Accounting information processing system, accounting information processing method, and accounting information processing program

Abstract

To reduce the effort required for tax officials to organize images of bankbooks. [Solution] This system comprises a bankbook image storage unit 1 that stores multiple bankbook images, a bankbook image identity determination unit 2 that determines whether multiple bankbook images are images of the same bankbook based on the multiple bankbook images stored in the bankbook image storage unit 1 and outputs the determination result, and a bankbook image identity storage unit 3 that stores the bankbook images and the determination result output by the bankbook image identity determination unit in association with each other.

Description

[Technical Field]

[0001] The present invention relates to an accounting information processing system, an accounting information processing method, and an accounting information processing program. [Background technology]

[0002] In some cases, clients may send images of bankbooks from financial institutions such as banks to tax accountant offices or accounting firms via email or communication software. At tax accountant offices, tax accountants look at the images of bankbooks sent by clients and enter the necessary tax information into accounting information processing systems. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] https: / / biz.moneyforward.com / tax_return / basic / 50737 / Summary of the Invention [Problem to be solved by the invention]

[0004] When a large number of bankbook images are sent from many customers, there is a problem in that it takes a lot of effort to sort out which bankbook images belong to which bankbook.

[0005] The object of the present invention is to provide an accounting information processing system, an accounting information processing method, and an accounting information processing program that can automatically determine which bankbook a bankbook image belongs to, thereby reducing the effort required by tax officials to organize bankbook images. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, an accounting information processing system according to an embodiment of the present invention comprises a bankbook image storage unit that stores multiple bankbook images, a bankbook identity determination unit that determines whether the multiple bankbook images stored in the bankbook image storage unit are bankbook images relating to the same bankbook and outputs the determination result, and a bankbook identity storage unit that stores the bankbook images and the determination result output by the bankbook identity determination unit in association with each other.

[0007] Furthermore, it is preferable that the system be provided with a transaction information output unit that outputs information regarding the date and balance of each transaction recorded in the multiple bankbook images, and that the bankbook identity determination unit determine whether the multiple bankbook images are bankbook images relating to the same bankbook based on the transaction date and balance output by the transaction information output unit, and output the determination result.

[0008] Furthermore, it is preferable that the device is provided with an order output unit that outputs the order of multiple deposit passbook images that the passbook identity determination unit has determined to be deposit passbook images relating to the same deposit passbook, based on the date of each transaction output by the transaction information output unit.

[0009] Furthermore, it is preferable that the bankbook identity determination unit determines whether the multiple bankbook images are bankbook images relating to the same bankbook based on the colors, item names and / or character format recorded in the multiple bankbook images, and outputs the determination result.

[0010] Furthermore, the system has a machine-learned learning model that, when a plurality of bankbook images are input, outputs whether the plurality of bankbook images are bankbook images relating to the same bankbook; It is preferable that the bankbook identity determination unit inputs multiple bankbook images stored in the bankbook image storage unit into the learning model, determines whether the multiple bankbook images are bankbook images relating to the same bankbook, and outputs the determination result.

[0011] Furthermore, if the bankbook image contains handwritten entries, it is preferable to have a memo recognition unit that applies character recognition technology to the handwritten entries to recognize the characters and store them in association with transactions near the handwritten entries.

[0012] In another embodiment of the present invention, an accounting information processing program causes a computer to execute the following steps: a bankbook image storage step for storing multiple bankbook images; a bankbook identity determination step for determining whether the multiple bankbook images stored by the bankbook image storage step are bankbook images relating to the same bankbook and outputting the determination result; and a bankbook identity storage step for storing the bankbook images and the determination result output by the bankbook identity determination step in association with each other.

