JP2015005100A - Information processor, template generation method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a template where a character region previously described in a document and a region on which a user can enter characters are segmented with high accuracy.SOLUTION: A character discrimination section 24 discriminates fixed characters previously described in a document and characters entered by a user from a plurality of pieces of image information of the document. A region discrimination section 25 discriminates an entry region where the user can enter characters on the document, and a non-entry region where the user cannot enter characters on the document on the basis of the fixed characters and the entered characters discriminated by the character discrimination section 24. A generation section 27 generates a template for the document on the basis of the entry region and the non-entry region discriminated by the region discrimination section 25.

Description

  The present invention relates to an electronic form technology.

2. Description of the Related Art Conventionally, there is a system for digitizing paper media using an OCR (Optical Character Reader) or the like. Thereby, for example, a paper medium form can be digitized, or a digitized form template can be generated.
Conventionally, an information processing system, an information processing apparatus, an information processing method for outputting description position information of an entry area described in a form such as position information and format information setting work, and format information of metadata relating to the entry area, And a program are provided (see, for example, Patent Document 1).

JP 2009-238217 A

  When an electronic form template is generated from a paper medium form using OCR or the like, the form template has a high character area preliminarily written in the form and an area in which the user can enter characters. It is desired to be classified by accuracy.

  SUMMARY OF THE INVENTION An object of the present invention is to generate a template in which a character area described in advance on a form and an area in which a user can enter a character are separated with high accuracy.

  The present application includes a plurality of means for solving at least a part of the above-described problems. Examples of such means are as follows. In order to solve the above-described problem, an information processing apparatus according to the present invention, from the image information of a plurality of forms, a character identification unit that identifies fixed characters previously written in the form and entry characters entered by the user, Based on the fixed character and the input character identified by the character identification unit, an area identification unit for identifying an entry area in which a user can enter a character in the form and a non-entry area in which the user cannot enter a character in the form; A generating unit that generates a template of the form based on the entry area and the non-entry area identified by the area identification unit.

  In the above information processing apparatus, the character identification unit identifies the character string as a fixed character when the same character string is included in the same position in a predetermined number of the same type of forms. It may be.

  In the information processing apparatus, the area identification unit identifies a character string area of the form identified as a fixed character by the character identification unit as a non-entry area, and the character identification unit identifies The identified character string area of the form may be identified as an entry area.

  In the information processing apparatus, the character identification unit further uses the item information including an item described in advance in the form and an entry entered by the user corresponding to the item. May be used to identify fixed characters and written characters described in the above.

  In the information processing apparatus, the character identification unit identifies the character string of the form corresponding to the item information item as a fixed character, and the character string of the form corresponding to the entry item of the item information. May be identified as an entry character.

  The information processing apparatus further includes a storage unit that stores a learning result related to identification of the character string of the form that is identified as a fixed character by the character identification unit, and the character identification unit includes the storage unit The fixed character described in the form may be identified using the learning result stored in the form.

  In the information processing apparatus, the learning result is a probability or a score of the character string of the form identified as a fixed character by the character identification unit, and the character identification unit is described in the form If the probability or score of a character string is greater than or equal to a predetermined value, it may be identified as a fixed character.

  In the information processing apparatus, the learning result is a probability or score of a character string of the form identified as a fixed character by the character identification unit, and the character identification unit stores the item information and the storage One or both of the probabilities and scores stored in the section may be used to identify fixed characters and written characters described in the form.

  In the above information processing apparatus, the image information of the form is grouped based on a predetermined method, and when fixed characters and entry characters belong to one group, they belong to different groups. And a grouping correction unit for dividing the group.

  Further, the template generation method according to the present invention includes a character identification step for identifying a fixed character previously written in the form and an entry character entered by the user from the image information of a plurality of forms, and the character identification step. Based on the identified fixed character and entry character, an area identification step for identifying an entry area in which the user can enter a character in the form and a non-entry area in which the user cannot enter the form, and an identification by the area identification step Generating a template for the form based on the filled-in area and the non-filled area.

  In addition, the program of the information processing apparatus according to the present invention includes a character identification step for identifying, from image information of a plurality of forms, a fixed character described in advance in the form and an entry character entered by the user, and the character identification An area identification step for identifying an entry area in which a user can enter a character in the form and a non-entry area in which the user cannot fill in the form based on the fixed character and the entry character identified in the step; and the area identification step And generating the template for the form based on the entry area and the non-entry area identified by the information processing apparatus.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is a figure explaining the outline | summary of the information processing apparatus which concerns on one Embodiment of this invention. It is the figure which showed an example of the functional block of the information processing apparatus of FIG. It is the figure which showed the hardware structural example of the information processing apparatus of FIG. 3 is a flowchart illustrating an operation example of the information processing apparatus in FIG. 1. It is the figure which showed the example of the form data-converted by a data conversion part. It is the figure which showed an example of the character string process TB. It is the figure which showed an example of area | region process TB. It is the figure which showed an example of supplementary data. It is the figure which showed an example of the character string process TB to which supplementary data was added. It is a figure explaining the character identification by several form data. It is the figure which showed an example of learning information TB. It is the figure which showed an example of the character string process TB based on learning information TB. It is the figure which showed an example of the character string process TB. It is the figure which showed an example of the determination process TB. It is the figure which showed an example of the determination process TB after grouping correction | amendment. It is a figure explaining a variable-length area | region. FIG. 17 is a diagram illustrating an example of a form determination process TB in FIG. 16. FIG. 17 is a diagram illustrating an example of a determination process TB for the form of FIG. 16 that is grouped by mistake. It is the figure which showed an example of the determination process TB after grouping correction | amendment. It is the figure which showed the example of the electronic form. It is the figure which showed the example of the form template. It is the figure which showed the example of template information.

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

  FIG. 1 is a diagram illustrating an overview of an information processing apparatus according to an embodiment of the present invention. In FIG. 1, in addition to the information processing apparatus 1, a form 11 and supplemental data 12 input to the information processing apparatus 1 are shown. FIG. 1 also shows an electronic form 13, form template 14, and template information 15 generated by the information processing apparatus 1.

  The form 11 is, for example, a form such as a transportation expense application slip or a delivery slip. The form 11 is, for example, a paper medium form.

  The supplementary data 12 is item information regarding items described in the form 11. For example, the supplementary data 12 includes items previously described in the form 11 and item contents entered by the user corresponding to the items.

  Specifically, the item included in the supplementary data 12 is a character string such as “application date”, “applicant”, or “application fee” previously described (for example, printed) in the form 11. The item content is, for example, the application date entered by the user corresponding to the item “application date”, the user name entered by the user corresponding to the item “applicant”, or the item “application cost” The amount of transportation expenses entered by the user in response to “

  The supplementary data 12 is described by the user in XML (Extensible Markup Language), for example. Alternatively, the supplemental data 12 may be CSV (Comma Separated Values), a text file, or the like.

  In the following, an item preliminarily described in the form 11 may be referred to as a fixed character, and an item content entered in the form 11 by the user corresponding to the item may be referred to as an entry character.

  The information processing apparatus 1 reads a plurality of forms 11 using, for example, the OCR function. Further, supplementary data 12 corresponding to the plurality of read forms 11 is input to the information processing apparatus 1. The information processing apparatus 1 generates a computerized form 13, a form template 14, and template information 15 based on the plurality of read forms 11 and the supplementary data 12 corresponding to each read form 11.

