JP7714930B2 - Character matching device and program - Google Patents

Description

The present invention relates to a character matching device and program.

In the past, information processing devices have been disclosed that identify character strings representing names, company names, store names, etc. from among character strings printed on specific media (for example, Patent Document 1). When printed characters are read using an optical character recognition device (OCR), incorrect recognition can occur, resulting in incorrect text being output. Patent Document 1 presents a method for matching names, etc., using information from email addresses and network addresses.

Patent No. 4991407

For example, during the review process for various operations, such as account opening, card issuance, and insurance contract processing, it is confirmed that the name and other information written on application documents matches the information on identification documents such as driver's licenses. Performing these checks visually by humans increases the workload and can ultimately lead to errors. For this reason, in recent years, computers have been used to compare text and determine whether it matches or does not. However, even if the comparison results in a mismatch, there are cases where it is acceptable under certain conditions, and this determination was not possible.

The present invention aims to provide a character matching device and program that, when a character matching result is a mismatch, notifies the user based on the nature of the mismatch.

The present invention solves the above problems by the following means.
A first invention is a character matching device comprising: a comparison means for comparing first text for a predetermined item obtained from a first document with second text for the predetermined item obtained from a second document; a determination means for determining the content of the mismatch when the comparison by the comparison means results in a mismatch; and a notification means for issuing a notification corresponding to the content of the mismatch determined by the determination means.
A second invention is the character matching device of the first invention, wherein the determination means determines that the mismatched content is a variant character when the mismatched pair of characters is registered in a variant character dictionary that stores variant character variants, and the notification means notifies the user that the mismatched content is a variant character.
A third invention is a character matching device according to the first or second invention, wherein the determination means calculates a similarity between a pair of mismatched characters, and when the calculated similarity is equal to or greater than a threshold, determines that the mismatch is a variant character or a mistranslation, and the notification means notifies the user that the mismatch is a variant character or a mistranslation.
A fourth aspect of the present invention is the character matching device of the third aspect of the present invention, wherein the determination means divides an image area showing a pair of mismatched characters into a plurality of small areas, and compares the small areas at the same position to calculate the similarity.
A fifth invention is a character matching device according to any one of the first to fourth inventions, further comprising a character count confirmation means for confirming the number of characters in the first text and the second text, and the comparison means, when the character count confirmation means confirms that the numbers of characters match, compares the first text and the second text one character at a time.
A sixth invention is the character matching device of the fifth invention, further comprising a similarity calculation means for, when the character number confirmation means confirms a mismatch in the number of characters, identifying each mismatched character and calculating a similarity between the identified combination of multiple characters and a paired character, and the notification means notifies the user that the conversion has been erroneous when the similarity calculated by the similarity calculation means is equal to or greater than a threshold value.
A seventh aspect of the present invention is the character matching device according to any one of the first to sixth aspects of the present invention, wherein the predetermined items include at least one of a name and an address.
An eighth invention is a character matching device according to any one of the first to seventh inventions, comprising: first data acquisition means for acquiring first data obtained by imaging the first document; second data acquisition means for acquiring second data obtained by imaging the second document; first text acquisition means for acquiring the predetermined items from the first data acquired by the first data acquisition means, and acquiring values corresponding to the acquired predetermined items as the first text; and second text acquisition means for acquiring the predetermined items from the second data acquired by the second data acquisition means, and acquiring values corresponding to the acquired predetermined items as the second text.
A ninth invention is the character matching device of the eighth invention, wherein the first data and/or the second data is image data of a document including handwritten characters, and the first text and/or the second text is text obtained by performing character recognition processing on the image data.
A tenth aspect of the present invention is a program for causing a computer to function as any one of the character matching devices according to the first to ninth aspects of the present invention.

The present invention provides a character matching device and program that, when a character matching result is a mismatch, notifies the user based on the nature of the mismatch.

