JP2013030040A - Information processing program, information processor, and character recognition method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently perform character recognition processing.SOLUTION: A detection section 1a detects identification information 2a showing that the layout of a form is changed from image information 2 obtained by imaging the form. When the detection section 1a detects the identification information 2a, a region specifying section 1b detects region identification information 2b for identifying a processing object region 3 for performing prescribed character recognition processing which is included in the image information 2 to specify the processing object region 3 on the basis of the region identification information 2b. A processing section 1c performs prescribed character recognition processing on the specified processing object region 3.

Description

  The present invention relates to an information processing program, an information processing apparatus, and a character recognition method for performing character recognition.

  2. Description of the Related Art Conventionally, information processing apparatuses that perform recognition (character recognition) of characters (or character strings) entered on paper and acquire them as data have been used. For example, the information processing apparatus performs character recognition as follows. First, a paper entry surface is imaged by an imaging device to obtain image information on the paper surface. Next, the character image information included in the image information on the paper is compared with a predefined character pattern, and the character image information corresponds to any character pattern based on the degree of coincidence with the character pattern. To decide. Then, the character code corresponding to the determined character pattern is acquired as the character data entered on the page. This process is repeated for each character to obtain character string data.

  In character recognition, a character string written on paper is sometimes obtained in association with a predetermined data item. For example, a form used for transactions such as deposits and withdrawals at a financial institution is provided with an entry field that allows a customer to enter character strings for data items such as name, account number, and transaction amount. Each entry field is arranged at a fixed position for each type of form. In this case, if the position (layout) of the entry field in the form is defined in advance in association with the data item, the information processing apparatus associates the data item with character string data based on the form image information. You can get it. As described above, a character recognition method for extracting a character string for a data item based on layout definition information may be referred to as character recognition by layout recognition.

  Here, the form layout may be changed. In that case, in the layout recognition, a work load for updating the definition information of the layout every time a change occurs. Therefore, a method of character recognition that makes it easy to change the layout by determining the correspondence between the entered character string and the data item without depending on the layout definition information has been considered.

  For example, a character string (keyword) indicating a predefined item name (such as “amount” or “transfer amount”) is extracted from a form, and based on the positional relationship between the position of the keyword and the entered character string, There is a proposal for associating an item name indicated by the keyword with a written character string (for example, see Patent Document 1). As described above, the character recognition method for detecting the keyword indicating the item name and determining the correspondence with the entered character string may be called character recognition by keyword recognition.

  In addition, for example, there is a proposal to describe an identification code corresponding to an item at a predetermined position with respect to a field on a form provided with a field for entering a character string (see, for example, Patent Document 2). In this proposal, the item of the field is recognized by detecting the identification code.

  Further, for example, there is a proposal for defining syntax rule information including a rule for a character string for each item (see, for example, Patent Document 3). In this proposal, the recognized character string is analyzed based on the syntax rule information, and the recognized character string is associated with the item.

JP 2010-3155 A JP 2004-164376 A Japanese Patent Laid-Open No. 2004-199529

When a predetermined character recognition process is performed on a form after the layout is changed, processing efficiency may be poor if the entire area on the form is a processing target.
For example, depending on the form, an area that does not have to be a character recognition target (for example, a column used on the business side or a notification column for transmitting information to a customer) may be included. If such a region is also subject to character recognition processing, it takes extra processing time and processing efficiency is poor.

  In addition, for example, there may be a case where existing character recognition (for example, character recognition by layout recognition) may be performed on a non-target area to be changed. In that case, for example, if a non-target area to be changed is also a target of character recognition processing as exemplified in Patent Documents 1 to 3, an area that can be handled by existing character recognition is overlapped. become. Then, it takes extra processing time and processing efficiency is poor.

  Therefore, when the layout on the form is changed, there is a problem of how to implement a mechanism for efficiently performing the character recognition process by making it possible to easily specify a target area for a predetermined character recognition process. .

  The present invention has been made in view of these points, and an object thereof is to provide an information processing program, an information processing apparatus, and a character recognition method capable of efficiently performing character recognition processing.

  A processing target area for performing a predetermined character recognition process included in the image information when the identification information indicating that the layout of the form has been changed is detected from the image information obtained by capturing the form and the identification information is detected. Information processing for detecting a region identification information for identifying a region, specifying a processing target region based on the region identification information, performing a predetermined character recognition process on the specified processing target region, and causing a computer to execute the processing A program is provided.

  An information processing apparatus having a detection unit, a region specifying unit, and a processing unit is provided. The detection unit detects identification information indicating that the layout of the form has been changed from image information obtained by capturing the form. Upon detecting the identification information, the area specifying unit detects area identification information included in the image information for identifying a processing target area for performing a predetermined character recognition process, and determines the processing target area based on the area identification information. Identify. The processing unit performs a predetermined character recognition process on the specified processing target area.

  A character recognition method executed by the information processing apparatus is also provided. In this character recognition method, when identification information indicating that the layout of the form has been changed is detected from image information obtained by capturing the form, a processing target area for performing a predetermined character recognition process included in the image information The region identification information for identifying the region is detected, and the processing target region is specified based on the region identification information. A predetermined character recognition process is performed on the identified processing target area.

  Character recognition processing can be performed efficiently.

It is a figure which shows the information processing apparatus of 1st Embodiment. It is a figure which shows the information processing system of 2nd Embodiment. It is a figure which shows the hardware of the form reading apparatus of 2nd Embodiment. It is a figure which shows the example of the form before revision of 2nd Embodiment. It is a figure which shows the example of the form after revision of 2nd Embodiment. It is a block diagram which shows the function structure of the form reading apparatus of 2nd Embodiment. It is a figure which shows the example of the layout definition table of 2nd Embodiment. It is a flowchart which shows the form reading process of 2nd Embodiment. It is a figure which shows the example of the form image after the dropout process of 2nd Embodiment. It is a flowchart which shows the keyword recognition process of 2nd Embodiment. It is a figure which shows the example of the keyword recognition object area | region of 2nd Embodiment. It is a figure which shows the example of form ID of 2nd Embodiment. It is a figure which shows the example of the form after revision of 3rd Embodiment. It is a flowchart which shows the keyword recognition process of 3rd Embodiment. It is a figure which shows the example of the form before revision of 4th Embodiment. It is a figure which shows the example of the form after revision of 4th Embodiment. It is a flowchart which shows the form reading process of 4th Embodiment. It is a flowchart which shows the keyword recognition process of 4th Embodiment. It is a figure which shows the 1st example of the object area | region of each character recognition of 4th Embodiment. It is a figure which shows the 2nd example of the object area | region of each character recognition of 4th Embodiment. It is a figure which shows the example of the keyword recognition object area | region of 5th Embodiment. It is a flowchart which shows the specific process of the data part of 5th Embodiment. It is a figure which shows the example of the identification method of the data part of 5th Embodiment. It is a figure which shows the other example of the identification method of the data part of 5th Embodiment.

Hereinafter, the present embodiment will be described with reference to the drawings.
[First Embodiment]
FIG. 1 is a diagram illustrating the information processing apparatus according to the first embodiment. The information processing apparatus 1 performs a character recognition process on the image information 2 obtained by imaging a form having a predetermined layout. The layout is an arrangement of entry fields on a form, for example.

  The image information 2 includes identification information 2a and area identification information 2b. The identification information 2a is information indicating that the layout on the form has been changed. The area identification information 2b is information for identifying the processing target area 3 for performing a predetermined character recognition process. Information corresponding to the identification information 2a and the area identification information 2b is printed in advance on the form.

The information processing apparatus 1 includes a detection unit 1a, a region specifying unit 1b, and a processing unit 1c.
The detection unit 1a detects identification information 2a indicating that the layout on the form has been changed from the image information 2. For example, the identification information 2a is arranged at a predetermined position of the image information 2. The arrangement position is preset in the detection unit 1a. For example, the identification information 2a is represented by numerical values, characters, symbols, figures, combinations thereof, and the like. What information corresponds to the identification information 2a is preset in the detection unit 1a.

  The area specifying unit 1b detects the area identification information 2b from the image information 2 when the detection unit 1a detects the identification information 2a. The area specifying unit 1b specifies the processing target area 3 based on the area identification information 2b. For example, the area identification information 2b is represented as a line surrounding a predetermined area. In this case, the area specifying unit 1b can specify the area inside the area identification information 2b as the process target area 3. Note that what type of line corresponds to the area identification information 2b, such as the thickness of the line of the area identification information 2b and the type of line (eg, solid line, dotted line, and alternate long and short dash line), is preset in the area specifying unit 1b. The figure shown by the area specifying unit 1b may be an arbitrary polygon or an arbitrary figure that is not limited to a polygon.

  Note that the area identification information 2b can also be expressed by a method other than the line surrounding the predetermined area. For example, numerical values, characters, symbols, figures, and combinations thereof may be used. In that case, for example, it is set in advance in the region specifying unit 1b that the processing target region 3 is formed in a rectangular shape. In that case, the region specifying unit 1b can specify the processing target region 3 by detecting two pieces of region identification information provided at a pair of diagonal corners. The processing target area 3 (a graphic indicated by the area identification information) is not limited to a rectangle, and may be an arbitrary polygon or graphic. For example, if region identification information is provided at each position corresponding to a vertex of a polygon, the region identification unit 1b can identify a vertex by the region identification information and identify a polygon surrounded by the vertex. Also in this case, what information corresponds to the region identification information is preset in the region specifying unit 1b.

  Here, the detection unit 1a, the region specifying unit 1b, and the processing unit 1c may be implemented as programs that are executed using a CPU (Central Processing Unit) and a RAM (Random Access Memory).