  The electronic form 13 is an electronic form of the form 11. The digitized form 13 is generated, for example, as a document file or an image file. By replacing the paper medium form 11 with the electronic form 13, the paper medium form 11 can be electronically managed.

  The form template 14 is an electronic template of the form 11. For example, the form template 14 is a file in which user entry characters are removed from the form 11 and the removed entry character area can be edited by the user. The form template 14 can be displayed on a display device of a terminal device, for example, and the user can enter a predetermined item in the form template 14 on the terminal device.

  Template information 15 is format information of the form template 14. The template information 15 includes format information such as a character string, position, format, page margin, paper size, or ruled line format described in the form template 14. The template information 15 can be used as a system development resource for the form 11 together with the form template 14, for example.

  In the above, the information processing apparatus 1 generates the electronic form 13, the form template 14, and the template information 15 based on the plurality of forms 11 and the supplementary data 12, but without inputting the supplementary data 12, Based on a plurality of forms 11, an electronic form 13, form template 14, and template information 15 can also be generated. That is, the information processing apparatus 1 can generate the digitized form 13, the form template 14, and the template information 15 from the plurality of forms 11 even if the supplementary data 12 is not input.

  FIG. 2 is a diagram illustrating an example of functional blocks of the information processing apparatus in FIG. As illustrated in FIG. 2, the information processing apparatus 1 includes a data conversion unit 21, an input unit 22, a character correction unit 23, a character identification unit 24, a region identification unit 25, a grouping correction unit 26, a generation unit 27, and a storage unit 28. have.

  The data conversion unit 21 identifies characters described in the form 11. For example, the data conversion unit 21 has an AD (Analog-Digital) conversion function, scans information such as character strings and ruled lines described in the form 11, and converts them into digital data.

  The data conversion unit 21 has an OCR function and recognizes character formats such as characters, character positions, or character font type and size of the form 11 converted into digital data. Further, the data conversion unit 21 has a general-purpose image identification processing function, recognizes the page margin, paper size, paper orientation, ruled line, ruled line position, and ruled line format of the form 11, and has an area of the ruled line as a reference. Perform grouping processing.

  Supplemental data 12 is input to the input unit 22. The information processing apparatus 1 can generate a high-quality digitized form 13, form template 14, and template information 15 even if the supplementary data 12 is not input to the input unit 22, but the supplementary data 12 is input. Then, it is possible to generate a digitized form 13, a form template 14, and template information 15 with higher quality.

  When the data conversion unit 21 erroneously converts a character, the character correction unit 23 corrects the erroneously converted character to a correct character. The character correction unit 23 detects and corrects erroneously converted characters from the data of the plurality of forms 11 subjected to OCR processing. Alternatively, when the supplementary data 12 is input to the input unit 22, the character correction unit 23 uses the supplementary data 12 to detect and correct erroneously converted characters.

  The character identification unit 24 determines whether the character string written in the OCR-processed form 11 is a fixed character (item of the form 11) or a character entered by the user (item content of the form 11). Identify. The character identifying unit 24 can identify, for example, a fixed character and a written character by the following three methods. In addition, although the character identification part 24 can identify a fixed character and an entry character only by the method 1, you may identify a fixed character and an entry character using all the methods 1-3. In addition, the character identification unit 24 may identify a fixed character and a written character by Method 1 and Method 2 or Method 1 and Method 3.

1. Identification using data of a plurality of forms 11 A fixed character described in a form 11 is generally written in the same position at the same position across the plurality of forms 11. On the other hand, since the entry characters described in the form 11 are entered by the user, the contents may be different across a plurality of forms 11. Thereby, the character identification part 24 can identify a fixed character and an entry character from the data of a plurality of forms 11.

  For example, when the same character string is written at the same position over a predetermined number of forms 11 of the same type, the character identification unit 24 identifies the character string at that position as a fixed character. In addition, when different character strings are written at the same position across a predetermined number of forms 11 of the same type, the character identification unit 24 identifies the character string at that position as an entry character. The predetermined number may be two or more, for example, but the larger the number, the better.

2. Identification by Learning The character identification unit 24 learns the recognized fixed character and identifies the fixed character and the entry character of the form 11.

For example, the character identification unit 24 stores a learning result related to identification of a character string of a form identified as a fixed character in the storage unit 28. More specifically, the character identification unit 24 stores, in the storage unit 28, a learning table indicating the probability that the OCR-processed character string has been identified as a fixed character in the past. The character identification unit 24 refers to the learning table stored in the storage unit 28 and identifies whether the character string of the form 11 subjected to the OCR process is a fixed character or an input character.
In addition, although learning of said fixed character is demonstrated using the probability, a score may be sufficient. For example, the learning result obtained by the learning algorithm may be indicated by a score. The same applies to the learning described below.

3. Identification Using Supplementary Data 12 As described with reference to FIG. 1, the supplementary data 12 includes information on items (fixed characters) of the form 11 and item contents (entry characters). The character identifying unit 24 refers to the supplementary data 12 including information on fixed characters and entry characters, and identifies whether the character string subjected to OCR processing is a fixed character or an entry character.

  For example, the character identification unit 24 compares the character example of the form 11 that has been subjected to the OCR processing with the input supplemental data 12. When the character string of the form 11 that has been subjected to the OCR processing matches the item of the supplementary data 12, the character identification unit 24 identifies the character string as a fixed character. In addition, when the character string of the OCR-processed form 11 matches the item content of the supplementary data 12, the character identification unit 24 identifies the character string as an entry character.

  The character identification unit 24 identifies whether the character is a fixed character or an input character by each of the above three methods, and integrates the three identification results to identify the final fixed character and the input character. I do. Note that the character identification unit 24 may identify a fixed character and a written character only by the result of the method 1 without integrating the three identification results. Moreover, the character identification part 24 may integrate the result of the method 1 and the method 2, and may identify a fixed character and a written character. Moreover, the character identification part 24 may integrate the result of the method 1 and the method 3, and may identify a fixed character and an entry character.

  The area identifying unit 25 identifies an entry area in which the user can fill in the form 11 and a non-filled area in which the user cannot fill in the form 11 based on the fixed character and the entered character identified by the character identifying unit 24.

  For example, the area identifying unit 25 identifies the area of the character string of the form 11 identified as a fixed character by the character identifying unit 24 as a non-filled area. Further, the area identifying unit 25 identifies the area of the character string of the form 11 identified as the entered character by the character identifying unit 24 as the entered area.

  The grouping correction unit 26 corrects the grouping processed by the data conversion unit 21.

  For example, as described above, the data conversion unit 21 performs a grouping process of regions based on ruled lines. In the grouping performed by the data conversion unit 21, a fixed character and an entry character may belong to one group. In this case, the grouping correction unit 26 divides the group so that the fixed character and the entered character are in separate groups. In other words, the grouping correction unit 26 causes the area in which the user can enter characters and the area in which characters cannot be entered in the generated form template 14 to be different.

  The generation unit 27 generates the electronic form 13 of the form 11 based on the data converted by the data conversion unit 21.

  The generation unit 27 generates the form template 14 of the form 11 based on the data converted by the data conversion unit 21. At that time, the generation unit 27 deletes the character string in the entry area of the form 11 and generates the form template 14 based on the entry area and the non-entry area identified by the area identification unit 25.

  Further, the generation unit 27 generates template information 15 of the form template 14 to be generated. That is, the generation unit 27 generates format information of the form template 14.