[0013] In another embodiment of the present invention, an accounting information processing method is executed by a computer and includes a bankbook image storage step for storing multiple bankbook images, a bankbook identity determination step for determining whether the multiple bankbook images stored by the bankbook image storage step are bankbook images relating to the same bankbook and outputting the determination result, and a bankbook identity storage step for storing the bankbook images and the determination result output by the bankbook identity determination unit in association with each other. [Effects of the Invention]

[0014] According to the present invention, it is possible to automatically determine which bankbook an image belongs to, thereby reducing the effort required of tax officials to sort through bankbook images. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a diagram illustrating the configuration of a transaction information processing system according to an embodiment of the present invention. [Figure 2] FIG. 1 is a diagram showing the configuration of a computer used to implement the present invention. [Figure 3] FIG. 1 is a diagram illustrating information processing according to a first embodiment. [Figure 4A] FIG. 10 is a diagram showing a bankbook image A. [Figure 4B] FIG. 10 is a diagram showing a bankbook image B. [Figure 4C] FIG. 10 is a diagram showing a bankbook image C. [Figure 5] FIG. 10 is a diagram showing a bankbook image D. [Figure 6] FIG. 10 is a diagram showing a portion of data obtained by reading a bankbook image D using OCR. DETAILED DESCRIPTION OF THE INVENTION

[0016] Figure 1 is a diagram showing the configuration of an accounting information processing system according to an embodiment of the present invention. This system comprises a bankbook image storage unit 1 that stores multiple bankbook images, a bankbook identity determination unit 2 that determines whether the multiple bankbook images stored in the bankbook image storage unit 1 are images of the same bankbook and outputs the determination result, and a bankbook identity storage unit 3 that stores the bankbook images and the determination result output by the bankbook identity determination unit in association with each other.

[0017] In this specification, the term "bankbook" refers to a bankbook prepared by a financial institution or the like for the purpose of continuously verifying the receipt and payment of deposits and savings between a depositor or saver, and includes savings bankbooks. Figures 4A to 4C are examples of bankbook images. The items, font, font size, layout, background color, etc. of a bankbook vary depending on the financial institution that issues the bankbook.

[0018] A "bankbook image" is an image of a bankbook obtained by photographing the bankbook with a digital camera or scanning it with a scanner.

[0019] In this specification, the bankbook image identity determination unit may not only determine that multiple bankbook images are images relating to the same bankbook, but may also determine that there is a high possibility that they are images relating to the same bankbook. The unit may display on a display that a certain bankbook image is highly likely to be an image relating to the same bankbook as other bankbook images, and may determine that the images are images relating to the same bankbook upon input of a user's decision.

[0020] Information processing by this system is performed by a computer. FIG. 2 shows the configuration of a computer 20 used in this system. The memory unit 21 stores bankbook images and functions as a bankbook memory unit. The calculation unit 22 applies character recognition technology (OCR: Optical Character Recognition / Reader) to the bankbook images to read transaction information recorded in the bankbook images, such as the date of each transaction, payment amount, deposit amount, and balance. The calculation unit 22 also functions as a bankbook image identity determination unit that determines whether multiple bankbook images belong to the same bankbook based on the read transaction information. The output unit 24 is a device that outputs information stored in the memory unit 21, such as a display or speaker. The input unit 23 is a device that allows a user to input information into the computer, such as a keyboard, mouse, camera, or microphone. The communication unit 25 is a device that communicates information with devices outside the computer. [Example]

[0021] A first embodiment of the present invention will be described below with reference to Figure 3. First, a bankbook image sent by a customer is input into the accounting information processing system and stored in the bankbook image storage unit. Bankbook images A to C in Figures 4A to 4C are examples of bankbook images.

[0022] Next, the bankbook image is scanned using AI-OCR, and transaction information such as the transaction date, summary, payment amount, deposit amount, and balance for each transaction recorded in the bankbook image is recognized and output (S11). Specifically, the bankbook image is input to a learning model that has been trained using training data to output information corresponding to fields such as "transaction date," "summary," "payment amount," "deposit amount," and "balance" when the bankbook image is input, and the learning model outputs information corresponding to these fields. For example, when AI-OCR is applied to bankbook image A (Figure 4A), information is output that a withdrawal of 10,4000 yen was made from an ATM on January 4, 2024, leaving a balance of 9,000,000 yen. Information is output that bankbook image A records transactions from January 4, 2024 to February 22, 2024. Furthermore, when AI-OCR is applied to bankbook image B (Figure 4B), information is output that indicates that transactions from February 20, 2024 to March 31, 2024 are recorded in bankbook image B.