  According to the information processing apparatus 1, the identification information 2 a is detected from the image information 2 by the detection unit 1 a. Then, the region identification information 1b included in the image information 2 is detected by the region identification unit 1b, and the processing target region 3 is identified based on the region identification information 2b. A predetermined character recognition process is performed on the processing target area 3 by the processing unit 1c.

  Thereby, a character recognition process can be performed efficiently. Specifically, an area where a predetermined character recognition process is performed is set as a process target area 3, and the predetermined character recognition process is not performed for other areas. That is, it is possible to prevent the predetermined character recognition processing from being performed on an extra area, and an extra processing time is not required.

  In the image information 2 in FIG. 1, the case where one processing target area 3 is specified by the area identification information 2 b is illustrated. On the other hand, a plurality of process target areas may be specified by a plurality of area identification information. More specifically, a plurality of region identification information expressed as lines surrounding the region may be provided so that a plurality of processing target regions can be specified. Alternatively, the first area identification information is provided at the vertex position of the first processing target area so that the first processing target area can be specified, and the first area identification information is set at the vertex position of the second processing target area. Second region identification information different from that may be provided so that the second processing target region can be specified.

Hereinafter, an example will be described in which the information processing apparatus 1 is applied to a form reading apparatus installed at a financial institution or the like and performing character recognition processing on image information of a form entered by a customer.
[Second Embodiment]
FIG. 2 illustrates an information processing system according to the second embodiment. This information processing system supports the window business of financial institutions. This information processing system includes a form reading device 100 and a server device 200. The form reading device 100 and the server device 200 are connected via the network 10. The network 10 is an intranet provided in the financial institution. The form reading device 100 and the server device 200 may be installed in separate bases. The route of the network 10 may include a dedicated line network laid for the information processing system, the Internet, and an IP (Internet Protocol) network of a communication carrier. When the Internet or IP network is included, communication security is ensured using a VPN (Virtual Private Network) or the like.

  The form reading apparatus 100 is an information processing apparatus installed at a financial institution window. The form reading device 100 performs character recognition processing on image information of a form (hereinafter, also referred to as a form image) acquired by imaging a form, acquires information entered in the form by a customer, and serves as a server To device 200.

  There are a plurality of types of forms read by the form reading apparatus 100 depending on the transaction. For example, forms corresponding to transactions such as deposits, withdrawals and new applications can be considered. It is assumed that the form reading apparatus 100 uses layout recognition for character recognition processing for a form having an existing layout.

  The server device 200 is an information processing device that manages customer account information. The server device 200 receives information on the form read by the form reading device 100 and executes a transaction process based on the information. For example, it is a process for confirming transactions such as cash withdrawal, cash deposit, transfer of deposits from one account to another.

In addition to the form reading device 100, a plurality of form reading devices may be provided at the counter of the financial institution.
Financial institutions sometimes change the layout of forms (hereinafter, sometimes referred to as revision of forms). For example, it may be possible to add a new data item entry field or to reduce an existing data item entry field. In layout recognition, a work load for changing the layout definition information of a form occurs when the layout is changed. Therefore, the form reading apparatus 100 can easily cope with the layout change by performing character recognition processing by keyword recognition for the revised portion. Hereinafter, a configuration for efficiently performing character recognition processing by keyword recognition in this case will be described.

  FIG. 3 is a diagram illustrating hardware of the form reading apparatus according to the second embodiment. The form reading apparatus 100 includes a CPU 101, a ROM (Read Only Memory) 102, a RAM 103, an HDD (Hard Disk Drive) 104, a graphic interface 105, an input interface 106, a scanner interface 107, a disk drive 108, and a communication interface 109.

The CPU 101 executes an OS (Operating System) program and an application program to control the entire form reading apparatus 100.
The ROM 102 stores a predetermined program such as a BIOS (Basic Input / Output System) program executed when the form reading apparatus 100 is activated. The ROM 102 may be a rewritable nonvolatile memory.

  The RAM 103 temporarily stores at least part of an OS program and application programs executed by the CPU 101. The RAM 103 temporarily stores at least a part of data used for the processing of the CPU 101.

  The HDD 104 stores an OS program and application programs. The HDD 104 stores data used for the processing of the CPU 101. Instead of the HDD 104 (or in combination with the HDD 104), other types of nonvolatile storage devices such as an SSD (Solid State Drive) may be used.

The graphic interface 105 is connected to the monitor 11. The graphic interface 105 displays an image on the monitor 11 in accordance with a command from the CPU 101.
The input interface 106 is connected to input devices such as the keyboard 12 and the mouse 13. The input interface 106 outputs an input signal sent from the input device to the CPU 101.

  The scanner interface 107 is connected to the image scanner 14. The image scanner 14 is an imaging device that captures a form and generates a form image. The scanner interface 107 outputs the form image acquired from the image scanner 14 to the CPU 101, the RAM 103, the HDD 104, and the like.

  The disk drive 108 is a reading device that reads data stored in the recording medium 15. For example, a program to be executed by the form reading apparatus 100 is recorded on the recording medium 15. The form reading apparatus 100 can realize functions as described later by executing a program recorded in the recording medium 15, for example. That is, the program can be recorded on a computer-readable recording medium 15 and distributed.

  As the recording medium 15, for example, a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory can be used. Examples of the magnetic recording device include an HDD, a flexible disk (FD), and a magnetic tape. Optical disks include CD (Compact Disc), CD-R (Recordable) / RW (ReWritable), DVD (Digital Versatile Disc), DVD-R / RW / RAM, and the like. Magneto-optical recording media include MO (Magneto-Optical disk). Semiconductor memory includes flash memory such as USB (Universal Serial Bus) memory.

The communication interface 109 is connected to the network 10. The communication interface 109 can perform data communication with the server device 200 via the network 10.
The form reading device 100 transmits the character string data of each data item read by character recognition to the server device 200.

The server apparatus 200 can also be realized by the same hardware configuration as the form reading apparatus 100.
FIG. 4 is a diagram illustrating an example of a form before revision according to the second embodiment. The form 300 exemplifies the (existing) form before revision. The form 300 has a layout determined by, for example, a standard. A form whose layout is defined by a standard or the like may be called an enacted form. On the form 300, a form ID (IDentifier) 310 and entry column groups 320 and 330 are printed.

  The form ID 310 is identification information for identifying the type of form. “200100” is printed as the form ID 310. The first four-digit numerical value “2001” of the form ID 310 indicates that the form 300 is a form for making an application for opening an account or performing a transaction for depositing into its own account. The last two digits “00” of the form ID 310 indicate that the form 300 has not been revised. Here, in the following, the last two digits of the form ID may be referred to as a branch number of the form ID.

  The entry column groups 320 and 330 are areas in which a plurality of entry columns to be entered by the customer are printed. Here, the entry field is a heading part of the entry field (for example, a part in which a character string such as “place” or “name” is printed in advance) and a part in which a customer enters a character string using a writing instrument. It is a column containing.

  The entry column group 320 includes entry columns 321, 322, 323, 324, 325, 326, 327, 328, and 329. The entry column 321 is a column for entering an application date (for example, date). The entry column 322 is a column for entering a number (store number) for identifying a store of a financial institution. The entry column 323 is a column for entering an account number. The entry column 324 is a column for entering a zip code and an address. The entry column 325 is a column for entering the name of the account holder. The entry column 326 is a column for entering the transaction amount. The entry column 327 is a column for selecting the type of financial product provided by the financial institution. The entry column 328 is a column for selecting the type of passbook to be used. The entry column 329 is a column for selecting the type of cash card to be used.

  The entry field group 330 includes entry fields 331, 332, and 333. The entry column 331 is a column for entering the account holder's telephone number. The entry column 332 is a column for selecting the gender of the account holder. The entry column 333 is a column for entering the date of birth of the account holder.

  A predetermined color is given to a frame line of each column of the entry column groups 320 and 330 and a part of the form 300. This color is a color that can be dropped out (erased) from the image by obtaining a color image (colored form image) of the form 300 and performing a predetermined dropout process on the image. Such a color may be referred to as a dropout color.

  FIG. 5 is a diagram illustrating an example of a revised form according to the second embodiment. The form 300a exemplifies a revised form for the form 300. A form ID 310a, entry column groups 320a and 330a, and area identification information 340 are printed on the form 300a.

  The form ID 310a is identification information for identifying the type of form. “200101” is printed as the form ID 310a. The first four-digit numerical value “2001” of the form ID 310a indicates that the form 300a is a form for making an application for opening an account or performing a transaction for depositing into its own account. The last two digits “01” of the form ID 310a indicate that the form 300a is a revised version “01”. That is, when the numerical value of the last two digits is other than “00”, it indicates that the form is the revised version, and the numeric value indicates the revised version number.

The entry column groups 320a and 330a are areas in which a plurality of entry columns to be entered by the customer are printed.
The entry column group 320a includes entry columns 321, 322, 323, 324, 325, 326, 327a, 328a, and 329a. Compared with the form 300, entry fields 327a, 328a, and 329a are different.

  The entry fields 327a, 328a, and 329a are fields for entering the same contents as the entry fields 327, 328, and 329. The entry fields 327a, 328a, and 329a differ from the entry fields 327, 328, and 329 in that new selection items are added. As a result, in the entry fields 328a and 329a, the area occupied by the fields is expanded downward toward the paper surface of FIG.

  The entry column group 330a includes entry columns 331a, 332a, and 333a. The entry fields 331a, 332a, 333a are fields for entering the same contents as the entry fields 331, 332, 333. Compared with the form 300, the areas occupied by the entry fields 328a and 329a are expanded, so that the entry fields 331a, 332a and 333a are moved downward toward the paper surface of FIG.