1 is a diagram showing the overall configuration of a character matching system according to an embodiment of the present invention and functional blocks of a character matching server. FIG. 2 is a diagram showing an example of a variant character DB according to the embodiment; 10 is a flowchart showing a character discrimination process of the character matching server according to the embodiment; 10 is a flowchart showing a matching target acquisition process of the character matching server according to the present embodiment. 10 is a flowchart showing a process of distinguishing variant characters, etc., performed by the character matching server according to the present embodiment. 10A and 10B are diagrams illustrating an example of a method for calculating similarity in the character variant discrimination process according to the embodiment. 10 is a flowchart showing an erroneous conversion determination process of the character matching server according to the present embodiment. 10A and 10B are diagrams showing specific examples of character discrimination in the character matching system according to the present embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, this is merely an example, and the technical scope of the present invention is not limited to this example.
(Embodiment)
<Overall Configuration of Character Matching System 100>
FIG. 1 is a diagram showing the overall configuration of a character matching system 100 according to this embodiment and the functional blocks of a character matching server 1.
FIG. 2 is a diagram showing an example of the variant character DB 4 according to this embodiment.

The character matching system 100 shown in FIG. 1 is a system that matches the contents of items to be matched (predetermined items) between, for example, an application document (first document) and an identification document (second document), and outputs the matching results. The character matching system 100 compares the text of the items to be matched in both documents, and if there is a mismatch, issues a notification according to the content of the mismatch. In this way, the character matching system 100 can reduce the visual inspection workload of the user who checks the documents (hereinafter, the person who checks the documents will also be referred to as the operator).
The character matching system 100 includes a character matching server 1 (character matching device), a character variant database (DB) 4 (character variant dictionary), and a terminal 5. The character matching server 1, the character variant DB 4, and the terminal 5 are communicatively connected via a communication network N.

<Character matching server 1>
The character matching server 1 is a server that can be used for various screening operations. For example, the character matching server 1 compares application document data (first data) obtained by imaging an application document with identity verification data (second data) obtained by imaging an identity verification document used in the screening of the application document.
The character matching server 1 includes a control unit 10 , a storage unit 20 , and a communication interface unit 29 .
The control unit 10 is a CPU (Central Processing Unit) that controls the entire character matching server 1. The control unit 10 executes various functions in cooperation with the above-mentioned hardware by appropriately reading and executing the OS (Operating System) and various application programs stored in the storage unit 20.

The control unit 10 includes an application data acquisition unit 11 (first data acquisition means), a first text acquisition unit 12 (first text acquisition means), a personal identification data acquisition unit 13 (second data acquisition means), a second text acquisition unit 14 (second text acquisition means), a comparison unit 15 (character count confirmation means, comparison means), a determination unit 16 (determination means, similarity calculation means), and a notification processing unit 17 (notification means).

The application data acquisition unit 11 acquires application data, which is an image of an application document. An application document may, for example, be an application form for opening a bank account or an application form for various insurance policies such as life insurance. An application document may contain handwritten characters in part. For example, if a requester submits a document via the web, the application data acquisition unit 11 accepts the application data from a client terminal (not shown). For example, if a requester submits a document by mail, the application data acquisition unit 11 accepts the application data after the paper-based application document has been converted into an image. The application data acquisition unit 11 then stores the acquired application data in the document data storage unit 22.

The first text acquisition unit 12 performs OCR (Optical Character Recognition) processing (character recognition processing) on the application data to obtain text data. The first text acquisition unit 12 then acquires items to be compared from the text data and acquires values corresponding to the acquired items to be compared as first text. Here, the items to be compared refer to items to be compared in the screening process, such as name (first and last name), address, etc. The first text acquisition unit 12 then stores the acquired items and their values (first text) in the comparison data storage unit 23.
In addition, when documents are submitted via the web, if the application data acquired by the application data acquisition unit 11 includes text data, the process of acquiring text data by the first text acquisition unit 12 is not necessary.

The personal identification data acquisition unit 13 acquires personal identification data, which is an image of a personal identification document. A personal identification document is used to verify the applicant, and includes, for example, a driver's license, insurance card, or passport. A personal identification document may include handwritten characters. For example, if a requester submits a document via the web, the personal identification data acquisition unit 13 accepts the personal identification data from a client terminal (not shown). For example, if a requester submits a document by mail, the personal identification data acquisition unit 13 accepts the personal identification data after an image of a copy of the mailed personal identification document has been created. The personal identification data acquisition unit 13 then associates the acquired personal identification data with the application data to be matched, which has also been accepted, and stores it in the document data storage unit 22.

