JP7685233B2 - Cardboard document management system - Google Patents

Description

This invention relates to a cardboard voucher management system, and in particular to a cardboard voucher management system that uses character recognition to register information written on vouchers, such as cardboard order forms, from image data of the vouchers.

Note that a voucher is a general term for documents (written documents) that serve as evidence of a transaction or contract, and examples of vouchers include purchase orders, delivery notes, invoices, and receipts.

Conventionally, there is a known system that receives image data of a cardboard order form and registers the order information using character recognition (see, for example, Patent Document 1).

The above-mentioned Patent Document 1 discloses a receiving fax management system that includes a fax (facsimile) receiving means and a destination recognition means having an OCR (optical character recognition) function that reads information such as a destination code from a fax document received by the fax receiving means. The receiving fax management system automatically recognizes the destination of a received fax document using the sender's fax number, information recorded in the fax header, or the OCR function, and stores the information in an image database. When displaying a received order fax, the original image data is displayed on the right side of the screen, and the order data read by the OCR function is displayed on the left side of the screen. The operator checks whether the order data on the left side of the screen has been entered correctly by appropriately enlarging or scrolling the image data on the right side of the screen, and if it is incorrect, corrects it using a mouse, keyboard, etc. In the above-mentioned Patent Document 1, in order to improve the accuracy of OCR recognition, it is proposed that customers use dedicated standard documents that require the use of character fonts for OCR, and for image data of non-standard documents, the operator is to manually input the information without using the OCR function.

JP 2002-165076 A

Reading image data using character recognition (OCR function) as in Patent Document 1 above can significantly improve work efficiency compared to when an operator manually inputs the evidence (or image data of the evidence) while visually checking it. On the other hand, with evidence such as purchase orders, input errors can directly lead to contractual problems such as incorrect orders, so input errors must be prevented as much as possible. Since reading image data using character recognition often does not provide sufficient accuracy, confirmation and correction work by the operator is essential in the registration process of evidence data. Therefore, it is desirable to improve the recognition accuracy of character recognition in order to improve the work efficiency of the registration process and reduce the workload of the operator.

The above-mentioned Patent Document 1 proposes that trading partners use dedicated standard documents that require the use of character fonts for OCR, but since trading partners usually have been using document vouchers in their own formats continuously for many years, it is difficult for them to switch to dedicated standard documents, which creates practical problems.

This invention has been made to solve the above problems, and one object of the invention is to provide a cardboard voucher management system that can improve processing efficiency by improving character recognition accuracy in the information registration process from image data of vouchers.

A cardboard voucher management system according to one aspect of the present invention includes an image acquisition unit that acquires image data of a voucher related to a cardboard box, a display unit that displays the image data, an input unit that accepts input operations, an information processing unit that performs a registration process of information included in the image data, and a storage unit that pre-records format data of the voucher for each issuer of the voucher, and the information processing unit is configured to select format data from the storage unit according to the issuer of the voucher, and perform character recognition processing of the image data with parameters based on the selected format data to acquire recognition character strings for each item of the voucher, and execute a process of inputting the acquired recognition character strings for each item of an input screen displayed on the display unit, and the format data includes row information including the maximum number of rows of the voucher and column information related to items included in one row of the voucher, and the information processing unit controls the image display screen that displays the image data and the input screen. The display unit is configured to display the document on the display screen, recognize characters on a line-by-line basis in the image data displayed on the image display screen, and input the recognized character strings of each item included in the recognized line into the input screen for each item, sequentially executing the process according to the maximum number of lines, the line information including, in addition to the maximum number of lines, information on the top and bottom end positions of any line and information on line spacing, the column information including information on the left and right end positions of each item, the information processing unit is configured to display a first frame line indicating each line of the document superimposed on the image data displayed on the image display screen based on the information on the top and bottom end positions, line spacing, and the maximum number of lines, and to display a second frame line indicating each item of the document superimposed on the image data displayed on the image display screen based on the information on the left and right end positions of each item, and to obtain the character strings included in the second frame line as the recognition character strings for each item, with the area within the first frame line being regarded as one line. With this configuration, it is possible to accurately identify the area of each line and each item included in the document, without setting the positions and ranges of all lines and items one by one. In addition, since the first frame line and the second frame line are displayed superimposed on the image data, the operator can visually check whether the format data used for character recognition matches the format of the evidence captured as the image data. Therefore, even if the wrong format data is selected, the operator can easily understand.

In the cardboard voucher management system according to the above aspect, the format data preferably includes information on the reference line included in the voucher, and the information processing unit is configured to perform a process of correcting the magnification rate and rotation angle of the image data during character recognition processing so that the reference line in the image data matches the reference line set in the format data. In this way, by setting the information on the reference line included in the voucher in advance in the format data and matching the reference line in the image data with the reference line in the format data, elements such as character size, inclination, line spacing, and item (column) width in the image data can be matched with the sample (template) used to set the format data. As a result, the size and direction of the voucher in the image data can be made constant, and character recognition processing optimized by the format data can be applied, so high recognition accuracy can be maintained.

In the cardboard voucher management system according to the above aspect, the format data preferably includes character spacing information for each item included in the voucher, and the information processing unit is configured to execute a process of adjusting character spacing in the image data based on the character spacing information for each item of the image data during character recognition processing. Here, depending on the font and character pitch used when the issuer issues the voucher, it may be difficult to separate and recognize adjacent characters, which causes a decrease in recognition accuracy. Therefore, in the above configuration, a pre-processing can be performed to adjust the character spacing to a level suitable for character recognition processing. In addition, since the character type (numbers, alphabet, kana/kanji, etc.) used is determined for each item of the voucher, the character spacing can be adjusted to an optimal level according to the character type by setting the character spacing information for each item. As a result, the accuracy of character recognition can be further effectively improved.

In the cardboard voucher management system according to the above aspect, the storage unit is preferably configured to store format data for each type of cardboard, including cardboard sheets and cardboard cases, and the information processing unit is configured to select format data from the storage unit according to the type of voucher. With this configuration, even if each type of cardboard has its own unique items, character recognition and information input for each item can be performed appropriately.

In the cardboard voucher management system according to the above aspect, preferably, the information processing unit is configured to, when receiving an input operation of correction via the input unit for a recognized character string input in any of the items, record the pre-correction character string and the post-correction character string in association with each other in the storage unit, and, when the recognized character string obtained by character recognition corresponds to the pre-correction character string recorded in the storage unit, perform a process of replacing it with the corresponding post-correction character string. With this configuration, even if a misrecognition occurs in the character recognition process and the operator corrects the input information, if the same pre-correction character string is recognized in the subsequent character recognition process, it can be automatically replaced with the post-correction character string. This effectively reduces the operator's workload. In addition, since certain character strings (e.g., product name, delivery location, etc.) are frequently written in the voucher issued for each transaction, the simple process of replacing the pre-correction character string with the post-correction character string can effectively suppress erroneous input.

As described above, the present invention provides a cardboard voucher management system that can improve processing efficiency by improving character recognition accuracy in the information registration process from image data of vouchers.

1 is a schematic diagram showing the overall configuration of a cardboard voucher management system according to one embodiment; 13A and 13B are diagrams showing examples of display (case order form) on an image display screen and an input screen. 13A and 13B are diagrams showing examples of display (sheet order form) on an image display screen and an input screen; FIG. 2 is a diagram for explaining format data. FIG. 13 is a diagram showing an example of a format data setting and registration screen. 11A and 11B are diagrams for explaining adjustment processing of character spacing. FIG. 13 is a diagram for explaining a method for setting information on a reference line. 13A shows image data that is the basis of the format data, (B) newly acquired image data, and (C) image data for explaining an example of adjustment using a second reference line. FIG. 1A and 1B are schematic diagrams for explaining character recognition processing for image data. 11 is a schematic diagram showing an example of a notification process when an input value includes an abnormal value (error). FIG. 1A is a schematic diagram illustrating a substitution table showing a character string at the time of initial recognition (A), a substitution table (B), and a character string at subsequent recognition times (C). FIG. 13 is a flowchart showing a registration process of order information. FIG. 13 is a schematic diagram showing an example of a received fax list screen. 13 is a flowchart showing the flow of the OCR registration process in FIG. 12 . FIG. 11 is a schematic diagram for explaining a verification process. FIG. 13 is a schematic diagram showing an example of an order confirmation image.

The following describes an embodiment of the present invention with reference to the drawings.

First, the configuration of the cardboard evidence management system 1 according to this embodiment will be described with reference to Figs. 1 to 11. As shown in Fig. 1, the cardboard evidence management system 1 is a system for managing the information on the evidence 6 by performing character recognition processing on image data 50 of the evidence 6 received from a trading partner, thereby registering the information on the evidence 6 in a database.

The vouchers 6 managed by the cardboard voucher management system 1 are documents such as order forms, delivery notes, invoices, and receipts related to transactions involving case-shaped or sheet-shaped cardboard as products, but may be other vouchers. Below, an order form for case-shaped or sheet-shaped cardboard received from a customer is given as an example of a voucher, and a configuration is described for performing a process of registering the information written on the order form (order contents) in the database of the management server 4.

The cardboard voucher management system 1 includes a receiving unit 2, an information processing terminal 3, and a management server 4. The cardboard voucher management system 1 also includes an image reading unit 5. The receiving unit 2 and the image reading unit 5 are an example of the "image acquisition unit" in the claims.

(Receiving section)
The receiving unit 2 acquires image data 50 of a voucher (order form for cardboard boxes) related to the cardboard boxes. Specifically, the receiving unit 2 includes a fax modem that receives image data 50 transmitted by fax from a customer over a telephone line. The receiving unit 2 converts a fax signal received over the telephone line to generate fax image data, and outputs the fax image data to the management server 4. Furthermore, when the receiving unit 2 acquires fax image data to be transmitted from the management server 4 or the information processing terminal 3, it converts the acquired fax image data into a fax signal and transmits the fax signal over the telephone line.