The area identification information 340 is a line surrounding the entry field group 330a and forms each side of a rectangle.
Similar to the form 300, dropout colors are given to the frame lines of the respective fields of the entry field groups 320a and 330a and a part of the form 300a. The region identification information 340 is assigned a color other than the dropout color. This is so that the form reading apparatus 100 can appropriately perform the detection processing of the area identification information 340 (described later).

  FIG. 6 is a block diagram illustrating a functional configuration of the form reading apparatus according to the second embodiment. The form reading apparatus 100 includes a storage unit 110, a form ID reading unit 120, a character recognition processing unit 130, and an area specifying unit 140. Each of these functions is realized on the form reading apparatus 100 by the CPU 101 executing a predetermined program, for example. All or some of these functions may be implemented by dedicated hardware.

The storage unit 110 stores a layout definition table that defines the layout of a form.
The form ID reading unit 120 reads the form ID included in the color form image received from the image scanner 14. At this time, the form ID reading unit 120 performs a dropout process on the form image. Which color is used as the dropout color is preset in the storage unit 110 in association with the form ID, for example. The form ID is printed at a predetermined position on the form, for example. Therefore, if the position (for example, coordinate value on the form image) is set in advance in the form ID reading unit 120, the form ID reading unit 120 can read the form ID from the coordinate position in the form image. Alternatively, the form ID reading unit 120 may read a character string that matches the format as the form ID based on the format (for example, a numerical value or the number of digits) of the form ID. The form ID reading unit 120 outputs the result of reading the form ID to the character recognition processing unit 130.

  The form ID reading unit 120 stores the form image received from the image scanner 14 and the image information after the dropout process is performed on the form image in a predetermined area on the RAM 103. In the subsequent processing, the character recognition processing unit 130 and the region specifying unit 140 refer to the region on the RAM 103 and execute processing of each unit for the form image (including image information after the dropout processing).

  The character recognition processing unit 130 executes character recognition processing by layout recognition on the form. When the form ID indicates that the form ID has been revised, the character recognition processing unit 130 performs character recognition by layout recognition and keyword recognition on the form image. The character recognition processing unit 130 displays the result of character recognition on the monitor 11. This is to prompt the operator to check the correctness of the result. If the reading result (data item and character string data contents) is appropriate, the operator can operate the keyboard 12 and the mouse 13 to instruct the character recognition processing unit 130 to transmit the reading result to the server device 200. . When the character recognition processing unit 130 receives this instruction, the character recognition processing unit 130 transmits the reading result to the server device 200 and requests execution of processing related to the transaction.

The character recognition processing unit 130 includes a layout recognition processing unit 131 and a keyword recognition processing unit 132.
The layout recognition processing unit 131 performs layout recognition processing (character recognition by layout recognition) on the form image based on the form ID and the layout definition table stored in the storage unit 110.

  The keyword recognition processing unit 132 performs keyword recognition processing (character recognition processing by keyword recognition) on the form image based on the form ID and the layout definition table stored in the storage unit 110. At that time, the keyword recognition processing unit 132 requests the region specifying unit 140 to specify the processing target region on the form image.

  The area specifying unit 140 reads area specifying information from the form image in response to a request from the keyword recognition processing unit 132. The area specifying information is printed on the form with a predetermined thickness and line type. Therefore, if the thickness and line type are set in advance in the area specifying unit 140, the area specifying unit 140 can read the area specifying information from the form image based on the setting contents.

  The area specifying information may be information indicating a vertex of the polygon. In that case, a numerical value, a character, a symbol, a figure, a combination thereof, and the like to be recognized as the area specifying information may be set in advance in the area specifying unit 140. The area specifying unit 140 outputs information indicating the specified area (for example, coordinate values indicating the area) to the keyword recognition processing unit 132.

  FIG. 7 is a diagram illustrating an example of a layout definition table according to the second embodiment. The layout definition table 111 is stored in the storage unit 110. The layout definition table 111 includes items such as form ID, data item name, entry column coordinates, category, and manual activity classification. The information arranged in the horizontal direction of each item is associated with each other, and indicates information for specifying one content to be entered on the form.

A form ID (excluding branch numbers) is set in the form ID item. The data item name is set in the data item name item.
The coordinate value of the entry field is set in the coordinate field of the entry field. Here, taking the form 300 shown in FIG. 4 as an example, the coordinate value of one entry column is the upper left coordinate value of the entry column in FIG. 4 and the lower right side of the entry column. This is indicated by specifying the coordinates of. The coordinate values are defined by defining the upper left vertex of the form 300 in FIG. 4 as the origin, the right direction as the positive direction of the x coordinate, and the downward direction as the positive direction of the y coordinate. However, the position of the entry field may be expressed in a different coordinate system.

  In the category item, it is set which type of character string to be entered in the entry field is numeric, katakana, kanji, or symbol. Information indicating whether the character string to be entered in the entry field is handwritten or printed as a type is set in the item of hand type.

  For example, in the layout definition table 111, the form ID is “2001”, the data item name is “application date”, the coordinates of the entry column are “S (35, 20), E (60, 30)”, and the category is “numeric value”. ", The information that the hand activity classification is" handwritten "is set. Note that the character “S” is assigned to the coordinate value on the upper left side of the entry field for setting the coordinates in the entry field. Similarly, the letter “E” is attached to the coordinate value on the lower right side of the entry field.

  Therefore, in this record, the entry field corresponding to the data item name “application date” is the form image (x, y) = (35, 20) for the form image of the form 300 whose first four digits are “2001”. ) And (x, y) = (60, 30) indicates a rectangular region having diagonal lines. The character string entered in the entry field is a numerical value, and indicates that the character string is handwritten by the customer. Further, this record indicates that the same entry field can be included for the form image of the form 300a whose first four digits of the form ID are “2001”.

Next, a processing procedure of the form reading apparatus 100 having the above configuration will be described.
FIG. 8 is a flowchart illustrating a form reading process according to the second embodiment. In the following, the process illustrated in FIG. 8 will be described in order of step number.

  (Step S <b> 11) The form ID reading unit 120 stores the color form image received from the image scanner 14 in a predetermined area on the RAM 103. The form ID reading unit 120 performs a dropout process on the form image, and stores the form image after the dropout process in another predetermined area on the RAM 103.

  (Step S12) The form ID reading unit 120 recognizes the form ID from the form image after the dropout process. For example, the form ID reading unit 120 can read the form ID from a predetermined position of the form image. The form ID reading unit 120 outputs the read form ID to the character recognition processing unit 130.

  (Step S <b> 13) The character recognition processing unit 130 delegates the processing to the layout recognition processing unit 131 and causes the layout recognition processing to be executed. The layout recognition processing unit 131 refers to the layout definition table 111 stored in the storage unit 110 on the basis of the form ID acquired from the form ID reading unit 120, and performs character recognition by layout recognition on the form image after the dropout process. Perform recognition processing.

  (Step S14) The character recognition processing unit 130 determines whether or not the last two digits of the form ID acquired from the form ID reading unit 120 is “01” or more. If it is “01” or more (that is, a revised form), the process proceeds to step S15. If it is not “01” or more (that is, if it is not a revised form), the process proceeds to step S16.

(Step S15) The character recognition processing unit 130 delegates the processing to the keyword recognition processing unit 132 and causes the keyword recognition processing to be executed. The keyword recognition process will be described later.
(Step S16) The character recognition processing unit 130 outputs a character recognition result for the form image and causes the monitor 11 to display the result.

  As described above, the character recognition processing unit 130 executes the layout recognition process on the form image after the dropout process. Further, the character recognition processing unit 130 determines whether or not the keyword recognition process is necessary according to the form ID included in the form image. The character recognition processing unit 130 performs the keyword recognition process on the revised form, and does not execute the keyword recognition process on the unrevised form.

Next, the form image after the dropout process related to the form 300a in step S11 is illustrated.
FIG. 9 is a diagram illustrating an example of a form image after the dropout process according to the second embodiment. The form image 300b is image information obtained by performing dropout processing on the form image captured by the image scanner 14.

  In the form image 300b, the frame line printed on the form 300a in the dropout color is deleted. The color of the character string entered by the customer in the frame line is entered in a color other than the dropout color and is not a dropout target (it is not erased by the dropout process). The same applies to the area identification information 340.

  The layout recognition processing unit 131 refers to the layout definition table 111 and uses a character string written at predetermined coordinates in the form image 300b (the first four digits of the form ID is “2001”) as an account number, a name, and the like. Can be acquired in association with data items. For example, when extracting the data of the data item name “account number” from the form image 300b, the layout recognition processing unit 131 performs the following processing.

  The layout recognition processing unit 131 identifies a frame 323 a corresponding to “account number” based on the layout definition table 111. Specifically, the coordinates (x, y) = (140, 20) of the form image 300b are set as the upper left vertex toward the paper surface of FIG. 9, and the coordinates (x, y) = (160, 30) are set as the lower right vertex. A rectangle to be displayed is a frame 323a. Note that the frame line 323a illustrated by the dotted line in FIG. 9 is not included in the form image 300b. The layout recognition processing unit 131 obtains a character image entered in the frame 323a. The layout recognition processing unit 131 collates the character image with the character pattern stored in the storage unit 110 in advance, and extracts a character string “9999999”. At this time, according to the layout definition table 111, the category of the data at this position is a numerical value. Therefore, the layout recognition processing unit 131 only has to collate numerical character patterns with character images.

In this way, the layout recognition processing unit 131 acquires the character string data “9999999” for the data item “account number”.
The area identification information 340 is printed in a color other than the dropout color. Therefore, the area identification information 340 is not a dropout target and is included in the form image 300b. In the form image 300b, the other frame lines are erased by the dropout process, so that the area specifying unit 140 can easily detect the area identification information 340.