  Data TB (TB: Table) is temporarily stored in the storage unit 28. Each unit of the information processing apparatus 1 generates data TB in the storage unit 28 or refers to the generated data TB when executing processing.

  Next, a hardware configuration example for realizing the functions of the functional block example shown in FIG. 2 will be described.

  FIG. 3 is a diagram illustrating a hardware configuration example of the information processing apparatus of FIG. As shown in FIG. 3, the information processing apparatus 1 includes a CPU (Central Processing Unit) 31, a RAM (Random Access Memory) 32, a ROM (Read Only Memory) 33, an HDD (Hard Disk Drive) 34, a scanner 35, and a drive 36. , A communication interface 37, and a bus 38.

  The CPU 31 controls the information processing apparatus 1 as a whole. A RAM 32, a ROM 33, an HDD 34, a scanner 35, a drive 36, and a communication interface 37 are connected to the CPU 31 via a bus 38.

  The RAM 32 and the ROM 33 temporarily store at least a part of an OS (Operating System) program to be executed by the CPU 31 and an application program for generating the computerized form 13, the form template 14, and the template information 15. . The RAM 32 stores various data necessary for processing by the CPU 31.

  The ROM 33 and HDD 34 store the OS, application programs, and various data.

  The scanner 35 is a device that reads information using light or magnetism. For example, the scanner 35 reads characters, ruled lines, and the like described in the form 11 with light.

  The drive 36 is a device that writes data to an external storage medium and reads data stored in an external storage medium. The drive 36 is, for example, a CD (Compact Disc) drive or a DVD (Digital Versatile Disc) drive.

  The communication interface 37 is connected to a network and communicates with an external device in accordance with an instruction from the CPU 31. The communication interface 37 is, for example, a LAN (Local Area Network) adapter.

  The configuration of the information processing apparatus 1 shown in FIG. 2 and FIG. 3 is classified according to main processing contents in order to facilitate understanding. The present invention is not limited by the way of classification and names of the constituent elements. For example, the configuration of the information processing apparatus 1 can be classified into more components depending on the processing content. Moreover, it can also classify | categorize so that one component may perform more processes. The processing of each component may be executed by one hardware, or may be executed by a plurality of hardware. Further, the processing of each component may be realized by one program or may be realized by a plurality of programs.

  The processing function of the information processing apparatus 1 can be realized by a computer. In that case, a program describing the processing contents of the functions of the information processing apparatus 1 is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic storage device, an optical disk, a magneto-optical recording medium, and a semiconductor memory. Examples of the magnetic storage device include a hard disk drive, a flexible disk (FD), and a magnetic tape. Examples of the optical disc include Blu-ray (registered trademark), DVD (Digital Versatile Disc), DVD-RAM, CD-ROM (Compact Disc Read Only Memory) / RW (ReWritable), and the like. Examples of the magneto-optical recording medium include an MO (Magneto-Optical disk).

  When distributing the program, for example, a portable recording medium such as a DVD or a CD-ROM in which the program is recorded is sold. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. In addition, each time a program is transferred from a server computer connected via a network, the computer can sequentially execute processing according to the received program.

  Further, at least a part or all of the above processing functions can be realized by an electronic circuit such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or a PLD (Programmable Logic Device).

  The operation of the information processing apparatus 1 will be described using a flowchart.

  FIG. 4 is a flowchart illustrating an operation example of the information processing apparatus of FIG. In the flowchart of FIG. 4, for example, processing is started in response to an instruction from the user.

  [Step S1] The data converter 21 converts the paper medium form 11 into electronic data, and generates character string processing TB and area processing TB described below. Before describing the character string process TB and the area process TB, the form 11 converted into electronic data will be described.

  FIG. 5 is a diagram showing an example of a form whose data is converted by the data conversion unit. A form 11 shown in FIG. 5 shows an example of a transportation expense application form.

  “Transportation cost application form”, “application date”, “application name”, “applicant”, “application cost”, and “reason:” shown in FIG. That is, these characters are items and fixed characters.

  “October 10, 2012”, “Application for Transportation Expenses for XX Prefecture”, “Taro XX”, “Jiro XX”, “¥ 1,000”, and “ “For” is a character entered by the user in the form 11. That is, these characters are item contents and entry characters. The data conversion unit 21 converts the form 11 as shown in FIG.

  The character string processing TB will be described. The data conversion unit 21 recognizes the character string of the form 11 and acquires various format information. For example, the data conversion unit 21 recognizes the character string of the form 11 by the OCR function, and acquires the character format such as the position, font type, and size of the character string. The data conversion unit 21 generates a character string process TB based on the acquired character string information.

  FIG. 6 is a diagram illustrating an example of the character string process TB. For example, the data conversion unit 21 recognizes the character string of the form 11 shown in FIG. 5 and generates a character string process TB41 as shown in FIG.

  The number assigned to the recognized character string is stored in the “No.” field of the character string processing TB41.

  The character string recognized from the form 11 is stored in the “character string” column. For example, in the form 11 shown in FIG. 5, a character string “application name” or “application for XX prefecture inspection transportation expenses” is described. Therefore, these character strings are stored in the column of the character string processing TB41.

  The “position” column stores position information of the recognized character string. The “position” column is further divided into a “Y” column and an “X” column. In the “Y” column, position information in the vertical direction of the recognized character string is stored. In the “X” column, position information in the horizontal direction of the recognized character string is stored.

  The area process TB will be described. The data conversion unit 21 has an image identification processing function, recognizes the page margin, paper size, paper orientation, ruled line, ruled line position, and ruled line format of the form 11, and performs a grouping process of areas based on the ruled line. There are various methods for grouping. Here, for example, the data conversion unit 21 groups an area surrounded by a ruled line into one area. The data conversion unit 21 generates a region process TB based on the grouping process.

  FIG. 7 is a diagram illustrating an example of the area process TB. For example, the data converter 21 recognizes the area of the form 11 shown in FIG. 5 and generates the area process TB 42 as shown in FIG.

  The number assigned to the grouped character string is stored in the “No.” column of the area process TB42.

  In the “character string” column, the character strings grouped in the form 11 are stored. For example, in the case of the example in FIG. 5, the “application name” is surrounded by one ruled line frame and grouped in the column “No. 1” in FIG. Further, “Reason: For XX prefecture inspection” in FIG. 5 is grouped in the column of “No. 7” in FIG. 7 because it is described in a grouped area outside the ruled line frame.

  The “start position” field stores the start position of the grouped area. The “end position” column stores the end position of the grouped area. The grouping area is, for example, a quadrangular shape, and the “start position” and “end position” are indicated by vertices on a square diagonal line. In the example of FIG. 7, “application names” are grouped as one group (area), the start position of the area is (Y, X) = (20, 10), and the end position is (X, Y) = (30, 40).

  In this manner, the data conversion unit 21 scans the paper medium form 11 and generates the character string process TB 41 and the area process TB 42.

  The data conversion unit 21 converts a plurality of the same type of forms 11 into electronic data. For example, the data conversion unit 21 performs data conversion on a plurality of forms 11 having the same format (for example, items and ruled lines are the same) as the form 11 shown in FIG. Then, the data conversion unit 21 generates a character string process TB41 and an area process TB42 for each of the plurality of forms 11.

  Returning to the description of FIG.

