JP2000194775A - Method for generating client data and computer readable recording medium recording program for computer to execute the method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To read various financial tables prepared by a client through an OCR and to automatically convert these tables into client data processible for another analytic program. SOLUTION: Document data are read by an OCR 10 and converted into image data, a character description area and a numeral description area in the image data are discriminated based on the layout data of an LC file 62, the image data of a character string and a numeral string respectively existent in the discriminated character description area and numeral description area are converted into the character data of character string and numeral string, the converted character data of character string and numeral string are converted into 'item' and 'amount' based on the word data of a word file 64, the converted 'item' is summarized into 'course' based on the summary data of a dictionary file, the 'amount' corresponding to the non-summarized 'item' is tabulated for each summarized 'course' and the document data are stored in a client data file 66 as client data composed of the set of summarized 'item' and tabulated 'amount'.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of preparing customer data for reading data of a financial statement prepared by a customer with an OCR and converting the data into customer data that can be processed by another analysis program, and a method thereof. The present invention relates to a computer-readable recording medium on which a program to be executed by a computer is recorded.

[0002]

2. Description of the Related Art Conventionally, many banks analyze the contents of financial statements prepared by customers and make a decision as to whether or not to make a loan based on the result of the analysis. The analysis of the contents of the financial statements can now be performed at high speed by a computer.

[0003]

However, although the analysis of the contents of financial statements has been speeded up, it takes a great deal of time to input financial statements prepared by a customer into a computer.

[0004] One of the causes is that the description format of financial statements prepared by customers is not uniform.

[0005] In other words, even if financial statements are described, not only the description format is different for each customer, but also the description items are also different. Therefore, before inputting the financial statements to the computer, the operator determines which item of the financial statements should be entered as which computer account, one item at a time.
The account and amount must be entered.

For example, “Cash”… 2,652,000 yen “Current account”… 11,365,000 yen “Other deposits”… 568,000 yen “Prepayment”… 152,400 yen If you enter this into a computer, the operator will make the items "Cash", "Checking account" and "Other deposits" into "Cash / deposit" accounts, And enter
In the item of "prepayment", the amount is entered after the account of "other current assets".

[0007] As described above, the operator must determine which account item should be entered and input the stated amount, so that it takes time to input.

[0008] The next conceivable cause is that it is necessary to confirm whether there is a mistake in the input data.

[0009] Even if the operator pays careful attention as described above, it cannot be said that there is no input error at all. Since financial statements are very important data about customers, there must be no typing errors. For this reason, it is absolutely necessary to confirm the data once entered.

As described above, it takes a great deal of time to input items described in the financial statements into a computer, and some banks secure human resources only for this input. Yes. If this input operation can be automated, the burden of the input operation on the bank side can be drastically reduced.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and has been made to reduce the burden on an operator and to enable data input without input errors.
A financial statement prepared by a customer is read by an OCR, and a method of creating customer data for automatically converting the financial statement into customer data that can be processed by another analysis program and a program for causing a computer to execute the method are recorded. It is intended to provide a computer-readable recording medium.

[0012]

According to the first aspect of the present invention, there is provided a first step in which document data is read by an OCR and converted into image data, and the converted image data is stored in an image data file; A second step of determining, based on the layout data of the LC file, a character description area and a number description area in the image data stored in the data file; and a character description area and a number description determined based on the layout data. A third step of converting image data of a character string and a number string respectively existing in the area into character data of a character string and a number string based on the conversion data of the character data conversion file, and converting the image data based on the conversion data Character string and numeric string character data into “items” and “amounts” based on word data in word files A fourth step of converting the “items” converted based on the word data into “subjects” based on the consolidated data of the dictionary file; and a “subject” consolidated based on the consolidated data. A sixth step of tallying the “amount” corresponding to the “item” before aggregation for each “
A seventh step of storing the document data in a customer data file as customer data including a set of “subjects” aggregated in the fifth step and “amount” aggregated in the sixth step. This is a method for creating customer data.

