Automatic examination program, document examination server apparatus, and document examination system

Abstract

To provide an automatic examination program or the like for making examination work more efficient.SOLUTION: A determination processing unit for each determination rule, which is at least one determination rule set for an examination object including examination items of an application document (40, 41) and an identity verification document (42, 43, 44), the determination processing unit determining the examination object in accordance with the determination rule; and an output processing unit outputting a determination result for the examination object by the normalization unit and / or the determination processing unit set as operation in an examination setting unit setting operation or non-operation of each of the normalization unit and the determination processing unit in accordance with the examination object.SELECTED DRAWING: Figure 9

Description

[Technical Field]

[0001] The present invention relates to an automatic review program for reviewing documents, a document review server device, and a document review system. [Background technology]

[0002] Systems have been developed to support document processing via a network. For example, Patent Document 1 discloses a business support system that includes a business workflow system that manages a business workflow in which image data optically read from an original document is used as an attachment, a collation workflow system that manages a collation workflow that proceeds independently of the progress of the business workflow and collates the image data with the original document, and a notification unit that notifies parties involved in the business workflow according to the progress of the collation workflow managed by the collation workflow system. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-156104 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in BPO (Business Process Outsourcing) businesses, which are service providers or service agents for applications, investigations, account opening, applications for various services, questionnaire surveys, etc., the application documents requested for review by customers and the required content of the review vary widely. Therefore, simply notifying relevant parties in the business workflow according to the progress of the collation workflow, as in conventional technology, makes it difficult to set up reviews that meet customer requests and improve the efficiency of the review work.

[0005] The present invention has been made in consideration of the above problems, and an object of the present invention is to provide an automatic examination program that makes examination work more efficient. [Means for solving the problem]

[0006] In order to solve the above problem, the invention described in claim 1 is an automatic screening program that performs screening using application documents and identification documents, comprising: a normalization means that normalizes text strings among the screening items of the application documents and the identification documents using a data normalization program that performs normalization processing as preprocessing on the text strings; a judgment processing means for each judgment rule that sets at least one judgment rule for a screening object including the screening items of the application documents and / or the identification documents and makes a judgment on the screening object in accordance with the judgment rule; and an output processing means that outputs the judgment result for the screening object by the normalization means and / or the judgment processing means that is set as the operation in an screening setting means that sets the operation or non-operation of the normalization means and each judgment processing means according to the screening object.

[0007] Furthermore, the invention described in claim 2 is characterized in that, in the automatic review program described in claim 1, the data normalization program includes a database for normalization, and the normalization means has an application programming interface function for exchanging data with the data normalization program.

[0008] Furthermore, the invention described in claim 3 is characterized in that, in the automatic screening program described in claim 1 or claim 2, the data normalization program normalizes the address of the screening item, and one of the multiple judgment processing means makes a mismatch judgment when a specified address category is normalized in the normalized address string.

[0009] Furthermore, the invention described in claim 4 is characterized in that, in the automatic review program described in claim 1 or claim 2, the data normalization program performs normalization of the name of the review item, and one of the multiple judgment processing means performs judgment of the normalized name.

[0010] In addition, the invention described in claim 5 is characterized in that, in the automatic review program described in claim 1 or claim 2, one of the multiple judgment processing means determines the boundary between the surname and given name in the kanji name of the review item, and performs a furigana match judgment by comparing the surname reading and given name reading of the kanji name with the furigana of the name of the review item.

[0011] Furthermore, the invention described in claim 6 is characterized in that, in the automatic review program described in claim 1 or claim 2, one of the multiple judgment processing means converts the date into a certain format and performs a date match judgment after the conversion.

[0012] Furthermore, the invention described in claim 7 is characterized in that, in the automatic review program described in claim 1 or claim 2, it functions as a judgment result output means that outputs the judgment results of the multiple judgment processing means and the normalized parts of the character string converted into text by the data normalization program to a display means of a document review system.

[0013] Furthermore, the invention described in claim 8 is characterized in that, in the automatic review program described in claim 7, the output processing means outputs the judgment result by the document review system and the judgment result by the judgment processing means to the outside, and the data normalized by the data normalization program is not output to the outside.

[0014] The invention described in claim 9 is characterized in that it comprises a document review server device that performs review using application documents and identification documents, a normalization means that normalizes text strings among the review items of the application documents and the identification documents using a data normalization program that performs preprocessing to normalize the text strings, a judgment processing means for each judgment rule that is set for an examination object that includes the review items of the application documents and / or the identification documents and that makes a judgment on the examination object in accordance with the judgment rule, and an output processing means that outputs the judgment result on the examination object by the normalization means and / or the judgment processing means that is set as the operation in an examination setting means that sets the operation or non-operation of the normalization means and each judgment processing means according to the examination object.

[0015] The invention described in claim 10 is a document review system that performs review using application documents and identification documents, and includes: a normalization means that normalizes text strings from the review items of the application documents and the identification documents using a data normalization program that performs preprocessing to normalize the text strings; a judgment processing means for each judgment rule that sets at least one judgment rule for a review object that includes the review items of the application documents and / or the identification documents and makes a judgment on the review object in accordance with the judgment rule; a review setting means that sets the operation or non-operation of the normalization means and each judgment processing means according to the review object; and an output processing means that outputs the judgment result for the review object by the normalization means and / or the judgment processing means that is set as the operation in the review setting means. [Effects of the Invention]

[0016] According to the present invention, there is provided a normalization means that performs normalization using a data normalization program that performs pre-processing to normalize the text strings of the screening items on application documents and identification documents; a judgment processing means for each judgment rule that has at least one judgment rule set for a screening object including the screening items on application documents and / or identification documents and that makes a judgment on the screening object in accordance with the judgment rule; and an output processing means that outputs the judgment result for the screening object by the normalization means and / or judgment processing means, which are set as operating in a screening setting means that sets the operation or non-operation of the normalization means and each judgment processing means according to the screening object.This allows the judgment rules to be applied to various screening objects for each business to be pre-set as screening settings in accordance with customer requests, i.e., for each business, and the screening settings can be applied to application documents and identification documents from applicants to perform automatic screening, thereby making the screening work more efficient. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a schematic diagram illustrating an example of a general configuration of a document examination system according to an embodiment of the present invention. [Figure 2A] FIG. 1 is a schematic diagram illustrating an example of a document. [Figure 2B] FIG. 1 is a schematic diagram illustrating an example of a document. [Figure 3] FIG. 1 is a schematic diagram illustrating an example of a document. [Figure 4] FIG. 1 is a schematic diagram illustrating an example of a document. [Figure 5] FIG. 1 is a schematic diagram illustrating an example of a document. [Figure 6] 2 is a block diagram showing an example of a schematic configuration of the document examination server device of FIG. 1. FIG. [Figure 7] FIG. 2 is a block diagram showing an example of a schematic configuration of the examination collaboration server device of FIG. 1. [Figure 8] 2 is a block diagram showing an example of a schematic configuration of the terminal device of FIG. 1. FIG. [Figure 9] 2 is a block diagram showing an example of a functional block of the document examination server device of FIG. 1. FIG. [Figure 10]2 is a flowchart showing an example of an operation for review setting in the document review system of FIG. 1. [Figure 11] FIG. 10 is a schematic diagram showing an example of an examination setting. [Figure 12] 2 is a flowchart showing an example of an examination operation in the document examination system of FIG. 1. [Figure 13] FIG. 10 is a schematic diagram illustrating an example of an examination question screen. [Figure 14] FIG. 10 is a schematic diagram showing an example of an examination result. [Figure 15] FIG. 10 is a schematic diagram showing an example of an examination result. [Figure 16] FIG. 10 is a schematic diagram showing an example of an examination result. [Figure 17] FIG. 10 is a schematic diagram showing an example of an examination result. DETAILED DESCRIPTION OF THE INVENTION

[0018] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the embodiment described below is an embodiment in which the present invention is applied to a document examination system.

