is subjected to contour detection to obtain the document text.

[0046] Another objective of this invention is to provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0047] Another objective of this invention is to provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method.

[0048] This invention, through text recognition on commercial insurance claim forms, can automatically extract text from these forms. Data cleaning of the text improves its accuracy. Data conversion of the cleaned text achieves a structured processing effect. Review and verification of the structured information further enhances its accuracy. Data reduction of the reviewed and verified structured information effectively removes invalid data, further improving its accuracy. This embodiment automatically reviews and verifies commercial insurance claim forms, eliminating the need for manual review, thus improving trademark review efficiency and reducing user waiting time and service costs. Attached Figure Description

[0049] Figure 1 This is a flowchart of the commercial insurance review method provided in the first embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the commercial insurance claim form provided in the first embodiment of the present invention;

[0051] Figure 3 This is a flowchart of the commercial insurance review method provided in the second embodiment of the present invention;

[0052] Figure 4This is a schematic diagram of the commercial insurance review system provided in the third embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of the structure of the terminal device provided in the fourth embodiment of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0055] To illustrate the technical solution described in this invention, specific embodiments are described below.

[0056] Example 1

[0057] Please see Figure 1 This is a flowchart of the commercial insurance review method provided in the first embodiment of the present invention. The commercial insurance review method can be applied to any terminal device or system, and the commercial insurance review method includes the following steps:

[0058] Step S10: Obtain the commercial insurance claim form and perform text recognition on the commercial insurance claim form to obtain the text of the form;

[0059] The commercial insurance claim form can be transmitted as an image; for example, please refer to [link / reference]. Figure 2 , Figure 2 The "Patient Expense List" is a commercial insurance claim form. By performing text recognition on the commercial insurance claim form, the text in the form can be obtained. The text includes the name of the commercial insurance item and the content of the commercial insurance item corresponding to the name.

[0060] Optionally, the text recognition of the commercial insurance claim document includes:

[0061] The commercial insurance claim form is binarized to obtain a binarized image, and the connected regions in the binarized image are obtained; wherein, by obtaining the connected regions in the binarized image, the accuracy of subsequent average pixel value calculation is improved;

[0062] The average pixel value of each connected region is calculated, and the denoising process is performed on each connected region based on the average pixel value to obtain a denoised image; wherein, by calculating the average pixel value of each connected region, the denoising effect of each connected region can be effectively achieved to remove noise pixels in each connected region.

[0063] The denoised image is dilated to obtain an expanded image, and the pixel value of the expanded image is detected. The dilation process on the denoised image effectively facilitates the pixel value detection operation of the expanded image. In this step, the pixel value of each pixel in the expanded image is obtained by detecting the pixel value of the expanded image.

[0064] The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line; wherein, pixels with pixel values ​​greater than the pixel threshold are connected to achieve the connection effect of text in the denoised image to obtain the text lines, and the angle between each text line and the horizontal baseline is calculated to obtain the rotation angle.

[0065] The denoised image is corrected according to the rotation angle to obtain a corrected image, and contour detection is performed on the corrected image to obtain the document text; wherein, by correcting the denoised image by rotating the angle, the accuracy of the text angle in the denoised image is improved.

[0066] Step S20: Clean the text of the document and convert the cleaned text to obtain structured information of the document.

[0067] In this process, data cleaning of the document text improves its accuracy, and data transformation of the cleaned document text achieves a structured processing effect.

[0068] Optionally, in this step, data extraction, transformation, and loading (ETL) algorithms can also be used to clean and transform the text in the document.

[0069] Optionally, the data cleaning of the document text includes:

[0070] Obtain the names of commercial insurance items from the text of the document, and determine the content of the commercial insurance items based on the position of each commercial insurance item name;

[0071] In this process, the content selection box corresponding to each commercial insurance item name is obtained. Starting from the position of each commercial insurance item name, the corresponding content selection box is drawn. The content within the content selection box is determined as the content of that commercial insurance item. The size and shape of the content selection box corresponding to each commercial insurance item name can be set according to the requirements. In this step, each commercial insurance item name is matched with the selection box lookup table to obtain the content selection box. The selection box lookup table stores the correspondence between different commercial insurance item names and corresponding content selection boxes.