According to a second aspect of the present invention, in the method for creating customer data according to the first aspect, the second step includes a step of determining a vertical position for distinguishing a character description area and a number description area from the image data. Inserting a ruled line as needed, registering new layout data in the LC file as needed, and calling layout data registered in the LC file. I do.

According to a third aspect of the present invention, in the method for creating customer data according to the first aspect, the fourth step uses the character data of the character string converted based on the converted data as new word data. The method may further include registering the word file.

According to a fourth aspect of the present invention, the document data is
The first method of reading with a CR, converting the image data into image data, and storing the converted image data in an image data file.
A second step of attaching attribute data for each customer to the image data stored in the image data file; and a layout data for each attribute based on the attribute data for each customer attached to the image data. A third step of calling from the LC file; a fourth step of determining a character description area and a number description area in the image data stored in the image data file based on the called layout data; Based on the conversion data of the character data conversion file, the image data of the character string and the number string existing respectively in the character description area and the number description area determined based on the character data conversion area are converted into the character data of the character string and the number string. The character data of the character string and the number string converted based on the conversion data A sixth step of converting “items” and “amounts” based on the word data of the file, and a step of integrating the “items” converted based on the word data into “subjects” based on the aggregated data of the dictionary file. Seventh step, and an eighth step of tallying "amount" corresponding to "item" before aggregation for each "subject" aggregated based on the aggregated data,
A ninth step of storing the document data in a customer data file as customer data including a set of “subjects” aggregated in the seventh step and “amount” aggregated in the eighth step. This is a method for creating customer data.

According to a fifth aspect of the present invention, in the method for creating customer data according to the fourth aspect, the fourth step is a step of making the image data include a vertical area for distinguishing between a character description area and a number description area. Inserting a ruled line as needed, registering new layout data in the LC file as needed, and calling layout data registered in the LC file. I do.

According to a sixth aspect of the present invention, in the method for creating customer data according to the fourth aspect, the sixth step uses the character data of the character string converted based on the converted data as new word data. The method may further include registering the word file.

According to a seventh aspect of the present invention, the document data is
A procedure for reading by a CR, converting the image data into image data, and storing the converted image data in an image data file, and converting a character description area and a number description area in the image data stored in the image data file into an LC file. A procedure for determining based on the layout data;
The image data of the character string and the number string respectively present in the character description area and the number description area determined based on the layout data are converted into the character data of the character string and the number string based on the conversion data of the character data conversion file. To convert the character data of the character string and the number string converted based on the conversion data into “item” and “amount” based on the word data of the word file; and For integrating the “items” converted into “subjects” based on the aggregated data of the dictionary file, and “subjects” aggregated based on the aggregated data.
A procedure for totalizing “amounts” corresponding to “items” before aggregation for each, and storing the document data in the customer data file as customer data including a set of aggregated “subjects” and aggregated “amounts” And a computer-readable recording medium on which a program for causing a computer to execute the steps described above is recorded.

According to an eighth aspect of the present invention, the document data is
A step of reading with a CR, converting the image data into image data, and storing the converted image data in an image data file; a step of attaching attribute data for each customer to the image data stored in the image data file; Based on the attached customer-specific attribute data, a corresponding attribute-specific layout data is called from the LC file, and a character description area and a number description area in the image data stored in the image data file are called. And a character data conversion file conversion process for converting character string and numeral string image data respectively present in the character description area and the number description area determined based on the layout data into a character data conversion file. Convert to character string and numeric string character data based on data Order, a procedure of converting the character data of the character string and the number string converted based on the conversion data into “items” and “amounts” based on the word data of the word file, and converted based on the word data. A procedure for integrating the “items” into “subjects” based on the aggregated data of the dictionary file, and tabulating “amounts” corresponding to the “items” before the aggregation for each “subject” aggregated based on the aggregated data. Computer-readable recording of a program for causing a computer to execute a step of storing the document data and a step of storing the document data in a customer data file as customer data including a set of an aggregated “subject” and an aggregated “amount”. It is a possible recording medium.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention reads out financial statements (document data) created by a customer with an OCR and automatically converts the financial statements into customer data that can be processed by another analysis program (a program for analyzing the business condition of a customer). It is a thing that can be converted. By using the method of the present invention, the operator completes the inputting operation to the computer only by reading the financial statements into the OCR and making some necessary modifications to the data of the read financial statements. Will be able to do it.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing a schematic configuration of a system for implementing the method of the present invention.