[0019] [1. Document Review System Configuration and Functionality Overview] (1.1 Configuration and Functions of the Document Review System) First, the configuration and general functions of a document examination system according to one embodiment of the present invention will be described with reference to FIGS. 1 to 4. FIG.

[0020] Fig. 1 is a schematic diagram showing an example of the general configuration of a document examination system according to an embodiment of the present invention, Fig. 2A to Fig. 4 are schematic diagrams showing examples of documents.

[0021] As shown in Figure 1, the document review system 1 comprises a document review server device 10 that reviews documents related to applications, investigations, account opening, etc., a plurality of review collaboration server devices 20 that provide various functions related to the document review and perform processing, and a plurality of terminal devices 30 on which operators perform document review work.

[0022] The document examination server device 10 and the examination linkage server device 20 are servers of a BPO company that is a service provider or service agent for applications, investigations, account opening, applications for various services, questionnaire surveys, etc.

[0023] The document review server device 10 works in conjunction with the review collaboration server device 20 to perform automatic reviews on sets of documents received on paper or documents received via the web, and receives the results of operator reviews by workers from the terminal device 30.

[0024] The review collaboration server device 20 is a server device that performs information processing such as sorting document types, a server device that performs processing to cut out partial images of specific review items such as names from documents, a server device that performs OCR (Optical Character Recognition) processing, a server device that performs image matching processing, a server device that performs processing to normalize text strings, etc.

[0025] The terminal device 30 is, for example, a mobile terminal such as a personal computer, a portable wireless telephone including a smartphone, or a tablet terminal. The terminal device 30 is installed according to each worker who is an operator performing the inspection. The terminal device 30 displays images of documents, etc.

[0026] The document examination server device 10, the examination linkage server device 20, and the terminal device 30 are capable of transmitting and receiving data to and from each other via the network 3, using, for example, a communication protocol such as TCP / IP. The network 3 is, for example, a local area network. Note that the network 3 may also be constructed using the Internet, a dedicated communication line (for example, a CATV (Community Antenna Television) line), a mobile communication network, a gateway, or the like.

[0027] One example of a service provided by a BPO provider is an agency service that investigates the purpose of transactions by sending questionnaires or confirmation letters to customers who have opened accounts with financial institutions and already have transactions, depending on the transaction content and circumstances. This investigation is required because financial institutions are required to continuously verify customer information in order to strengthen measures against money laundering and terrorist financing. Customers must respond by mailing documents or accessing a website via email. Verification is required to be conducted periodically. For example, a BPO provider may print documents to be sent and mail them to customers or send them via email or other communication methods. Customers may handwrite the necessary information on questionnaires or other documents, seal them in an envelope, and send them to the financial institution along with a copy of their identification document or an image taken with a mobile device. Alternatively, they may enter the necessary information on a website and submit it. The BPO provider then receives the information mailed or entered on a website, extracts personal information from it, and conducts identity verification and other screening.

[0028] Next, we will provide an overview of the overall work flow of the services provided by BPO providers.

[0029] (1) Submitting documents In the case of paper-based reception, return envelopes containing a set of documents including a questionnaire, application forms for opening a bank account, etc., and a copy of identification documents are collected at a collection center, where workers open the envelopes and separate the application documents from the identification documents. The set of documents is scanned and converted into image data, and the application documents and identification documents are then submitted as image data. In the case of online reception of questionnaires, etc., users who hold bank accounts and are the subjects of the survey receive an email or an envelope. The email or envelope contains a URL for accessing the questionnaire to confirm the purpose of the transaction, etc., and users access it to complete the survey. A photograph of the identification documents is taken using a smartphone camera or similar, and the image of the identification documents is sent along with the questionnaire. The questionnaire is submitted as text data, and the identification documents are submitted as image data. (2) Pre-examination work The image data of each document in the set is sorted (determined by document type) by an automatic sorting system, and after determining the document type, a portion of the image is cut out to extract partial images of review items such as the name depending on the document type. (3) Examination In the document examination system 1, examination questions related to application information are created for examination items of documents such as application documents and identity verification documents, and examination is carried out. (4) Delivery to the customer Delivery data is prepared that includes text information that has been converted from application information and identity verification information, screening questions, screening results corresponding to the screening questions, details of defects corresponding to the screening results (for example, handwritten characters that are difficult to read, illegible due to dirt, illegible due to glare, illegible due to finger obscuring the screening items and illegible, etc.), and images of the identity verification documents.

[0030] In the present application, as an example, the embodiment of the examination work (3) above will be mainly described.

[0031] Next, the documents include questionnaires for confirming the purpose of the transaction, application documents such as an application form for opening a bank account, identity verification documents, contracts, delivery notes, etc. A set of documents, which is an example of multiple related documents, is, for example, in the case of confirming the purpose of the transaction, multiple documents such as questionnaires for confirming the purpose of the transaction, identity verification documents, etc. The set of documents may include a copy of the person's photograph and multiple identity verification documents.

[0032] As shown in Figures 2A and 2B, the questionnaire (customer information confirmation form) has columns for each item, including name, gender, date of birth, address, telephone number, place of employment, occupation, purpose of transaction, assets, etc., and often consists of multiple pages, such as questionnaire document 40 on the first page and questionnaire document 41 on the second page.

[0033] Identification documents include driver's licenses, resident registration cards, family register cards, health insurance cards, passports, My Number cards, and receipts (such as receipts for utility bills such as electricity and gas, which include the person's name and address).

[0034] Examples of document types include questionnaires, bank account opening application forms, identity verification documents, blank sheets, and other documents. More specifically, questionnaires include the first, second, and third pages of a questionnaire, while identity verification documents include driver's licenses, health insurance cards, passports, My Number cards, resident registration cards, receipts, and so on.

[0035] Documents are also classified into standard documents, in which the positions of items written on the document are fixed for each type of document, and non-standard documents, in which the positions of items written on the document are not fixed for each type of document. Standard documents include driver's licenses, passports, and My Number cards. Non-standard documents include resident registration cards, whose formats vary depending on the city, town, or village, health insurance cards, receipts, and so on, whose formats vary depending on the issuer. As shown in Figure 3, the format of a driver's license 42 is uniform nationwide, and the driver's license 42 is classified as a standard document.

[0036] On the other hand, as shown in Figure 4, the format of a health insurance card varies depending on the health insurance association to which one belongs, so it is classified as an atypical document, as in health insurance card 43. As shown in Figure 5, the format of a resident's certificate is not standardized, as in resident's certificate 44, so it is classified as an atypical document.

[0037] Regarding a questionnaire document, if the location of each item on the questionnaire document is known in advance by the entity that determines the document type, the questionnaire document may be considered to be a standard document.

[0038] (1.2 Configuration and Function of Document Review Server Device 10) Next, the configuration and functions of the document examination server device 10 will be described with reference to FIG.

[0039] FIG. 6 is a block diagram showing an example of a schematic configuration of the document examination server device 10. As shown in FIG.