The second text acquisition unit 14 performs OCR processing on the personal identification data to obtain text data. The second text acquisition unit 14 then acquires items to be compared from the text data and acquires values corresponding to the acquired items to be compared as second text. The second text acquisition unit 14 then associates the acquired values (second text) with the items and values (first text) obtained from the application data to be compared, and stores the values in the comparison data storage unit 23.
In addition, in the case of submitting documents via the Web, if the personal identification data acquired by the personal identification data acquisition unit 13 includes text data, the process of acquiring the text data by the second text acquisition unit 14 is not necessary.

The comparison unit 15 checks the number of characters in the first text and the second text for the item to be matched stored in the comparison data storage unit 23 .
Furthermore, when the comparison unit 15 confirms that the number of characters matches, it compares the first text with the second text character by character.

When the comparison unit 15 confirms that the number of characters does not match, the determination unit 16 identifies the mismatched characters and calculates the similarity between the combination of the identified characters and the paired characters.
Furthermore, if the comparison of the characters by the comparison unit 15 results in a mismatch, the determination unit 16 determines the content of the mismatch.
More specifically, the determination unit 16 determines whether or not both of the mismatched pair of characters are registered in the character variant DB 4. When the mismatched pair of characters is registered in the character variant DB 4, the determination unit 16 determines that the mismatch content is a character variant.
The determination unit 16 may also calculate the similarity between a pair of mismatched characters. If the calculated similarity is equal to or greater than a threshold, the determination unit 16 determines that the mismatch is a variant character or a mistranslation. Here, the determination unit 16 calculates the similarity by, for example, dividing an image area showing a pair of mismatched characters into multiple small areas and comparing the small areas at the same position.

When the determination unit 16 determines that there is a mismatch, the notification processing unit 17 issues a notification corresponding to the mismatch. More specifically, when the determination unit 16 determines that the mismatch is a variant character, the notification processing unit 17 issues a notification to that effect. Furthermore, when the determination unit 16 determines that the mismatch is a variant character or a misconversion, the notification processing unit 17 issues a notification to that effect. Furthermore, when the similarity between the combination of multiple characters that constitutes the mismatch calculated by the determination unit 16 and the paired character is equal to or greater than a threshold, the notification processing unit 17 issues a notification to the effect of a misconversion.

The storage unit 20 is a storage device such as a hard disk or semiconductor memory element for storing programs, data, etc. required for the operation of the character matching server 1 .
The storage unit 20 includes a program storage unit 21 , a document data storage unit 22 , and a comparison data storage unit 23 .
The program storage unit 21 is a storage area for storing programs, and stores a character matching program 21a (program).
The character matching program 21 a is a program for executing each function of the control unit 10 .

The document data storage unit 22 is a storage area that stores application data and personal identification data to be checked against the application data in association with each other.
The comparison data storage unit 23 stores the items to be matched and their values (first text) obtained from the application data, in association with the values (second text) corresponding to the items obtained from the personal identification data.
It should be noted that document data storage unit 22 and comparison data storage unit 23 may be used temporarily when performing character matching processing.
The communication interface unit 29 is an interface unit for communicating with the variant character DB 4 and the terminal 5 via the communication network N.

It should be noted that a computer refers to an information processing device that includes a control unit, a storage device, etc., and the character matching server 1 is an information processing device that includes a control unit 10, a storage unit 20, etc., and is included in the concept of a computer.
Furthermore, there is no limit to the number of pieces of hardware that make up the character matching server 1. It may be configured with one or more pieces of hardware as needed. Furthermore, the hardware of the character matching server 1 may be configured to include various servers such as a web server, a DB (database) server, and an application server as needed, and may be configured with a single server or with separate servers for each. Furthermore, the character matching server 1 may be, for example, a cloud.

<Variant Character DB4>
The variant character DB 4 is a database that stores variant characters.
As shown in Figure 2, the variant character DB 4 stores characters and variant characters in association with each other. A character is a character in a standard font. A variant character is a Chinese character that differs from the standard font but has the same meaning and pronunciation and is widely used. The variant character DB 4 may store multiple variant characters for one character.
The character variant DB 4 is created and registered in advance, and may be manually updated as needed.
Although not shown, the character variant DB 4 includes a control unit, a storage unit, and a communication interface unit. Note that the character variant DB 4 may not include a control unit, and the control unit 10 of the character matching server 1 may control the character variant DB 4.