This system has a plurality of general-purpose computers, a scanner (OCR) for reading financial statements, and a server having data common to all computers.

FIG. 2 is a block diagram showing elements necessary for implementing the method of the present invention in this system.

The scanner (OCR) 10 outputs, for example, financial statements shown in FIG. 4 to FIG. 16 created by a customer as image data, and is generally used.

The keyboard 20 is used for inputting attribute data to be described later and correcting the input data, and can be replaced with a mouse or the like.

The display device 30 is, specifically, a CRT, a liquid crystal display, or a plasma display, and the output device 40 is, specifically, a printer.

The CPU 50 is a general term for the CPUs of the computers and servers shown in FIG. 1, and controls the system as a whole.

The storage device 60 is a general term for the storage units of the system.

The image data file 61 is a file for storing image data read by the scanner 10.

The LC file 62 is a file in which layout data described later is registered. In this file, layout data in a standard format as described later, layout data for each customer (company), and layout data arbitrarily created by an operator are registered.

The character data conversion file 63 is a file storing conversion data for converting image data into character data.

The word file 64 is a file in which word data for converting character data into words of a character string or a number string is registered.

The dictionary file 65 is a file that stores aggregated data for integrating words converted into character strings and numeric strings into “subjects” that can be processed by another analysis program.

The customer data file 66 is a file for storing finally generated customer data. Therefore, the customer data used for another analysis program is the one stored in the customer data file 66.

The input / output controller 100 controls input / output of data between the scanner 10, the keyboard 20, the display device 30, the output device 40, and the CPU 50 based on a command from the CPU 50.

The present invention is carried out as follows by the device configured as described above. Hereinafter, the procedure of the method of the present invention will be described together with the description of the functions and the like of each element constituting the apparatus. First, the operator reads the financial statements (a plurality of sheets), which are the document data shown in FIGS.
Let R10 read. There are five types of financial statements: balance sheet, profit and loss statement, selling expenses, manufacturing cost report, and profit appropriation.

The financial statements read by the OCR 10 are converted into image data and temporarily stored in the image data file 61. OCR10
(Up to 50 sheets) is possible, so that all financial statements read continuously are temporarily stored in the image data file 61.

A file name ("number") is automatically assigned to each piece of read image data. The attribute definition operator of each financial statement adds attribute data to each of the financial statements stored in the image data file 61. Input of the attribute data is performed using the keyboard 20.

As attribute data, “branch number”,
"Company number", "Fiscal year", "Type of financial statements (one of balance sheet, profit and loss statement, selling expenses, manufacturing cost report, profit appropriation)", "Input unit (yen, thousand Yen unit,
In millions of yen) and “Financial Statement Format (Character and Number Arrangement Format)”.

By inputting this attribute data,
It will be specified which financial statements are for which company, for which financial period and for which financial statements. Conversion to character data of each financial statement Character data area and number data area in the image data are discriminated and the image data of the character string and number string existing in both areas are converted to character data of the character string and number string. I do.

It should be noted that the distinction between the character description area and the number description area is made based on the vertical ruled lines existing in the image data (regardless of the horizontal ruled lines).