[0040] As shown in Figure 6, the document review server device 10, which is a computer, includes a control unit 11 that controls the document review server device 10, a memory unit 12 that has various databases, a communication unit 13 that communicates with terminal devices 30, etc., and an output unit 14 that displays system management information, etc.

[0041] The control unit 11 includes, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The control unit 11 may also include a computing chip dedicated to AI calculations, such as a GPU (Graphics Processing Unit). The CPU 11 reads and executes various programs, including the automatic examination program of the present application, stored in the ROM or memory unit 12. The control unit 11 controls each unit (memory unit 12, communication unit 13, output unit 14, etc.) of the document examination server device 10. The control unit 11 may also read and execute these programs from a recording medium or the like that stores them.

[0042] The memory unit 12 (an example of a storage means) is configured with, for example, a hard disk drive, a silicon disk drive, etc. The memory unit 12 stores a database of data to be examined, such as examination questions created from the applicant's receipt data. The memory unit 12 stores an application of examination judgment rules for each examination method. The application of judgment rules is used in the form of a plug-in for each case.

[0043] The subject of the review is the review items such as the name and address written on the document, as well as the entire document. The review items include review items that can be converted to text, such as the name in kanji, the name in kana, the address, and the date (date of birth, date of entry), as well as review items that are images such as photographs and official seals.

[0044] The review methods include exact match review, partial match review, exact content match review, date comparison review, pseudonym name review, document sorting review, presence or absence review, facial photo comparison review, card thickness review, official seal comparison review, format judgment, combination review, and list review.

[0045] An exact match review is a review that compares two review items to determine whether they match exactly, etc. A partial match review is a review that compares two review items to determine whether one of the review items or the review subject contains the character string of the other review item. An exact content match review is a review that compares a review item with a pre-set character string to determine whether they match exactly, etc. A date comparison review is a review that compares the dates of two review items, or the date of a review item with a pre-set date, to determine whether they match, or whether the dates are larger or smaller, etc. A kana name review is a review that determines whether the reading of a kana name and a kanji name match, etc. A description presence review is a review that compares a review item with pre-set judgment conditions to determine whether or not there are characters that meet the conditions.

[0046] The document sorting inspection is an inspection to determine the type of document. The facial photo matching inspection is an inspection to calculate the similarity between two facial images and determine whether it is above a predetermined threshold. The card thickness inspection is an inspection to determine whether the thickness (inspection item) of a driver's license or other card photographed from an angle is appropriate. The official seal matching inspection is an inspection to determine whether there are any missing official seals (inspection item) on a driver's license or other card.

[0047] A combined audit is an audit that judges the results of multiple automated audits using the logical expressions "OR" and "AND." For example, when the combined condition is "OR," if "Automated Audit Result 1" is audited OK (passed) and "Automated Audit Result 2" is audited NG (failed), the combined audit result will be audited OK. When the combined condition is "OR," if "Automated Audit Result 1" is audited NG and "Automated Audit Result 2" is audited NG, the combined audit result will be audited NG. When the combined condition is "AND," if "Automated Audit Result 1" is audited OK and "Automated Audit Result 2" is audited NG, the combined audit result will be audited NG. When the combined condition is "AND," if "Automated Audit Result 1" is audited OK and "Automated Audit Result 2" is audited OK, the combined audit result will be audited OK.

[0048] Furthermore, a judgment can be made by combining logical expressions "OR" and "AND." For example, in the combined condition "Automatic Review Result 1" AND ("Automatic Review Result 2" OR "Automatic Review Result 3"), if "Automatic Review Result 1" is a review OK and "Automatic Review Result 2" is a review NG, and "Automatic Review Result 3" is a review OK, the combined review result will be a review OK.

[0049] List verification is a verification that compares the strings contained in a predetermined list file created in advance and makes a judgment. For example, if there is an exact match with any of the strings contained in the predetermined list file, the verification is judged to be OK.

[0050] Furthermore, strings in the list file may contain wildcards such as "*" (any string of 0 or more characters) and "?" (any single character). Furthermore, original wildcards may be set. Specifically, "···" may be set to any string of one or more characters, and "····" may be set to any string of two or more characters. For the string "A···BC", if the string being reviewed has one or more characters between "A" and "BC" (for example, ABBC, AABC, etc.), the review is judged to be OK. For the string "A····BC", if the string being reviewed has two or more characters between "A" and "BC" (for example, ABBBC, AAABC, etc.), the review is judged to be OK. For the string "···ABC", if the string being reviewed has one or more characters before "ABC" (for example, BABC, AABC, etc.), the review is judged to be OK. In the case of the string "ABC...", if the string being reviewed has one or more characters of any character string after "ABC" (for example, ABCD, ABCA, etc.), the review will be judged as OK.

[0051] A specific example of list review is shown below. If the contents of the list file are the string "ABCDE" and the string "ABC···" and the review target is "ABCDE Co., Ltd.", the review will be successful because it matches the string "ABC···". If the contents of the list file are the string "···ABC" and the string "A····BC" and the review target is "ABCDE", the review will be unsuccessful because there is no matching string. If the contents of the list file are the string "···ABC" and the string "A····BC" and the review target is "ABCBC", the review will be successful because it matches the string "A····BC".

[0052] The examination questions vary depending on the combination of the examination method, the document to be examined, and the examination items. For example, there are questions to examine whether the kanji name on the questionnaire document 40 matches the kanji name on the driver's license 42, whether the address on the driver's license 42 matches the address on the health insurance card 43, whether the reading of the kanji name on the questionnaire document 40 matches the kana name, whether the address on the questionnaire document 40 matches the address on the resident registration card 44, whether the face photo on the driver's license 42 matches a selfie, and whether the card thickness of the driver's license 42. The examination questions may include information on answer options for the examination questions (such as "match," "mismatch," and "unreadable").

[0053] They may be managed by the same database of the same document examination server device 10, or each database may be stored in a database of a different server device.

[0054] The storage unit 12 may also store various programs such as the automatic examination program of the present application, an operating system, and a server program. Note that the various programs may be acquired, for example, from another server or the like via the network 3, or may be recorded on a recording medium and read via a drive device.

[0055] The communication unit 13 is connected to the network 3 and controls communication with the examination linkage server device 20 and the terminal device 30 .

[0056] The output unit 14 is configured by, for example, a liquid crystal display element or an organic EL (Electro Luminescence) element.

[0057] (1.3 Configuration and Function of Examination Collaboration Server Device 20) Next, the configuration and functions of the examination linkage server device 20 will be described with reference to FIG.

[0058] FIG. 7 is a block diagram showing an example of the general configuration of the examination collaboration server device 20. As shown in FIG.

[0059] As shown in Figure 7, the review collaboration server device 20, which is a computer, includes a control unit 21 that controls the review collaboration server device 20, a memory unit 22 that has various databases, a communication unit 23 that communicates with the document review server device 10, etc., and an output unit 24 that displays system management information, etc.

[0060] The control unit 21 has, for example, a CPU, a ROM, and a RAM. The control unit 21 may have a calculation chip such as a GPU that is dedicated to performing AI calculations. The control unit 21 reads and executes various programs stored in the ROM or the memory unit 22 by the CPU, which reads and executes various control programs stored in the ROM or the RAM. The control unit 21 controls each unit (such as the memory unit 22, the communication unit 23, and the output unit 24) of the examination collaboration server device 20. Note that the control unit 21 may read and execute these programs from a recording medium or the like that stores them.