  [Step S2] Next, the input unit 22 receives supplementary data 12 corresponding to the form 11 to be converted. When the supplementary data 12 is input, the input unit 22 adds the information of the input supplemental data 12 to the character string process TB 41 generated by the data conversion unit 21. Before explaining the character string processing TB to which the information of the supplementary data 12 is added, the supplementary data 12 will be explained.

  FIG. 8 is a diagram showing an example of supplementary data. Supplementary data 12 shown in FIG. 8 shows an example of supplementary data of the form 11 described in FIG.

  The supplementary data 12 includes information on items and item contents. The item is described between tags <item name> and </ item name>, for example, and the item content is described between tags <content> and </ content>, for example.

  Specifically, the “application name” shown in FIG. 5 is an item, and “XX prefecture inspection transportation fee application” is the item content. Accordingly, in the supplementary data 12 in the example of FIG. 5, as shown in FIG. 8, “application name” is described between <item name> and </ item name>, and <content> and </ content Between “and” is the “Application for transportation inspection for XX prefecture”.

  The supplementary data 12 only needs to identify the item and the item content. Accordingly, the supplementary data 12 is created by, for example, XML, CVS, a text file, or the like.

  The supplementary data 12 is for generating a digitized form 13, a form template 14, and template information 15 with higher quality. Therefore, the information processing apparatus 1 generates the computerized form 13, form template 14, and template information 15 of the form 11 from a plurality of the same type of forms 11 even if supplementary data is not input to the input unit 22. Can do.

  Such supplementary data 12 for improving quality may include items that are fixed characters of the form 11 and item contents that are written characters entered by the user. Therefore, the supplementary data 12 may not include information on the absolute position / relative position of items depending on the layout of the form 11 or information on the character type depending on the system environment. Of course, the supplementary data 12 may include other additional information in order to improve the quality of information to be generated.

  The character string process TB to which the supplementary data 12 information is added will be described. As described above, when the supplementary data 12 is input, the input unit 22 adds the information of the supplementary data 12 to the character string process TB generated by the data conversion unit 21.

  FIG. 9 is a diagram showing an example of a character string process TB to which supplementary data is added. As shown in FIG. 9, the character string processing TB43 to which the supplementary data 12 is added has a “conversion result” column and a “supplemental data information” column.

  The information in the “conversion result” column corresponds to the information on the character string processing TB generated by the data conversion unit 21. For example, the information in the “conversion result” column of FIG. 9 corresponds to the information of the character string processing TB 41 described in FIG.

  In the “supplemental data information” field, information on the supplementary data 12 input to the input unit 22 is stored. The “supplemental data information” field includes a “character string” field, an “item name” field, an “item / content category” field, and a “fixed character probability” field.

  In the “character string” column, items of supplementary data 12 and character strings of item contents are stored. For example, in the example of FIG. 8 described above, the character string of the item is “application name”, and the character string of the item content is “XX prefecture inspection transportation fee application”. Therefore, in the “Character string” column of the “Supplementary data information” column in FIG. 9, “Application name” and “Application for transportation expenses for XX prefecture inspection” are stored. The character string stored in the “character string” field in the “supplemental data information” field is stored in correspondence with the character string in the “character string” field in the “conversion result” field.

  In the “item name” column, names of items and item contents are stored. For example, the item name of the item “application name” is “application name”. In addition, the item name of “Application for Expenses for Exploring XX Prefecture” is “Application Name”.

  In the “item / content category” column, information indicating whether the character string stored in the “character string” column in the “supplemental data information” column is an item or item content is stored. For example, the “application name” is an item from the example of the supplementary data 12 shown in FIG. 8, and therefore, an “item” indicating an item is stored in the “item / content category” column. In addition, “XX prefecture inspection transportation fee application” is an item content from the example of the supplementary data 12 shown in FIG. 8, and therefore, in the “item / content category” column, “content” indicating the item content is displayed. Is stored.

  In the “fixed character probability” column, the probability that the character string described in the “character string” column in the “supplemental data information” column is a fixed character is stored. For example, since “application name” is an item, it is a fixed character. Therefore, the information “100%” indicating that the fixed character rate is 100% is stored in the “fixed character probability” column corresponding to the character string “application name”. In addition, “XX prefecture inspection transportation fee application” is an entry character, because it is an item content. Therefore, the information “0%” indicating that the fixed character rate is 0% is stored in the “fixed character probability” column corresponding to the character string “application for XX prefecture inspection and transportation expenses”.

  As described above, when the supplementary data 12 is input, it is possible to accurately know whether the character string of the form 11 converted by the data converter 21 is a fixed character or an input character.

  Returning to the description of FIG. The processing order of the above steps S1 and S2 may be reversed or simultaneous.

  [Step S3] Next, the character correction unit 23 detects an erroneous conversion of the character string by the data conversion unit 21 and corrects it. The character correction unit 23 can detect and correct the erroneously converted character string by the following two methods. The character correction unit 23 may detect and correct a character string erroneously converted only by the method 1. However, when the supplementary data 12 is input, the character correction unit 23 detects the erroneously converted character string by the method 2. And correction may be performed. In addition, when the supplementary data 12 is input, the character correction unit 23 may detect and correct the erroneously converted character string by the method 1 and the method 2.

  1. Correction by multiple forms

  The character correction unit 23 refers to the character string processing TB 41 of the plurality of forms 11 generated by the data conversion unit 21 and corrects erroneous conversion of the data conversion unit 21.

  For example, as described above, the data conversion unit 21 converts a plurality of the same type of forms 11 and generates a plurality of character string processes TB41 as shown in FIG. The character correction unit 23 compares a plurality of generated character string processes TB41 and corrects them to a character string that seems to be the most correct, for example, by a method based on probability theory or rule method.

  Specifically, consider a case where six character string processes TB41 of A, B, C, D, E, and F are generated. Of these six character string processes TB41, for example, a character string included in A, D, and F is “date”. Further, it is assumed that the character string included in B and C among these six character string processes TB41 is “dete”. Further, it is assumed that the character string included in E among these six character string processes TB41 is “dot”. In this case, the appearance probability of “date” is the highest at 3/6. Accordingly, in this case, the character correction unit 23 corrects the character strings “dete” and “date” to “date”.

  Alternatively, the character correction unit 23 may collect and correct the accuracy of character identification by the OCR function of the data conversion unit 21. For example, when the accuracy of “date” in A is 90% and the accuracy of “dete” in B is 70%, “date” may be corrected as a correct character.

  2. Correction by supplementary data

  When the supplementary data 12 is input, the character correction unit 23 can correct erroneous conversion of the data conversion unit 21 using the input supplemental data 12.

  For example, in the supplementary data 12, the user directly specifies whether the character string of the form 11 is a fixed character or an entry character. Therefore, the character correction unit 23 can detect and correct the character string of the form 11 that has been erroneously converted by comparing the supplementary data 12 and the character string of the form 11 that has been converted.

  Specifically, the character correction unit 23 detects the erroneous conversion of the data conversion unit 21 with reference to the character string process TB 43 (character string process TB to which the supplementary data 12 is added) described in FIG. To do. For example, in the case of the example in FIG. 9, it is assumed that “application name” is erroneously converted to “subject name” as shown in the “character string” column of the “conversion result” column. In this case, the character correction unit 23 can detect an erroneous conversion in comparison with the “application name” described in the “character string” column of the “supplemental data information” column of the character string processing TB 43. Can be corrected to the correct “Sen” character.

  Returning to the description of FIG.