[0072] Obtain the number of items in each item column of each commercial insurance item, and determine whether the number of items in each item column is equal to the preset number;

[0073] The preset quantity can be set according to needs. By checking whether the number of items in each column equals the preset quantity, it can determine whether there are any missing items in each commercial insurance item. For example, please refer to... Figure 2 The number of items corresponding to the commercial insurance project name "Nine Fragrant Insects" is 8, namely the items corresponding to the project column "Secondary Category", "Commercial Insurance Project Name", "Specification", "Unit", "Unit Price", "Quantity", "Pricing Amount" and "Amount Payable".

[0074] If the number of items in any of the project columns is less than the preset number, the commercial insurance project items corresponding to the number of items in the project column will be marked as abnormal content.

[0075] If the number of items in a project column is less than the preset number, it is determined that the commercial insurance project content corresponding to the number of items in that project column is missing. By marking the commercial insurance project content as abnormal content, it is convenient for users to view abnormal data.

[0076] If the number of items in each item column is equal to the preset number, the commercial insurance items corresponding to the number of items in that item column are normal. If the number of items in each item column is greater than the preset number, the commercial insurance items corresponding to the number of items in that item column are marked as abnormal content.

[0077] The text type of each item in the commercial insurance project content is obtained, and the text type of each item is compared with the text type of the adjacent item. The text type includes numerical values, text, letter codes, or units of measurement. By comparing the text type of each item with the text type of the adjacent item, it is determined whether there are any errors in the text of each item.

[0078] If the text type of any item column text fails to match the text type of an adjacent item column text, the commercial insurance item content corresponding to the item column text is marked as abnormal content; wherein, if the text type of any item column text fails to match the text type of an adjacent item column text, it is determined that the item column text has an error.

[0079] If a deletion instruction is received for the abnormal content, then the abnormal content is deleted.

[0080] Furthermore, the data conversion of the document text after data cleaning includes:

[0081] Entity recognition is performed on the text in the item column of each commercial insurance project to obtain the text entity type, and the entity mapping type of the item column in each commercial insurance project is queried.

[0082] The text entity types are matched with the entity mapping types of each item column, and the text in each item column is mapped according to the type matching results to obtain the document structure information.

[0083] Specifically, the mapping relationship between the text entity type and the entity mapping type of each item column is determined by matching the text type with the entity mapping type of each item column, and the text of each item column is mapped based on the determined mapping relationship.

[0084] Step S30: Review and verify the structured information of the document, and reduce the data in the reviewed and verified structured information of the document.

[0085] Among them, an audit engine can be used for automated audit and verification. Based on the regional code and the corresponding audit point, the machine can automatically perform audit. Based on the sorted three-dimensional data, the deduction of categories B and C can be completed, which effectively avoids the problems of misjudgment or omission that occur in manual audit and verification, and improves the accuracy and efficiency of audit.

[0086] Optionally, the step of reviewing and verifying the structured information of the document includes:

[0087] Query the medical insurance deduction ratio for each commercial insurance item name in the structured information of the document, and calculate the target amount based on the medical insurance deduction ratio and the pricing amount in the structured information of the document;

[0088] For each commercial insurance item name, determine whether the amount payable in the structured information of the document is equal to the target amount;

[0089] If the amount payable is not equal to the target amount, then the name of the commercial insurance project corresponding to the amount payable will be marked as incorrect in the document structure information;

[0090] The audit engine receives the structured information of the invoice and matches the commercial insurance item names with the deduction catalog one by one. Once a match is found, the corresponding medical insurance deduction ratio is extracted. The pricing amount is multiplied by the medical insurance deduction ratio to obtain the target amount. The target amount is calculated for all items in the structured information of the invoice in turn, and it is determined whether the amount payable is equal to the target amount. This automatically determines whether the amount payable in the structured information of the invoice is correct. The deduction catalog stores the correspondence between different commercial insurance item names and their corresponding medical insurance deduction ratios.

[0091] Step S40: Generate commercial insurance audit results based on the structured information of the document after data reduction;

[0092] In this process, by reducing the data in the structured information of the documents after review and verification, invalid data in the structured information of the documents can be effectively removed, thereby improving the accuracy of the structured information of the documents. In this step, the commercial insurance review results can be displayed directly, including information such as deduction items (names of commercial insurance items for which amounts can be deducted) and deduction amounts. Commercial insurance companies and users can know the review results at any time, avoiding inconvenience and disputes caused by waiting for the review results.