[0023] The transaction information in bankbook image A and bankbook image B is then compared, and it is automatically determined whether bankbook image A and bankbook image B are images relating to the same bankbook. Specifically, since the transaction period recorded in bankbook image A and the transaction period recorded in bankbook image B overlap between February 20, 2024 and February 22, 2024, and the balance of bankbook image A on February 20, 2024 is 9,616,000 yen and the balance of bankbook image B on February 20, 2024 is 1,026,259 yen, and the balances for the same day do not match, it is determined that bankbook image A and bankbook image B are bankbook images relating to different bankbooks.

[0024] When AI-OCR is applied to bankbook image C (Figure 4C), the output information indicates that bankbook image C records transactions from February 25, 2024, to March 22, 2024. There is no overlap between the transaction periods of bankbook image A and bankbook image C, and the last transaction date recorded in bankbook image A is February 22, 2024, which is close to the first transaction date recorded in bankbook image C, February 25, 2024. Furthermore, the balance of bankbook image A on February 22, 2024, is 9,571,000 yen, and in the transaction of bankbook image B on February 25, 2024, a payment of 1,000,000 yen was made, leaving a balance of 8,571,000 yen. Looking at the payment amount and balance, bankbook image A and bankbook image B are presumed to be images of the same bankbook and to be images of consecutive pages of the same bankbook. Therefore, it is determined that the bankbook A and the bankbook B are images relating to the same bankbook (S12).

[0025] Based on these determination results, multiple bankbook images relating to the same bankbook are grouped and stored in a storage unit (S13). For example, bankbook image A and bankbook image C are stored with a label such as bankbook number, "bankbook 001." Furthermore, "bankbook 001" may be labeled as being in the name of "A Corporation" and related to "Bank A." The order of the bankbook images may be determined based on the transaction date, and each bankbook image may be assigned an identification number indicating the page number, such as "bankbook 001-001" for bankbook image A and "bankbook 001-002" for bankbook image C, and the images may be displayed in chronological order for each bankbook. [Example]

[0026] The processing procedure of the second embodiment of the present invention will be described below, and the description common to the first embodiment will be omitted. The font of bankbook image A (Figure 4A) is "MS Mincho," and the item names are only "Date," "Application," "Payment Amount," "Deposit Amount," and "Balance," with the background of "Summary" and "Balance" being pink. Note that in this specification, the font and size of characters are referred to as the "format" of the characters.

[0027] On the other hand, the font of bankbook image B (Figure 4B) is "MS Gothic," the item names contain symbols that are not present in bankbook image A, and the background of the "balance" is light green. Therefore, bankbook image A and bankbook image B differ in font, item, background color, etc., and are therefore known to be issued by different financial institutions. Therefore, bankbook image A and bankbook image B are determined to be images of different bankbooks. Note that if information about each financial institution's font, item, background color, etc. is recorded in advance, the name of the financial institution that issued the bankbook can be identified from the bankbook image. The name of the financial institution that issued the bankbook may also be identified based on information about the layout, such as the dimensions (width, height, etc.) of each item, and the position, etc.

[0028] Bankbook image A and bankbook image B have in common the font "MS Mincho," the field names "Date," "Summary," "Payment amount," "Deposit amount," and "Balance," and the background of "Summary" and "Balance" is pink. Therefore, bankbook image A and bankbook image B are presumed to be bankbook images related to bankbooks issued by the same financial institution. Bankbook images related to bankbooks issued by the same financial institution do not necessarily mean that they are bankbook images related to the same bankbook, but this information is useful for determining the identity of the bankbooks. [Example]

[0029] The processing procedure of the third embodiment of the present invention will be described below, and explanations common to the first and second embodiments will be omitted.

[0030] In this embodiment, a learning model is constructed that is trained to input multiple bankbook images, for example, two bankbook images, and output whether the two input bankbook images are bankbook images relating to the same bankbook.

[0031] Then, bankbook image A and bankbook image C are input into the constructed learning model. As bankbook image A and bankbook image C share the same characteristics, that is, telephone bills are deducted on the 15th of each month, electricity bills are deducted on the 20th of each month, water bills are deducted on the 22nd of each month, and salaries are paid on the 25th of each month, the learning model outputs and determines that bankbook image A and bankbook image C are bankbook images related to the same bankbook. [Example]

[0032] The processing procedure of the fourth embodiment of the present invention will be described below, and explanations common to the embodiments described so far will be omitted.