  [Step S <b> 4] Next, the character identifying unit 24 identifies a fixed character previously written in the form 11 and a character entered by the user in the form 11. The character identification unit 24 identifies whether the character is a fixed character or an entry character using the following three methods.

  1. Identification with multiple forms

  In general, in the same type of form 11, fixed characters such as a headline and a note are written in the same position at the same position even in another form 11. On the other hand, even if the entry character entered by the user is the same form 11, the character string may be different if it is another form 11.

  Therefore, the character identification unit 24 determines the difference between the character strings among the plurality of forms 11, identifies those that have no difference as fixed characters, and identifies those that have a difference as entry characters. Note that the greater the number of forms whose data is converted by the data conversion unit 21, the higher the identification accuracy of the character identification unit 24. This is because if the number of forms 11 that have undergone data conversion is small, the user entry may happen to be the same character string.

  FIG. 10 is a diagram for explaining character identification using a plurality of form data. FIG. 10 shows character string processing TBs 44a, 44b, and 44c. FIG. 10 shows character string comparison TBs 45a and 45b generated when identifying fixed characters and entry characters. FIG. 10 shows a character string process TB46 showing the identification result of the fixed character and the entry character generated based on the character string comparison TBs 45a and 45b.

  For example, it is assumed that the data conversion unit 21 performs data conversion on three forms A, B, and C, and generates character string processes TB 44a, 44b, and 44c. As shown in FIG. 10, the character string “application name” is stored in the “No. 1” column of the character string processing TB 44 a. In the column 2, a character string “travel expense application” is stored.

  In the “No. 1” column of the character string processing TB 44 b, the “application name” character string is stored, and in the “No. 2” column, the “equipment purchase application” character string is stored.

  In the “No. 1” column of the character string processing TB 44 c, the “application name” character string is stored, and in the “No. 2” column, the “equipment purchase application” character string is stored.

  The character identification unit 24 collects a character string and its position for each “No.” in the character string processing TBs 44a to 44c. Then, the character identifying unit 24 generates character string comparison TBs 45a and 45b.

  For example, the character identifying unit 24 acquires the character string “application name” in the “No. 1” field of the character string processing TBs 44 a to 44 c and the position thereof, and generates a character string comparison TB 45 a. Here, the character string “application name” in the “No. 1” column of the character string processes TB44a to 44c is not different between the character string processes TB44a to 44c. That is, the character string “application name” has no difference among the plurality of forms 11 of A to C. Therefore, the character identification unit 24 can identify the character string “application name” as a fixed character, and stores “fixed character” in the determination result column of the character string comparison TB 45a.

  In addition, the character identification unit 24 acquires the character string in the “No. 2” column of the character string processing TBs 44a to 44c and the position thereof, and generates a character string comparison TB 45b. Here, the character string in the column “No. 2” of the character string processing TBs 44a to 44c is different between the character string processing TBs 44a to 44c. That is, the character string in the column “No. 2” is different among the plurality of forms 11 of A to C. Therefore, the character identification unit 24 can identify the character strings “travel application” and “equipment purchase application” as input characters, and stores “entry characters” in the determination result column of the character string comparison TB 45 b.

  When the character identification unit 24 generates the character string comparison TBs 45a and 45b for each character string stored in the “No.” column of the character string processing TBs 44a to 44c, the character string comparison TBs 45a and 45b are used. A character string process TB46 is generated. For example, the character identification unit 24 generates a character string process TB 46 indicating the probability of a fixed character of the character string in each region.

  For example, a character string for each area of the form 11 is stored in the “character string” field of the character string processing TB 46 shown in FIG. In the “position” column, the position where the character string of the form 11 is written is stored. In the “judgment result” column, the probability of being a fixed character of the character string stored in the “character string” column is stored.

  Note that due to erroneous conversion by the data conversion unit 21 and fluctuation of the character string for each form 11, the character strings may differ between the forms 11 even though they are fixed characters. In this case, the character identification unit 24 may evaluate the similarity of the character shape and the word meaning with a numerical value, and recognize it with a width based on a threshold value.

  Further, it is necessary to compare character strings in a plurality of forms 11 in the same item. The comparison in the same item can be performed on the basis of the position of the character string described in the some form 11, for example. At that time, a minute shift or error in the position of the character string between the plurality of forms 11 becomes a problem, but it can be dealt with by providing a threshold regarding the error range.

  In addition, the character identification unit 24 may recognize a character string object to be compared from a structural feature such as a ruled line of the form 11 as another method for identifying a comparison object.

  2. Identification by learning

  There is a characteristic difference between the fixed character and the written character with respect to the word used, the number of characters, the size of the area to be described, and the like. Therefore, the character identification unit 24 accumulates information that evaluates whether the character is a fixed character or a written character, and performs identification.

  For example, the character identification unit 24 generates learning information TB for a fixed character from the determination result of “1. Identification by a plurality of form data” described above.

  FIG. 11 is a diagram illustrating an example of the learning information TB. As shown in FIG. 11, the number assigned to the learned character string is stored in the “No.” field of the learning information TB47.

  In the “character string” column, the learned character string is stored. For example, in the “character string” column, a character string that has been determined whether or not it is a fixed character by “1. Identification by a plurality of form data” described above is stored. For example, it can be seen that “application name” and “applicant” shown in FIG. 11 have been determined whether or not they are fixed characters in the past by “1. Identification by a plurality of form data”. .

  The “fixed character probability” field stores the probability that the character string stored in the “character string” field is a fixed character. For example, it is understood that the character string “application name” has been identified as a fixed character with a probability of 91% in the past.

  The character identifying unit 24 refers to the generated learning information TB47, and acquires the probability that the character string converted by the data converting unit 21 is a fixed character.

  FIG. 12 is a diagram illustrating an example of the character string process TB based on the learning information TB. The character identification unit 24 refers to the learning information TB47 described in FIG. 11 and generates a character string process TB48 based on learning.

  The information in the “No.”, “character string”, and “position” fields of the character string processing TB48 is stored by the data conversion unit 21 as in the character string processing TB41 described with reference to FIG. .

  The “learning result fixed character probability” field stores the probability that the character string corresponding to the “character string” field is a fixed character.

  For example, the character identification unit 24 refers to the learning information TB47 described with reference to FIG. 11 and acquires the probability that the character string “application name” is a fixed character. In the case of the example in FIG. 11, the probability that the character string “application name” is a fixed character is 91%. The character identification unit 24 stores the acquired probability as shown in the column of learning result fixed character probability in FIG.

  As described above, the character identifying unit 24 identifies whether the character string is a fixed character or an input character by learning the past character string.

  3. Identification with supplementary data

  When the supplementary data 12 is input, the character identification unit 24 identifies whether the supplementary data 12 is a fixed character or a written character.

  For example, in the supplementary data 12, the user directly specifies whether the character string of the form 11 is a fixed character or an entry character. Therefore, by comparing the supplementary data 12 and the character string of the data-converted form 11, it is possible to identify whether the data-converted character string is a fixed character or an input character.

  Specifically, when the character string converted into data matches the item (fixed character) described in the supplementary data 12, the character identification unit 24 converts the converted character string into a 100% fixed character. Is identified. In addition, when the character string after data conversion matches the item content (entry character) described in the supplementary data 12, the character identification unit 24 uses the data converted character string as a 0% fixed character. Identify.