Next, the procedure of the keyword recognition process described in step S15 in FIG. 8 will be described. Hereinafter, it is assumed that the following information is acquired on the RAM 103.
(1) “Form image of form 300a”
(2) “Form image 300b” obtained by executing dropout processing on “Form image of form 300a”
The “form image of the form 300a” indicates image information obtained in color for all areas included in the form 300a. In the case of “form image 300b”, “form image of form 300a” indicates image information obtained by executing the dropout process. In any form image, the region identification information 340 is indicated by the same symbol.

FIG. 10 is a flowchart illustrating keyword recognition processing according to the second embodiment. In the following, the process illustrated in FIG. 10 will be described in order of step number.
(Step S21) The keyword recognition processing unit 132 saves the result of the layout recognition processing by the layout recognition processing unit 131 in a predetermined area on the RAM 103 or the HDD 104.

  (Step S22) The keyword recognition processing unit 132 requests the region specifying unit 140 to specify the target region for the keyword recognition processing. The area specifying unit 140 detects area identification information 340 from the form image 300b. For example, information related to the region identification information 340 (type of shape, line thickness, line type such as solid line / broken line, etc.) is preset in the region specifying unit 140. Here, the type of shape is a rectangular frame line, an elliptical frame line, a symbol attached to the vertex of a polygon (including what the symbol is), etc. Information. The area specifying unit 140 can detect the area identification information 340 from the form image 300b based on the information. The area specifying unit 140 outputs area information indicating the specified processing target area to the keyword recognition processing unit 132. The area information is, for example, the coordinate value of each vertex of the processing target area. When the region specifying unit 140 detects a plurality of processing target regions, the region specifying unit 140 outputs a plurality of region information to the keyword recognition processing unit 132.

  (Step S23) The keyword recognition processing unit 132 extracts a color image of one unprocessed area from the form image of the form 300a based on the area information acquired from the area specifying unit 140. For example, the keyword recognition processing unit 132 extracts a color image of an area (processing target area) in the area identification information 340.

  (Step S24) The keyword recognition processing unit 132 extracts one keyword from the processing target area from which the color image has been acquired in Step S23. For example, the keyword recognition processing unit 132 detects a frame line included in the processing target region, and tries to extract a keyword for each frame surrounded by the frame line. For example, the keyword recognition processing unit 132 extracts the keyword “phone number” included in the entry field 331a.

  (Step S25) The keyword recognition processing unit 132 refers to the form ID of the form 300a and the layout definition table 111 stored in the storage unit 110, and determines whether or not the keyword matches a data item name that can be included in the form 300a. To determine. If they match, the process proceeds to step S26. If they do not match, the keyword is discarded and the process proceeds to step S29. For example, the keyword recognition processing unit 132 detects that the keyword “phone number” is set in the layout definition table 111 in association with the first four digits “2001” of the form ID “200101” of the form 300a. . In this case, it is determined that the keyword “phone number” matches the data item name that can be included in the form 300a. A cell in which the keyword exists is a heading cell. Here, the cell indicates one frame surrounded by a drop-out color frame line.

  (Step S26) The keyword recognition processing unit 132 specifies the position (data part) of the entry field for the data item. For example, the position of the data part is set in advance in the keyword recognition processing unit 132 as “a cell adjacent to the right side of the cell in which the keyword exists”. Alternatively, the data portion corresponding to the keyword may be specified based on the keyword or the table structure on the form 300a. As a method for associating the keyword with the data portion, for example, a method described in JP 2010-3155 A can be used.

  (Step S <b> 27) The keyword recognition processing unit 132 analyzes the characteristics of the data part, removes frame lines, noise, and the like from the image of the data part, and processes the image into an image suitable for character recognition.

  (Step S <b> 28) The keyword recognition processing unit 132 acquires a character image entered in the data portion. The keyword recognition processing unit 132 collates the character image with a character pattern stored in the storage unit 110 in advance, and acquires character string data entered in the data unit. For example, the keyword recognition processing unit 132 acquires character string data “000-0000-0000” for the data item “phone number”.

  (Step S <b> 29) The keyword recognition processing unit 132 determines whether or not there is a portion where no keyword has been extracted in the current processing target area. If there is a keyword-unextracted portion, the process proceeds to step S24. If there is no keyword-unextracted portion, the process proceeds to step S30.

  (Step S30) The keyword recognition processing unit 132 determines whether there is an unprocessed processing target area other than the current processing target area. If there is an unprocessed processing target area, the process proceeds to step S23. If all the process target areas have been processed, the process proceeds to step S31.

  (Step S31) The keyword recognition processing unit 132 merges the temporarily saved layout recognition result and the keyword recognition result. Specifically, among the data items of the form 300a defined in the layout definition table 111, those that could not be acquired by the layout recognition process are supplemented with the contents acquired by the keyword recognition process.

In this way, the keyword recognition processing unit 132 performs character recognition by keyword recognition for the changed area of the form 300a.
FIG. 11 is a diagram illustrating an example of a keyword recognition target area according to the second embodiment. The keyword recognition target area is an area inside the area identification information 340. In the area, keywords 351, 352, and 353 are also shown.

  The keyword 351 indicates the character string “telephone number”, and according to the layout definition table 111 stored in the storage unit 110, matches the data item name “telephone number” that can be included in the form image of the form 300a. .

The keyword 352 indicates the character string “sex” and matches the data item name “sex” that can be included in the form image of the form 300a.
The keyword 353 indicates the character string “date of birth” and matches the data item name “date of birth” that can be included in the form image of the form 300a.

  When the keyword recognition processing unit 132 detects the keyword 351, the keyword recognition processing unit 132 identifies the cell 351a adjacent to the right side of the cell in which the keyword 351 exists as the data unit. Then, the keyword recognition processing unit 132 acquires character string data “000-0000-0000” for the data item “phone number”. Note that the cell including the keyword 351 and the cell 351a correspond to the entry field 331a shown in FIG.

  Further, when the keyword recognition processing unit 132 detects the keyword 352, the keyword recognition processing unit 132 identifies the cell 352a adjacent to the right side of the cell in which the keyword 352 exists as the data unit. Then, the keyword recognition processing unit 132 acquires a selection symbol (check mark) indicating that “male” is selected for the data item “sex”. Such a selection symbol is also included in the character string data. Note that the cell including the keyword 352 and the cell 352a correspond to the entry field 332a shown in FIG.

  Furthermore, when the keyword recognition processing unit 132 detects the keyword 353, the cell 353a adjacent to the right side of the cell in which the keyword 353 exists is specified as the data unit. Then, the keyword recognition processing unit 132 acquires the character string data “1985” “11” month “15” day for the data item “birth date”. Note that the cell including the keyword 353 and the cell 353a correspond to the entry field 333a shown in FIG.

  The area includes other characters such as “male”, “female”, “year”, “month”, and “day”. However, none of these are defined as data item names in the form 300a in the layout definition table 111. Therefore, the keyword recognition processing unit 132 does not regard these other characters as data item names.

FIG. 12 is a diagram illustrating an example of a form ID according to the second embodiment. The printing method of the form ID 310a is not limited to the above method. For example, you may print as follows.
In FIG. 12A, a form ID 310b is shown. The form ID 310b indicates that the form 300a has been revised by a combination of the form ID 310 and the branch number 311 printed below the form ID 310. The branch number 311 is represented by a numerical value such as “01”.

  In FIG. 12B, a form ID 310c is shown. The form ID 310c indicates that the form 300a has been revised by a combination of the form ID 310 and the branch number 311a printed on the lower right side from the position of the form ID 310 toward the paper surface. The branch number 311a is represented by a numerical value such as “01”.

  In FIGS. 12A and 12B, when the form is not revised, it is conceivable that the branch numbers 311 and 311a are set to “00” or the branch numbers 311 and 311a are not printed. .

  In FIG. 12C, a form ID 310d is shown. The form ID 310d indicates that the form 300a has been revised by a combination of the form ID 310 and the branch number 311b printed on the right side from the position of the form ID 310 toward the paper surface. The branch number 311b is represented by a character (or character string) such as “R”.

  In FIG. 12C, when the form is not revised, it is conceivable that the branch number 311b is set to another character (or character string) or the branch number 311b is not printed.

  As described above, various methods can be used for the form ID. The above example is one example, and the form ID may be expressed by a position, numerical value, symbol, character (or character string), etc. other than the above. Which position or numerical value indicates the form ID is preset in the form ID reading unit 120 as described above.

  As described above, according to the form reading apparatus 100 of the second embodiment, character recognition processing can be performed efficiently. Specifically, the area for performing the keyword recognition process is specified based on the area identification information 340 for the revised form 300a. The form reading apparatus 100 performs keyword recognition processing on the area and does not perform keyword recognition processing on the other areas. That is, it is possible to prevent the keyword recognition process from being performed on an extra area, and an extra processing time is not required.

  In addition, a region that is not subjected to keyword recognition processing can be recognized by layout recognition processing based on the existing layout definition table 111. Therefore, it is particularly effective when used in combination with layout recognition. Specifically, since the layout recognition process reads information at a predetermined position, the processing time is generally shorter than the keyword recognition process. Therefore, by narrowing down the target area for keyword recognition processing as in the form reading device 100, the layout of the form can be flexibly changed while enjoying the good processing efficiency by layout recognition.