[0033] Figure 5 is bankbook image D. In the column for February 28, 2024, "Deposit by K" is handwritten. Figure 6 shows part of the data scanned by AI-OCR from bankbook image D. In a transaction on February 28, 2024, 2,500,000 yen was deposited, and the "Handwritten notes" column reads: K Since "Deposit" is entered, the deposit of 2,500,000 yen is K Is it a payment from a business partner such as K It is assumed that the payment is for a product, and the user will have a clue as to the reason for the payment. can be obtained.

[0034] The embodiments and examples of the present invention described above are merely exemplifications of the present invention, and the technical scope of the present invention is not limited to the above examples, etc. The configuration of the present invention can be implemented in various modified forms within the scope of the claims. [Explanation of symbols]

[0035] 1. Bankbook image storage unit 2. Bankbook Identity Verification Unit 3. Bankbook Identity Memory Unit 20 Computer 21 Memory section 22 Arithmetic section 23 Input section 24 Output section 25 Communications Department

Claims

1. a bankbook image storage unit that stores a plurality of bankbook images; a transaction information output unit that outputs information about the date and balance of each transaction recorded in the plurality of bankbook images; a bankbook identity determination unit that determines whether or not a plurality of bankbook images are bankbook images relating to the same bankbook based on the information on the transaction date and balance output by the transaction information output unit, and outputs the determination result; An accounting information processing system comprising: a bankbook identity memory unit that, when the bankbook identity determination unit determines that multiple bankbook images are bankbook images relating to the same bankbook, stores information indicating that the multiple bankbook images are bankbook images relating to the same bankbook.

2. An accounting information processing system as described in claim 1, further comprising an order output unit that outputs the order of multiple deposit passbook images that the deposit passbook identity determination unit has determined to be deposit passbook images relating to the same deposit passbook based on the date of each transaction output by the transaction information output unit.

3. The accounting information processing system of claim 1, wherein the bankbook identity determination unit determines whether the multiple bankbook images are bankbook images relating to the same bankbook based on the color, item name and / or character format recorded in the multiple bankbook images, and outputs the determination result.

4. a learning model that has undergone machine learning to output whether or not a plurality of bankbook images are bankbook images relating to the same bankbook when the bankbook images are input; The accounting information processing system of claim 1, wherein the bankbook identity determination unit inputs multiple bankbook images stored in the bankbook image storage unit into the learning model, determines whether the multiple bankbook images are bankbook images relating to the same bankbook, and outputs the determination result.

5. The accounting information processing system of claim 1 has a memo recognition unit that, if there is handwritten writing in the bankbook image, applies character recognition technology to the handwritten writing to recognize the characters and stores them in association with transactions near the handwritten writing.

6. On the computer, a bankbook image storage step for storing a plurality of bankbook images; a transaction information output step of outputting information about the date and balance of each transaction recorded in the plurality of bankbook images; a bankbook identity determination step of determining whether or not the plurality of bankbook images are bankbook images relating to the same bankbook based on the transaction dates and balances output in the transaction information output step, and outputting a determination result; a bankbook identity storage step of storing information indicating that the plurality of bankbook images are bankbook images relating to the same bankbook when the bankbook identity determination step determines that the plurality of bankbook images are bankbook images relating to the same bankbook; An accounting information processing program that executes the above.

7. The computer executes a bankbook image storage step for storing a plurality of bankbook images; a transaction information output step of outputting information about the date and balance of each transaction recorded in the plurality of bankbook images; a bankbook identity determination step of determining whether or not the plurality of bankbook images are bankbook images relating to the same bankbook based on the transaction dates and balances output by the transaction information output step, and outputting a determination result; a bankbook identity storage step of storing information indicating that the plurality of bankbook images are bankbook images relating to the same bankbook when the bankbook identity determination step determines that the plurality of bankbook images are bankbook images relating to the same bankbook; An accounting information processing method comprising:

Images