[0061] The storage unit 22 is configured by, for example, a hard disk drive, a silicon disk drive, or the like.

[0062] In each memory unit 22, various databases etc. are constructed for each function of each examination linkage server device 20.

[0063] For example, the storage unit 22 of the review linkage server device 20, which is used to set up the review of review items, etc., stores a business ID that identifies the business, a document type ID of the review target, an automatic review item ID of the review target, a review method ID, a flag indicating whether normalization processing is performed, a flag indicating whether automatic review is performed, etc. If there is a comparison target for the review target, the document type ID of the comparison target and the automatic review item ID of the comparison target are also associated and stored as review setting data in this storage unit 22. In the review setting, answer value patterns for answer options such as match, mismatch, and unreadable for the review answers may also be associated and set.

[0064] In addition, the predetermined storage unit 22 stores, as the applicant's receipt data, associations such as business IDs, applicant IDs identifying the applicants, document type IDs for each document, image data or text data for each document, each review item ID, partial image data corresponding to the review item or text data for the review item, etc. Here, the review item ID is an ID assigned to each review item in the workflow. The automatic review item ID is an ID assigned to each setting of automatic review in the document review system 1. By setting an automatic review item ID corresponding to a review item ID on the workflow side, i.e., on the document review system 1 side, the automatic review results for the automatic review item ID can be imported. In this way, the worker (operator) can refer to the results of the automatic review. Alternatively, the worker can perform operator review while referring to the results of the automatic review.

[0065] A database required for OCR processing is constructed in the storage unit 22 of the examination collaboration server device 20 that performs OCR processing. A database required for image matching processing is constructed in the storage unit 22 of the examination collaboration server device 20 that performs image matching processing.

[0066] The storage unit 22 of the review linkage server device 20, which performs the normalization process for addresses and names, contains databases necessary for the normalization process (examples of databases for normalization), including a surname dictionary database, a given name dictionary database, and an address dictionary database. This storage unit 22 stores a data normalization program that performs preprocessing to normalize text strings from the review items on application documents and identity verification documents. The data normalization program standardizes variations in the external character notation of names, complements prefectures and other information in addresses, updates administrative divisions, and standardizes the notation of chome, number, and street addresses.

[0067] Furthermore, a template database storing template images of each document is constructed in the memory unit 22 of the review collaboration server device 20, which determines the document type. This memory unit stores multiple sorting engine programs, such as a sorting engine capable of sorting standard documents by document type, a sorting engine capable of sorting non-standard documents by document type, and a sorting engine capable of sorting specific documents with high efficiency. A sorting engine is an algorithm or information processing method for determining the document type based on document image data.

[0068] The storage unit 22 of the examination linkage server device 20, which manages the business, has a business management database that manages the progress of the examination, the examination results, etc.

[0069] Furthermore, various trained machine learning models are constructed in the storage unit 22 for performing processes such as sorting documents by type and detecting document review items using AI. The trained machine learning models are, for example, deep learning models with a multilayer neural network structure, machine learning models for classification such as linear SVM (Support Vector Machine), and gradient boosted trees. In particular, in the case of AI that performs the process of detecting review items, training data in which information tags are added to each review item on the document through review by an operator may be created in annotation, and the AI ​​may be trained to perform machine learning to construct a machine learning model for detecting review items.

[0070] Various programs such as an operating system and a server program may also be stored in the storage unit 22. The various programs may be acquired, for example, from another server or the like via the network 3, or may be recorded on a recording medium and read via a drive device.

[0071] The communication unit 23 is connected to the network 3 and controls communication with the document examination server device 10 and the terminal device 30 .

[0072] The output unit 24 is configured by, for example, a liquid crystal display element or an organic EL element.

[0073] (1.3 Configuration and Function of Terminal Device 30) Next, the configuration and functions of the terminal device 30 will be described with reference to Fig. 8. Fig. 8 is a block diagram showing an example of the general configuration of the terminal device 30.

[0074] 3, the terminal device 30 functioning as a computer includes a communication unit 31, a storage unit 32, a display unit 33, an operation unit 34, a control unit 35, and an input / output interface unit 36. The control unit 35 and the input / output interface unit 36 ​​are connected via a system bus 37.

[0075] The communication unit 31 is connected to the network 3 and controls the state of communication with the document examination server device 10.

[0076] The storage unit 32 is configured by, for example, a hard disk drive, a silicon disk drive, etc., and stores various programs such as an operating system and a server program.

[0077] The display unit 33 (an example of a display means of the document examination system 1) is configured, for example, by a liquid crystal display element or an organic EL element. The operation unit 34 is configured, for example, by a keyboard and a mouse. The input / output interface unit 36 ​​performs interface processing between the communication unit 31 and the control unit 35.

[0078] The control unit 35 is composed of a CPU, a ROM, a RAM, and the like.

[0079] (1.4 Functional Blocks of Document Review Server Device 10) Next, the functional blocks of the document examination server device 10 will be described with reference to the diagram. Fig. 9 is a block diagram showing an example of the functional blocks of the document examination server device 10.

[0080] As shown in FIG. 9, the document examination server device 10 has an examination target data creation function, an automatic examination function, an examination result output function, an image matching system cooperation function, an OCR system cooperation function, and a normalization processing system cooperation function.

[0081] The review target data creation function applies the review setting information set for each business to the applicant's received data to generate review target data for automatic review. The review target data creation function may also generate review target data for operator review.

[0082] The automatic review function performs review processing on the review items of the review target data in accordance with the review judgment rules in accordance with the review settings. The automatic review function has a function that allows various judgment rules to be used in plug-in format. The automatic review function is an example of a judgment processing means for each judgment rule that makes a judgment on the review target in accordance with at least one judgment rule set for the review target including the review items of the application document and / or the identification document.

[0083] As a pre-processing for the screening process, if the data for the screening items is image data, the document screening server device 10 utilizes the OCR system linkage function to request OCR processing from the screening collaboration server device 20, which performs OCR processing, and acquires a character string that converts the image data of the screening items into text. For standard documents such as application forms and My Number cards, the position of the screening items is identified based on the position coordinates of the image, and for non-standard documents such as health insurance cards and resident registration cards, the position of the screening items is identified by AI, and then OCR is performed for each screening item.

[0084] When training the AI ​​to identify the location of the inspection items, the operator specifies the location where the inspection items are written. When creating training data for the AI ​​to identify the location of the inspection items, annotation to specify the location of the inspection items is required, but this is a separate task from the inspection work. Therefore, it is preferable to perform "semi-automated OCR annotation" that combines the annotation work with automatic inspection.

[0085] In this semi-automated OCR annotation, before automation, the operator performs the review process by specifying the location of review items, indicating the range of review items, on a screen that displays the entire document, such as an application form or identification document, using a mouse or other device, and then a segmented image is generated from the review items and OCR processing is performed. Of the series of tasks, the task of specifying the location of the review items is used for annotation. Since the document type and review items to be reviewed by the operator are known, learning data is created by linking the information on the document type and review items to be reviewed with the information on the location of the review items identified by the operator. During the operator's review process, information on the location of review items across various documents and review items is collected.

[0086] By using learning data that links information on the document type and review items to be reviewed with information on the location of the review items specified by the operator, for example, in the case of an insurance card, by specifying the location of the name only for the insurance card and having it learn, it is possible to create an AI that extracts the review item for the name on the insurance card.In this way, annotation work and review work can be performed together, reducing the amount of work required compared to performing each work separately.