(Image reading unit)
The image reading unit 5 acquires image data 50 of a voucher 6 (order form for cardboard) relating to the cardboard box. Specifically, the image reading unit 5 is configured to generate image data 50 of the order form by optically reading the printed order form and converting it into digital data (image information). The image reading unit 5 is an image scanner. The image reading unit 5 outputs the read image data 50 to the information processing terminal 3. The read image data 50 is transmitted to the management server 4 via the information processing terminal 3. The receiving unit 2 and the image reading unit 5 may be configured integrally as, for example, a multifunction machine (an office machine having the functions of a fax, scanner, copy, and printer).

(Management Server)
The management server 4 accepts and manages the registration of information described on the evidence 6 in the cardboard evidence management system 1. The management server 4 is connected to the information processing terminal 3 via a network. The management server 4 is connected to the receiving unit 2. The management server 4 includes a memory unit 4a.

The memory unit 4a stores a registration database 41 that records registered information. The memory unit 4a stores an image database 42 that records image data 50 of the evidence 6 (order form) generated by each of the receiving unit 2 and the image reading unit 5.

In the image database 42, a separate folder is created for each issuer (transaction partner) of the evidence 6. The management server 4 is configured to sort and record the image data 50 in the folder for the corresponding issuer (transaction partner).

When the image data 50 of the order form is acquired, the cardboard voucher management system 1 performs a registration process to register the order information R (see FIG. 2) described on the order form in the registration database 41 using the information processing terminal 3. In the registration process, the management server 4 transmits the image data 50 of the order form recorded in the image database 42 to the information processing terminal 3 in response to an access request from the information processing terminal 3. The person in charge of registration (operator) uses the information processing terminal 3 to perform the registration process of the order information R described on the order form. The management server 4 receives the order information R (order content data R2 described later) input by the registration process from the information processing terminal 3. The management server 4 is configured to record the order registration data 41a that associates the received order information R with the order number Ab assigned to the order information R in the registration database 41. The order number Ab is a unique identification information assigned to one piece of order information R.

In this embodiment, prior to the registration process, the storage unit 4a pre-records the format data 43 of the evidence for each issuer of the evidence 6. The format data 43 is data that defines the description format (format) of the evidence, and includes various parameters for performing character recognition processing (OCR processing) on the image data 50 of the evidence.

In addition, the management server 4 is configured to record in the image database 42 an order confirmation image 70 that is generated by adding information indicating that the order registration has been completed to the image data 50 of the order form. The management server 4 is configured to return the generated order confirmation image 70 to the sender (customer) of the order form via the receiving unit 2. The management server 4 is capable of sending multiple order confirmation images 70 all at once, and sending each order confirmation image 70 individually with comments from the person in charge of registration, etc. added to it.

The management server 4 is composed of a computer including a processor, memory, a non-volatile storage device, a communication interface, etc. The non-volatile storage device constitutes the storage unit 4a. The computer functions as the management server 4 by the processor executing the order registration management software stored in the storage device.

(Information processing terminal)
The information processing terminal 3 is configured to perform a registration process of the order information R described in the image data 50 of the order form received by the management server 4. The cardboard voucher management system 1 includes at least one information processing terminal 3. In the example of Fig. 1, a plurality of (three) information processing terminals 3 are provided.

Each information processing terminal 3 includes, as its basic configuration, a display unit 31 that displays image data 50, an input unit 32 that accepts input operations, and an information processing unit 33 that performs registration processing of information included in the image data 50. As with the information processing terminal 3 on the right side of FIG. 1, it may be equipped with multiple display units 31 (first display unit 31a, second display unit 31b). The information processing terminal 3 on the right side is connected to an image reading unit 5. Each of the other information processing terminals 3 may also be provided with an image reading unit 5. Since the basic configuration of the information processing terminals 3 is the same, an explanation will be given by taking one of the multiple information processing terminals 3 as an example.

The information processing terminal 3 is composed of a PC (Personal Computer). The information processing unit 33 is composed of a PC main body including a CPU (Central Processing Unit), HDD (Hard Disc Drive), memory, and a communication interface. The CPU executes a control program stored in the HDD to perform the functions of the information processing unit 33. The display unit 31 is composed of an image display device such as a liquid crystal monitor, and displays an image on the screen based on the image output of the information processing unit 33. The input unit 32 is composed of a keyboard, mouse, etc. that accept operational inputs from the person in charge of registration.

In this embodiment, the information processing unit 33 is configured to, during the registration process of the order information R, (1) select format data 43 from the memory unit 4a according to the issuer of the voucher, (2) execute character recognition processing of the image data 50 using parameters based on the selected format data 43 to obtain a recognition character string 85 (see FIG. 2) for each item of the voucher, and (3) execute processing to input the obtained recognition character string 85 for each item of the input screen 62 (see FIG. 2) displayed on the display unit 31.

The character recognition process is a process of converting the image of a character contained in the image data 50 into character data (character code), and is known as OCR (Optical character recognition) processing. A program (character recognition engine) that performs the character recognition process is installed in the information processing unit 33.

In this embodiment, the registration clerk does not need to manually register each piece of information contained in the image data 50 of the order form. The registration clerk checks whether the data for each item automatically entered into the input screen 62 by character recognition processing matches the information written in the image data 50 of the order form, and manually corrects any inconsistencies. Note that in this embodiment, the information processing unit 33 can also switch modes to execute a mode in which all registration work is performed by the registration clerk through manual input.

<Image display screen and input screen>
When performing the registration process, the information processing section 33 displays an image display screen 61 for displaying the image data 50 and an input screen 62 on the display section 31, as shown in FIG.

The information processing unit 33 acquires image data 50 of the order form selected by the registration clerk from the management server 4 and displays it on the image display screen 61.

The input screen 62 is a screen that has input areas (63a to 63f) for inputting the order information R written on the order form. The registration process is performed by inputting the contents (order information R) written on the image data 50 of the order form displayed on the image display screen 61, item by item, into the corresponding input areas (63a to 63f) on the input screen 62.

The information processing unit 33 displays the image display screen 61 and the input screen 62 side by side. At this time, the information processing unit 33 may display both the image display screen 61 and the input screen 62 on one display unit 31. As in the information processing terminal 3 on the right side of FIG. 1, the information processing unit 33 may display the image display screen 61 and the input screen 62 on the first display unit 31a and the second display unit 31b, respectively, in a multi-display format.

In FIG. 2, the image display screen 61 includes a display area for the image data 50, a type selection button 61a, a format selection button 61b, a setting button 61c, an OCR button 61d, and an end button 61e. The registration clerk can select the type of order using the type selection button 61a, select the format data 43 using the format selection button 61b, and input (click) the OCR button 61d to execute character recognition processing on the image data 50 displayed on the image display screen 61. As will be described later, the information processing unit 33 can automatically select the format data 43 corresponding to the image data 50 based on an image element (a reference line 83 described later) in the image data 50, without relying on the operation input of the format selection button 61b.

The image display screen 61 in Figure 2 shows an example of image data 50 of an order form. An order form contains one or more pieces of order information R. As a business practice in the cardboard industry, an order form displays one piece of order information R per unit line. In the example in Figure 2, one piece of order information R is listed per line. Note that if there are many items in the order information R, one piece of order information R may be listed in multiple lines (for example, two lines, one above the other) per line.

Case Order Form
2 shows an example of a case order for a case-shaped cardboard product. In the image data 50 of the order, the words "Case Order" are placed at the top and in the center in the horizontal direction. A case-shaped cardboard product is a cardboard box on which necessary printing (such as the company name and the name of the product to be packaged) is performed and on which cuts and fold lines are provided so that the cardboard box can be assembled into a case.

In addition to order information R, the purchase order contains issuer information R1 (such as a company name) to identify the issuer (orderer) who issued the purchase order. Order information R is information that describes the contents of the order from the customer. One piece of order information R consists of multiple items of information. In the example of Figure 2, order information R includes the following items: "Order number" item 51a, "Product number" item 51b, "Product name" item 51c, "Quantity" item 51d, "Delivery date" item 51e, and "(Delivery) location" item 51f.

Item 51a's "Order number" is a number used by the purchaser to identify the order. Item 51b's "Product number" is an identification number uniquely assigned to the cardboard product. Item 51c's "Product name" is the name of the cardboard product. Product numbers and product names basically have a one-to-one correspondence. Therefore, one piece of order information R may contain only one of items 51b and 51c. Item 51d's "Quantity" is the number of cardboard products ordered. Item 51e's "Delivery date" specifies the deadline for delivering the ordered product to the purchaser. Item 51f's "(Delivery) location" specifies the location where the product is delivered to the purchaser.

In the example of FIG. 2, the image data 50 of the order form lists the issuer information R1, "XXX Cardboard Co., Ltd." For example, the first order information R (first line) has the text strings "201703" in item 51a, "101" in item 51b, "New Product Tato Type" in item 51c, "100" in item 51d, "December 16th, 8:00" in item 51e, and "XX Factory" in item 51f. In the example of FIG. 2, the case order form is written in a format that can contain up to 10 pieces of order information R (10 lines), and the first to eighth pieces of order information R are listed. Depending on the purchaser, the case order form may also have "remarks" and "shipping notes" fields.

When "Case Order" is selected with the type selection button 61a, the case order input screen 62 is displayed. On the case order input screen 62, multiple input areas 63a to 63f corresponding to each item 51a to 51f included in the order information R are displayed. The number of input areas is set in the format data 43 (see FIG. 1) for each purchaser according to the number of items on the case order. The multiple input areas 63a to 63f are displayed aligned vertically on the input screen 62. After the order is registered as the order registration data 41a, the input screen 62 also functions as a display screen for viewing and confirmation by displaying the registered order contents. The input screen 62 shows one order content (order registration data 41a) per page, and includes as many pages as the number of order information R listed on the order form. The top right corner of the input screen 62 displays which line (which item) of the order content is on the order form (in FIG. 2, "first line" is displayed).