[0093] In this embodiment, text recognition is performed on commercial insurance claim forms to automatically extract text. Data cleaning of the text improves its accuracy. Data conversion of the cleaned text achieves a structured processing effect. Review and verification of the structured information further enhances its accuracy. Data reduction of the reviewed and verified structured information effectively removes invalid data, further improving its accuracy. This embodiment achieves automated review, improving review efficiency and accuracy, shortening review time, reducing user waiting time and service costs, enhancing the competitiveness and service quality of commercial insurance companies, and providing convenient and high-quality services for both commercial insurance companies and users.

[0094] Example 2

[0095] Please see Figure 3 This is a flowchart of the commercial insurance review method provided in the second embodiment of the present invention. This embodiment is used to further refine step S30 in the first embodiment, including the following steps:

[0096] Step S31: Obtain the field values ​​of each row in the document structure information, and perform field validation on the document structure information based on the field values ​​to obtain the field validation results;

[0097] The process involves matching the field values ​​of each row with a preset range of field values ​​to determine whether the field values ​​of each row are within the preset range. This achieves the field validation effect on the structured information of the document and can effectively determine whether there is any missing data in each row. The preset range of field values ​​can be set according to requirements.

[0098] Step S32: Obtain the document value of each number in the document structured information, and perform numerical verification on each document value according to the preset numerical range to obtain the numerical verification result.

[0099] The preset value range can be set according to requirements. It can be used to determine whether the document value of each number is within the preset value range, so as to achieve the effect of verifying the value of the document structure information and effectively determine whether there is any abnormality in the document value of the number in the document structure information.

[0100] Step S33: Obtain the text of each line in the structured information of the document, and perform text verification on the text of each line according to a preset dictionary to obtain the text verification result;

[0101] The text in the preset dictionary can be set according to requirements. By matching the text in each line with the preset dictionary, it can be determined whether the text in each line exceeds the specified character range, so as to achieve the text verification effect of the structured information of the document and effectively determine whether there are text anomalies in the structured information of the document.

[0102] Step S34: Reduce the data in the document structured information based on the field validation result, the numerical validation result, and the text validation result;

[0103] The process involves identifying abnormal rows, numbers, and text based on field validation results, numerical validation results, and text validation results. An error message is then generated based on the identified abnormal rows, numbers, and text. If a confirmation instruction for the error message is received, the abnormal row, number, and text are deleted, thus achieving the effect of reducing the amount of data in the structured information of the document.

[0104] In this embodiment, it can automatically determine whether there is missing data in each row, whether there are abnormal numerical values ​​in the document structure information, and whether there are abnormal text in the document structure information. It eliminates the need for manual review for anomaly detection, thus improving the efficiency of commercial insurance review.

[0105] Example 3

[0106] Please see Figure 4 This is a schematic diagram of the structure of the commercial insurance review system 100 provided in the third embodiment of the present invention, including: a text recognition module 10, a data conversion module 11, a review and verification module 12, and a result generation module 13, wherein:

[0107] The text recognition module 10 is used to obtain commercial insurance claim documents and perform text recognition on the commercial insurance claim documents to obtain the document text.

[0108] Optionally, the text recognition module 10 is further configured to: perform binarization processing on the commercial insurance claim form to obtain a binarized image, and obtain the connected regions in the binarized image;

[0109] Calculate the average pixel value of each connected region, and perform denoising processing on each connected region based on the average pixel value to obtain a denoised image;

[0110] The denoised image is dilated to obtain a dilated image, and the pixel values ​​of the dilated image are detected.

[0111] The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line.

[0112] The denoised image is corrected according to the rotation angle to obtain a corrected image, and the corrected image is subjected to contour detection to obtain the document text.

[0113] The data conversion module 11 is used to clean the text of the document and convert the cleaned text of the document to obtain the document structure information.

[0114] Optionally, the data conversion module 11 is also used to: obtain the names of commercial insurance items in the document text, and determine the content of the commercial insurance items according to the position of each commercial insurance item name;

[0115] Obtain the number of items in each item column of each commercial insurance item, and determine whether the number of items in each item column is equal to the preset number;

[0116] If the number of items in any of the project columns is less than the preset number, the commercial insurance project items corresponding to the number of items in the project column will be marked as abnormal content.