<Terminal 5>
The terminal 5 is a terminal used by an operator involved in the screening work. The terminal 5 can be configured, for example, as a personal computer (PC), a tablet terminal, etc. The terminal 5 may be a terminal dedicated to the work, or may be a terminal shared with other work. Although not shown, the terminal 5 includes a control unit, a memory unit, an input unit, a display unit, a communication interface unit, etc.

<Processing Description>
Next, the processing performed by the character matching server 1 will be described.
FIG. 3 is a flowchart showing the character discrimination process of the character matching server 1 according to this embodiment.
FIG. 4 is a flowchart showing the matching target acquisition process of the character matching server 1 according to this embodiment.

In step S (hereinafter, "step S" will be simply referred to as "S") 11 in FIG. 3, the control unit 10 of the character matching server 1 performs a matching target acquisition process.
Here, the matching target acquisition process will be described with reference to FIG.
4, the control unit 10 (application data acquisition unit 11) acquires application data obtained by converting application documents into images. The control unit 10 then stores the acquired application data in the document data storage unit 22.
In S22, the control unit 10 (first text acquisition unit 12) performs OCR processing on the application data to acquire items to be matched and values (first text) corresponding to the items from the character string (text data). Here, the items to be matched may be multiple items, and in that case, the control unit 10 acquires each value corresponding to each item to be matched.

In S23, the control unit 10 (personal identification data acquisition unit 13) acquires personal identification data, which is an image of the personal identification document to be compared with the application document. The control unit 10 then associates the acquired personal identification data with the application data to be compared, and stores the data in the document data storage unit 22.
In S24, the control unit 10 (second text acquisition unit 14) performs OCR processing on the personal identification data, and acquires items to be matched and values (second text) corresponding to the items from the character string (text data).
In S25, the control unit 10 associates the item to be compared with the first text and second text corresponding to the item and stores them in the comparison data storage unit 23. Thereafter, the control unit 10 proceeds to S12 in FIG.

In S12 of FIG. 3, the control unit 10 (comparison unit 15) acquires each text corresponding to the item to be matched from the comparison data storage unit 23, and compares the texts with each other.
In S13, the control unit 10 (comparison unit 15) determines whether the number of characters matches. If the number of characters matches (S13: YES), the control unit 10 proceeds to S14. On the other hand, if the number of characters does not match (S13: NO), the control unit 10 proceeds to S16. Note that a case where the number of characters does not match includes a case where there is a possibility of a conversion error when converting image data to text.

In S14, the control unit 10 (comparison unit 15) determines whether the characters match exactly. Here, a complete match means that the text of the item to be compared written on the application document matches the text of the same item written on the personal identification document without a single character difference. If the characters match exactly (S14: YES), the control unit 10 proceeds to S17. On the other hand, if the characters do not match exactly (S14: NO), the control unit 10 proceeds to S15.

In S15, the control unit 10 performs a process of determining variant characters, etc., which will be described later with reference to Fig. 5. After that, the control unit 10 moves the process to S17.
In S16, the control unit 10 performs an erroneous conversion determination process, which will be described later with reference to Fig. 7. After that, the control unit 10 moves the process to S17.

In S17, the control unit 10 determines whether all matching targets have been processed. If all data stored in the comparison data storage unit 23 has been processed, the control unit 10 can determine that all matching targets have been processed. If all matching targets have been processed (S17: YES), the control unit 10 proceeds to S18. On the other hand, if all matching targets have not been processed (S17: NO), the control unit 10 proceeds to S12 and processes the remaining matching targets stored in the comparison data storage unit 23.

In S18, the control unit 10 (notification processing unit 17) performs a result output process. Specifically, if a message has been generated in the processing up to this point, the control unit 10 outputs that message to the terminal 5. On the other hand, if no message has been generated in the processing up to this point, the control unit 10 outputs a message to the terminal 5 indicating that there is no problem with the matching result. Note that if the text to be matched is a perfect match, no message has been generated by the time the processing of S18 is reached. The control unit 10 then terminates this processing.

Next, the process of determining variant characters will be described.
FIG. 5 is a flowchart showing the process of distinguishing variant characters, etc., performed by the character matching server 1 according to this embodiment.
FIG. 6 is a diagram showing an example of a method for calculating the similarity in the character variant discrimination process according to this embodiment.