The order registration data 41a (see FIG. 1) is registration information indicating the order contents of the cardboard box that is registered based on the order information R described in the order form image data 50. Specifically, the order registration data 41a includes issuer information R1, order content data R2, and annotation information R3. The order content data R2 is the order information R described in the order form image data 50 that is entered by the registration process, and matches the order information R. The annotation information R3 is annotation (comment) data entered by the person in charge of registration for the order contents. In addition, the order registration data 41a includes information Aa about the person in charge of registration, and an order number Ab assigned to each order content data R2.

In addition, the input screen 62 displays in advance information Aa about the person in charge of registration based on the login process. Similarly, the input screen 62 displays issuer information R1 (customer number R11). If the order form is received by fax, issuer information R1 (customer number R11) is identified based on the sender fax number or fax header information and is automatically entered. If the issuer cannot be identified from the sender fax number, etc., or if the printed order form is imaged by the image reading unit 5, issuer information R1 (customer number R11) can be entered by the person in charge (registration person) manually entering it, by the person in charge selecting it from a customer database (a database of customers stored in the management server 4), or by reading it from the image data 50 using character recognition processing.

<Example of sheet order form>
Fig. 3 shows an example of a sheet order form for a sheet-like cardboard product. In the image data 50 of the order form shown in Fig. 3, the words "sheet order form" are located at the top of the page and in the center in the left-right direction of the page. In the image data 50 of the order form, issuer information R1 ("XXX Cardboard Co.") and order information R are located. A sheet-like cardboard product is cardboard in a sheet state before folding lines are formed and cutting is performed.

In the example of FIG. 15, the order information R includes the following items: "Order number" item 52a, "Product number" item 52b, "Flute" item 52c, "Paper quality" item 52d, "Dimensions (width x flow)" item 52e, "Quantity (number of sheets)" item 52f, "Line" item 52g, "Delivery date" item 52h, and "(Delivery) location" item 52i. The "Order number", "Product number", "(Delivery) location", and "Delivery date" items are the same as those of the case order form. "Flute" in item 52c indicates the specifications of the flutes formed in the cardboard, such as flute A, flute B, and flute AB, and "Paper quality" in item 52d indicates the type of cardboard base paper used. "Dimensions" in item 52e are the dimensions of one sheet of cardboard. "Line" in item 52g indicates the position of the fold line for box making. Although not shown in Figure 3, other items on the sheet order form may include, for example, the number of sheets (the number that can be punched out from one cardboard sheet), remarks, shipping notes, and purchasing notes.

When "Sheet Order" is selected with the type selection button 61a, the sheet order form input screen 62 is displayed. The sheet order form input screen 62 displays multiple input areas 64a-64i corresponding to each of the items 52a-52i included in the order information R. As mentioned above, the number of input areas is set in the format data 43 for each purchaser according to the number of items on the sheet order form. Other details are basically the same as for case orders.

For the sake of convenience, the following explanation will focus on the registration process for image data 50 of the case order form shown in FIG. 2, and the explanation of the sheet order form shown in FIG. 3 will be omitted since it is generally similar.

(Format data)
The character recognition process is performed using format data 43 corresponding to the voucher (order form) shown in the image data 50. Fig. 3 is a schematic diagram showing the concept of the data structure of the format data 43 stored in the storage unit 4a of the management server 4.

<Type of format data>
As shown in Fig. 3, the storage unit 4a prestores voucher (order form) format data 43 for each voucher issuer. The issuer of the voucher is, for example, a business entity (corporation or individual) that is a trading partner, and is specified by issuer information R1 (see Fig. 2). Since the order form format may differ for each ordering person even for the same business entity, multiple format data 43 can be registered for the same business entity.

The memory unit 4a also stores format data 43 for each type information CL corresponding to the cardboard type information CL, which includes cardboard sheets and cardboard cases. The format data 43 is created separately for each order type (type information CL) and stored in the memory unit 4a. In this embodiment, the type information CL includes two types for orders of cardboard: cardboard cases (case orders) and cardboard sheets (sheet orders). Furthermore, the type information CL can include various types such as delivery notes, invoices, and receipts for orders of cardboard materials (such as base paper).

In this way, the memory unit 4a registers format data 43 for each issuer in a number corresponding to the number of types of type information CL. In the example of Figure 4, two format data 43 are registered for issuer "Company A", "Format: 1" corresponding to the order type of cardboard cases and "Format: 2" corresponding to the order type of cardboard sheets. If the order form format differs for each ordering clerk, multiple format data 43 such as Format: 3 and Format: 4 may be registered for the same type (for example, cardboard cases). Similarly, format data 43 is registered in a number corresponding to the type information CL for issuers "Company B" and "Company C".

<Contents of format data>
Next, the contents and setting method of the format data 43 will be described. Fig. 5 shows a setting registration screen 65 that is displayed by inputting the setting button 61c (see Fig. 2). The setting registration screen 65 is a screen for setting the format data 43 for the image data 50 displayed on the image display screen 61. Below, Figs. 4 and 5 will be referred to.

As shown in FIG. 4, the format data 43 includes row information 71, column information 72, and reference line information 73. Note that the reference line information 73 is an example of "reference line information included in the evidence" in the claims.

<Line information>
The line information 71 is format information for each line of the voucher. The line information 71 includes a maximum number of lines 71a of the order form. The maximum number of lines 71a corresponds to the upper limit of the number of orders (i.e., the number of lines) that can be written on one order form. The maximum number of lines 71a is set by inputting a numerical value in the number of lines input field 65a of the setting and registration screen 65.

Line information 71 includes top and bottom end information 71b of one of the lines, and line spacing information 71c. Top and bottom end information 71b and line spacing information 71c are registered by the person in charge of registration specifying a position on image data 50 on setting registration screen 65. Top and bottom end information 71b is an example of "information on the top and bottom end positions of a line" in the claims.

Specifically, the information processing unit 33 (see FIG. 1) displays the image data 50 on the setting registration screen 65 of FIG. 5, and accepts, via the input unit 32 (see FIG. 1), a specification (click) of the position of the top end position P1 (top left corner) of any line, and a specification (click) of the position of the bottom end position P2 (bottom left corner). The information processing unit 33 acquires the position coordinates of the specified top end position P1 and bottom end position P2 as top and bottom end information 71b.

The top and bottom positions P1 and P2 of a line are preferably set as the top and bottom positions of the character writing area in that line. In other words, the top position P1 is specified as the position between the top of the characters in the line and the horizontal line 55 on the upper side of the line. The bottom position P2 is specified as the position between the bottom of the characters in the line and the horizontal line 55 on the lower side of the line. By setting position coordinates inside the horizontal line 55 of the line as the top and bottom information 71b, it is possible to prevent the horizontal line 55 from being included in the target area for character recognition processing, and therefore prevent the horizontal line 55 from being erroneously recognized as part of a character.

The information processing unit 33 (see FIG. 1) also acquires the distance between vertically adjacent horizontal ruled lines 55 in the image data 50 as line spacing information 71c. When setting the reference line information 73, the position coordinates of a reference line 83 (described later) are specified, as shown in FIG. 7. The information processing unit 33 then accepts input of position P3 of the horizontal ruled line 55 adjacent to the lower side of the reference line 83. The information processing unit 33 (see FIG. 1) acquires the vertical distance between the reference line 83 and position P3 (the distance between the start position P6 of the reference line 83 and position P3) as line spacing information 71c.

As a result, the information processing unit 33 displays the first frame line 81, with the line height specified by the top and bottom edge information 71b, superimposed on the image data 50 for the maximum number of lines 71a, at the interval specified by the line spacing information 71c. The position of the right end of each line is automatically set based on the position coordinates of the right end position P5 of the rightmost column in the column information 72. The vertical ruled line 56 at the right end of the image data 50 may also be detected.

The first frame lines 81 displayed in this manner indicate each line of the voucher (order form). Therefore, when the line information 71 is set appropriately, the area in which the character strings of each line in the image data 50 are written falls within the range of each of the first frame lines 81 that are superimposed.

<Column information>
4 is information about items included in one row of the voucher (i.e., one piece of order information R). The column information 72 includes information about the position and range of each item included in one row. The column information 72 is made up of a collection of individual item data Ka to Kf, the number of which is the same as the number of items in the voucher.

Specifically, the column information 72 (individual item data Ka to Kf) first includes item name information 72a for each item. Item name information 72a is the name of the item written on the voucher (order form), such as the order number, product number, product name, quantity, delivery date, (delivery) location, remarks, and special shipping notes. The items to be registered in the management server 4 are predetermined, so the person in charge of registration selects or inputs which item from the registered items to set.

Column information 72 includes left and right edge information 72b, which is information on the left and right edge positions of each item. Left and right edge information 72b is registered by the person in charge of registration specifying (clicking) a position on image data 50 on setting registration screen 65. Left and right edge information 72b is an example of "information on the left and right edge positions of each item" in the claims.

Specifically, the information processing unit 33 (see FIG. 1) accepts, via the input unit 32 (see FIG. 1), a specification (click) of the left end position P4 (top left corner) and a specification (click) of the right end position P5 (top right corner) of each item in the image data 50 displayed on the setting registration screen 65.