  Here, there is a case where a financial institution wants to change the layout of a part of a predetermined form temporarily only for a short period such as a campaign. Specifically, there may be a case where a wide campaign notification column is secured to affect the layout of an existing layout, or a case where a column for requesting a customer to fill in is provided only during the campaign. In that case, it is particularly desirable that the layout change can be easily handled by suppressing the influence of the system change on the work load and the processing efficiency of character recognition as much as possible as compared to the case of performing a permanent layout change. According to the form reading apparatus 100, the keyword recognition process is performed only on a part of the revision target as described above. As a result, it is possible to realize a character recognition process that suppresses the influence on the layout definition changing work and the character recognition processing efficiency.

[Third Embodiment]
Hereinafter, a third embodiment will be described. Differences from the second embodiment will be mainly described, and description of similar matters will be omitted.

  In the second embodiment, it has been described that a plurality of area identification information may exist on a form. In this case, the same data item name (keyword) may be included in a plurality of processing target areas indicated by a plurality of area identification information. At this time, even the same data item name does not necessarily have the same use. For example, a form corresponding to a plurality of transactions may be used in one form. For example, a form that can handle two transactions, that is, cash withdrawal from an account and transfer from the account to another account, can be considered. More specifically, (1) withdrawing cash from a first account indicated by a first account number, and (2) a second account indicated by a second account number from the first account. It is a transfer to.

  In that case, the first account number and the second account number can be entered in the form, but when these characters are recognized by keyword recognition, which account number is the withdrawal source account number and which account number is transferred It may be difficult to distinguish whether it is a previous account number. Specifically, this is a case where all the keywords printed on the form are “account numbers”. In this case, even if the keyword is simply compared with the data item name in the layout definition information, it is difficult to distinguish the usage of each “account number”.

Therefore, in the third embodiment, even if the same data item name is included in a plurality of processing target areas, a function of acquiring character string data by distinguishing the use of each data is provided.
Here, the configuration of the information processing system of the third embodiment is the same as the configuration of the information processing system of the second embodiment described in FIG. Further, the hardware and functional configuration of the form reading apparatus according to the third embodiment is the same as the hardware and functional configuration of the form reading apparatus 100 according to the second embodiment described with reference to FIGS. Hereinafter, the form reading apparatus according to the third embodiment also indicates each component using the same reference numeral and name as the form reading apparatus 100.

  FIG. 13 is a diagram illustrating an example of a revised form according to the third embodiment. The form 400 exemplifies a case where there are a plurality of target areas for keyword recognition processing indicated by the area identification information. On the form 400, a form ID 410, an entry column group 420, an entry column group 430, area identification information 441, 442, 443, 444, 460 and an entry column group 450 are printed.

  The form ID 410 is identification information for identifying the type of form. “300101” is printed as the form ID 410. A numerical value “3001” in the first four digits of the form ID 410 indicates that the form 400 is a form for performing cash withdrawal and transfer transactions between accounts. The last two digits “01” of the form ID 410 indicate that the form 400 is a revised version “01”.

The entry column groups 420, 430, and 450 are areas in which a plurality of entry columns to be input by the customer are printed.
The entry column group 420 includes entry columns 421, 422, and 423. The entry column 421 is a column for entering the application date. The entry column 422 is a column for entering the name of the account holder. The entry column 423 is a column for imprinting a delivery seal.

  The entry column group 430 is provided with an entry column for entering information of the withdrawal source account. The entry field group 430 includes entry fields 431, 432, 433, and 434. The entry column 431 is a column for entering a store number. The entry column 432 is a column for entering the account number of the withdrawal source account. The entry column 433 is a column for selecting a deposit item of the withdrawal source account. The entry column 434 is a column for entering the amount to be withdrawn.

  The area identification information 441, 442, 443, 444 is key-type area identification information surrounding the entry field group 430. Four region identification information 441, 442, 443, and 444 indicate one processing target region for keyword recognition in one set. Specifically, an area surrounded by a rectangle having the area identification information 441, 442, 443, 444 as a vertex is the processing target area.

  The entry field group 450 is provided with an entry field for entering information of the transfer destination account. The entry column group 450 includes entry columns 451, 452, 453, and 454. The entry column 451 is a column for selecting a deposit item of the transfer destination account. The entry column 452 is a column for entering a store number. The entry column 453 is a column for entering the account number of the transfer destination account. The entry column 454 is a column for entering the account holder of the transfer destination account.

The area identification information 460 is a line surrounding the entry column group 450 and forms each side of a rectangle.
Dropout colors are assigned to the frame lines of the fields of the entry field groups 420, 430, and 450 and a part of the form 400. The region identification information 441, 442, 443, 444, 460 is assigned a color other than the dropout color.

  According to the form 400, cash can be withdrawn from the account designated in the entry field group 430 by filling in only the entry field groups 420 and 430. Furthermore, by designating the transfer destination account in the entry column group 450, the deposit can be transferred from the account designated in the entry column group 430 to the account designated in the entry column group 450.

  In this way, key-type area identification information 441, 442, 443, and 444 is preprinted on the form 400 for the entry column group 430 for the withdrawal source account. Further, in the form 400, the area identification information 460, which is a line forming each side of a quadrilateral (having a shape different from the key shape), is preprinted in the entry field group 450 for the transfer destination account.

  Next, a processing procedure of the form reading apparatus 100 having the above configuration will be described. Note that the procedure of the form reading process of the third embodiment is the same as the procedure of the form reading process of the second embodiment described with reference to FIG. Next, the procedure of the keyword recognition process described in step S15 of the form reading process will be described. Hereinafter, it is assumed that the following information is acquired on the RAM 103.

(1) “Form image of form 400”
(2) “Form image after dropout processing” obtained by executing dropout processing on “form image of form 400”
In any form image, the region identification information 441, 442, 443, 444, 460 is indicated by the same reference numeral.

FIG. 14 is a flowchart illustrating keyword recognition processing according to the third embodiment. In the following, the process illustrated in FIG. 14 will be described in order of step number.
(Step S41) The keyword recognition processing unit 132 saves the result of the layout recognition processing by the layout recognition processing unit 131 in a predetermined area on the RAM 103 or the HDD 104.

  (Step S42) The keyword recognition processing unit 132 requests the region specifying unit 140 to specify the target region for the keyword recognition processing. The area specifying unit 140 detects area identification information 441, 442, 443, 444, and 460 from the form image after the dropout process. For example, information related to the region identification information 441, 442, 443, 444, 460 (type of shape, line thickness, line type such as solid line / broken line, etc.) is preset in the region specifying unit 140. Based on this information, the area specifying unit 140 can detect the area identification information 441, 442, 443, 444, and 460 from the form image after the dropout process. The area specifying unit 140 outputs area information indicating the specified processing target area to the keyword recognition processing unit 132. The area information is, for example, the coordinate value of each vertex of the processing target area. In addition, the area specifying unit 140 outputs the type of detected area identification information (whether it is rectangular or keyed) together with the area information to the keyword recognition processing unit 132.

  (Step S <b> 43) The keyword recognition processing unit 132 extracts a color image of one unprocessed area from the form image of the form 400 based on the area information acquired from the area specifying unit 140. For example, the keyword recognition processing unit 132 extracts a color image in an area surrounded by the area identification information 441, 442, 443, 444.

  (Step S44) The keyword recognition processing unit 132 extracts one keyword from the processing target area from which the color image has been acquired in Step S43. For example, the keyword recognition processing unit 132 detects a frame line included in the processing target region, and tries to extract a keyword for each frame surrounded by the frame line. For example, the keyword recognition processing unit 132 extracts the keyword “account number” from the entry field 432.

  (Step S 45) The keyword recognition processing unit 132 refers to the form ID of the form 400 and the layout definition table 111 stored in the storage unit 110, and determines whether or not the keyword matches a data item name that can be included in the form 400. To determine. If they match, the process proceeds to step S46. If they do not match, the keyword is discarded and the process proceeds to step S52. For example, it is assumed that the keyword “account number” matches a data item name that can be included in the form 400.

  (Step S46) The keyword recognition processing unit 132 specifies the position (data portion) of the entry field for the data item. For example, the position of the data part is set in advance in the keyword recognition processing unit 132 as “a cell adjacent to the right side of the cell in which the keyword exists”. Alternatively, based on the keyword and the table structure on the form 400, the data portion corresponding to the keyword may be specified.

  (Step S <b> 47) The keyword recognition processing unit 132 analyzes the characteristics of the data portion, removes frame lines and noise from the image of the data portion, and processes the image into an image suitable for character recognition.

  (Step S48) The keyword recognition processing unit 132 acquires a character image written in the data portion. The keyword recognition processing unit 132 collates the character image with a character pattern stored in the storage unit 110 in advance, and acquires character string data entered in the data unit. For example, the keyword recognition processing unit 132 acquires character string data “9999999” for the keyword “account number”.

  (Step S49) The keyword recognition processing unit 132 determines the type of region identification information corresponding to the current processing target region. If the shape type of the area identification information is a key type, the process proceeds to step S50. If the shape type of the area identification information is rectangular, the process proceeds to step S51. For example, if the region is surrounded by the region identification information 441, 442, 443, 444, the type of the region identification information is “key type”.

  (Step S50) The keyword recognition processing unit 132 acquires the character string data acquired in step S48 as the data of the withdrawal source. Thereby, for example, “account number” “9999999” acquired from the processing target area surrounded by the area identification information 441, 442, 443, 444 can be handled as the account number of the withdrawal source account. Then, the process proceeds to step S52.

  (Step S51) The keyword recognition processing unit 132 acquires the character string data acquired in step S48 as transfer destination data. Thereby, for example, “account number” “88888888” acquired from the processing target area surrounded by the area identification information 460 can be handled as the account number of the transfer destination account. Then, the process proceeds to step S52.

  (Step S <b> 52) The keyword recognition processing unit 132 determines whether or not there is a portion where no keyword has been extracted in the current processing target area. If there is a keyword-unextracted portion, the process proceeds to step S44. If there is no keyword-unextracted portion, the process proceeds to step S53.