For example, in a description form in which character strings and numeric strings are arranged as in the balance sheet shown in FIG. 4, the first area surrounded by vertical ruled lines is "subject", and The second area surrounded by the vertical ruled line is “money”, the next third area and the next fourth area are also “subjects”,
Since the arrangement follows “amount”, “family + amount + family + amount” is designated as “financial statement format” by layout data registered in LC file 62 described later. By this designation, the first area and the third area surrounded by the vertical ruled line are determined to be character areas, and the second area and the fourth area surrounded by the vertical ruled line are numbers. It will be determined that the area is the description area. Such designation enables efficient character recognition (subject and amount).

Sixteen types of such "financial statement format" are prepared as layout data in a standard format so as to be compatible with any description format, and are registered in advance in an LC (layout control) file.

The 16 types are as follows. In addition,
The operator can freely set any other "financial statement format", and can register this as layout data. "Course + Amount""Course + Amount + □""Course + Amount + □ + □ + □""Course + □ + Amount""Course + □ + □ + Amount + □" + □ + Amount + Amount ”“ Section + Amount + □ + Amount ”“ Section + Amount + Section + Amount ”“ Section + □ + Amount + Section + □ + Amount ” "Course + □ + Amount + □ + Department □ + Amount + □""Course + Amount + □ + □ + Department + Amount + □ + □""Course + Amount + Amount + □ + Department + Amount + Amount"" Department + □ + Amount + Amount + Department + □ + Amount + Amount “Department + Amount + □ + Amount + Department + Amount + □ + Amount” In the description format as shown in FIG. amount"
Is called from the LC file 62 and designated as "financial statement format".

In the case of the description format as shown in FIG. 11, since a ruled line is not included in the image data, it is not possible to distinguish between the character description area and the number description area. The operator draws a vertical ruled line for segmenting the area using a keyboard or the like in such a description format. Then, the layout data of “department + department + amount + amount” outside the standard format registered by the operator is called from the LC file 62 and designated as “financial statement format”.

According to the designated “financial statement format”, the computer converts a character string (excluding numbers) existing in the character description area and a number string (only numbers) existing in the number description area one by one into known characters. Convert using recognition processing. That is, based on the conversion data of the character data conversion file 63, the image data of the character string and the number string in the image data is converted into the character data of the character string and the number string.

The range for character recognition as described above is ultimately designated by the operator (usually, the entire range of image data is designated by a frame), but this range can be limited. For example, if it is desired to perform character recognition on only the left half of a financial statement (document data) in a description format as shown in FIG. 4, a first area surrounded by vertical lines and a second area surrounded by vertical lines Specify the area by surrounding it with a frame. In addition, if the “financial report format” of “department + amount + □ + □” is registered as layout data in the LC file 62, the “financial report format” is designated and the image data Can be limited to only the left half even if the entire range is surrounded by a frame.

The layout data of the LC file 62 is stored for each "company number", that is, for each customer. Therefore, based on the company number attached to the image data, the layout data corresponding to that company is
It can be called from the C file 62. In this way, layout data that conforms to the format of the financial statements commonly used by customers can be automatically selected without any operator's selection operation, and input efficiency is dramatically improved. become.

For example, suppose that the customer A prefers to use the description format of the financial statement “family + □ + amount”, and the customer B has the description format of the financial statement “family + □ + amount + amount”. You prefer to use When the operator performs character recognition of such image data having a different layout, it is usually necessary to select suitable layout data from the LC file 62. However, if layout data is stored for each customer, Based on the attribute data (“customer's“ company number ”)”, layout data suitable for the image data is automatically called up.

That is, when the image data of the customer A is to be character-recognized, the image data corresponding to the "company number" in the LC file 62 is determined based on the "company number" of the customer A attached to the image data. The layout data called "Department + □ + Amount" is automatically called. When character recognition is performed on the image data of the customer of B, based on the “company number” of the customer B attached to the image data, registration is made in the LC file 62 corresponding to the “company number”. "Department + □"
The layout data “+ amount + amount” is automatically called. The character data of the aggregate conversion string and numeric string to the "subjects" on the basis of the word data of word file 64 "item" and "amount of money"
And converts the converted “items” into “subjects” based on the consolidated data of the dictionary file 65.