Here, it is preferable that the left end position P4 and the right end position P5 of an item (column) are set as the left end position and the right end position of the character writing area within that item (column). In other words, the left end position P4 is specified as the position between the left end of the character in the item and the vertical line 56 on the left side of the item. The right end position P5 is specified as the position between the right end of the character in the item and the vertical line 56 on the right side of the item. By setting position coordinates inside the vertical lines 56 on both sides of the item as the left/right end information 72b, it is possible to prevent the vertical lines 56 from being included in the target area for character recognition processing, and therefore prevent the vertical lines 56 from being erroneously recognized as part of a character.

In the example of FIG. 5, six items are listed, so the information processing unit 33 acquires the position coordinates (12 locations) of the left end position P4 and the right end position P5 for each of the six items (individual item data Ka to Kf) as left and right end information 72b.

As a result, the information processing unit 33 displays the second frame line 82, whose position and item width (column width) are specified by the left and right edge information 72b for each item, superimposed on the image data 50 for each item. Note that the position of the bottom edge of each item is automatically set based on the position of the last row (the 10th row in the case of FIG. 5) specified by the row information 71.

The second frame lines 82 displayed in this manner indicate each column of the evidence. Therefore, when the column information 72 is set appropriately, the area in which the character strings of each column (each item) in the image data 50 are written falls within the range of each superimposed second frame line 82.

With this configuration, the information processing unit 33 is configured to obtain the region of the character string for each item included in the image data 50 as the region where the mutually intersecting first frame line 81 and second frame line 82 overlap in the character recognition process, and to vary the parameters of the character recognition process for each item, as described below. Note that FIG. 2 illustrates a state in which the first frame line 81 and the second frame line 82 are superimposed, and FIG. 3 illustrates the state before they are superimposed.

As shown in FIG. 4, column information 72 (individual item data Ka to Kf) includes, for each item, information 72c on the character type used for that item. Since the character type used for each item of the voucher is determined according to the content of the item, the character type information 72c used for that item is set in advance as format data 43. Character type information 72c can be defined, for example, by three types of characters: numbers, English letters, kana (hiragana and katakana), and kanji, and combinations of these. The information processing unit 33 accepts input of character type information 72c for each item from the registration clerk via the input unit 32 on a setting screen not shown, and stores the accepted character type information 72c in the memory unit 4a in association with the item name.

As a result, when performing character recognition processing, the information processing unit 33 is configured to recognize, for each item, the character string for that item within the range of character types specified by the character type information 72c. By narrowing down the character types when performing character recognition processing, it is possible to improve the recognition accuracy.

In addition, in this embodiment, the column information 72 includes character spacing information 72d for each item. The character spacing (character spacing) of the character string written on the evidence is generally constant for each item. If the character spacing is too narrow, adjacent characters will be difficult to recognize as independent characters. For this reason, character spacing information 72d for each item is set in advance as format data 43. The information processing unit 33 accepts input of character spacing information 72d for each item from the registration clerk via the input unit 32 (see FIG. 1) on a setting screen not shown, and stores the accepted character type information 72c in the memory unit 4a in association with the item name.

Thus, the information processing unit 33 is configured to execute a process of adjusting the character spacing in the image data 50 based on the character spacing information 72d for each item of the image data 50 during character recognition processing. For example, as shown in FIG. 6A, assume that the character spacing t1 of a character string of a certain item ("XX Factory" in the example of FIG. 6) is too small before adjusting the character spacing. In this case, the registration staff sets the character spacing information 72d to increase the character spacing. As a preprocessing prior to the character recognition processing, the information processing unit 33 adjusts the character spacing in the image data 50 by image processing according to the setting value of the character spacing information 72d. FIG. 6B shows the character string after the character spacing has been adjusted to t2. There is no particular limit to the way in which the character spacing information 72d is set. The character spacing information 72d may be set, for example, to the minimum value of the character spacing amount, or may be set to a magnification factor for the character spacing amount before adjustment. If the character spacing in the image data 50 is too large, the character spacing is reduced based on the character spacing information 72d during preprocessing.

As shown in FIG. 4, the above item name information 72a, left/right margin information 72b, character type information 72c, and character spacing information 72d are included in one piece of individual item data Ka-Kf. The individual item data Ka-Kf are set separately for the number of items included in the voucher (order form). Therefore, in the example of format data 43 created from image data 50 shown in FIG. 2, column information 72 includes a total of six individual item data Ka-Kf corresponding to the order number item 51a, product number item 51b, product name item 51c, quantity item 51d, delivery date item 51e, and (delivery) location item 51f.

In addition, the information processing unit 33 can perform binarization and smoothing (blurring) processing for each item as preprocessing for character recognition processing.

The binarization process is set to be performed when the printed evidence (order form) is imaged by the image reading unit 5. Therefore, the individual item data Ka to Kf may include the threshold value 72e for the binarization process.

Smoothing (blurring) processing is image processing that smooths out jagged parts of the outlines of characters in image data 50. Individual item data Ka to Kf may include a setting value for the strength of the smoothing processing (smoothing parameter 72f).

<Baseline Information>
Next, the reference line information 73 is information on a line (i.e., a reference line) that is a reference for the position and range of each character string in the entire evidence (the entire image data 50). The reference line information 73 includes a start position P6 and an end position P7 of the reference line 83 (see FIG. 7).

As shown in FIG. 7, the reference line 83 is one of the lines (ruled lines) included in the image data 50 that extends in the image width direction. A line having a length of 60% or more (preferably 70% or more) of the paper width of the voucher is used as the reference line 83. Typically, the reference line 83 is the longest line in the image data 50 that extends in the image width direction, and corresponds to the horizontal ruled line 55 (line surrounded by a dotted line) that constitutes the upper part (upper side) of the frame line that surrounds each piece of order information R shown on the order form.

The information processing unit 33 (see FIG. 1) accepts the specification (click) of the start position P6 of the reference line 83 and the specification (click) of the end position P7 via the input unit 32 (see FIG. 1) on the setting registration screen 65. The information processing unit 33 acquires the position coordinates of the specified start position P6 and end position P7 as reference line information 73, and stores them in the memory unit 4a.

The information processing unit 33 is configured to perform a process of correcting the magnification ratio and rotation angle of the image data 50 so that the reference line 83 in the image data 50 coincides with the reference line 83 set in the format data 43 during the character recognition process. For example, assume that FIG. 8(A) is the image data 50a on which the format data 43 is based, and FIG. 8(B) is the newly acquired image data 50b. The image data 50b in FIG. 8(B) shows the order form smaller than the image data 50a in FIG. 8(A) and rotated within the plane. Therefore, as a preprocessing step for the character recognition process, the information processing unit 33 performs magnification/reduction and rotation of the image data 50b so that the start point and end point of the reference line 83b shown in the image data 50b coincide with the start point position P6 and end point position P7 of the reference line 83a defined in the format data 43 (i.e., the length L and angle θ of the reference line 83b of the image data 50b are matched). As a result, when character recognition processing is performed, the size, rotation angle, and arrangement of each row and column of the voucher (order form) shown in image data 50b match those of image data 50a, which is the basis of format data 43.

As shown in FIG. 7, the reference line information 73 may further include information on a second reference line 84 in addition to the first reference line 83. During character recognition processing, the information processing unit 33 may perform keystone correction of the image data 50 so that both the first reference line 83 and the second reference line 84 in the image data 50 coincide with the first reference line 83 and the second reference line 84 set in the format data 43, as shown in FIG. 8(C).

The second reference line 84 is a line having a length of 60% or more (preferably 70% or more) of the paper width of the voucher, but is selected to be located as far away as possible above or below the first reference line 83. Typically, the second reference line 84 corresponds to the line that constitutes the lower part (lower side) of the frame that surrounds each piece of order information R shown on the order form.

Keystone correction is particularly useful when imaging a voucher (order form) using an overhead type image reading unit 5. The overhead type is a method in which, as shown in FIG. 1, a voucher 6 (order form) placed on a surface is imaged non-contact by a reading unit (camera) located above the surface. In an overhead type image reading unit 5, the paper is placed at an angle relative to the reading unit, so the image is prone to be distorted into a trapezoid. Therefore, when both the first reference line 83 and the second reference line 84 are used, it is possible to perform accurate keystone correction (correction to make the lengths L1 and L2 match the settings of the format data 43) on the image data 50c distorted into a trapezoid as shown in FIG. 8(C). Since the distortion of the character string is corrected, it is possible to effectively improve the accuracy of character recognition.

The procedure for setting the first frame line 81 and the second frame line 82 on the setting registration screen 65 will now be described. The person in charge of registration first sets the reference line information 73 (start position P6 and end position P7 in FIG. 7). Next, the person in charge of registration sets the line spacing information 71c by specifying the position P3 (see FIG. 7) of the horizontal ruled line 55 adjacent to and below the reference line 83. In addition, the maximum number of lines 71a is set in the number of lines input field 65a in FIG. 5.

Next, the left and right edge information 72b (each of the left edge positions P4 and right edge positions P5) for each item in FIG. 5 is set. As a result, the second frame line 82 for each item is displayed superimposed on the position set by the left and right edge information 72b so as to extend to the bottom edge position of the item calculated from the maximum number of lines 71a and line spacing information 71c.

Next, the registration clerk sets the top and bottom end information 71b (top end position P1 and bottom end position P2) for one of the lines. This causes a first frame line 81, which has the height set by the top and bottom end information 71b and extends in the row direction to the right end position P5 of the rightmost line, to be superimposed and displayed so that it is lined up for the maximum number of lines 71a at a distance defined by the line spacing information 71c. As a result, the format data 43 is set so that each character string of the order information R written in the image data 50 fits within the area where the intersecting first frame line 81 and second frame line 82 overlap.

<Other information on the form data>
The format data 43 shown in Fig. 4 is mainly composed of the above-mentioned information. The format data 43 also includes orientation setting information 74 indicating whether the voucher (image data 50) is in portrait or landscape orientation. The information processing unit 33 automatically changes the orientation (rotation angle) of the image data 50 based on the format data 43 so that the orientation of the voucher is appropriate.