  The character identification unit 24 identifies whether the character is a fixed character or an input character by the above three methods, and combines the identification results to perform final identification. For example, the character identification unit 24 averages the probabilities of fixed characters identified (calculated) by the above three methods.

  FIG. 13 is a diagram illustrating an example of the character string process TB. The character identification unit 24 combines the identification results of the above three methods to generate a character string process TB49 shown in FIG.

  The data converted by the data converter 21 is stored in the “No.”, “character string”, and “position” fields of the character string processing TB49.

  The “fixed character probability” column has a “multiple images” column, a “supplemental data” column, a “learning” column, and an “average” column.

  In the “multiple images” column, the fixed character probabilities calculated in “1. Identification by a plurality of form data” are stored. For example, as described with reference to FIG. 10, the character identification unit 24 generates a character string process TB 46 from a plurality of forms 11. The character identification unit 24 stores the “determination result” of the generated character string processing TB 46 in the “multiple images” field. Note that the TB contents in FIG. 10 and FIG. 13 do not match.

  In the “supplemental data” column, the fixed character probabilities calculated in “3. Identification by supplementary data” are stored. For example, when the character string “application name” is described as an item in the supplementary data 12, the character identifying unit 24 identifies the character string “application name” as a 100% fixed character. The character identifying unit 24 stores the identified fixed character probability 100% in the “supplemental data” field. Further, for example, when the character string “XX prefecture inspection transportation expense application” is described as the item content in the supplementary data 12, the character identification unit 24 fixes the character string “XX prefecture inspection transportation expense application” at 0%. Identifies it as a character. Then, the character identification unit 24 stores the identified fixed character probability 0% in the “supplemental data” field.

  In the “learning” column, the fixed character probability calculated in “2. Identification by learning” is stored. For example, as described with reference to FIGS. 11 and 12, the character identification unit 24 generates the character string process TB48 from the learning information TB47. The character identifying unit 24 stores the “learning result fixed character probability” of the generated character string processing TB48 in the “learning” column.

  The character identification unit 24 calculates an average of fixed character probabilities stored in the “multiple images” column, the “supplemental data” column, and the “learning” column, and stores the average in the “final result” column. For example, the fixed character probability of a plurality of images of the character string “application name” is 100%, the fixed character probability of supplementary data is 100%, and the fixed character probability of learning is 91%. In the column, 97% is stored.

  In the “judgment result” column, a result indicating whether or not the character string stored in the “character string” column is finally a fixed character is stored. For example, the character identification unit 24 compares the probability stored in the “final result” column with a predetermined threshold, and the character string stored in the “character string” column is finally fixed. Judge whether it is a character or not. For example, when the probability stored in the “average” column is 80% or more, the character identification unit 24 determines that the character string is a fixed character.

  In the above description, the character identification unit 24 performs final recognition (determination) of fixed characters based on the results of identification by three methods. And 2. The final fixed character may be recognized based on the result identified by the above method. Alternatively, the character identification unit 24 is 1. And 3. The final fixed character may be recognized based on the result identified by the above method.

  In addition, the character identification unit 24 may determine the fixed character by weighting the result identified by the above three methods. For example, the supplementary data 12 is generated by the user and may contain errors. In this case, you may make it make the weight of the identification result by the supplementary data 12 small.

  Returning to the description of FIG.

  [Step S5] Next, the area identifying unit 25, based on the fixed character and the entered character identified by the character identifying unit 24, the entry area in which the user can write a character on the form 11, and the user placing a character on the form 11. Identify non-fillable areas that cannot be filled. The area identifying unit 25 generates a determination process TB described below, and identifies the entry area and the non-entry area.

  FIG. 14 is a diagram illustrating an example of the determination process TB. The area identifying unit 25 generates the determination process TB50 shown in FIG. 14 by combining the character string process TB41 described in FIG. 6 and the area process TB42 described in FIG. Further, the area identifying unit 25 stores the determination result of the character string process TB49 described in FIG. 13 in the generated determination process TB50 in the determination process TB50.

  For example, the “region processing table” column of the determination processing TB50 shown in FIG. 14 corresponds to the region processing TB42 shown in FIG. Further, the “No.” column and the “character string” column in the “processing table” column of the determination process TB50 are, for example, the “No.” column and the “character” of the character string process TB41 shown in FIG. Corresponds to the column.

  Information is written in the “determination result” column of the determination process TB50 based on the “determination result” column of the character string process TB49 described in FIG.

  For example, when the “determination result” of the character string processing TB49 is “fixed character”, “not filled” is stored in the “determination result” column. When the “determination result” of the character string process TB49 is “entry character”, “entry” is stored in the “determination result” column. Specifically, since the character string “application name” in FIG. 14 is determined as a fixed character by the character string processing TB49 in FIG. 13, the “determination result” corresponding to the “application name” in FIG. In the column, “not filled” is stored.

  “Non-entry” in the “judgment result” column indicates that the corresponding character string area is a fixed character area. “Entry” in the “judgment result” field indicates that the area of the corresponding character string is an entry character area.

  “Non-entry / entry” in the “judgment result” column indicates that the corresponding character string area is a fixed character area and an entered character area. In other words, “not filled / filled” indicates that it is a filled area and a non-filled area. For example, it can be seen that the character string “reason: for XX prefecture visit” shown in FIG. 14 is an entry area and a non-entry area (“reason:” is a fixed character, and “XX” "For prefectural inspection" is a written character.)

  The reason why the entry area and the non-entry area are mixed is that when the data conversion unit 21 performs grouping, the fixed character and the entry character may be recognized as one group. For example, in FIG. 5, the character string “reason: for XX prefecture visit” is described in a blank area outside the ruled line frame, and the data conversion unit 21 groups the blank area outside the ruled line frame as one group. Because there are cases.

  Information on whether or not to divide the grouping including the entry area and the non-entry area is stored in the “necessity of division” column. For example, the area of “reason: for visiting XX prefecture” is an entry area and a non-entry area, and is grouped into one group. In this case, the area identifying unit 25 stores information “necessary” indicating that it is necessary to divide the grouping in the “necessity of division” column.

  Returning to the description of FIG.

  [Step S6] Next, the grouping correction unit 26 corrects the grouping performed by the data conversion unit 21. The grouping correction unit 26 performs grouping correction by the following three methods. The grouping correction unit 26 performs grouping correction using all or any of the methods 1 to 3.

  1. Correction using determination processing TB50

  The grouping correction unit 26 performs grouping correction with reference to the determination process TB50 described in FIG.

  FIG. 15 is a diagram illustrating an example of the determination process TB after the grouping correction. The determination process TB51 illustrated in FIG. 15 is substantially the same as the determination process TB51 described with reference to FIG. 14, but the determination process TB50 illustrated in FIG.

  The grouping correction unit 26 refers to the “division necessity” column of the determination process TB50 described in FIG. When the “necessary” information is stored in the “necessity of division” column, the grouping correction unit 26 corrects the grouping so that the character string corresponding to the column is in another group.

  For example, in the case of the determination process TB50 of FIG. 14, the character string "" reason: "," for visiting XX prefecture "" indicates whether or not the division is necessary. Therefore, the grouping correction unit 26 corrects the grouping so that the character strings ““ reason: ”and“ for XX prefecture inspection ”” are in separate groups. Specifically, as shown in the determination process TB51 of FIG. 15, the grouping correction is performed so that the character string of “reason:” and the character string of “for XX prefecture inspection” are in separate groups. Then, the grouping correction unit 26 corrects the “determination result” field to be appropriate information.