This processing includes the steps of converting the converted character string and character string character data into "items" and "amounts" based on the word data in the word file 64, and converting the converted character string character data into character strings. If the corresponding word data does not exist in the word file 64, the step of registering new word data in the word file 64 as necessary, and the step of converting the converted “items” into the aggregated data of the dictionary file 65. On the basis of the subject.

A word file 64 and a dictionary file 65 are provided for this aggregation. Based on the word data and the aggregated data registered in these files, the “items” of the read financial statements are aggregated into “subjects” that can be read by other analysis programs.

The word file 64 is a dictionary in which word data for converting character data of a converted character string into “items” and converting character data of a converted numeric string into “money” are registered. Yes, for example, "this + seat +
Character data for converting character data of one character of "deposit + money" into one word of "current account" is registered.

If an erroneous reading is made at the input stage of the OCR 10, the operator cannot correct the word because the word cannot be recognized correctly. Whether or not the word has been corrected is automatically registered in the word file 64.

For example, if the converted character data is “Xiao + Za + Deposit + All”, it is searched whether or not the word “XiaoZaDezen” exists in the word data of the word file 64. However, such words are stored in word file 6
4, the operator corrects the current account to "Checking Account". In such a case, the word file 64 is automatically registered with word data for changing “Xiao Zhao Zhen” to “Checking Account”.

After this registration, even if the converted character data is "Xiao + Za + Deposit + All", it will be converted to the word "Current Account". The fact that such registration can be performed is very useful for improving the reading reliability of the OCR 10 which is considered to be frequently misread.

The dictionary file 65 is a file in which consolidated data for conversion into “subjects” that can be processed by another analysis program is registered.
Specifically, as shown in Tables 1 and 2 below, 130
It is a subject. The “items” described in the financial statements read by the OCR 10 are collected into any of the 130 kinds of “subjects”.

[0059]

[Table 1]

[0060]

[Table 2]

For example, as described in "Problems to be Solved by the Invention", "Cash" ... 2,652,000 yen "Current account" ... 11,365,000 yen "Other deposits" ... 568 000 yen "Prepayment" ... If you read the financial statement with 152,400 yen in OCR10, all items of "Cash", "Current account" and "Other deposits" are "Cash and deposit" , "Prepayment", and "other current assets", the dictionary file 65 indicates that "cash" is "cash / deposit" and "current account" is "cash / deposit". In addition, consolidated data for consolidating “other deposits” into “cash / deposits” and “prepayment” into “other current assets” are registered.

Based on the data registered in both files as described above, for example, the character data converted to “current + za + deposit + money” is changed to the item “current account” based on the registered data in the word file 64. The item “current account” is converted into an item “cash / deposit” that can be processed by another analysis program based on the aggregated data of the dictionary file 65.

In addition, due to misreading of OCR10, “Xiao + za +
Even if the character data is converted to “deposit + all”, if this character data is newly registered in the word file 64 as registration data, the character data is changed to the item “Checking Account” based on the registration data in the word file 64. Converted,
The item “current account” is converted to an item “cash / deposit” that can be processed by another analysis program based on the aggregated data in the dictionary file 65. Aggregation of Amounts Next, the amounts of the numeric strings in the number entry area corresponding to the "items" before the aggregation are totalized for each "subject". This processing simply sums up the amounts corresponding to the “subjects” collected in the above processing.

The amount is recognized based on the character data of the digit string converted earlier, and it is known which "item" before the aggregation corresponds to each amount. Therefore, "Cash" ... 2,652,000 yen "Checking account" ... 11,365,000 yen "Other deposits" ... 568,000 yen "Prepayment" ... 152,400 yen Is read by OCR10, the above processing will result in “Cash / Deposit”… 2,652,000 yen “Cash / Deposit”… 11,365,000 yen “Cash / Deposit”… 568,000 yen “Other current assets” "... 152,400 yen. Therefore, this description is tabulated as "cash / deposit": 14,585,000 yen "Other current assets": 152,400 yen.