[0087] The normalization processing system linkage function is an example of a normalization means that normalizes text strings among the review items of application documents and identification documents using a data normalization program that performs preprocessing to normalize the text strings. The normalization means has an application programming interface function that exchanges data with the data normalization program. When the review items are name and address, the document review server device 10 uses the normalization processing system linkage function to request the review collaboration server device 20, which performs normalization processing using the data normalization program, to normalize the name and address strings and receives the results.

[0088] When the examination item is a facial photograph and a facial photograph matching examination is performed, the document examination server device 10 utilizes the image matching system linkage function to request image matching from the examination collaboration server device 20 that processes image matching, and receives the results of facial photograph matching. When an official seal matching examination is performed, the document examination server device 10 utilizes the image matching system linkage function to request official seal matching from the examination collaboration server device 20 that processes image matching, and receives the results of official seal matching.

[0089] The screening result output function outputs the screening result corresponding to the screening question (screening OK or NG, reason for NG screening, etc.). The screening result output function may also output image data obtained by scanning a document or received image data, text information obtained by converting the screening items into text, and delivery data summarizing each screening question and the screening results corresponding to each screening question. The screening result output function is an example of an output processing means that outputs the judgment result for the screening object by the normalization means and / or the judgment processing means that is set as the operation in a screening setting means that sets the operation or non-operation of each of the normalization means and each judgment processing means according to the screening object.

[0090] [2. Example of document review system 1 operation] (2.1 Example of review settings in document review system 1) Next, an example of the operation of review setting in the document review system 1 will be described with reference to Fig. 10 and Fig. 11. Fig. 10 is a flowchart showing an example of the operation of review setting in the document review system 1. Fig. 11 is a schematic diagram showing an example of review setting.

[0091] As shown in FIG. 10, the document examination system 1 sets examination items (step S1). For each requested business, the examination target, comparison target, examination method, etc. are set in advance. Specifically, as shown in FIG. 11, an examination setting screen 50 for setting examination items is displayed on the display unit 23 of the examination collaboration server device 20. On the examination setting screen 50, the business, the type of document to be examined (e.g., "Document A" for application documents) and the examination items of that document (e.g., "Kanji name"), the type of document to be compared (e.g., "Driver's license" for personal identification documents) and the items, the examination method (e.g., "Exact match examination"), whether or not normalization processing is performed, etc. are input. In the case of comparing the date of birth between a driver's license and a health insurance card, the examination target is "Driver's license (date of birth)", the comparison target is "Health insurance card (date of birth)", the examination method is "Date comparison examination", and normalization is "Yes".

[0092] In addition, answer options for operator review (for example, "match," "mismatch," "unreadable," etc.) may be set on the review setting screen 50. In the case of automatic review, the judgment results corresponding to the answer options (for example, the respective answer values ​​corresponding to "match," "mismatch," "unreadable," etc.) are programmed in the application of the review judgment rules.

[0093] The review linkage server device 20 for setting review items functions as an example of a review setting means for setting the operation or non-operation of the normalization means and each of the judgment processing means according to the review subject.

[0094] Next, the document review system 1 registers the review settings (step S2). Specifically, the review collaboration server device 20 for setting review items associates the setting number, the automatic review item ID of the set review target, the automatic review item ID of the comparison target, the review method ID, a flag indicating whether normalization is performed, etc. with the business ID in the storage unit 22. Usually, multiple review targets are set for one business.

[0095] The example of the operation of the examination setting may be performed by the document examination server device 10.

[0096] (2.2 Example of the review process in Document Review System 1) Next, an example of the screening operation in the document screening system 1 will be described with reference to the drawings. Fig. 12 is a flowchart showing an example of the screening operation in the document screening system 1. Fig. 13 is a schematic diagram showing an example of a screening question screen. Figs. 14 to 17 are schematic diagrams showing examples of screening results.

[0097] First, as for the above-mentioned document submission (1), in the case of an application via the web, the document submission examination collaboration server device 20 acquires a series of image data of a set of documents, such as questionnaire documents and identity verification documents, from the applicant's user terminal device and stores them in the memory unit 22 together with the applicant ID for each applicant and the document type ID issued for each document. The identity verification documents are image data photographed with a mobile terminal device equipped with a camera. The set of documents may also include image data of the applicant's face.

[0098] In addition, in the case of questionnaire documents such as questionnaire documents 40 and 41, they may be acquired as text data instead of image data. The document submission examination collaboration server device 20 acquires data from the applicant's user terminal device that associates a document type ID, an item ID indicating each item, and text data entered in the field for each item displayed on the applicant's user terminal device.

[0099] In the case of paper applications, the envelopes that have been mailed are opened. For example, the delivered cardboard box is opened, and an envelope containing a set of documents, such as a questionnaire with necessary information and copies of identification documents, is removed. The envelope is then opened, and the set of documents is removed. An image input device scans each document in the set, and each document is converted into image data. The image input device includes a scanner or digital camera with an imaging element such as a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary Metal Oxide Semiconductor) image sensor.

[0100] Next, as pre-examination work (2), the document examination system 1 performs tasks such as sorting documents (determining the document type), and after determining the document type, extracting partial images of the applicant for examination items such as name according to the document type. These tasks are performed by a dedicated examination linkage server device 20, and associations such as business IDs, applicant IDs that identify the applicants, document type IDs for each document, image data or text data for each document, each examination item ID, partial image data corresponding to the examination items or text data for the examination items, etc. are stored in a database in a predetermined storage unit 22 as receipt data for the applicant.

[0101] Next, the document examination system 1 performs the following process as examination work (3).

[0102] 12, the document review system 1 acquires information on review settings (step S10). Specifically, the document review server device 10 acquires information on the review settings that have been set in advance, based on the business ID, from the memory unit 22 of the review collaboration server device 20 for setting the review items, such as the setting number (automatic review ID), the automatic review item ID of the set review target, the automatic review item ID of the comparison target, the review method ID, the flag indicating whether normalization is performed, and the pattern of answer options.

[0103] For example, in the case of an exact match review between a kanji name on a questionnaire document and a kanji name on an identity verification document, the review setting information includes the document type ID of the review target (document type ID of the questionnaire document), the automatic review item ID of the review target (automatic review item ID of the kanji name on the questionnaire document), the document type ID of the comparison target (document type ID of the identity verification document), the automatic review item ID of the comparison target (automatic review item ID of the kanji name on the identity verification document), the review method ID for the exact match review, etc. In the case of an exact match review between the address on a driver's license which is an identity verification document and the address on a health insurance card which is an identity verification document, the review setting information includes the document type ID of the review target (document type ID of the driver's license), the automatic review item ID of the address on the driver's license, the document type ID of the comparison target (document type ID of the health insurance card), the automatic review item ID of the address on the health insurance card, the review method ID for the exact match review, etc. In the case of a review to check whether the readings of the kanji name and kana name on a questionnaire document match, the review setting information includes the document type ID of the review target (document type ID of the questionnaire document), the automatic review item ID of the kanji name on the questionnaire document, the document type ID of the comparison target (document type ID of the questionnaire document), the automatic review item ID of the kana name on the questionnaire document, and the review method ID of the kana name review.In the case of reviewing the expiration date of an identification document, the review setting information includes the document type ID of the review target (document type ID of the driver's license), the automatic review item ID of the driver's license expiration date, the automatic review item ID of the comparison target (a text value indicating the date of receipt, etc.), and the review method ID of the date comparison review.In the case of comparing the facial photograph on an identification document with a selfie, the review setting information includes the document type ID of the review target (document type ID of the driver's license), the document type ID of the comparison target (document type ID of a copy of the selfie), and the review method ID of the facial photograph matching review. In the case of verifying and judging the issue number of a personal identification document, the information in the review settings includes the document type ID to be reviewed (document type ID of driver's license), the automatic review item ID of the driver's license issue number, and the review method ID for format judgment.