  (Step S53) The keyword recognition processing unit 132 determines whether there is an unprocessed processing target area other than the current processing target area. If there is an unprocessed processing target area, the process proceeds to step S43. If all the process target areas have been processed, the process proceeds to step S54.

  (Step S54) The keyword recognition processing unit 132 merges the temporarily recognized layout recognition result and the keyword recognition processing result. Specifically, among the data items of the form 400 defined in the layout definition table 111, those that could not be acquired by the layout recognition process are complemented with the contents acquired by the keyword recognition process.

  In this manner, the keyword recognition processing unit 132 distinguishes recognized character string data according to the type of region identification information. Which area identification information corresponds to which transaction is set in the keyword recognition processing unit 132 in advance. The association between the type of region identification information exemplified in steps S49 to S51 and the transaction (use of data) is an example, and various other associations are conceivable. For example, the following contents can be considered as a method for distinguishing the region identification information.

(A1) Line type of area identification information (dotted line, broken line, one-dot chain line, etc.).
(A2) The shape of the area identification information (rectangle, ellipse, etc.).
(A3) Line thickness of area identification information (thick, thin, etc.).

(A4) Color of area identification information (dark, light, etc.).
(A5) The shape of the region identification information to be placed at the vertex of the processing target region (such as a key shape).
The area identification information may be distinguished based on any one of the above (A1) to (A5), or may be distinguished by combining (A1) to (A5). Further, it is conceivable that each identified area identification information is associated with one of the following uses.

(B1) Information related to the deposit account when depositing by cash or transfer.
(B2) Information on the withdrawal source account when withdrawing money or transferring money.
The keyword recognition processing unit 132 can distinguish which transaction the character string data is used for (ie, use of the character string data) by distinguishing the region identification information. You may match with uses other than the said use.

  In this way, the form reading apparatus 100 according to the third embodiment is capable of handling each data item even if the same data item name is included in a plurality of processing target areas in a form that can handle a plurality of transactions. Character string data can be acquired by distinguishing usage.

This makes it possible to more efficiently perform character recognition processing for a form whose layout has been changed.
[Fourth Embodiment]
Hereinafter, a fourth embodiment will be described. Differences from the second and third embodiments will be mainly described, and description of similar matters will be omitted.

  In the second and third embodiments, the necessity of the keyword recognition process is determined after the layout recognition process. On the other hand, the layout recognition process does not have to be performed on the area that is the target of the keyword recognition process. In this way, the processing efficiency can be further improved. Therefore, in the fourth embodiment, a function of suppressing the layout recognition process is provided for the area that is the target of the keyword recognition process.

  Here, the configuration of the information processing system of the fourth embodiment is the same as the configuration of the information processing system of the second embodiment described in FIG. The hardware and functional configuration of the form reading apparatus according to the fourth embodiment is the same as the hardware and functional configuration of the form reading apparatus 100 according to the second embodiment described with reference to FIGS. Hereinafter, the form reading apparatus according to the fourth embodiment also indicates each component using the same reference numerals and names as the form reading apparatus 100.

  FIG. 15 is a diagram illustrating an example of a form before revision according to the fourth embodiment. A form 500 illustrates an (existing) form before revision. A form ID 510 and entry column groups 520 and 530 are printed on the form 500.

  The form ID 510 is identification information for identifying the type of form. “100100” is printed as the form ID 510. The first four-digit numerical value “1001” of the form ID 510 indicates that the form 500 is a form for performing a transaction for depositing cash into the account. The last two digits “00” of the form ID 510 indicate that the form 500 has not been revised.

  The entry column group 520 is an area in which a plurality of entry columns to be entered by the customer are printed. The entry column group 520 includes entry columns 521, 522, 523, 524, 525, 526. The entry column 521 is a column for entering the application date. The entry column 522 is a column for entering the name of the account holder. The entry column 523 is a column for entering a store number. The entry column 524 is a column for entering an account number. The entry column 525 is a column for selecting a deposit item of the deposit destination account. The entry column 526 is a column for entering the amount to be deposited.

The entry column group 530 is a column for entering information used by the financial institution staff for business. The entry column group 530 is a non-target area for layout recognition processing.
Drop-out colors are assigned to the frame lines of each column of the entry column groups 520 and 530 and a part of the form 500.

  FIG. 16 is a diagram illustrating an example of a revised form according to the fourth embodiment. The form 500a illustrates a revised form for the form 500. A form ID 510a, entry field groups 520a and 530a, a notification field 540, and area identification information 550 are printed on the form 500a.

  The form ID 510a is identification information for identifying the type of form. “100101” is printed as the form ID 510a. The first four-digit numerical value “1001” of the form ID 510a indicates that the form 500a is a form for performing a transaction for depositing cash into the account. The last two digits “01” of the form ID 510a indicate that the form 500a is the revised version number “01”.

  The entry column group 520a is an area in which a plurality of entry columns to be input by the customer are printed. The entry column group 520a includes entry columns 521a, 522a, 523a, 524a, 525a, and 526a. The entry column 521a is a column for entering the application date. The entry column 522a is a column for entering the name of the account holder. The entry column 523a is a column for entering a store number. The entry column 524a is a column for entering the account number of the deposit destination account. The entry column 525a is a column for selecting a deposit item of the deposit destination account. The entry column 526a is a column for entering the amount to be deposited.

The entry column group 530a is a column for entering information used by the financial institution staff for business.
The notification column 540 is an area where a message to be notified to the customer is printed.
The entry column group 530a and the notification column 540 are non-target areas for layout recognition processing.

  Dropout colors are given to the frame lines of the fields of the entry field groups 520a and 530a and the notification field 540 and to some portions on the form 500a. The region identification information 550 is assigned a color other than the dropout color.

Next, a processing procedure of the form reading apparatus 100 having the above configuration will be described.
FIG. 17 is a flowchart illustrating a form reading process according to the fourth embodiment. In the following, the process illustrated in FIG. 17 will be described in order of step number.

  (Step S 61) The form ID reading unit 120 stores the form image received from the image scanner 14 in a predetermined area on the RAM 103. The form ID reading unit 120 performs a dropout process on the form image, and stores the form image after the dropout process in another predetermined area on the RAM 103.

  (Step S62) The form ID reading unit 120 recognizes the form ID from the form image after the dropout process. The form ID reading unit 120 outputs the read form ID to the character recognition processing unit 130.

  (Step S63) The character recognition processing unit 130 determines whether or not the last two digits of the form ID acquired from the form ID reading unit 120 is “01” or more. If it is “01” or more (that is, a revised form), the process proceeds to step S65. If it is not “01” or more (that is, not a revised form), the process proceeds to step S64.

  (Step S <b> 64) The character recognition processing unit 130 delegates the processing to the layout recognition processing unit 131 and causes the layout recognition processing to be executed. The layout recognition processing unit 131 refers to the layout definition table 111 stored in the storage unit 110 on the basis of the form ID acquired from the form ID reading unit 120, and performs character recognition by layout recognition on the form image after the dropout process. Perform recognition processing.

(Step S <b> 65) The character recognition processing unit 130 delegates the processing to the keyword recognition processing unit 132 and causes the keyword recognition processing to be executed. The keyword recognition process will be described later.
(Step S66) The character recognition processing unit 130 outputs a character recognition result for the form image and causes the monitor 11 to display the result.

As described above, the character recognition processing unit 130 determines whether the form is revised before performing layout recognition. Then, the keyword recognition process in step S65 is performed as follows.
Hereinafter, it is assumed that the following information is acquired on the RAM 103.

(1) “Form image of form 500a”
(2) “Form image after dropout processing” obtained by executing dropout processing on “form image of form 500a”
In any form image, the region identification information 550 is indicated by the same symbol.

FIG. 18 is a flowchart illustrating keyword recognition processing according to the fourth embodiment. In the following, the process illustrated in FIG. 18 will be described in order of step number.
(Step S71) The keyword recognition processing unit 132 requests the region specifying unit 140 to specify the target region for the keyword recognition processing. The area specifying unit 140 detects the area identification information 550 from the form image after the dropout process. For example, information related to the region identification information 550 (shape type, line thickness, line type such as solid line / broken line, etc.) is preset in the region specifying unit 140. Based on the information, the area specifying unit 140 can detect the area identification information 550 from the form image after the dropout process. The area specifying unit 140 outputs area information indicating the specified processing target area to the keyword recognition processing unit 132. The area information is, for example, the coordinate value of each vertex of the processing target area.

  (Step S <b> 72) The keyword recognition processing unit 132 refers to the region information acquired from the region specifying unit 140 and the layout definition table 111 stored in the storage unit 110 to determine whether or not there is a layout recognition target region. Determine whether. If it exists, the process proceeds to step S73. If not, the process proceeds to step S75. Here, whether or not a layout recognition target area exists can be determined as follows, for example. First, the keyword recognition processing unit 132 identifies a region that is a keyword recognition target based on the region information. Next, the keyword recognition processing unit 132 specifies a layout recognition target area of the form 500 based on the layout definition table 111. Then, the keyword recognition processing unit 132 superimposes the keyword recognition target area and the layout recognition target area, and determines whether or not the layout recognition target area protrudes from the keyword recognition target area. If it protrudes, it can be determined that there is a target area for layout recognition. If it does not protrude, the target area for layout recognition is included in the target area for keyword recognition, so that it can be determined that there is no target area for layout recognition.

  (Step S73) The keyword recognition processing unit 132 outputs information indicating the protruding region (for example, the coordinate value of the vertex of the region) to the layout recognition processing unit 131, and causes the layout recognition processing to be executed. The layout recognition processing unit 131 refers to the layout definition table 111, executes layout recognition processing for the entry fields included in the area, and outputs the result to the keyword recognition processing unit 132.