In S31 of FIG. 5, the control unit 10 (determining unit 16) searches the character variant DB 4 for the mismatched pair of characters.
In S32, the control unit 10 (determination unit 16) determines whether the mismatched pair of characters is registered in the character variant DB 4. If both of the mismatched pair of characters are registered as characters or character variants in the same record in the character variant DB 4, the control unit 10 determines that the mismatched pair of characters is registered in the character variant DB 4.
For example, in the case of "taka" shown in Fig. 2, if the mismatched pair of characters is both the character and a variant character, the control unit 10 determines that the mismatched pair of characters is registered in the variant character DB 4. Also, in the case of "hen" shown in Fig. 2, the mismatched pair of characters may be both the character and a variant character, or both may be variant character characters.
If the mismatched pair of characters is registered in the character variant DB 4 (S32: YES), the control unit 10 proceeds to S33. On the other hand, if the mismatched pair of characters is not registered in the character variant DB 4 (S32: NO), the control unit 10 proceeds to S34.

In S33, the control unit 10 (determining unit 16) generates a notification message indicating that the character is a variant character, and the process proceeds to S17 in FIG.
In S34, the control unit 10 (determination unit 16) calculates the similarity between the mismatched characters. As an example of a method for calculating the similarity between characters, the control unit 10 divides the image area representing the characters into multiple small areas. Then, the control unit 10 calculates the similarity between the characters by comparing the small areas at the same position.

Here, a method for calculating the similarity between characters will be described with reference to FIG.
6A shows a calculation of the similarity between characters shown in an image area 61 and characters shown in an image area 62. The control unit 10 divides the image area 61 into n×n small areas. The control unit 10 then vectorizes the image area 61 and represents each small area with a value of 0 or 1. Similarly, the control unit 10 vectorizes the image area 62.
The control unit 10 then uses the vectorized values to calculate the sum of square differences (SSD). The SSD is used to calculate the degree of difference between the image area 61 and the image area 62.
When the vectorized value of the image area 61 is represented by x(a, b) and the vectorized value of the image area 62 is represented by y(a, b), the SSD can be calculated by the following formula.

According to this (Equation 1), the number of small areas with different values among small areas at the same position can be calculated.
Next, the control unit 10 calculates Sim, which indicates the similarity between the image area 61 and the image area 62, using the SSD according to the following formula.

Here, an explanation will be given using an example of FIG. 6B, which is easier to understand.
When calculating the similarity between image area 63 and image area 64, the control unit 10 calculates the SSD by applying the vectorized value 63a of image area 63 and the vectorized value 64a of image area 64 to (Equation 1).

Next, the control unit 10 applies the calculated SSD to (Equation 2) to calculate Sim, and uses the calculated Sim value as the similarity between image area 63 and image area 64.

5, the control unit 10 (determination unit 16) determines whether the calculated similarity is equal to or greater than a threshold. If the similarity is equal to or greater than the threshold (S35: YES), the control unit 10 proceeds to S36. On the other hand, if the similarity is not equal to or greater than the threshold (S35: NO), the control unit 10 proceeds to S37.
For example, if the threshold value is 0.8, in the case of FIG. 6B, the control unit 10 determines that the similarity is equal to or greater than the threshold value.

In S36, the control unit 10 (determination unit 16) generates a notification message indicating that the character is a variant or a mistranslation, and proceeds to S17 in Fig. 3. For example, in the case of Fig. 6(A), if the calculated similarity is equal to or greater than the threshold, the process proceeds to S36, but in this example, the character is a variant. In actual processing, the character illustrated in Fig. 6(A) is registered in the variant DB 4, so the result of S32 becomes YES. Therefore, when this process is reached and the character is a variant, it corresponds to a case where the character is a variant that is not registered in the variant DB 4.
Examples of incorrect conversions include "tsuki" and "me", "iri" and "hito", and "ga" and "ke".
In S37, the control unit 10 (determining unit 16) generates a warning message indicating a mismatch, and the process proceeds to S17 in FIG.

Next, the erroneous conversion determination process will be described.
FIG. 7 is a flowchart showing the erroneous conversion determination process of the character matching server 1 according to this embodiment.
In S41 of FIG. 7, the control unit 10 (determination unit 16) identifies the mismatched characters.
In S42, the control unit 10 (determination unit 16) calculates the similarity between the identified combination of multiple characters and the paired character. The calculation of the similarity here can be performed using the same process as the calculation of the similarity between characters described in FIG. 6. However, the character region of the combination of multiple characters is twice as large horizontally as the paired character region. Therefore, for example, when dividing the character region of the combination of multiple characters into small regions, the control unit 10 doubles the horizontal size of the character to make the width of one region. This makes it possible to generate the same number of small regions as the paired characters, so that the processes from vectorization onwards can be performed in the same way.