[0117] Obtain the text type of the text in the item column of each commercial insurance item, and compare the text type of each item column text with the text type of the adjacent item column text;

[0118] If the text type of any item column text fails to match the text type of the adjacent item column text, the commercial insurance item content corresponding to the item column text will be marked as the abnormal content.

[0119] If a deletion instruction is received for the abnormal content, then the abnormal content is deleted.

[0120] Furthermore, the data conversion module 11 is also used to: perform entity recognition on the text in the item column of each commercial insurance item to obtain the text entity type, and query the entity mapping type of the item column in each commercial insurance item.

[0121] The text entity types are matched with the entity mapping types of each item column, and the text in each item column is mapped according to the type matching results to obtain the document structure information.

[0122] The audit and verification module 12 is used to audit and verify the structured information of the document, and to reduce the data in the audited and verified structured information of the document.

[0123] Optionally, the audit and verification module 12 is also used to: query the medical insurance deduction ratio of each commercial insurance item name in the document structure information respectively, and calculate the target amount based on the medical insurance deduction ratio and the pricing amount in the document structure information;

[0124] For each commercial insurance item name, determine whether the amount payable in the structured information of the document is equal to the target amount;

[0125] If the amount due is not equal to the target amount, then the name of the commercial insurance project corresponding to the amount due will be marked as incorrect in the document structure information.

[0126] Furthermore, the audit and verification module 12 is also used to: obtain the field values ​​of each row in the document structured information respectively, and perform field verification on the document structured information according to the field values ​​to obtain the field verification result;

[0127] The document values ​​of each number in the document structure information are obtained respectively, and the document values ​​are verified according to the preset value range to obtain the value verification results.

[0128] The text of each line in the structured information of the document is obtained respectively, and the text of each line is validated according to a preset dictionary to obtain the text validation result.

[0129] The data in the document's structured information is reduced based on the field validation results, the numerical validation results, and the text validation results.

[0130] Result generation module 13 is used to generate commercial insurance audit results based on the document structured information after data reduction.

[0131] In this embodiment, text recognition is performed on commercial insurance claim forms to automatically extract the text. Data cleaning of the text improves its accuracy. Data conversion of the cleaned text achieves a structured processing effect. Review and verification of the structured information further enhances its accuracy. Data reduction of the reviewed and verified structured information effectively removes invalid data, further improving its accuracy. This embodiment automatically reviews and verifies commercial insurance claim forms, eliminating the need for manual review, thus improving trademark review efficiency and reducing user waiting time and service costs.

[0132] Example 4

[0133] Figure 5 This is a structural block diagram of a terminal device 2 provided in the fourth embodiment of this application. For example... Figure 5 As shown, the terminal device 2 in this embodiment includes: a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a program for a commercial insurance review method. When the processor 20 executes the computer program 22, it implements the steps in each embodiment of the above-described commercial insurance review method.

[0134] For example, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to complete this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, the processor 20 and the memory 21.

[0135] The processor 20 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0136] The memory 21 can be an internal storage unit of the terminal device 2, such as a hard disk or memory of the terminal device 2. The memory 21 can also be an external storage device of the terminal device 2, such as a plug-in hard disk, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard equipped on the terminal device 2. Furthermore, the memory 21 can include both internal and external storage units of the terminal device 2. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 can also be used to temporarily store data that has been output or will be output.

[0137] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0138] If an integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer-readable storage medium can be non-volatile or volatile. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the contents of a computer-readable storage medium may be appropriately added to or subtracted from the contents as required by the legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, a computer-readable storage medium may not include electrical carrier signals and telecommunication signals.

[0139] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A commercial insurance review method, characterized in that, The method includes: Obtain the commercial insurance claim form and perform text recognition on the commercial insurance claim form to obtain the text of the form; The document text is cleaned, and the cleaned document text is converted to obtain the document structure information. The structured information of the document is reviewed and verified, and data is reduced from the reviewed and verified structured information of the document. The commercial insurance audit result is generated based on the structured information of the documents after data reduction. The data cleaning of the document text includes: Obtain the names of commercial insurance items from the text of the document, and determine the content of the commercial insurance items based on the position of each commercial insurance item name; Obtain the number of items in each item column of each commercial insurance item, and determine whether the number of items in each item column is equal to the preset number; If the number of items in any of the project columns is less than the preset number, the commercial insurance project items corresponding to the number of items in the project column will be marked as abnormal content. Obtain the text type of the text in the item column of each commercial insurance item, and compare the text type of each item column text with the text type of the adjacent item column text; If the text type of any item column text fails to match the text type of the adjacent item column text, the commercial insurance item content corresponding to the item column text will be marked as the abnormal content. If a deletion instruction is received for the abnormal content, then the abnormal content is deleted.