The information processing unit 33 is also configured to perform a line removal process to erase lines (horizontal and vertical lines that divide table cells) contained in the image data 50 as a pre-processing step for the character recognition process. This is because, in the character recognition process, if a line is close to a character, the line may be mistakenly recognized as part of the character. Since the thickness of the lines used in the evidence varies for each purchaser, the format data 43 includes a thickness setting value 75 for the line to be removed as a parameter for the line removal process. The person in charge of registration adjusts the thickness setting value 75 on a setting screen (not shown) to set an appropriate value that removes only the lines without removing the characters.

(Registration process)
Next, the registration process will be described in detail.

The information processing unit 33 is configured to select format data 43 from the memory unit 4a according to the type information CL of the voucher. That is, during the registration process, when image data 50 is displayed on the image display screen 61 shown in FIG. 2, the issuer information R1 of the order form shown in the image data 50 is identified. Then, the type of order (case order, sheet order) is selected with the type selection button 61a, and the value of the type information CL (for example, cardboard case) is identified.

If there are multiple format data 43 that correspond to the identified issuer and type, the information processing unit 33 detects the reference line 83 in the image data 50, and automatically selects the format data 43 in which the reference line information 73 that is most similar to the start position P6 and end position P7 of the detected reference line 83 is set as the format data 43 that corresponds to the image data 50. The information processing unit 33 displays the automatically selected format data 43 in the format selection button 61b (in FIG. 2, "Format: 1" is displayed). In case the automatically selected format data 43 is not appropriate, the information processing unit 33 can accept the selection of format data 43 from the format selection button 61b in the form of a pull-down list.

When the format data 43 is selected, the information processing unit 33 arranges the image data 50 in portrait or landscape orientation on the image display screen 61 according to the orientation setting information 74, and adjusts the magnification rate and rotation angle of the image data 50 based on the reference line information 73. This causes the positions and ranges of each row and column displayed in the image data 50 to match the positions and ranges of the first frame line 81 and second frame line 82 defined in the format data 43.

The information processing unit 33 performs a line removal process on the image data 50 based on the format data 43 (line thickness setting value 75). This removes the lines that separate each row and each column (item) from the image data 50.

Then, the information processing unit 33 displays a first frame line 81 indicating each line of the voucher superimposed on the image data 50 displayed on the image display screen 61 based on the top and bottom end information 71b, the line spacing (line spacing information 71c), and the maximum number of lines 71a, and displays a second frame line 82 indicating each item of the voucher superimposed on the image data 50 displayed on the image display screen 61 based on the left and right end information for each item 72b.

As a result, the ruled lines are deleted from the image data 50, but the first frame lines 81 and the second frame lines 82 are superimposed and displayed on the image display screen 61 in place of the ruled lines. If appropriate format data 43 has been selected, the displayed first frame lines 81 and second frame lines 82 will approximately match the ruled lines (the removed ruled lines) that were written on the image data 50. The person in charge of registration will verify that appropriate format data 43 has been selected based on the positional relationship between the first frame lines 81 and second frame lines 82 and the character strings on the image display screen 61.

After each of the above processes, the information processing unit 33 is configured to sequentially execute a process of performing character recognition on a line-by-line basis for the image data 50 displayed on the image display screen 61 and inputting the recognition character string 85 for each item contained in the recognized line into each item on the input screen 62 according to the maximum number of lines 71a.

That is, as shown in FIG. 9, the information processing unit 33 is configured to regard the area within the first frame line 81 displayed superimposed on the image data 50 as one line, and to acquire the character string contained within the second frame line 82 as the recognition character string 85 for each item. As shown by the circled numbers in FIG. 9, each of the first frame lines 81 is the first line to the Mth line (M is the setting value of the maximum number of lines 71a) from the top to the bottom of the image data 50.

The information processing unit 33 first performs character recognition processing on the first frame line 81 located on the first row. Specifically, the information processing unit 33 performs character recognition processing on the individual areas 86a to 86f within the first frame line 81 based on the individual item data Ka to Kf included in the column information 72 of the format data 43.

That is, the information processing unit 33 reads the individual item data Ka for the item 51a of the order number. Based on the left and right edge information 72b of the individual item data Ka, the information processing unit 33 identifies the individual area 86a in the first frame line 81 that corresponds to the item 51a. Based on the character spacing information 72d of the individual item data Ka, the information processing unit 33 adjusts the character spacing in the individual area 86a. The information processing unit 33 converts the image of the characters in the individual area 86a into character data by character recognition processing based on the character type information 72c of the individual item data Ka. As a result, the information processing unit 33 obtains the recognized character string 85 that corresponds to the item 51a in the first row (first item).

In the same manner, the information processing unit 33 obtains the recognition character string 85 in the individual area 86b by character recognition processing based on the individual item data Kb in the product number item 51b. The information processing unit 33 obtains the recognition character string 85 in the individual area 86c by character recognition processing based on the individual item data Kc in the product number item 51c. The information processing unit 33 obtains the recognition character string 85 in the individual area 86d by character recognition processing based on the individual item data Kd in the product number item 51d. The information processing unit 33 obtains the recognition character string 85 in the individual area 86e by character recognition processing based on the individual item data Ke in the product number item 51e. The information processing unit 33 obtains the recognition character string 85 in the individual area 86f by character recognition processing based on the individual item data Kf in the product number item 51f.

As a result, in the example of FIG. 9, the recognition character strings 85 for each of the items 51a to 51f are obtained for the first line (first item) of order information R. The information processing unit 33 then executes a process of inputting the obtained recognition character strings 85 for each item on the input screen 62.

As a result, "201703", which is the recognition character string 85 of individual area 86a, is entered into input area 63a for the order number corresponding to item 51a, "101", which is the recognition character string 85 of individual area 86b, is entered into input area 63b for the product number corresponding to item 51b, "New Product Tato Type", which is the recognition character string 85 of individual area 86c, is entered into input area 63c for the product name corresponding to item 51c, "100", which is the recognition character string 85 of individual area 86d, is entered into input area 63d for the quantity corresponding to item 51d, "December 16th, 8:00", which is the recognition character string 85 of individual area 86e, is entered into input area 63e for the delivery date corresponding to item 51e, and "XX Factory", which is the recognition character string 85 of individual area 86f, is entered into input area 63f for the delivery location corresponding to item 51f.

In this way, the first line (first item) of order information R is input. After the information is input, if there are no errors (abnormal values) in the information input in each of the input areas 63a to 63f, the information processing unit 33 automatically completes the registration. That is, the information processing unit 33 transmits the first line (first item) of order information R to the management server 4. The management server 4 assigns and records an order number Ab (see FIG. 1) to the received order information R, and transmits the assigned order number Ab to the information processing unit 33. When the information processing unit 33 receives the order number Ab, it inputs the order number Ab into the input area 62d of the input screen 62. Furthermore, the information processing unit 33 displays the order number Ab (see FIG. 10) at the position on the image display screen 61 corresponding to the corresponding order information R, and notifies the person in charge of registration that the registration process for the order information R has been completed.

After displaying the order number Ab, the information processing unit 33 updates the input screen 62 to the next page and proceeds to the registration process for the next row (row 2). The information processing unit 33 performs the registration process for rows 2 to M in the same manner as for row 1.

Error Handling
During the registration process, errors may occur, for example, when there is a clerical error on the voucher (order form) or when a misrecognition occurs during character recognition processing.

When an abnormal value (error) is included in the input value on the input screen 62, the information processing unit 33 is configured to temporarily suspend character recognition processing for the subsequent lines and to execute processing to notify the input screen 62 that an abnormal value is included.

Figure 10 illustrates a case in which an abnormal value is included in the quantity item 51d during character recognition processing of the third line of order information R. That is, the correct numerical value for the quantity item 51d on the third line of the image data 50 is "300," but "3" has been entered in the input area 63d of the input screen 62. The quantity input area 63d is set with lower and upper limits for order acceptance, and the input value "3" is an abnormal value that falls below the lower limit.

In this case, the information processing unit 33 pauses the character recognition process at the third line containing the abnormal value, and notifies the user on the input screen 62. Specifically, the process of notifying the user that an abnormal value is included includes a process of displaying an indicator 87 for identifying the item containing the abnormal value on the input screen 62, and a process of moving the cursor 88 to the input area of the item containing the abnormal value.

10, the information processing unit 33 displays the sign 87 at a position corresponding to the input area 63d containing the abnormal value so that the sign 87 is aligned side by side. The sign 87 is displayed in a different display mode from other input information. In the example of FIG. 10, each piece of information on the input screen 62 is displayed in a preset basic color (e.g., black), while the sign 87 is displayed in a specific color (e.g., red, yellow, etc.) different from the basic color. Also, in FIG. 10, the sign 87 includes text explaining the contents of the abnormality. That is, since the abnormality is in the quantity item, the sign 87 displays the text "Quantity Error" in white characters (i.e., characters with the text part filled in white on a specific color background). Note that the text explaining the contents of the abnormality differs depending on the item (input area). For example, if an order number that has already been received is entered in the order number input area 63a, "Order Registered" is displayed. If an unregistered number is entered in the product number input area 63b, "Product Error" is displayed. If a past date or the like is entered in the delivery date input area 63e, "Delivery date error" is displayed. If an unregistered location is entered in the delivery location input area 63f, "Location error" will be displayed.

In addition, in FIG. 10, the information processing unit 33 highlights the input area 63d itself by displaying the background of the input area 63d containing the abnormal value in a specific color. Furthermore, the information processing unit 33 displays a specific color indicator 89 on the image display screen 61 at the position (right end of the row) where the order number Ab of the row containing the abnormal value (here, the third row) is displayed.