In S43, the control unit 10 (determination unit 16) determines whether the similarity is equal to or greater than a threshold. If the similarity is equal to or greater than the threshold (S43: YES), the control unit 10 proceeds to S44. On the other hand, if the similarity is not equal to or greater than the threshold (S43: NO), the control unit 10 proceeds to S45.
In S44, the control unit 10 (determination unit 16) generates a notification message indicating that the conversion is incorrect, and the process proceeds to S17 in Fig. 3. Examples of cases that proceed to the process of S44 include the erroneous conversion of one character "好" (good) into two characters "女" (woman) and "子" (child), the erroneous conversion of one character "明" (light) into two characters "日" (sun) and "月" (month), and the erroneous conversion of one character "飯" (meal) into two characters "食" (food) and "反" (anti-food).
In S45, the control unit 10 (determination unit 16) generates a warning message indicating a mismatch, and the process proceeds to S17 in FIG.

Next, a series of processing steps will be described with a specific example.
FIG. 8 is a diagram showing a specific example of character discrimination in the character matching system 100 according to this embodiment.
First, application data obtained by imaging an application document 7 and personal identification data obtained by imaging an identification document 8 are input to the character matching server 1 .
In the character matching server 1, the control unit 10 extracts items and values to be matched from the application data and stores them in the comparison data storage unit 23. The control unit 10 also extracts the items and values from the personal identification data and stores them in the comparison data storage unit 23.
Next, in the character matching server 1, the control unit 10 extracts the value of the name as the matching target item from the comparison data storage unit 23 and performs a comparison.
If the comparison results in a mismatch and the character is determined to be a variant, the control unit 10 of the character matching server 1 generates a message indicating that the character is a variant, and outputs the message to the terminal 5 as a pop-up message 51 .

As described above, according to this embodiment, the character matching server 1 has the following effects.
(1) A first text, which is the value for the item to be matched obtained from the application documents, is compared with a second text, which is the value for the item obtained from the personal identification documents. If the comparison results in a mismatch, the details of the mismatch are determined, and a notification corresponding to the determined mismatch is sent.
Therefore, if the result of comparing the two texts is a mismatch, a notification corresponding to the mismatch can be given, making it possible to output a notice that is easy to understand to the operator who checks the matched content.

(2) The character variant DB 4 is referenced, and if the mismatched pair of characters is registered in the character variant DB 4, it is determined that the mismatched content is a character variant, and a notification is given to that effect.
Therefore, whether or not the character is a variant is determined based on the variant character DB 4, and if so, the operator is notified that the character is a variant character, thereby alerting the operator.

(3) The similarity between the mismatched characters is calculated, and if the calculated similarity is equal to or greater than a threshold, the mismatched content is determined to be a variant character or a mistranslation, and a notification to that effect is sent.
Therefore, even if there is a variant character that is not registered in the variant character DB 4, the operator can be alerted by the notification. Also, if the characters look similar, there is a possibility of mistranslation, so the operator can be alerted by the notification.
(4) The image area showing the mismatched pair of characters is divided into a plurality of small areas, and the small areas at the same position are compared to calculate the similarity, so that the similarity can be calculated by simple processing.

(5) The number of characters in the first text and the second text is checked, and if the number of characters is confirmed to match, the characters are compared one by one in order. Therefore, by first excluding cases where the number of characters does not match, processing can be performed efficiently.
(6) If the number of characters does not match, the mismatched characters are identified, the similarity between the identified combination of characters and the paired characters is calculated, and if the calculated similarity is equal to or greater than a threshold, it is determined to be a misconversion. Thus, it is possible to alert the operator to a misconversion such as when one character is converted into two characters.

(7) The predetermined items include at least one of a name and an address, and therefore, it can be used for items that tend to include variant characters as values for the items.
(8) Items to be matched and their values are obtained from application data that is an image of an application document, and values corresponding to the same items to be matched are obtained from identity verification data that is an image of an identity verification document, and these values are compared. Thus, character matching for items to be matched can be performed automatically based on the image data.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Furthermore, the effects described in the embodiments are merely a list of the most favorable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments. The above-described embodiments and the modified forms described below can also be used in appropriate combinations, but detailed explanations will be omitted.