[0104] A single business operation usually has multiple audit items as audit targets, and even for a single audit item, there can be multiple audit methods.

[0105] Next, the document review system 1 generates review target data (step S11). Specifically, the document review server device 10 reads the applicant's receipt data based on the business ID and applicant ID, applies the review setting information to the applicant's receipt data, and generates review target data for review. More specifically, the document review server device 10 generates review target data for each review question for automatic review using an automatic review data creation function. The generated review target data includes, for each text data of review items in the received data, an automatic review ID, an automatic review item ID of the review target, an automatic review item ID of the comparison target, a review method ID, and a flag indicating whether or not normalization is performed. Note that if the received data is an image and specific items in the image are used as review items, the review collaboration server device 20, which performs OCR processing, requests OCR processing on the image data, receives text data of character strings that have been converted from the specific items in the image data, assigns an automatic review item ID, and uses the text data as the review target. Note that if the comparison target is a document included in the received data, text data corresponding to the review items to be compared may also be included in the review target data. When the examination method is an image matching examination such as a facial photograph examination or an official seal examination, the examination item specifies a document type ID indicating full or partial image data.

[0106] The data to be screened may include data to be screened for operator screening. In the case of data to be screened for operator screening, data to be screened for a screening question format for a screening question image 51 as shown in FIG. 13 is generated. The data to be screened may include screening questions for operator screening (for example, "Does the name on document A match the name on text A?"), image data of the identification document, and answer options such as "match," "mismatch," and "unreadable." Note that if "match," the screening may be OK, and if "mismatch" or "unreadable," the screening may be NG.

[0107] In addition, the examination linkage server device 20 for generating examination data may generate the examination target data.

[0108] Next, the document examination system 1 performs a data normalization process (step S12). Specifically, the document examination server device 10 performs a normalization process on the examination target data for which normalization is set to "yes". More specifically, the document examination server device 10, using the normalization processing system linkage function, requests the examination linkage server device 20 for normalization processing to perform a normalization process on the text data of the examination target examination items (e.g., address, name) for which normalization is set to "yes", and receives a response of the normalization process for each of the target examination items.

[0109] Here, when the examination item is an address, the following normalization process is performed by the data normalization program of the examination linkage server device 20 for normalization process. Here, the predetermined category of the address is, for example, prefecture, city, ward, town, village, house number, etc. (1) In the case of "○○ 1-chome, 2-banchi, Urawa-ku, Saitama-shi, Saitama Prefecture," the notation of address numbers is standardized as "○○ 1-chome, 2-banchi, Urawa-ku, Saitama-shi, Saitama Prefecture." The address "○○ 1-chome, 2-banchi, 3-go" is an example of a predetermined address division. (2) When the prefecture is not specified, such as "1-2-3 ○○, Urawa-ku, Saitama-shi," the address is normalized with the prefecture included, such as "1-2-3 ○○, Urawa-ku, Saitama-shi, Saitama-ken." The prefecture of an address is an example of a predetermined address category. (3) In the case of the name of an old administrative district, such as "○○ 1-2-3, Urawa City, Saitama Prefecture," it is normalized to the name of the latest administrative district, such as "○○ 1-2-3, Urawa Ward, Saitama City, Saitama Prefecture." The city, ward, town, or village of an address is an example of a predetermined division of an address. (4) When the name of a part of an administrative district, such as "Urawa Ward," is missing, such as "2-3 XX 1-chome, Saitama City, Saitama Prefecture," normalization processing is performed to add "Urawa Ward." The city, ward, town, or village of an address is an example of a predetermined division of an address.

[0110] The document review server device 10 receives information on the normalized parts and information on which type of normalization was performed (1) to (4) above from the review collaboration server device 20 for normalization processing as a response to the normalization processing.

[0111] In addition, the document review server device 10 may allocate the review target to automatic review or review by an operator according to the types of the normalization processes (1) to (4) above. For example, when the normalization processes (1) and (2) above are performed on the review target, the automatic review process is performed. On the other hand, when the character string completely changes as in (3) and (4) above, review by an operator may be performed. Thus, the document review server device 10 has a function that the data normalization program normalizes the address of the review item, and one of the plurality of determination processing means makes a non-conformity determination when a predetermined section of the address is normalized in the normalized address character string.

[0112] When the review item is a name, the data normalization program of the review cooperation server device 20 for normalization processing unifies the notation fluctuations of foreign characters. For example, Chinese characters such as “斉”, “斎”, and “齋” are unified as “斉”.

[0113] Thus, the document review server device 10 functions as an example of normalization means that performs normalization by a data normalization program that normalizes the texturized character string among the review items of the application documents and the identity confirmation documents as a preprocess. The document review server device 10 functions as an example of normalization means that has an application programming interface function for transferring data with the data normalization program.

[0114] Next, the document review system 1 determines the review target according to the determination rules (step S13). Specifically, the document review server device 10 reads out the application or program of the review determination rules of the review method based on the review method ID of each review method set for each review item of the review target data, and makes a determination on the text data of each review item of the review target data.

[0115] In the case of an exact match inspection between the kanji name on the questionnaire and the kanji name on the identity verification document, the document examination server device 10 uses a plug-in for the judgment rules for exact match inspection to judge whether the normalized text data of the kanji name on the questionnaire document of the data to be examined matches the normalized text data of the kanji name on the identity verification document, and outputs a response value corresponding to "match" or "mismatch." The document examination server device 10 may also output information (such as the reason for NG) for the worker to confirm the automatic examination results as shown in Figure 14.

[0116] In the case of an exact match inspection between the address on a driver's license, which is an identification document, and the address on a health insurance card, which is also an identification document, the document inspection server device 10 uses a plug-in for the judgment rules for exact match inspection to judge whether the normalized text data of the address on the driver's license and the normalized text data of the address on the health insurance card of the data to be inspected match exactly, and outputs a response value corresponding to "match" or "mismatch." The document inspection server device 10 may also output information (such as the reason for NG, inspection items before normalization) for an operator to confirm the automatic inspection results as shown in Figure 15.

[0117] In the case of an exact match review between the furigana name on the questionnaire document and the furigana name on another document, the document review server device 10 uses a plug-in for the judgment rules for exact match review to judge whether the text data of the furigana name on the questionnaire document of the data to be reviewed matches the text data of the furigana name on another document, and outputs a response value corresponding to "match" or "mismatch." The document review server device 10 may also output information (such as the reason for NG) for the worker to confirm the automatic review results as shown in Figure 16.