  (Step S <b> 74) The keyword recognition processing unit 132 saves the layout recognition processing result by the layout recognition processing unit 131 in a predetermined area on the RAM 103 or the HDD 104.

  (Step S75) The keyword recognition processing unit 132 extracts a color image of one unprocessed area from the form image of the form 500a based on the area information acquired from the area specifying unit 140.

  (Step S76) The keyword recognition processing unit 132 extracts one keyword from the processing target area from which the color image has been acquired in Step S75. For example, the keyword recognition processing unit 132 detects a frame line included in the processing target region, and tries to extract a keyword for each frame surrounded by the frame line.

  (Step S77) The keyword recognition processing unit 132 refers to the form ID of the form 500a and the layout definition table 111, and determines whether or not the keyword matches a data item name that can be included in the form 500a. If they match, the process proceeds to step S78. If not, the process proceeds to step S81.

  (Step S78) The keyword recognition processing unit 132 specifies the position (data portion) of the entry field for the data item. For example, the position of the data part is set in advance in the keyword recognition processing unit 132 as “a cell adjacent to the right side of the cell in which the keyword exists”. Alternatively, based on the keyword and the table structure on the form 500a, the data portion corresponding to the keyword may be specified.

  (Step S79) The keyword recognition processing unit 132 analyzes the characteristics of the data portion, removes frame lines and noise from the image of the data portion, and processes the image into an image suitable for character recognition.

  (Step S80) The keyword recognition processing unit 132 acquires a character image entered in the data portion. The keyword recognition processing unit 132 collates the character image with a character pattern stored in the storage unit 110 in advance, and acquires character string data entered in the data unit.

  (Step S <b> 81) The keyword recognition processing unit 132 determines whether or not there is a portion where no keyword has been extracted in the current processing target area. If there is a keyword-unextracted portion, the process proceeds to step S76. If there is no keyword-unextracted portion, the process proceeds to step S82.

  (Step S82) The keyword recognition processing unit 132 determines whether there is an unprocessed processing target area other than the current processing target area. If there is an unprocessed processing target area, the process proceeds to step S75. If processing has been completed for all processing target areas, the process proceeds to step S83.

  (Step S83) The keyword recognition processing unit 132 merges the temporarily saved layout recognition result and the keyword recognition result. Specifically, among the data items of the form 500a defined in the layout definition table 111, those that could not be acquired by the layout recognition process are complemented with the contents acquired by the keyword recognition process.

In this way, the layout recognition processing unit 131 performs character recognition by layout recognition on an area that is not subject to keyword recognition processing.
FIG. 19 is a diagram illustrating a first example of target areas for character recognition according to the fourth embodiment. FIG. 19A shows a form 500 before revision. FIG. 19B shows a revised form 500a.

The form 500 includes a layout recognition target area 560. The layout recognition target area 560 is the same as the area occupied by the entry column group 520.
The form 500a includes a keyword recognition target area 550a. The keyword recognition target area 550a is the same as the area surrounded by the area identification information 550. Here, the form 500 a also shows a frame line of the layout recognition target area 560 in the form 500. In the form 500a, the layout recognition target area 560 is included in the keyword recognition target area 550a. Therefore, the form reading apparatus 100 does not execute the layout recognition process on the form image of the form 500a. The form reading apparatus 100 executes keyword recognition processing on the keyword recognition target area 550a in the form image of the form 500a.

  FIG. 20 is a diagram illustrating a second example of target areas for character recognition according to the fourth embodiment. FIG. 20A shows a form 600 before revision. FIG. 20B shows a revised form 600 a for the form 600.

The form 600 includes a layout recognition target area 610.
The form 600a includes a layout recognition target area 610a and a keyword recognition target area 620. Here, the layout recognition target area 610a is equal to the area obtained by excluding the area overlapping the keyword recognition target area 620 from the layout recognition target area 610 when the forms 600 and 600a are overlapped. The form reading apparatus 100 performs layout recognition on a layout recognition target area 610 a that protrudes from the keyword recognition target area 620 in the layout recognition target area 610. Further, the form reading apparatus 100 executes keyword recognition for the keyword recognition target area 620.

  As described above, according to the form reading apparatus 100 of the fourth embodiment, layout recognition is performed on the keyword recognition processing target area and the non-covered area. Thereby, character recognition can be performed more efficiently.

  Note that there may be a case where a plurality of keyword recognition target areas exist after the revision. In this case, it is determined whether or not there is an overlap between the plurality of keyword recognition target areas and the layout recognition target area before revision. Then, the layout recognition process may be performed on the non-overlapping areas for the revised form.

[Fifth Embodiment]
Hereinafter, a fifth embodiment will be described. Differences from the second to fourth embodiments will be mainly described, and description of similar matters will be omitted.

  In the second to fourth embodiments, in the keyword recognition process, the position of the data part with respect to the detected keyword is specified as a predetermined position, or specified based on the table structure of the form, etc. .

  On the other hand, it is desirable that the data part can be arbitrarily arranged to cope with layout change with a higher degree of freedom. For example, you may want to mix an entry field that associates the lower cell of the cell in which the keyword exists with the data part and an entry field that associates the lower right cell of the cell in which the keyword exists with the data part. It is done. Further, it is desirable that the data part can be easily specified when the data part is arranged at an arbitrary position with respect to the cell in which the keyword exists. This is because, for example, when analyzing the table structure of a form, the load due to the analysis process may increase.

Therefore, in the fifth embodiment, a function is provided for making it possible to cope with a layout change with a higher degree of freedom while making it possible to easily specify the data portion.
Here, the configuration of the information processing system of the fifth embodiment is the same as the configuration of the information processing system of the second embodiment described in FIG. The hardware and functional configuration of the form reading apparatus according to the fifth embodiment is the same as the hardware and functional configuration of the form reading apparatus 100 according to the second embodiment described with reference to FIGS. Hereinafter, the form reading apparatus according to the fifth embodiment also indicates each component using the same reference numeral and name as the form reading apparatus 100.

  FIG. 21 is a diagram illustrating an example of a keyword recognition target area according to the fifth embodiment. The keyword recognition target area is an area inside the area identification information 340. The area includes entry fields 331b, 332b, and 333b. The entry fields 331b, 332b, and 333b correspond to the entry fields 331a, 332a, and 333a described in FIG.

A keyword 351 and support information 351b are printed in the cell of the heading portion of the entry field 331b.
The keyword 351 indicates the character string “phone number”. This corresponds to the data item name that can be included in the form image of the form. The support information 351b is a symbol indicating the position of the data part with respect to the cell in which the keyword 351 exists. The support information 351b is the number “1”, and indicates that the right cell toward the page of FIG. 21 adjacent to the cell in which the keyword 351 exists is the data part. The keyword recognition processing unit 132 can specify the data part for the keyword 351 based on the support information 351b. For example, the keyword recognition processing unit 132 refers to a cell adjacent to the right side with respect to the data item “phone number”, and acquires character string data “000-0000-0000”.

A keyword 352 and support information 352b are printed in the cell of the heading portion of the entry column 332b.
The keyword 352 indicates the character string “sex”. This corresponds to the data item name that can be included in the form image of the form. The support information 352b is a symbol indicating the position of the data part with respect to the cell in which the keyword 352 exists. The support information 352b is a number “2”, and indicates that the cell adjacent to the cell in which the keyword 352 exists and the lower cell toward the page of FIG. 21 is the data part. The keyword recognition processing unit 132 can specify the data part for the keyword 352 based on the support information 352b. For example, the keyword recognition processing unit 132 refers to a cell adjacent on the lower side with respect to the data item “sex”, and acquires a selection symbol (check mark) indicating that “male” is selected.

A keyword 353 and support information 353b are printed in the cell of the heading portion of the entry column 333b.
The keyword 353 indicates the character string “date of birth”. This corresponds to the data item name that can be included in the form image of the form. The meaning of the support information 353b is the same as that of the support information 352b. The keyword recognition processing unit 132 can specify the data part for the keyword 353 based on the support information 353b. Thereby, the keyword recognition processing unit 132 refers to the cell adjacent to the lower side with respect to the data item “birth date”, and acquires the character string data “1985” “11” month “15” day.

  Next, a processing procedure of the form reading apparatus 100 having the above configuration will be described. Note that the procedure of the form reading process of the fifth embodiment is the same as the procedure of the form reading process of the second embodiment described with reference to FIG.

  Further, the procedure of the keyword recognition process of the fifth embodiment is the same as the procedure of the keyword recognition process of the second embodiment described with reference to FIG. However, the process for specifying the data part in step S26 of the keyword recognition process is different. Below, the procedure of the process for specifying this data part is demonstrated.

FIG. 22 is a flowchart illustrating the data portion specifying process according to the fifth embodiment. In the following, the process illustrated in FIG. 22 will be described in order of step number.
(Step S91) The keyword recognition processing unit 132 determines whether or not support information exists in the vicinity of the extracted keyword. If support information exists, the process proceeds to step S92. If support information does not exist, the process proceeds to step S93. Here, whether or not support information exists in the vicinity of the keyword can be determined by the following method, for example. More specifically, the relative position of the support information with respect to the keyword and the format of the support information are set in the keyword recognition processing unit 132 in advance. The keyword recognition processing unit 132 determines whether or not the support information exists by determining whether a character string or the like that matches the format is printed at a predetermined position with respect to the keyword.