Then, the information processing unit 33 moves the cursor 88 to the input area 63d that contains the abnormal value, and accepts correction operations by the person in charge of registration via the input unit 32.

At this time, the information processing unit 33 performs processing to pop up a display field (text box) 90 that displays the input value into the input area 63d in a position adjacent to the top, bottom, left or right of the individual area containing the abnormal value (individual area 86d in the third row in this case) on the image display screen 61. In FIG. 10, the display field 90 is displayed in a position adjacent to the bottom of the individual area 86d, and since "3" has been entered in the input area 63d, "3" is also displayed in the display field 90. When the person in charge of registration corrects the input value in the input area 63d, the information processing unit 33 also reflects the corrected value in the text in the display field 90.

As a result, when the registration clerk refers to the image data 50 to correct the input values, he or she can easily and reliably determine whether the contents of the individual area 86d on the order form match the values entered in the input area 63d, making the correction process easier and reducing correction errors.

If the error is a clerical error on the voucher (order form), the person in charge of registration enters notes in the input area for annotation information R3 to be used by the customer to confirm the abnormal values. After correction, when the user clicks the registration button 62a, the information processing unit 33 sends the corrected input to the management server 4.

Substitution Table
For example, for items that include kanji or kana, such as product name item 51c (input area 63c), recognition accuracy is lower than for numbers and alphabets, and it is difficult to uniformly distinguish between normal and abnormal values.

As shown in FIG. 11, particularly for items that include at least one of kanji and kana, the information processing unit 33 is configured to record in the memory unit 4a a pre-correction string 85a and a post-correction string 85b in association with each other when it receives a correction input operation via the input unit 32 for a recognized string 85 entered in any item. The memory unit 4a of the management server 4 stores a replacement table 44 that associates the pre-correction string 85a with the post-correction string 85b. The information processing unit 33 is configured to perform a process of replacing the recognized string 85 obtained by character recognition with the corresponding post-correction string 85b when the recognized string 85 corresponds to the pre-correction string 85a recorded in the memory unit 4a.

For example, as shown in FIG. 11(A), in the initial character recognition process, the information processing unit 33 acquires a recognized character string 85 of "New Product Tatowa Style" and inputs the acquired recognized character string 85 into the input screen 62 (see "Input Character String"). In response, it is assumed that the information processing unit 33 receives an input from the person in charge of registration to amend the character string to "New Product Tato Style".

In this case, as shown in FIG. 11(B), the information processing unit 33 records the recognized character string 85, "New Product Tatowa Style," as a pre-correction character string 85a, and the corrected and input character string, "New Product Tato Style," as a post-correction character string 85b in the substitution table 44 in one-to-one correspondence.

After that, as shown in FIG. 11(C), when the information processing unit 33 performs character recognition processing, it checks whether the acquired recognized character string 85 matches any pre-correction character string 85a recorded in the replacement table 44. If there is a pre-correction character string 85a that matches the acquired recognized character string 85, the information processing unit 33 acquires a post-correction character string 85b that corresponds to that pre-correction character string 85a. The information processing unit 33 inputs the acquired post-correction character string 85b into the input screen 62 instead of the recognized character string 85. In this way, the information processing unit 33 learns the correction content of the recognized character string 85 and replaces the character string so that the same correction does not occur again.

(Flowchart of registration process)
Next, the flow of processing by the information processing unit 33 in the registration process will be described with reference to FIG. 12 and FIG.

In step S1 of FIG. 12, the information processing unit 33 accepts the selection of unprocessed image data 50 and the input of the enlarged display button 60a on the received fax list screen 60 shown in FIG. 13 (or the image management screen (not shown) read by the image reading unit 5), thereby displaying the image display screen 61 and the input screen 62. The selected image data 50 is displayed on the image display screen 61. At this time, the information processing unit 33 obtains the issuer information R1 of the voucher (order form) from the sender information of the fax. When an image read by the image reading unit 5 is selected, the information processing unit 33 obtains the issuer of the voucher (order form) by the selection input of the registration staff. Note that for each part of the cardboard voucher management system 1, refer to FIG. 1. For the image display screen 61 and the input screen 62, refer to FIG. 2. For the format data 43, refer to FIG. 4.

In step S2 of FIG. 13, the information processing unit 33 displays the image display screen 61 and the input screen 62 side by side on the display unit 31.

In step S3, the information processing unit 33 selects type information CL by accepting input from the type selection button 61a.

In step S4, the information processing unit 33 selects format data 43. As described above, the information processing unit 33 selects format data 43 that corresponds to the identified issuer information R1 and type information CL. If there are multiple corresponding format data 43, the information processing unit 33 detects the reference line 83 appearing in the image data 50, automatically selects format data 43 in which reference line information 73 that is similar to the detected reference line 83 is set, and displays it in the format selection button 61b.

In step S5, the information processing unit 33 updates the input screen 62 so that an input area corresponding to the format data 43 is displayed on the input screen 62.

In step S6, the information processing unit 33 performs preprocessing of the image data 50 based on the selected format data 43. That is, the information processing unit 33 adjusts the orientation of the image data 50 to vertical or horizontal writing based on the orientation setting information 74, adjusts the magnification ratio and rotation angle based on the reference line information 73, and performs ruled line removal processing based on the thickness setting value 75.

In step S7, the information processing unit 33 superimposes the first frame line 81 and the second frame line 82 on the image data 50 based on the format data 43. The registration staff member checks the positions of the first frame line 81 and the second frame line 82 and confirms that the appropriate format data 43 has been selected. If the positions of the first frame line 81 and the second frame line 82 do not match the character string on the image data 50 and the format data 43 is inappropriate, the registration staff member reselects the format data 43 using the type selection button 61a.

In step S8, the information processing unit 33 accepts input from the OCR button 61d and executes OCR registration processing for the image data 50 displayed on the image display screen 61. When the OCR registration processing is completed, the registration processing for the image data 50 selected in step S1 is also completed.

(OCR registration process)
Next, the OCR registration process in step S8 in Fig. 12 will be described in detail with reference to Fig. 14. For the image display screen 61 and the input screen 62, refer to Fig. 9 and Fig. 10. For the format data 43, refer to Fig. 4.

In step S11, the information processing unit 33 starts the registration process for the Nth row of the image data 50. Here, the initial value of N is 1. Therefore, the first frame line 81 of the first row is selected.

In step S12, the information processing unit 33 performs character recognition processing for each item on the first frame line 81 of the Nth row. As described above, the information processing unit 33 performs character recognition processing on the individual areas 86a to 86f within the first frame line 81 based on the corresponding individual item data Ka to Kf. As a result of the character recognition processing, a recognized character string 85 is obtained for each of the individual areas 86a to 86f.

In step S13, the information processing unit 33 inputs each recognition character string 85 acquired for each of the individual areas 86a to 86f into the corresponding input areas 63a to 63f on the input screen 62. At this time, the information processing unit 33 compares each recognition character string 85 with the replacement table 44, and if there is a pre-correction character string 85a that matches the recognition character string 85, it replaces it with the corresponding post-correction character string 85b and inputs it.

In step S14, the information processing unit 33 determines whether the order is based on the product name in item 51c. If the product number in item 51b has not been entered in the input area 63b, the information processing unit 33 determines that the order is based on the product name. On the other hand, if the product number in item 51b has been entered in the input area 63b, the information processing unit 33 determines that the order is not based on the product name, and proceeds to step S15.

<Order by product number>
In step S15, the information processing unit 33 judges whether there is an error. That is, it judges whether any abnormal value is included in any of the input areas 63a to 63f of the input screen 62. If it is judged that no abnormal value is included and there is no error, the information processing unit 33 proceeds to step S22 and automatically executes the registration of the order detail data R2.

If any of the input areas 63a to 63f contains an abnormal value (see FIG. 10), the information processing unit 33 determines that there is an error and proceeds to step S16. In step S16, the information processing unit 33 performs the above-mentioned notification process, such as displaying a sign 87 in the input area of the input screen 62 that contains the abnormal value, and moves the cursor 88 to the corresponding input area. Then, in step S17, the information processing unit 33 accepts correction of the input value by the person in charge of registration via the input unit 32. For correction of a character string including kana or kanji, the information processing unit 33 records the pre-correction character string 85a and the corrected character string 85b in the replacement table 44. Although not shown in the flowchart, if the skip button 62b (see FIG. 10) on the input screen 62 is pressed, the registration of the order content data R2 can be skipped and the error can be left as it is and the process can proceed to the next step. After that, when the information processing unit 33 accepts an input to the registration button 62a on the input screen 62 in step S18, the process proceeds to step S22. The registration process will be paused until the skip button 62b or the registration button 62a is pressed.

<Orders by product name>
On the other hand, if it is determined in step S14 that the order is based on the product name, the information processing unit 33 proceeds to step S19. Since the product name contains both kana and kanji, the recognition accuracy is lower than that of the product number. Therefore, in step S19, the information processing unit 33 searches the product database (not shown) of the management server 4 for product names that completely or partially match the character string entered in the input area 63c, and displays a list of the corresponding product names as candidates. The product database records product names and product numbers in association with each other. In step S20, the information processing unit 33 accepts the selection of a product name by the registration staff from the candidates displayed in the list via the input unit 32. In step S21, the information processing unit 33 inputs the selected product name and the product number corresponding to the product name into the input areas 63c and 63b of the input screen 62, respectively, and proceeds to step S22.

<Processing after input>
In step S22, the information processing unit 33 registers the order details data R2. That is, the information processing unit 33 transmits the input contents (order details data R2) of the input screen 62 to the management server 4, and receives the order number Ab from the management server 4. The information processing unit 33 displays the received order number Ab (see FIG. 10) by superimposing it on a predetermined position (the right end of the line) of the corresponding line of the image data 50. Note that the order number Ab and the order details data R2 are associated with each other and registered as the order registration data 41a in the management server 4. At this time, an unchecked flag is recorded in the order registration data 41a.