[0118] In the case of a date comparison review between the date of birth on the questionnaire document and the date of birth on the identification document, the document review server device 10 uses a plug-in for the judgment rules of the date comparison review to judge whether the dates match and output a response value corresponding to "match / condition met" or "mismatch / condition not met." When comparing dates, the display format of dates such as "December 21, 2022," "December 21, 2022," and "2022 / 12 / 21" is converted to the "YYYYMMDD" format, such as "20221221," and after the conversion, a date comparison (date match / size comparison) is performed. Normalization, such as converting dates into a specific format or removing specific character strings, may be included in a date comparison and review plug-in. The document review server device 10 may output information (such as reasons for rejection) for the worker to confirm the automatic review results, as shown in Figure 17. The review target after normalization and before normalization may be displayed, such as "20221221 (December 21, 2022)" in Figure 17. It may also be possible to specify an additional date for the comparison target date. It may also be determined by comparing whether the review target date is within the "comparison target date + additional date." The unit of the additional date for the comparison target may be selectable from "days," "months," or "years." If the additional date changes year, such as by "year," the year of the comparison target may be aligned with the year of the review target for comparison.

[0119] When comparing the facial photograph on an identification document with a selfie, the document screening server device 10 uses a plug-in for the judgment rules for facial photograph matching screening to judge whether the facial photograph matches the full or partial image data of the photograph on the driver's license of the data to be screened with the image data of the selfie, and outputs a response value corresponding to "match" or "mismatch." Note that the plug-in for the judgment rules for facial photograph matching screening may cooperate with the screening collaboration server device 20, which performs the image matching process, and send the full or partial image data of the photograph on the driver's license of the data to be screened and the image data of the selfie to the screening collaboration server device 20, and receive the response value.

[0120] Furthermore, when determining whether the kanji name and the furigana name on the questionnaire match, the document examination server device 10 performs the following process: It is assumed that the kanji name has been normalized.

[0121] When the kanji name on the questionnaire is divided into a surname and a given name, such as "Kanji Taro," a surname dictionary is referenced for the surname, and a name dictionary is referenced for the given name, and a list of pronunciation candidates is created for each (surname: Kanji, given name: Taro / Tera / Ryutaro). The document screening server device 10 uses a plug-in for the kana name screening rules to output a response value for the match determination between one of the pronunciation candidate lists (surname: Kanji, given name: Taro / Tera / Ryutaro) created from the kanji name and the furigana name (surname: Kanji, given name: Taro).

[0122] If the kanji name on the questionnaire is written with the family name and given name connected, such as "Kanji Taro," the boundary between the family name and given name in the kanji name is determined as follows, and a match between the kanji name and the furigana name is determined.

[0123] The document examination server device 10 acquires the kanji name by adding one character at a time from the beginning, and each time refers to a family name dictionary to acquire a list of pronunciation candidates and determine whether the initial part matches the furigana name. Next, if there is no match, the kanji name is added one character at a time and the process is repeated until all kanji names have been referenced. If there is a match, the remaining part of the kanji name is considered to be the given name, and a name dictionary is referred to to acquire a list of pronunciation candidates and determine whether it completely matches the remaining part of the furigana name.

[0124] For example, in the case of "Yamada Taro" and "Yamada Taro," first, the surname candidate is set to "Yama" and a search is performed in the surname dictionary to obtain a list of pronunciation candidates for "Yama," "Yama." Then, a match is performed between the pronunciation candidate list "Yama" and the kana name "Yamada Taro" and the initials are determined to match up to "Yama." Then, "Tataro" is set as the given name candidate and a search is performed in the name dictionary, with no results for "Tataro."

[0125] If the name dictionary does not find a search, the search starts again from extracting surname candidates, and the surname candidate is set to "Yamada." A search of the surname dictionary results in a list of pronunciation candidates: "Yamagata," "Yamata," "Yamada," and "Yoda." A match is performed between the initials of these pronunciation candidates "Yamagata," "Yamata," "Yamada," or "Yoda" and the furigana name "Yamada Tarou," and since "Yamada" matches, the given name candidate becomes "Taro." A name dictionary search is performed on the given name candidate "Taro," and the pronunciation candidate list returns "Taro," "Tera," and "Ryutaro." These pronunciations, "Taro," "Tera," or "Ryutaro," match the given name "Taro" in furigana. Since the pronunciations of both the surname and given name match, the search results are determined to be a match.

[0126] In this way, the document examination server device 10 functions as an example of a judgment processing means for each judgment rule that makes a judgment on the examination object according to at least one judgment rule set for the examination object including the examination items of the application documents and / or the identification documents. Also, one of the multiple judgment processing means of the document examination server device 10 functions as an example of making a furigana match judgment by determining the boundary between the family name and given name in the kanji name of the examination item and comparing the reading of the family name and the reading of the given name in the kanji name with the furigana of the name of the examination item.

[0127] If the review result is determined to be not a "match" (step S14; NO), the document review system 1 performs an operator review (step S15). Specifically, the document review server device 10 transmits the results of the automatic review, together with the business ID, automatic review ID, etc., to the terminal device 30 of a predetermined operator as evidence of the automatic review. The terminal device 30 displays the review results of the automatic review on the display unit 33, as shown in Figure 14, Figure 15, Figure 16, or Figure 17. The operator compares the review target text with the comparison target text and inputs a response of "match" indicating a successful review or "mismatch" indicating a failed review. As shown in Figure 16, if the difference is the presence or absence of a semi-voiced consonant mark, the operator may determine the result as a "match." The terminal device 30 transmits the response value to the document review server device 10, together with the business ID, review item ID, etc.

[0128] If the inspection result is NG, the inconsistencies in the inspection target may be highlighted or the reason for NG may be displayed, as shown in Figures 14 to 17. As shown in Figure 15, the inspection items of the inspection target before and after normalization processing may be displayed.

[0129] In addition, the terminal device 30 may display the results of the automatic screening as shown in Figure 14, Figure 15, Figure 16 or Figure 17 on the display unit 33 together with the question image as shown in Figure 13, or may be configured to display the results of the automatic screening at the operator's selection.

[0130] If the results of the automatic screening and the results of the operator screening do not match, the results of the automatic screening and the results of the operator screening may be sent to the terminal device 30 of the operator in charge of the re-screening in order to conduct a re-screening.

[0131] In addition, in the normalization process of step S12, (3) if the name of the latest administrative district is normalized, or (4) if normalization processing is performed to complement the name of an administrative district smaller than a prefecture, the subject of review will also be sent for operator review.

[0132] When a combined screening rule plugin is used, the combined screening may output a final screening result for each of the screening results for multiple screening targets, such as name, address, and date, as well as the screening results for each document type. The combined screening allows multiple automated screening results to be linked by logical expressions for screening. For example, when verifying personal identification documents, if the exact match screening rule plugin is configured to check (1) a match between the driver's license and name and (2) a match between the My Number card and name, linking the results of (1) a match between the driver's license and name and (2) a match between the My Number card and name with an OR condition can be used to indicate that the screening is successful if either the driver's license or the My Number card is confirmed. If a photo ID is required, but either a driver's license or My Number card is acceptable, using the combined screening rule to link them with an OR condition allows the application to be processed regardless of the document submitted. The results of the operator screening may also be included in the combined screening.

[0133] Also, if the screening question is, "Is a delivery note attached?", it is effective to use a plugin with list screening rules. There are various titles that are considered to be delivery notes, and many possible character string patterns, so using a list screening plugin that sets the character strings of multiple titles that are considered to be delivery notes all at once makes things less complicated. In this way, for list screening, a text file is prepared that lists multiple titles that are considered to be delivery notes, and if a title matches any of them, the screening is judged to be OK.