  (Step S92) The keyword recognition processing unit 132 specifies a position corresponding to the support information as a data unit. In the example of FIG. 21, based on the support information 351b, the cell adjacent to the right side of the cell in which the keyword 351 exists is specified as the data part. Further, based on the support information 352b, the cell adjacent to the lower side of the cell in which the keyword 352 exists is specified as the data part. Which support information indicates what position is set in the keyword recognition processing unit 132 in advance. Then, the process ends.

  (Step S93) The keyword recognition processing unit 132 specifies a data part for the extracted keyword by a default method. For example, the keyword recognition processing unit 132 specifies a cell at a predetermined position (for example, “a cell adjacent to the right side of the cell in which the keyword exists”) as the data unit. Alternatively, the data portion corresponding to the keyword may be specified based on the keyword or the table structure on the form. Then, the process ends.

In this way, the keyword recognition processing unit 132 can efficiently identify the data part for the keyword by reading the support information printed together with the keyword.
21 illustrates the case where numerical values are used as support information, but other characters (or character strings), symbols, and the like may be used.

  FIG. 23 is a diagram illustrating an example of a data part specifying method according to the fifth embodiment. FIG. 23A illustrates an entry field 700. The entry field 700 includes cells 701 and 702. A cell 701 is a heading cell. In the cell 701, a keyword “transfer amount” and support information “◯” (circle) are printed. The support information “◯” indicates that the cell 702 adjacent to the right side of the cell 701 is a cell of the data portion for the keyword. Therefore, the keyword recognition processing unit 132 performs character recognition on the cell 702 based on the support information “◯” for the entry field 700 and acquires character string data “¥ 11000” for the keyword “transfer amount”.

  FIG. 23B illustrates an entry field 710. The entry field 710 includes cells 711 and 712. A cell 711 is a heading cell. In the cell 711, the keyword “transfer amount” and support information “Δ” (triangle mark) are printed. The support information “Δ” indicates that the cell 712 on the lower right side of the cell 711 is a cell of the data portion for the keyword. Therefore, the keyword recognition processing unit 132 performs character recognition on the cell 712 for the entry field 710 based on the support information “Δ”, and acquires character string data “¥ 11000” for the keyword “transfer amount”.

  FIG. 23C illustrates an entry field 720. The entry field 720 includes cells 721 and 722. A cell 721 is a heading cell. In the cell 721, a keyword “transfer amount” and support information “□” (square mark) are printed. The support information “□” indicates that the cell 722 adjacent to the lower side of the cell 721 is a cell of the data portion for the keyword. Therefore, the keyword recognition processing unit 132 performs character recognition on the cell 722 for the entry field 720 based on the support information “□”, and acquires character string data “¥ 11000” for the keyword “transfer amount”.

  The above is an example, and various numerical values, characters (or character strings), symbols, and the like can be used as the support information. Further, the keyword recognition processing unit 132 may detect support information as a frame line, a region within the frame, a keyword format, and the like, and specify the position of the data unit. Next, an example of such support information will be described.

  FIG. 24 is a diagram illustrating another example of the data portion specifying method according to the fifth embodiment. FIG. 24A illustrates an entry field 800. The entry field 800 includes cells 801 and 802. A cell 801 is a heading cell. In the frame of the cell 801, a ruled line parallel to the one side is printed on one side corresponding to the direction of the data portion from the center of the frame. Specifically, when the cell of the data part for the cell 801 is a cell 802 adjacent to the right side of the cell 801, the cell 801 side (however, the center of the cell 801 is located at the border between the cells 801 and 802). A ruled line 801a parallel to the frame line is printed on the cell 802 side. The keyword recognition processing unit 132 detects the ruled line 801a as support information. Then, the keyword recognition processing unit 132 identifies the cell 802 as a data portion based on the position of the ruled line 801a in the cell 801. Thus, the keyword recognition processing unit 132 acquires the character string data “¥ 11000” in the cell 802 for the keyword “transfer amount” in the cell 801.

  For example, when a cell adjacent to the lower side of the cell 801 is used as the data portion, a ruled line parallel to the lower frame line of the cell 801 is printed on the lower side in the cell 801. The keyword recognition processing unit 132 can specify a cell adjacent to the lower side of the cell 801 as a data portion based on the ruled line.

  FIG. 24B illustrates an entry field 810. The entry field 810 includes cells 811 and 812. A cell 811 is a heading cell. A cell 812 is a data portion cell corresponding to the keyword in the cell 811. A boundary line 811a between the cell 811 and the cell 812 is printed with a thickness different from that of the other frame lines. The keyword recognition processing unit 132 specifies the position of the data part based on the boundary line 811a having a different thickness. Specifically, the cell 812 adjacent in the direction of the boundary line 811a of the cell 811 is specified as the data portion. Accordingly, the keyword recognition processing unit 132 acquires the character string data “¥ 11000” in the cell 812 for the keyword “transfer amount” in the cell 811.

  FIG. 24C illustrates an entry field 820. The entry field 820 includes cells 821 and 822. A cell 821 is a heading cell. In the cell 821, “transfer amount” is printed as a character string in which keywords are arranged in a horizontal row. The character strings arranged in a horizontal row indicate that the cell adjacent to the right side of the cell 811 is a data portion corresponding to the keyword. The keyword recognition processing unit 132 detects the keyword “transfer amount” in the cell 821 and further detects that the keywords are arranged in a horizontal row. Then, the keyword recognition processing unit 132 specifies the cell 822 adjacent to the right side of the cell 821 as the data portion. Accordingly, the keyword recognition processing unit 132 acquires the character string data “¥ 11000” in the cell 822 for the keyword “transfer amount” in the cell 821.

  For example, when a cell adjacent to the lower side of the cell 821 is used as a data portion, the keywords in the cell 821 are printed in a vertical line. Further, for example, when the cell on the lower right side of the cell 821 is used as the data portion, the keywords are printed in a diagonal line from the upper left of the cell 821 to the lower right. The keyword recognition processing unit 132 specifies the cell of the data part for the keyword according to the arrangement of the characters included in the keyword.

  FIG. 24D illustrates the entry field 830. The entry field 830 includes cells 831 and 832. A cell 831 is a heading cell. Cell 832 is a cell of the data part corresponding to the keyword in cell 831. A predetermined color is given in the frame of the cell 831. In the entry field 830, the position of the cell 832 with respect to the cell 831 is indicated by this color. For example, in the image information, the color is expressed by a plurality of parameters. For example, in color representation by HSV, a color is represented by three parameters, hue (hue), saturation (saturation value), and lightness (value). In that case, it is defined in advance whether the cell adjacent to the upper, lower, left, or right side of the cell 831 is the data portion by any one of the parameters or a plurality of parameters. For example, it is conceivable to divide the hue range into four equal parts and assign each of the equally divided ranges to the top, bottom, left or right. The same applies when using saturation and lightness.

  The color in the cell 831 is preset in the keyword recognition processing unit 132 to the effect that the cell 832 adjacent to the right side of the cell 831 is a data part corresponding to the cell 831. The keyword recognition processing unit 132 identifies the cell 832 of the data part corresponding to the cell 831 by analyzing the color in the cell 831. Accordingly, the keyword recognition processing unit 132 acquires the character string data “¥ 11000” in the cell 832 for the keyword “transfer amount” in the cell 831.

If the image is binarized, the data portion may be specified according to the type of hatching, the dot density, and the like.
As described above, the form reading device 100 detects the support information printed together with the keyword, and specifies the cell of the data part. Thereby, it is possible to change the layout with a high degree of freedom for the form. In addition, in the keyword recognition process, the data portion for the keyword can be easily specified.

DESCRIPTION OF SYMBOLS 1 Information processing apparatus 1a Detection part 1b Area | region specific part 1c Processing part 2 Image information 2a Identification information 2b Area identification information 3 Processing object area | region

Claims (7)

Detecting identification information indicating that the layout of the form has been changed from image information obtained by imaging the form;
When the identification information is detected, area identification information included in the image information for identifying a processing target area for performing a predetermined character recognition process is detected, and the processing target area is specified based on the area identification information. ,
Performing the predetermined character recognition processing on the identified processing target area;
An information processing program that causes a computer to execute processing.   The information processing program according to claim 1, wherein the use of the data acquired by the predetermined character recognition process is distinguished according to the area identification information.   When the identification information is not detected, a character recognition process other than the predetermined character recognition process is performed on the entire area of the image information. When the identification information is detected, the other area is detected in an area other than the process target area. The information processing program according to claim 1, wherein character recognition processing is performed.   The other character recognition processing is performed based on the data item name and the character based on the layout definition information stored in a storage unit that stores layout definition information in which a data item name and a character recognition target position are defined in association with each other. The information processing program according to claim 3, which is a character recognition process by layout recognition that is acquired in association with data extracted from a recognition target position.   The predetermined character recognition process detects a predetermined keyword and support information for specifying a position of a data portion corresponding to the keyword in the image information from the image information, and based on the support information, the data portion The information processing program according to any one of claims 1 to 4, which is a process of identifying the data and acquiring data corresponding to the keyword from the data portion. A detection unit that detects identification information indicating that the layout of the form has been changed from image information obtained by imaging the form;
When the identification information is detected, area identification information included in the image information for identifying a processing target area for performing a predetermined character recognition process is detected, and the processing target area is specified based on the area identification information. An area identification unit;
A processing unit that performs the predetermined character recognition processing on the identified processing target area;
An information processing apparatus. A character recognition method executed by an information processing apparatus,
When identification information indicating that the layout of the form has been changed is detected from the image information obtained by imaging the form, the process includes a process for identifying a processing target area to be subjected to a predetermined character recognition process included in the image information. Detecting region identification information, specifying the processing target region based on the region identification information,
Performing the predetermined character recognition processing on the identified processing target area;
Character recognition method.

Images