2. The commercial insurance review method as described in claim 1, characterized in that, The text recognition process for the commercial insurance claim form includes: The commercial insurance claim form is binarized to obtain a binarized image, and the connected regions in the binarized image are obtained. Calculate the average pixel value of each connected region, and perform denoising processing on each connected region based on the average pixel value to obtain a denoised image; The denoised image is dilated to obtain a dilated image, and the pixel values ​​of the dilated image are detected. The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line. The denoised image is corrected according to the rotation angle to obtain a corrected image, and the corrected image is subjected to contour detection to obtain the document text.

3. The commercial insurance review method as described in claim 1, characterized in that, The data conversion of the document text after data cleaning includes: Entity recognition is performed on the text in the item column of each commercial insurance project to obtain the text entity type, and the entity mapping type of the item column in each commercial insurance project is queried. The text entity types are matched with the entity mapping types of each item column, and the text in each item column is mapped according to the type matching results to obtain the document structure information.

4. The commercial insurance review method as described in claim 1, characterized in that, The review and verification of the structured information in the document includes: Query the medical insurance deduction ratio for each commercial insurance item name in the structured information of the document, and calculate the target amount based on the medical insurance deduction ratio and the pricing amount in the structured information of the document; For each commercial insurance item name, determine whether the amount payable in the structured information of the document is equal to the target amount; If the amount due is not equal to the target amount, then the name of the commercial insurance project corresponding to the amount due will be marked as incorrect in the document structure information.

5. The commercial insurance verification method as described in any one of claims 1 to 4, characterized in that, The process of reducing data in the structured information of the document after review and verification includes: The field values ​​of each row in the document structure information are obtained respectively, and the field validation of the document structure information is performed based on the field values ​​to obtain the field validation results; The document values ​​of each number in the document structure information are obtained respectively, and the document values ​​are verified according to the preset value range to obtain the value verification results. The text of each line in the structured information of the document is obtained respectively, and the text of each line is validated according to a preset dictionary to obtain the text validation result. The data in the document's structured information is reduced based on the field validation results, the numerical validation results, and the text validation results.

6. A commercial insurance verification system, characterized in that, The system includes: The text recognition module is used to obtain commercial insurance claim documents and perform text recognition on the commercial insurance claim documents to obtain the document text; The data conversion module is used to clean the text of the document and convert the cleaned text to obtain structured information of the document. The review and verification module is used to review and verify the structured information of the document, and to reduce the data in the reviewed and verified structured information of the document. The result generation module is used to generate commercial insurance audit results based on the structured information of the document after data reduction. The data conversion module is also used to: obtain the names of commercial insurance items in the document text, and determine the content of commercial insurance items according to the position of each commercial insurance item name; Obtain the number of items in each item column of each commercial insurance item, and determine whether the number of items in each item column is equal to the preset number; If the number of items in any of the project columns is less than the preset number, the commercial insurance project items corresponding to the number of items in the project column will be marked as abnormal content. Obtain the text type of the text in the item column of each commercial insurance item, and compare the text type of each item column text with the text type of the adjacent item column text; If the text type of any item column text fails to match the text type of the adjacent item column text, the commercial insurance item content corresponding to the item column text will be marked as the abnormal content. If a deletion instruction is received for the abnormal content, then the abnormal content is deleted.

7. The commercial insurance verification system as described in claim 6, characterized in that, The text recognition module is also used for: The commercial insurance claim form is binarized to obtain a binarized image, and the connected regions in the binarized image are obtained. Calculate the average pixel value of each connected region, and perform denoising processing on each connected region based on the average pixel value to obtain a denoised image; The denoised image is dilated to obtain a dilated image, and the pixel values ​​of the dilated image are detected. The text in the denoised image is connected according to the pixel value detection results to obtain text lines, and the rotation angle is determined according to each text line. The denoised image is corrected according to the rotation angle to obtain a corrected image, and the corrected image is subjected to contour detection to obtain the document text.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images