In step S23, the information processing unit 33 determines whether the registration process has been completed for all lines of the image data 50. If the registration process has not been completed for each line up to the maximum number of lines 71a, the information processing unit 33 returns the process to step S11. At this time, the information processing unit 33 adds (increments) "1" to the value of the line number N that is the target of the registration process. As a result, the information processing unit 33 proceeds with the registration process for the second line onward.

When the process from step S11 to step S23 is repeated and the registration process for all lines is completed, the image data 50 on the image display screen 61 will display the same number of order numbers Ab as the number of orders (number of lines) in a predetermined location (the right end of the lines) in association with the order information R. The information processing unit 33 determines in step S23 that the registration process for all lines has been completed, and ends the registration process for the image data 50.

(Verification process)
The work subsequent to the registration process will be briefly described. After the registration process is performed, a second registration clerk (checking clerk) different from the first registration clerk who performed the registration process performs the task of verifying the registration contents. The second registration clerk accesses the management server 4 from the information processing terminal 3, and selects image data 50 including the order registration data 41a to which an unchecked flag has been added. The information processing section 33 displays the selected image data 50 on the image display screen 61, and displays the input screen 62 in which the order content data R2 has already been input in the "correction mode" shown in FIG. 15. An OK button 62c is displayed on the input screen 62.

The information processing unit 33 displays the first page of the input screen 62 on which the first order detail data R2 (first line) has been entered, and accepts input of the OK button 62c for each of the input areas 63a to 63f. At this time, the information processing unit 33 moves the cursor to the input area 63a, and displays a display field 90 on the image display screen 61 near the individual region 86a corresponding to the input area 63a where the cursor is located, and also displays the input value entered in the input area 63a in the display field 90.

The person checking confirms that the character string written in the individual area 86a (the character string in the image data 50) matches the character string displayed in the display field 90, and inputs (clicks) the OK button 62c. If the character strings do not match, the person checking inputs a correction. When the input of the OK button 62c is received, the information processing unit 33 performs the same process for the next input area 63b onwards. This verifies that the order information R in the image data 50 matches the order content data R2 on the input screen 62. When this process has been performed for all items of all order information R in the image data 50, the verification process is completed.

When the verification work is completed, the information processing unit 33 controls the rewriting of the unchecked flag in the order registration data 41a in the registration database 41 of the management server 4 to a check completed flag. When the verification process is completed and the check completed flag is recorded in the order registration data 41a, the management server 4 receives a fax transmission instruction from the information processing terminal 3 and transmits an order confirmation image 70 (see FIG. 16) to the customer (in the case of a fax order).

The order confirmation image 70 is an image of an order confirmation created based on the image data 50 of the order form. The order confirmation image 70 shown in FIG. 16 is an image in which a second image layer containing the order registration data 41a is superimposed on a first image layer containing the image data 50 of the order form. As shown in FIG. 16, the second image layer displays the person in charge information Aa (person in charge name), customer number R11, and order number Ab. The multiple order numbers Ab are arranged in positions adjacent to (in the same row as) the order information R to which each order number Ab is assigned. Furthermore, the second image layer is configured to write receipt information 53a indicating that the order has been received as text data in the second image layer. The receipt information 53a is, for example, the phrase "Thank you very much for your order." The information processing unit 33 is also configured to write image data 53b of the company seal "AAA Co., Ltd." in the second image layer. The area surrounded by a dotted line in FIG. 16 indicates the information written in the second image layer.

Although not shown in FIG. 16, if an error (abnormal value) occurs during the registration process and a note is entered in annotation information R3 to prompt the customer to confirm the abnormal value, this annotation information R3 is automatically entered in the second image layer.

In the order confirmation image 70, the first image layer and the second image layer can be displayed not only superimposed, but also individually using the viewer software. Even if the person in charge information Aa, customer number R11, and order number Ab are superimposed on the text of the order form in the order confirmation image 70 of FIG. 16, the second image layer can be temporarily hidden to check the text of the order form (original) in the first image layer.

The order confirmation image 70 is sent to the customer and is also stored in the image database 42 of the management server 4 (see FIG. 1). This allows the management server 4 to manage the image data 50 of the original voucher recorded in the first image layer and the order number Ab recorded in the second image layer in a state where they can be confirmed as a single file.

[Effects of this embodiment]
In this embodiment, the following effects can be obtained.

In this embodiment, as described above, the information processing unit 33 and the storage unit 4a in which the format data 43 of the voucher for each issuer of the voucher is pre-recorded are provided. The information processing unit 33 is configured to select the format data 43 from the storage unit 4a according to the issuer of the voucher, and execute character recognition processing of the image data 50 with parameters based on the selected format data 43 to obtain the recognition character string 85 for each item of the voucher, and execute processing to input the obtained recognition character string 85 for each item of the input screen 62 displayed on the display unit 31. As a result, the format data 43 according to the format of the voucher used by the issuer of the voucher (transaction partner) is set exclusively for each issuer, and the character recognition processing can be optimized by the format data 43 for each issuer. As a result, the character recognition accuracy can be improved even for vouchers in a format for each issuer of the voucher, without requiring the issuer of the voucher (transaction partner) to use a dedicated standard document or dedicated font. As a result, the character recognition accuracy can be improved in the registration processing of information from the image data 50 of the voucher, thereby improving processing efficiency. As a result, the burden of confirmation and correction work on the person in charge of registration can be reduced, and the time required for the information registration process can also be shortened. Note that in this embodiment, an example of the registration process for a purchase order has been specifically described, but the same can be applied to vouchers other than purchase orders, such as delivery notes, invoices, and receipts, and the same effects can be obtained.

In this embodiment, as described above, the format data 43 includes row information 71 including the maximum number of rows 71a of the voucher and column information 72 related to the items included in one row of the voucher, and the information processing unit 33 is configured to display the image display screen 61 displaying the image data 50 and the input screen 62 on the display unit 31, perform character recognition on a row-by-row basis for the image data 50 displayed on the image display screen 61, and input the recognized character string 85 of each item included in the recognized row for each item on the input screen 62 in sequence according to the maximum number of rows 71a. This allows character recognition to be performed on an order-by-order basis (one row), input processing to be performed for each item included in the order, and character recognition and information input for one order to be performed sequentially for each row. This allows the registration staff to check and correct the input information, and effectively suppresses erroneous input of order details.

In addition, in this embodiment, as described above, the information processing unit 33 is configured to suspend character recognition processing for subsequent lines when an abnormal value (error) is included in the input value on the input screen 62, and to execute processing to notify the input screen 62 that an abnormal value is included. As a result, when character recognition and input processing are executed on an order line (one item) basis, if an abnormal value (error) is found, processing is stopped at that line, and a notification is issued before processing proceeds to the next line. As a result, even when the image data 50 contains multiple orders, the person in charge of registration can reliably recognize which order the error occurred in and can make appropriate corrections or confirm with the customer. As a result, the possibility of erroneous registration of order data can be more reliably reduced.

In addition, in this embodiment, as described above, the process of notifying that an abnormal value is included includes a process of displaying a sign 87 on the input screen 62 to identify the item that contains the abnormal value, and a process of moving the cursor to the input area 63 of the item that contains the abnormal value. In this way, if an error (abnormal value) occurs during the character recognition process, the corresponding item on the input screen 62 can be recognized by the sign 87 by the registration clerk. Since the cursor automatically moves to that item, the operational burden on the registration clerk can be reduced. It is also possible to prevent human error in which the registration clerk mistakenly selects the input area 63 of an item that needs to be corrected. In addition, if annotation information R3 is entered to confirm with the customer about this abnormal value, annotation information R3 is automatically written in the order confirmation image 70, so that a clerical error on the voucher (an error on the customer's side) can also be appropriately dealt with.

In addition, in this embodiment, as described above, the information processing unit 33 displays the first frame line 81 indicating each line of the evidence on the image data 50 displayed on the image display screen 61 based on the top and bottom end information 71b, the line spacing, and the maximum number of lines 71a, and displays the second frame line 82 indicating each item of the evidence on the image data 50 displayed on the image display screen 61 based on the left and right end information 72b for each item, and acquires the character string contained in the second frame line 82 as the recognition character string 85 for each item, with the area within the first frame line 81 being regarded as one line. This makes it possible to accurately identify the area of each line and each item contained in the evidence without setting the position and range of all lines and items one by one. In addition, since the first frame line 81 and the second frame line 82 are displayed on the image data 50, the registration staff can visually confirm whether the format data 43 used for character recognition matches the format of the evidence shown as the image data 50. Therefore, even if the wrong format data 43 is selected, the person in charge of registration can easily understand.

In this embodiment, as described above, the format data 43 includes the reference line information 73 included in the evidence, and the information processing unit 33 is configured to perform a process of correcting the magnification rate and rotation angle of the image data 50 during character recognition processing so that the reference line 83 in the image data 50 matches the reference line 83 set in the format data 43. In this way, by setting the reference line information 73 included in the evidence in advance in the format data 43 and matching the reference line 83 in the image data 50 with the reference line 83 in the format data 43, elements such as character size, inclination, line spacing, and item (column) width in the image data 50 can be matched with the sample (template) used to set the format data 43. As a result, the size and direction of the evidence in the image data 50 can be made constant, and character recognition processing optimized by the format data 43 can be applied, so high recognition accuracy can be maintained.