  2. Multiple image correction

  In an area where there is no ruled line and where a variable-length character string exists, it is difficult to appropriately determine the end position of the area. However, the grouping correction unit 26 can detect an appropriate end position by the area processing TB 42 in the plurality of forms 11.

  FIG. 16 is a diagram for explaining the variable length region. The form 52 shown in FIG. 16 is almost the same as the form 11 shown in FIG. 5, but the character string described in the “reason:” column outside the ruled line frame is different. For example, in the form 11 in FIG. 5, the character string in the “reason:” column is one line, but in the form 52 in FIG. 16, the character string in the “reason:” column is three lines.

  “Reason:” in the form 52 is a fixed character. The three-line character string indicating the reason is an input character entered by the user. Originally, it is appropriate that the entry character area in which the user can enter a character string is recognized with the point 52a as the start position and the point 52b as the end position. However, in the case of the form 52, the point 52a is recognized as the start position and the point 52c is recognized as the end position.

  However, the data conversion unit 21 converts a plurality of forms 11 and generates a plurality of region processes TB42. Therefore, character string entry data of various users can be obtained. For example, there may be a user who entered a character string up to the point 52b. Therefore, the grouping correction unit 26 can acquire an appropriate end position of the variable length region by adopting the end position having the largest value. That is, it is possible to perform highly accurate area determination by converting data of a plurality of forms.

  FIG. 17 is a diagram showing an example of the form determination process TB in FIG. Assume that the form 52 in FIG. 16 is converted into data, characters are identified, areas are identified, and grouping correction is performed. Assume that the determination process TB53 shown in FIG. 17 is obtained.

  The end position of “No. 8” in the determination process TB53 is (Y, X) = (90, 120). This end position has a larger Y value than the end position of the determination process TB51 (FIG. 15) for the form 11 shown in FIG. Therefore, for example, the grouping correction unit 26 can correct the value of “Y” at the end position of “No. 8” of the determination process TB51 from “70” to “90”.

  3. Correction by supplementary data

  When the supplementary data 12 is input, the grouping correction unit 26 performs grouping correction using the input supplemental data 12.

  The supplementary data 12 is obtained by designating fixed characters and entry characters of the form 11 after data conversion. Therefore, the grouping correction unit 26 can recognize which character strings should be grouped into one group by referring to the input supplemental data 12.

  FIG. 18 is a diagram illustrating an example of the determination processing TB for the form of FIG. 16 that has been grouped by mistake. For example, it is assumed that the form 52 shown in FIG. 16 is converted into data, characters are identified, and areas are identified. Assume that the determination process TB54 shown in FIG. 18 is obtained.

  Originally, the character strings “hereinafter, for participation in training”, “• patent creation training”, and “• marketing theory” must be grouped into one group, but the data conversion unit 21 indicates to the determination process TB54. Thus, it is assumed that each character string is grouped as one group.

  Here, the supplementary data 12 is created by the user. In the case of the form 52 of FIG. 16, the character strings “hereinafter, for training participation”, “• patent creation training”, and “• marketing theory” are described in one entry area, It is described with the following tag.

  <Contents> In order to participate in the training, ¥ n ・ Patent creation training ¥ n ・ Marketing theory </ Contents>

  The grouping correction unit 26 can recognize from the above tags that the character strings of “for training participation”, “• patent creation training”, and “• marketing theory” must be grouped into one group. Thereby, the grouping correction | amendment part 26 can correct | amend the determination process TB53 shown in FIG. 18 to appropriate TB.

  FIG. 19 is a diagram illustrating an example of the determination process TB after the grouping correction. The grouping correction unit 26 corrects the grouping based on the input supplemental data 12 as in the determination process TB55 shown in FIG.

  For example, in the determination process TB54 of FIG. 18, the character strings “hereinafter, for training participation”, “• patent creation training”, and “• marketing theory” are grouped into separate groups. On the other hand, in the determination process TB55 of FIG. 19, these characters are corrected into one group by the grouping correction unit 26.

  Returning to the description of FIG.

  [Step S <b> 7] Next, the generation unit 27 generates the electronic form 13, the form template 14, and the template information 15 by using various types of TB generated in the storage unit 28.

  For example, the generation unit 27 generates the computerized form 13 from the data converted by the data conversion unit 21. The digitized form 13 is generated by, for example, a document file or an image file.

  Further, the generation unit 27 generates a form template 14 that can edit, for example, characters and ruled lines of the form 11 from the data converted by the data conversion unit 21. At this time, the generation unit 27 deletes the character string existing in the entry area of the form template 14 to be created based on the determination result (entry / non-entry result) of the determination result TB. The form template 14 is generated by, for example, a document file that can edit characters and ruled lines.

  Further, the generation unit 27 generates template information 15 of the form template 14. The template information 15 is a file including format information such as a character string, position, format, page margin, paper size, or ruled line format described in the form template 14.

  FIG. 20 is a diagram showing an example of an electronic form. An electronic form 13 in FIG. 20 shows an example in which the form 52 in FIG. 16 is digitized. A dotted line frame 56 shown in FIG. 20 indicates an entry area in which a user can enter characters. In the entry area, the user may be able to edit the characters.

  FIG. 21 is a diagram showing an example of a form template. The form template 14 in FIG. 21 shows an example in which the form in FIG. 16 is made into a template. A dotted line frame 57 shown in FIG. 20 indicates an entry area in which a user can enter characters. In the dotted line frame 57, the user can input characters using, for example, a terminal device.

  FIG. 22 is a diagram showing an example of template information. The template information 15 in FIG. 22 indicates the template information of the form template 14 in FIG.

  The template information 15 includes, for example, a fixed character, position, character format (font type, size), page margin, paper size, paper orientation, ruled line position, ruled line format, region information, and the like.

  As described above, the information processing apparatus 1 identifies, from the electronic data of the plurality of forms 11, the fixed characters described in advance on the form 11 and the entry characters entered by the user, and identifies the identified fixed characters and entry characters. Based on this, an entry area where the user can enter characters in the form 11 and a non-entry area where the user cannot enter characters in the form 11 are identified. The information processing apparatus 1 generates the form template 14 including the identified entry area and non-entry area. Accordingly, the information processing apparatus 1 can generate a form template 14 in which a fixed character area described in advance in a form and an entry area in which a user can enter a character are divided with high accuracy.

  Moreover, since the information processing apparatus 1 generates the form template 14 from the plurality of forms 11, it can save the user's trouble.

  For example, the information processing apparatus 1 can generate the form template 14 from the plurality of forms 11 without inputting the format information of the form 11 depending on the layout generated by the user. More specifically, the user can generate the form template 14 without creating various kinds of format information of the form 11 such as input character types such as “Hiragana” and “number” and the number of input characters. .

  Moreover, since the information processing apparatus 1 generates the form template 14, it can be used as a template of the form on, for example, a terminal device. Further, since the information processing apparatus 1 generates the template information 15, it can be used as a resource in system development together with the form template 14.

  Moreover, the character identification part 24 of the information processing apparatus 1 identifies a character string as a fixed character, when the same character string is contained in the same position in the predetermined number of forms of the same kind. Thereby, the information processing apparatus 1 can generate the form template 14 in which the area where the user enters the character and the area where the user does not enter the area are distinguished with high accuracy. In particular, the larger the predetermined number of values, the more accurate form template 14 can be generated.