Since it is conceivable to correct the amount at the time of this totaling, the amount before and after the totaling can be corrected by the operation of the operator. Creation of Customer Data The data finally obtained as described above is arranged according to “company + financial term” previously input as attribute data, and is used as customer data. This customer data is registered and stored in the customer data file 66.

The customer data conforms to a form that can be used as it is in another analysis program (a program for analyzing the business condition based on the customer data).

Next, a series of procedures of the method of the present invention will be described with reference to the flowchart of FIG.

This series of procedures will be described on the assumption that financial statements shown in FIGS. 4 to 16 are input.

First, the operator collectively sets the financial statements shown in FIGS. 4 to 16 in the OCR 10 and outputs a read instruction. According to this instruction, the OCR 10 reads the financial statements one by one and converts them into image data (S1).
Image data file 61 with file name
(S2). The above reading is shown in FIGS.
The process is continued until the reading of the thirteen financial statements shown in FIG. 6 is completed (S3).

Next, the operator calls out the financial statements stored in the image data file 61 one by one (S4) and adds attribute data (S5). That is, the image data is called one by one, and the “branch number”, “company number”, “financial term”, “financial statement type (balance sheet, profit and loss statement, selling expense statement, manufacturing cost report) , Profit appropriation), "input unit (any of yen, thousand yen, or million yen)", and "financial statement format (letter and number format)". If there is no ruled line in the called image data, a vertical ruled line is drawn in the image data (S6, S7).

In the input of the attribute data, the operator designates the “financial statement type” by the operator in the LC file 6.
2 is called and specified, or if there is no corresponding layout data in the LC file 62, the corresponding layout data is created. The created layout data is registered as new layout data in the LC file 62.

When the layout data registered with the attribute data for each company is present in the LC file 62, the layout data can be read without the operator calling the layout data as described above. Is automatically called.

For example, when the financial statements shown in FIG. 4 are called, “branch number” is changed to “xxxx” and “company number”
"XXX", "Fiscal term" as "H6 / 3", "Type of financial statements" as "Balance sheet", "Input unit" as "Thousands of yen", and "Format of financial statements" as Enter "Department + Amount + Department + Amount". When the profit and loss statement of FIG. 5 is called, “branch number” is “xxxxxx”, and “company number” is “xxxxx”.
○, “Fiscal term” is “H6 / 3”, “Financial statement type” is “Profit and loss statement”, “Input unit” is “Thousands of yen”, and “Financial statement format” is “Family + Amount + Amount ". FIG.
When the profit and loss statement is called, the “branch number” is “xxxx”, the “company number” is “xxxxxx”, the “financial term” is “H6 / 3”, "Type of statement" is "Balance sheet", "Input unit" is "Thousands of yen", "Format of financial statements"
Is input as "family + sum + sum + family + sum + sum". When the profit and loss statement in FIG. 14 is called, “branch number” is changed to “×
"XXX", "Company number" is "XXX", "Fiscal term"
For “H6 / 3”, “Type of financial statements” for “Profit and loss statement”,
Enter "input unit" as "thousands of yen" and "financial statement format" as "department + amount + amount + amount + amount".

Since the financial statements shown in FIG. 8 to FIG. 12 are not ruled, a vertical ruled line for separating the character description area and the number description area is drawn for such image data. become.

When the input of the attribute data as described above is completed, an area for converting the image data into the character data is designated. Specifically, the operator draws a frame indicating the area designation range on the image data displayed on the screen display screen. The image data in this frame is converted into character data (S8).

With the above input, preparations for automatically converting the items described in the financial statements prepared by the customer into customer data are completed.