[0134] If the review results are determined to be a match (step S14; YES), or after the operator review in step S15 is completed, the document review system 1 outputs delivery data (step S16). Specifically, the document review server device 10 combines the automatic review results, the item names and contents of the operator review, the target image data such as application documents and identification documents, and the results of the document type for each document, to generate and output delivery data as the final results. The delivery data may include, for example, scanned image data or received image data, text information in which the review items are converted into text, review questions, and review results corresponding to the review questions (review OK or review NG, reason for review NG, etc.). It is not necessary to include all processing results in the delivery data; for example, the delivery data may not include normalized character strings.

[0135] Note that even if the examination result is determined not to be a "match," the operator examination in step S15 may be omitted. In this case, the process is fully automated without operator examination, further improving the efficiency of the examination.

[0136] Furthermore, even if the inspection result is determined to be a "match," an operator inspection may be performed as in step S15. The double judgment of automatic inspection and operator inspection further improves the accuracy of the inspection.

[0137] According to this embodiment, a normalization means performs normalization using a data normalization program that performs pre-processing to normalize the text strings of the screening items on application documents and identification documents; a judgment processing means for each judgment rule that has at least one judgment rule set for a screening object including the screening items on application documents and / or identification documents and that makes a judgment on the screening object in accordance with the judgment rule; and an output processing means that outputs the judgment result for the screening object by the normalization means and / or judgment processing means, which are set as operating in a screening setting means that sets the operation or non-operation of the normalization means and each judgment processing means according to the screening object.This allows the judgment rules to be applied to various screening objects for each business to be pre-set as screening settings in accordance with customer requests, i.e., for each business, and the screening settings can be applied to application documents and identification documents from applicants to perform automatic screening, thereby making the screening work more efficient.

[0138] Furthermore, multiple judgment rules can be arbitrarily combined as appropriate depending on the case and the screening item. For example, by performing the judgment process after data normalization, the accuracy of automatic screening can be improved. Furthermore, by dividing the roles of data normalization and the judgment processing means, arbitrary combinations can be made as appropriate depending on the case. That is, for screening items where normalization is better performed before the judgment process by the judgment processing means, the normalization means and the judgment processing means can be combined, and for screening items where normalization is not necessary, only the appropriate judgment processing means can be selected. Flexible response using arbitrary combinations depending on the case prevents unnecessary processing and enables efficient processing.

[0139] If the data normalization program includes a database for normalization and the normalization means has an application programming interface function for transferring data to and from the data normalization program, the automated review program will have an API function, enabling it to exchange data with the data normalization program; and by having this data normalization program outside the automated review program, it will be easier to keep up with the updating of the database for normalization.

[0140] In addition, the data normalization program normalizes the address of the review item, and one of the multiple judgment processing means determines a mismatch when a specified section of the address is normalized in the normalized address string, thereby preventing erroneous judgments due to matches in parts other than those to be normalized when the item being reviewed is an address.

[0141] When the name of the review item is normalized and one of the plurality of judgment processing means judges the normalized name, if the review item to be judged is a name, an accurate judgment can be made even if the kanji in the name are external characters.

[0142] One of the multiple judgment processing means determines the boundary between the surname and given name in the kanji name of the review item, and compares the reading of the surname and given name in the kanji name with the furigana of the name of the review item, thereby making an accurate judgment when a furigana match judgment is made when the review item being reviewed is a comparison of the kanji name and the furigana match.

[0143] One of the multiple judgment processing means converts the date into a certain format and performs a date match judgment after conversion, and if the review item to be reviewed is a date, an accurate date judgment can be performed.

[0144] If the judgment results of multiple judgment processing means and the normalized parts of the character string converted into text by the data normalization program are output to the display means of the document review system, the operator can be informed of the judgment results of the automatic review.

[0145] If the judgment results by the document examination system 1 and the judgment results by the judgment processing means are output to the outside, and the data normalized by the data normalization means is not output to the outside, it can be used as delivery data in accordance with the customer's request. [Explanation of symbols]

[0146] 1: Document review system 10: Document review server device 20: Examination collaboration server device 30: Terminal device 40, 41: Questionnaire documents (application documents) 43: Driver's license (identification document) 44: Health insurance card (identification document)

Claims

1. An automated screening program that performs screening using application documents and identity verification documents, normalization means for normalizing character strings converted into text from among the examination items of the application documents and the personal identification documents using a data normalization program that performs normalization processing as preprocessing; a judgment processing means for each of the judgment rules, which is configured to make a judgment on the examination object according to at least one judgment rule set for the examination object including the examination items of the application documents and / or the personal identification documents; and an output processing means for outputting a judgment result for the subject of examination by the normalization means and / or the judgment processing means, which is set as the operation in an examination setting means for setting the operation or non-operation of each of the normalization means and each of the judgment processing means according to the subject of examination; An automated screening program that allows a computer to function as an

2. The automatic examination program according to claim 1, the data normalization program includes a database for normalization; The automatic examination program is characterized in that the normalization means has an application programming interface function for exchanging data with the data normalization program.

3. 3. The automated examination program according to claim 1, The data normalization program performs address normalization of the screening items, An automatic screening program characterized in that one of the multiple judgment processing means judges a mismatch when a specified section of the address is normalized in the normalized address string.

4. 3. The automated examination program according to claim 1, The data normalization program normalizes the names of the review items, An automatic screening program characterized in that one of the plurality of judgment processing means performs judgment of the normalized name.

5. 3. The automated examination program according to claim 1, An automatic review program characterized in that one of the multiple judgment processing means determines the boundary between the surname and given name in the kanji name of the review item, and performs a furigana match determination by comparing the surname reading and given name reading of the kanji name with the furigana of the name of the review item.

6. 3. The automated examination program according to claim 1, An automatic review program characterized in that one of the plurality of judgment processing means converts the date into a specific format and performs a date match judgment after the conversion.

7. 3. The automated examination program according to claim 1, An automatic review program characterized by functioning as a judgment result output means that outputs the judgment results of the multiple judgment processing means and the normalized parts of the character string converted into text by the data normalization program to a display means of a document review system.

8. The automatic examination program according to claim 7, An automatic review program characterized in that the output processing means outputs the judgment results by the document review system and the judgment results by the judgment processing means to the outside, and the data normalized by the data normalization program is not output to the outside.

9. A document review server device that performs review using application documents and identity verification documents, normalization means for normalizing character strings converted into text from among the examination items of the application documents and the personal identification documents using a data normalization program that performs normalization processing as preprocessing; a judgment processing means for each of the judgment rules, which is configured to make a judgment on the examination object according to at least one judgment rule set for the examination object including the examination items of the application documents and / or the personal identification documents; an output processing means for outputting a judgment result for the subject of examination by the normalization means and / or the judgment processing means, which is set as the operation in an examination setting means for setting the operation or non-operation of each of the normalization means and each of the judgment processing means according to the subject of examination; A document examination server device comprising:

10. In a document review system that performs review using application documents and identity verification documents, normalization means for normalizing character strings converted into text from among the examination items of the application documents and the personal identification documents using a data normalization program that performs normalization processing as preprocessing; a judgment processing means for each of the judgment rules, which is configured to make a judgment on the examination object according to at least one judgment rule set for the examination object including the examination items of the application documents and / or the personal identification documents; an examination setting means for setting whether the normalization means and each of the judgment processing means are operated or not in accordance with the examination target; an output processing means for outputting a judgment result for the subject of the examination by the normalization means and / or the judgment processing means, which is set as the operation in the examination setting means; A document examination system comprising:

Images