  In addition, the information processing apparatus 1 receives the supplementary data 12 including an item preliminarily described in the form 11 and an entry item entered by the user corresponding to the item, thereby allowing the user to enter characters. Then, it is possible to generate the form template 14 in which the areas that are not so are classified with higher accuracy.

  The supplementary data 12 includes items preliminarily described in the form 11 and entries entered by the user corresponding to the items, and does not include information depending on the layout. Therefore, the user can create supplementary data 12 in a simple format, and can create a form template 14 in which the user can enter the character entry area and the area where the user does not enter with high accuracy without taking time and effort. Can be generated.

  The information processing apparatus 1 further includes learning information TB47 that stores the probability of the character string of the form 11 that has been identified as a fixed character by the character identifying unit 24. The character identifying unit 24 is stored in the learning information TB47. The fixed character described in the form 11 is identified using the probability of Thereby, the information processing apparatus 1 can generate the form template 14 in which the area in which the user enters the character and the area in which the user does not enter the area are distinguished with high accuracy without taking time and effort for the user.

  Further, the information processing apparatus 1 can generate a form template 14 with higher accuracy by combining character identification with a plurality of forms 11, character identification with learning information TB47, and character identification with supplementary data 12.

  Furthermore, when the fixed character and the entry character grouped based on a predetermined method belong to one group, the information processing apparatus 1 divides the group so as to belong to different groups. Thereby, the information processing apparatus 1 can generate the form template 14 in which the area where the user enters the character and the area where the user does not enter the area are distinguished with high accuracy.

  In the above description, the form 11 is a paper medium. However, the form 11 of the paper medium may be converted into electronic data and input to the information processing apparatus 1 using the OCR function and the image identification processing function in advance. In this case, the data conversion unit 21 does not have to have an OCR function and an image identification processing function.

  Each TB described above is an example, and is not limited to the illustrated example. Further, each information is not limited to the TB structure, but may be another data structure.

  Further, the same position described above includes a slightly shifted position. For example, in a plurality of forms 11, a character string described at a position shifted within a predetermined range may be determined as the same position.

  Further, the processing unit of the flowchart shown in FIG. 4 is divided according to the main processing contents in order to make the processing of the information processing apparatus 1 easy to understand. The present invention is not limited by the way of dividing the processing unit or the name. The processing of the information processing apparatus 1 can be divided into more processing units according to the processing content. Moreover, it can also divide | segment so that one process unit may contain many processes. Further, the processing order of the above flowchart is not limited to the illustrated example.

  In the above description, the information processing apparatus 1 generates the computerized form 13, the form template 14, and the template information 15 from the plurality of forms 11. However, when supplementary data 12 is input, the information processing apparatus 1 generates the electronic form 13 from the one form 11. It is possible to generate the chemical form 13, the form template 14, and the template information 15. This is because the supplementary data 12 includes information on fixed characters and information on written characters. In addition, when learning is performed for all character strings of a certain form, the information processing apparatus 1 can identify whether the character is a fixed character or an entry character even if there is no data of a plurality of forms 11. . For example, it is assumed that the information processing apparatus 1 performs data conversion on a certain form 11. Then, it is assumed that all the character strings of the data-converted form 11 have been learned in the past. In this case, the information processing apparatus 1 can identify whether it is a fixed character or an input character from one data-converted form 11.

  Further, although the learning information TB47 described with reference to FIG. 11 stores the fixed character probability, it may store the entry character probability. That is, the character identification unit 24 may acquire the probability of being a written character.

  Furthermore, other arbitrary components and processes may be added to the present case.

  1: Information processing device, 11: Form, 12: Supplementary data, 13: Computerized form, 14: Form template, 15: Template information, 21: Data conversion section, 22: Input section, 23: Character correction section, 24: Character identification unit, 25: area identification unit, 26: grouping correction unit, 27: generation unit, 28: storage unit, 31: CPU, 32: RAM, 33: ROM, 34: HDD, 35: scanner, 36: drive, 37: Communication interface, 38: Bus, 41, 43, 44a to 44c, 45, 46, 48, 49: Character string processing TB, 42: Area processing TB, 47: Learning information TB, 50, 51, 53-55: Determination process TB, 52: form, 52a to 52c: point, 56, 57: dotted line frame

Claims (11)

A character identification unit for identifying fixed characters previously written in the form and entry characters entered by the user from image information of a plurality of forms;
Based on the fixed character and the input character identified by the character identification unit, an area identification unit for identifying an entry area in which a user can enter a character in the form and a non-entry area in which the user cannot enter a character in the form; ,
A generating unit that generates a template of the form based on the entry region and the non-entry region identified by the region identification unit;
An information processing apparatus comprising: The information processing apparatus according to claim 1,
In the information processing apparatus, the character identification unit identifies the character string as a fixed character when the same character string is included in the same position in a predetermined number of the same type of forms. The information processing apparatus according to claim 1, wherein:
The area identifying unit identifies a region of the character string of the form identified as a fixed character by the character identifying unit as a non-entry region, and a region of the character string of the form identified as a written character by the character identifying unit An information processing apparatus characterized by distinguishing from an entry area. The information processing apparatus according to any one of claims 1 to 3,
The character identification unit uses item information including an item preliminarily described in the form and an entry entered by the user corresponding to the item, and a fixed character and an input character described in the form. An information processing apparatus characterized by identifying. The information processing apparatus according to claim 4,
The character identification unit identifies the character string of the form corresponding to the item of the item information as a fixed character, and identifies the character string of the form corresponding to the entry of the item information as an input character. Information processing apparatus. The information processing apparatus according to any one of claims 1 to 5,
A storage unit that stores a learning result related to identification of the character string of the form identified as a fixed character by the character identification unit;
The information processing apparatus, wherein the character identification unit identifies fixed characters described in the form using the learning result stored in the storage unit. The information processing apparatus according to claim 6,
The learning result is a probability or score of a character string of the form identified as a fixed character by the character identification unit,
The character identification unit identifies a character as a fixed character when a probability or score of a character string described in the form is a predetermined value or more. The information processing apparatus according to claim 6,
The learning result is a probability or score of a character string of the form identified as a fixed character by the character identification unit,
The character identifying unit identifies fixed characters and written characters described in the form using one or both of the item information and the probability or score stored in the storage unit. Information processing apparatus. The information processing apparatus according to any one of claims 1 to 8,
The image information of the form is grouped based on a predetermined method, and when a fixed character and an entry character belong to one group, a grouping correction unit that divides the group so as to belong to different groups is further provided. An information processing apparatus comprising: A template generation method for an information processing device,
A character identification step for identifying, from image information of a plurality of forms, fixed characters previously written in the form and entry characters entered by the user;
Based on the fixed character and the entry character identified by the character identification step, an area identification step for identifying an entry area in which the user can enter a character in the form and a non-entry area in which the user cannot enter the form;
A template generation method comprising: generating a template of the form based on the entry area and the non-entry area identified by the area identification step. A program for an information processing apparatus that generates a template for a form,
The program is
A character identification step for identifying, from image information of a plurality of forms, fixed characters previously written in the form and entry characters entered by the user;
Based on the fixed character and the entry character identified by the character identification step, an area identification step for identifying an entry area in which the user can enter a character in the form and a non-entry area in which the user cannot enter the form;
Generating a template for the form based on the entry area and the non-entry area identified by the area identification step;
The information processing apparatus executes the program.

Images