Next, the image data of the designated area is converted into character data. This conversion is performed based on the conversion data stored in the character data conversion file 63. In other words, first, "Format of financial statements"
The "family" area designated as "
Words consisting of kanji such as "deposit", "accounts receivable" and "raw materials" are described. These words are converted into character data (words) of accurate character strings. Since there are only numbers in, this is converted to character data (numbers) of an accurate number string (S9).

Then, the converted character string and character string character data are converted into “items” and “amounts” based on the word data in the word file 64. If a word cannot be recognized correctly due to misreading of the OCR 10, the operator corrects it. When this correction is made, which character is corrected to which word is automatically registered in the word file 64.

For example, if the converted character data is “deposit + all”, it is searched whether the word “deposit all” exists in the word file 64. Usually, such a word is not registered in the word file 64, so the operator corrects it to "deposit".
After such correction, the word data for converting the character data "deposit + all" to "deposit" is automatically registered (S10).

Next, the converted "items" are converted into "subjects" which can be processed by another analysis program. The conversion from “items” to “subjects” is performed based on the dictionary file 65 consolidated data. The converted “subjects” are associated with one of the 130 subjects as shown in Tables 1 and 2 and are aggregated (S11).

Next, the amounts corresponding to the “items” before the aggregation are totaled for each of the aggregated “subjects”, and sorted by “company + financial term” previously input as the attribute data. It is stored in the data file 66 (S1
2).

[0082]

According to the first aspect of the present invention, there is provided the above procedure.
According to the method for creating customer data according to claim 6,
The operator can complete the input operation to the computer only by reading the financial statements into the OCR 10 and making a slight modification as necessary to the read financial statement data. Since the input operation is simplified and the input operation does not require skill as in the related art, anyone who can operate the computer can perform the input operation. In addition, since input errors can be reliably eliminated, reliability of input data can be ensured.

Further, according to a computer-readable recording medium on which a program to be executed by a computer according to claim 7 or 8 is recorded, the computer can read the program recorded on this recording medium. Thus, the same effects as those of the first to sixth aspects can be obtained.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a system for implementing a method of the present invention.

FIG. 2 is a block diagram showing the elements required to carry out the method of the present invention.

FIG. 3 is a flowchart showing a procedure of the method of the present invention.

FIG. 4 is a diagram showing an example of financial statements read by OCR.

FIG. 5 is a diagram showing an example of financial statements read by OCR.

FIG. 6 is a diagram showing an example of financial statements read by OCR.

FIG. 7 is a diagram showing an example of financial statements read by OCR.

FIG. 8 is a diagram showing an example of financial statements read by OCR.

FIG. 9 is a diagram showing an example of financial statements read by OCR.

FIG. 10 is a diagram showing an example of financial statements read by OCR.

FIG. 11 is a diagram showing an example of financial statements read by OCR.

FIG. 12 is a diagram showing an example of financial statements read by OCR.

FIG. 13 is a diagram showing an example of financial statements read by OCR.

FIG. 14 is a diagram showing an example of financial statements read by OCR.

FIG. 15 is a diagram showing an example of financial statements read by OCR.

FIG. 16 is a diagram showing an example of financial statements read by OCR.

[Explanation of symbols]

 10 scanner (OCR), 61 image data file, 62 LC file, 63 character data conversion file, 64 word file, 65 dictionary file, 66 customer data file.

Claims (8)