(Modified form)
(1) In the present embodiment, the character variant DB is described as being communicably connected to the character matching server, but this is not limiting. The character variant DB may be included in the character matching server.
(2) In the present embodiment, the method using the variant character DB and the method calculating the similarity between mismatched pairs of characters are used together. However, the present invention is not limited to this. Only one of the methods may be used. However, using both methods can achieve more reliable matching.

(3) In this embodiment, the character matching server determines whether a character is a variant or a mistranslation, but this is not limited to this. The processing performed by the character matching server may be incorporated into a device that processes applications, and the function may be used as part of various processes for applications.

(4) In this embodiment, the variant character DB has been described as storing variant character strings themselves, but this is not limited to this. For example, a variant character string storage unit (not shown) that stores character strings including variant character strings commonly used in personal names and place names may be used. In this case, the character matching server determines the character string to be matched based on whether the characters before and after the variant character are stored in the variant character string storage unit. The character matching server then determines whether the character string is a variant character by determining whether the character string, including its paired character string, is stored in the variant character string storage unit. This type of processing further improves the reliability of matching.

1. Character matching server 4. Variant character DB
5 Terminal 7 Application document 8 Personal identification document 10 Control unit 11 Application data acquisition unit 12 First text acquisition unit 13 Personal identification data acquisition unit 14 Second text acquisition unit 15 Comparison unit 16 Determination unit 17 Notification processing unit 20 Storage unit 21a Character matching program 23 Comparison data storage unit 51 Pop-up message 100 Character matching system N Communication network

Claims (10)

a comparison means for comparing a first text for a predetermined item obtained from a first document with a second text for the predetermined item obtained from a second document;
a determination means for determining the content of the mismatch, including variant characters and incorrect conversions , when the comparison by the comparison means results in a mismatch;
a notification means for notifying the content of the mismatch determined by the determination means;
A character matching device comprising: 2. The character matching device according to claim 1,
the determining means determines that the mismatched content is a variant character when the mismatched pair of characters is registered in a variant character dictionary that stores variant character;
The notifying means notifies that the character is a variant character. 3. The character matching device according to claim 1,
the determining means calculates a similarity between a pair of mismatched characters, and when the calculated similarity is equal to or greater than a threshold, determines that the mismatched content is a variant character or an incorrect conversion;
The notifying means notifies the user that the character is a variant character or a mistranslation. 4. The character matching device according to claim 3,
The determining means divides an image area showing a pair of mismatched characters into a plurality of small areas, and compares the small areas at the same position to calculate the similarity. 5. The character matching device according to claim 1,
character number confirmation means for confirming the number of characters of the first text and the second text;
The character matching device, wherein the comparison means compares the characters one by one when the character number confirmation means confirms that the numbers of characters match. 6. The character matching device according to claim 5,
a similarity calculation means for, when a mismatch in the number of characters is confirmed by the character number confirmation means, specifying a combination of a plurality of characters that is a mismatch and a pair of characters by comparing the first text with the second text by the comparison means , and calculating a similarity by dividing an image area showing the specified combination of a plurality of characters and an image area showing the specified pair of characters into a plurality of small areas of the same number and comparing the divided areas;
The notifying means notifies that the conversion is erroneous when the similarity calculated by the similarity calculation means is equal to or greater than a threshold value. 7. The character matching device according to claim 1,
The character matching device, wherein the predetermined items include at least one of a name and an address. 8. The character matching device according to claim 1,
a first data acquisition means for acquiring first data obtained by imaging the first document;
second data acquisition means for acquiring second data obtained by imaging the second document;
a first text acquisition means for acquiring the predetermined item from the first data acquired by the first data acquisition means, and acquiring a value corresponding to the acquired predetermined item as the first text;
second text acquisition means for acquiring the predetermined item from the second data acquired by the second data acquisition means, and acquiring a value corresponding to the acquired predetermined item as the second text;
A character matching device comprising: 9. The character matching device according to claim 8,
the first data and/or the second data are image data of a document including handwritten characters;
The character matching device, wherein the first text and/or the second text is text obtained by performing character recognition processing on the image data. A program for causing a computer to function as the character matching device described in any one of claims 1 to 9.