In addition, in this embodiment, as described above, the format data 43 includes character spacing information 72d for each item included in the evidence, and the information processing unit 33 is configured to perform a process of adjusting the character spacing in the image data 50 based on the character spacing information 72d for each item of the image data 50 during character recognition processing. This allows pre-processing to be performed to adjust the character spacing to a level suitable for character recognition processing. Also, since the character type (numbers, alphabet, kana/kanji, etc.) used is fixed for each item of the evidence, by setting the character spacing information 72d on an item-by-item basis, it is possible to adjust the character spacing to an optimal level according to the character type. This makes it possible to further effectively improve the accuracy of character recognition.

In addition, in this embodiment, as described above, the memory unit 4a stores format data 43 for each type information CL corresponding to the cardboard type information CL including cardboard sheets and cardboard cases, and the information processing unit 33 is configured to select format data 43 from the memory unit 4a according to the type information CL of the evidence. This makes it possible to appropriately perform character recognition and information input for each item, even when each type of cardboard has its own unique item.

In addition, in this embodiment, as described above, when the information processing unit 33 receives a correction input operation via the input unit 32 for the recognition character string 85 input in any of the items, the information processing unit 33 is configured to record the pre-correction character string 85a and the post-correction character string 85b in association with each other in the replacement table 44 of the storage unit 4a, and to perform a process of replacing the recognition character string 85 obtained by character recognition with the corresponding post-correction character string 85b when it corresponds to the pre-correction character string 85a recorded in the replacement table 44. As a result, even if a misrecognition occurs in the character recognition process and the input information is corrected by the registration staff, if the same pre-correction character string 85a is recognized in the subsequent character recognition process, it can be automatically replaced with the post-correction character string 85b. This effectively reduces the workload of the registration staff. In addition, since certain character strings (for example, product name, delivery location, etc.) are frequently written on the evidence issued for each transaction, the simple process of replacing the pre-correction character string 85a with the post-correction character string 85b can effectively suppress erroneous input.

[Modification]
It should be noted that the embodiments disclosed herein are illustrative and not restrictive in all respects. The scope of the present invention is indicated by the claims, not by the description of the embodiments above, and further includes all modifications (variations) within the meaning and scope of the claims.

For example, in the above embodiment, an example was shown in which the cardboard evidence management system 1 was provided with multiple information processing terminals 3 and a management server 4, but the present invention is not limited to this. It is sufficient that the cardboard evidence management system 1 is provided with at least one information processing unit 33 and a storage unit 4a. It is not necessary to provide a management server 4 separate from the information processing terminal 3. For example, in the case of a small-scale user, one information processing terminal 3 may be connected to the receiving unit 2, and the HDD provided in the information processing terminal 3 may be configured as the storage unit 4a. In this configuration, the information processing terminal 3 further performs processing equivalent to that of the management server 4 described above. Also, in the above embodiment, an example was shown in which the receiving unit 2 was provided separately from the information processing terminal 3, but the information processing terminal 3 may have the receiving unit 2 built in.

In addition, in the above embodiment, an example was shown in which character recognition and input processing was performed on a line-by-line basis of the image data 50, but the present invention is not limited to this. In the present invention, character recognition and input processing may be performed on all lines of the image data 50 at once (without receiving input from the person in charge of registration by pressing the registration button). Also, character recognition and input processing may be performed on a column-by-column basis (item-by-item basis) of the image data 50.

In addition, in the above embodiment, an example was shown in which character recognition processing was temporarily stopped when an abnormal value was included in character recognition and input processing on a column-by-column basis, but the present invention is not limited to this. In the present invention, character recognition processing may be continued without being stopped.

In the above embodiment, an example was shown in which a sign 87 for identifying an item containing an abnormal value is displayed on the input screen 62, and the cursor 88 is moved to the input area 63 of the item containing the abnormal value, but the present invention is not limited to this. When notifying that an abnormal value is included, only the sign 87 may be displayed and the cursor 88 may not be moved, or the sign 87 may not be displayed. In addition, the sign 87 may be, for example, a pop-up dialog box indicating that an abnormal value is present.

In addition, in the above embodiment, an example in which the first frame line 81 and the second frame line 82 are displayed is shown, but the present invention is not limited to this. In the present invention, the first frame line 81 and the second frame line 82 do not have to be displayed.

In the above embodiment, an example was shown in which pre-processing was performed to match the reference line 83 in the image data 50 with the reference line 83 set in the format data 43, but the present invention is not limited to this. In the present invention, various data (such as data on the first frame line 81 and the second frame line 82) set in the format data 43 may be changed to match the reference line in the image data 50. Also, the size and rotation angle of the image data 50 may be adjusted based on the size and arrangement direction of the characters in the image data 50.

In addition, in the above embodiment, an example was shown in which processing was performed to adjust character spacing in image data 50 based on character spacing information 72d, but the present invention is not limited to this. In the present invention, character spacing does not have to be adjusted. In this case, character spacing information 72d does not have to be included in format data 43.

In the above embodiment, when the recognized character string 85 obtained by character recognition corresponds to the pre-correction character string 85a recorded in the storage unit 4a (replacement table), a process of replacing it with the corresponding post-correction character string 85b is performed, but the present invention is not limited to this. In the present invention, the replacement process does not have to be performed. For example, if the program (character recognition engine) for performing the character recognition process has a learning function, the correction data may be fed back to the character recognition engine.

In the above embodiment, when an order is placed by product name in the OCR registration process (Yes in step S14), the product name is entered by searching for product name candidates and selecting from the candidates, but the present invention is not limited to this. If the recognition accuracy of the character recognition process is sufficiently high, the recognized character string 85 may be entered as is, even when an order is placed by product name.

In the above embodiment, in step S22 of the OCR registration process shown in FIG. 14, an example was shown in which an unchecked flag is recorded when the order registration data 41a is registered in the management server 4, and the unchecked flag is rewritten to a check completed flag when the verification work is completed, but the present invention is not limited to this. In the present invention, it is not necessary to record an unchecked flag when the order registration data 41a is registered. For example, a method may be adopted in which the registration clerk visually checks the order contents, and the cardboard voucher management system 1 does not process anything after registration. The order may be confirmed by a method such as having the registration clerk print out the registered order contents and visually check the contents.

In the above embodiment, for convenience of explanation, the processing of the information processing unit 33 is explained using a flow-driven flowchart, but the present invention is not limited to this. In the present invention, the processing operation of the information processing terminal may be performed by event-driven processing in which processing is executed for each event. In this case, the processing may be completely event-driven, or event-driven and flow-driven may be combined.

1. Cardboard document management system 2. Receiving unit (image acquisition unit)
4a Memory unit 6 Evidence 31 (31a, 31b) Display unit 32 Input unit 33 Information processing unit 43 Format data 50, 50a, 50b, 50c Image data 61 Image display screen 62 Input screen 63a to 63f, 64a to 64i Input area 71 Line information 71a Maximum number of lines 71b Top and bottom end information (information on the top and bottom end positions of a line)
71c Row spacing information 72 Column information 72b Left and right edge information (information on left and right edge positions)
72d Character spacing information 73 Baseline information (baseline information)
81 First frame line 82 Second frame line 83 Reference line 85 Recognized character string 85a Pre-correction character string 85b Corrected character string 87 Mark 88 Cursor CL Type information

Claims (5)

an image acquisition unit for acquiring image data of evidence relating to the cardboard box;
A display unit that displays the image data;
an input unit that accepts input operations;
an information processing unit that performs registration processing of information included in the image data;
a storage unit in which format data of the voucher for each issuer of the voucher is recorded in advance,
The information processing unit includes:
Selecting the format data from the storage unit according to the issuer of the voucher;
executing character recognition processing of the image data with parameters based on the selected format data to obtain a recognized character string for each item of the voucher;
The display unit is configured to execute a process of inputting the recognition character string for each of the acquired items for each item of the input screen displayed on the display unit,
The format data includes line information including a maximum number of lines of the voucher, and column information regarding items included in one line of the voucher,
The information processing unit includes:
displaying an image display screen for displaying the image data and the input screen on the display unit;
a process of performing character recognition on a line-by-line basis of the image data displayed on the image display screen, and inputting the recognized character strings of each item included in the recognized line into each item of the input screen in sequence according to the maximum number of lines;
the line information includes, in addition to the maximum number of lines, information on the top and bottom end positions of any line and information on line spacing;
The column information includes information on the left and right end positions of each item,
The information processing unit includes:
displaying a first frame line indicating each line of the voucher based on the information on the top and bottom end positions, the line spacing, and the maximum number of lines, superimposing the first frame line on the image data displayed on the image display screen;
based on the information on the left end position and the right end position of each item, a second frame line indicating each item of the voucher is displayed superimposed on the image data displayed on the image display screen;
A cardboard document management system configured to obtain the area within the first frame line as one line and the character string contained within the second frame line as the recognition character string for each item. The format data includes baseline information included in the evidence,
The cardboard evidence management system of claim 1, wherein the information processing unit is configured to perform a process of correcting the magnification ratio and rotation angle of the image data during the character recognition process so as to match the reference line in the image data with the reference line set in the format data . The format data includes character spacing information for each item included in the evidence,
The cardboard evidence management system of claim 1 or 2, wherein the information processing unit is configured to execute a process of adjusting character spacing in the image data based on the character spacing information for each item of the image data during the character recognition process. The storage unit stores the format data for each type of cardboard, the type of cardboard including a cardboard sheet and a cardboard case,
The cardboard voucher management system according to any one of claims 1 to 3 , wherein the information processing unit is configured to select the format data from the storage unit according to the type information of the voucher. The information processing unit includes:
when a correction input operation is received via the input unit for the recognized character string input in any of the items, a pre-correction character string and a corrected character string are associated with each other and recorded in the storage unit;
The cardboard document management system according to any one of claims 1 to 4, configured to perform a process of replacing the recognized character string obtained by character recognition with the corresponding corrected character string when the recognized character string corresponds to the pre-correction character string recorded in the memory unit.

Images