[Claims] 1. A first step in which document data is read by an OCR and converted into image data, and the converted image data is stored in an image data file; and a description of characters in the image data stored in the image data file is provided. A second step of determining an area and a number description area based on the layout data of the LC file; and a character string and a number string respectively existing in the character description area and the number description area determined based on the layout data. A third step of converting the image data of the character string into character string and numeric string character data based on the conversion data of the character data conversion file; and converting the character string and numeric string character data converted based on the conversion data into a word file. A fourth step of converting into “items” and “amounts” based on the word data of Fifth aggregated into "subjects" on the basis of the "item" that has been converted to Zui the aggregate data of the dictionary file
A step, a sixth step of tallying the "amount" corresponding to the "item" before the aggregation for each "subject" aggregated based on the aggregated data, and a "step" in which the document data is aggregated in the fifth step. And storing the customer data in a customer data file as customer data comprising a set of "items" and "money" totaled in the sixth step. 2. The method according to claim 1, wherein the second step includes inserting a vertical ruled line for discriminating between a character description area and a number description area in the image data as needed, and adding new layout data in the LC file. 2. The method according to claim 1, further comprising: registering as required, and calling layout data registered in the LC file. 3. The method according to claim 1, wherein the fourth step includes a step of registering, in the word file, character data of a character string converted based on the converted data as new word data. How to create customer data. 4. A first step in which document data is read by an OCR and converted into image data, and the converted image data is stored in an image data file. A second step of attaching attribute data; a third step of calling corresponding layout data of each attribute from the LC file based on the attribute data of each customer attached to the image data; and storing the layout data in the image data file. A fourth step of determining a character description area and a number description area in the image data based on the called layout data; and a character description area and a number description area determined based on the layout data, respectively. Character string and numeric string image data based on the conversion data of the character data conversion file. A fifth step of converting the character data of the character string and the number string converted based on the conversion data into “items” and “amounts” based on the word data of the word file. A sixth step of converting the “items” converted based on the word data into “subjects” based on the aggregated data of the dictionary file
A step, an eighth step of tallying the “amount” corresponding to the “item” before the aggregation for each “subject” aggregated based on the aggregated data, and “an aggregation of the document data in the seventh step” A ninth step of storing the customer data in the customer data file as customer data comprising a set of “subjects” and “amount” calculated in the eighth step. 5. The method according to claim 5, wherein the fourth step includes, if necessary, inserting a vertical ruled line for discriminating between a character description area and a number description area in the image data; The method according to claim 4, further comprising the steps of: registering as necessary, and calling layout data registered in the LC file. 6. The method according to claim 4, wherein the sixth step includes a step of registering character data of a character string converted based on the conversion data as new word data in the word file. How to create customer data. 7. A procedure for reading document data by an OCR, converting the document data into image data, storing the converted image data in an image data file, and a character description area in the image data stored in the image data file. A step of determining the number description area based on the layout data of the LC file; and a step of determining the image data of the character string and the number string existing in the character description area and the number description area, respectively, determined based on the layout data. A step of converting character data of a character string and a number string based on the conversion data of the character data conversion file into character data of a character string and a number string, based on the word data of the word file; A procedure for converting into “item” and “amount”, and a conversion based on the word data A procedure for integrating “items” into “subjects” based on the aggregated data in the dictionary file, and for each “subject” aggregated based on the aggregated data, “amount” corresponding to the “item” before the aggregation is totaled. Computer-readable recording a program for causing a computer to execute a procedure and a procedure for storing the document data in a customer data file as customer data including a set of an aggregated “subject” and an aggregated “amount” Recording medium. 8. A procedure for reading document data by an OCR, converting the document data into image data, storing the converted image data in an image data file, and storing attribute data for each customer in the image data stored in the image data file. , A procedure for calling layout data for each attribute from the LC file based on the attribute data for each customer attached to the image data, and a description of characters in the image data stored in the image data file. A procedure of determining the area and the number description area based on the called layout data; and an image of a character string and a number string respectively existing in the character description area and the number description area determined based on the layout data Character strings and numbers based on the conversion data of the character data conversion file Converting the character data of the character string and the number string converted based on the converted data into “items” and “amount” based on the word data of the word file; A procedure for integrating the “items” converted based on the data into “subjects” based on the aggregated data in the dictionary file, and corresponding to the “items” before the aggregation for each “subject” aggregated based on the aggregated data And a step of storing the document data in the customer data file as customer data including a set of the aggregated “subjects” and the totalized “amount”. A computer-readable recording medium on which a program is recorded.