JP2010072850A - Image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor for deciding the array rule of a desired character information group from among a plurality of pieces of character information included in original data. <P>SOLUTION: An address pertinent to a designated address retrieval keyword is retrieved from input name book data, and the calculation of a position and the discrimination of a character direction (horizontal writing/vertical wiring) is performed. An address array direction in the name book data using the address as a base is predicted based on the character direction of the address (S121). The characters arranged in the predicted array direction are extracted as address candidates for the address (S122), and whether or not the characters are the address is collated by referring to the database (S123). When the collation result is positive (S124; Yes), the positions of the characters (addresses) are calculated (S125). Thus, the positions of a plurality of addresses included in the name book data are searched, and the array of all the addresses included in the name book data is calculated based on the positional relation (interval) (S126). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image processing apparatus having a function of determining an arrangement rule of character information included in input document data.

  A document in which various character information (character strings) is listed is read with an optical character reader (OCR), and the desired character information is extracted from the data and synthesized into another document. A technique for printing out is generally known.

  For example, in the abandoned bicycle management system described in Patent Document 1, a crime prevention registration number, a registrant's name, and a registrant's address extracted by reading a document (security response registration list) containing information on abandoned bicycles It is automatically issued after being combined with a receipt notice for the bicycle owner. Also, with this technology, the format of the document to be read is determined in advance, and the desired character information can be obtained by reading only a predetermined range (the range in which the crime prevention registration number, registrant name, and registrant address are written). To extract.

Japanese Patent Laid-Open No. 2000-40101

  However, in the above technique, the document to be read is limited to a specific format, and it is not possible to cope with the case where the position and arrangement of character information to be extracted are unknown. For example, a general name list and the like have various formats, and positions and arrangements of names and addresses are various. Since it is not possible to extract only the names of all the members from a list of various formats, there is a problem of poor versatility.

  The present invention is intended to solve the above problem, and provides an image processing apparatus capable of determining an arrangement rule of a desired character information group from a plurality of character information included in document data. The purpose is that.

  The gist of the present invention for achieving the object lies in the inventions of the following items.

[1] An input unit for inputting document data;
The character information belonging to the type is used when determining the arrangement rule of the character information group of the same type from among the plurality of character information included in the first document data input from the input unit The first character information corresponding to the character information is searched, the second character information belonging to the type is specified by searching from the first character information in a predetermined direction, and the first character information and A control unit that determines the arrangement rule based on a positional relationship with the second character information;
An image processing apparatus comprising:

  In the above invention, character information of the same type as the character information group for which the arrangement rule is to be determined is selected from the plurality of character information included in the input document data (first document data). The first character information corresponding to the character information (search keyword) is searched.

  The type of character information group described above refers to a type (name, word, etc.) classified by content (information content), not a type (Arabic numerals, Roman characters, Chinese characters, etc.) classified by characters themselves. As the character information used for the search, for example, typical character information of the above type or character information designated by the user can be used.

  Then, the second character information belonging to the type is specified by searching from the first character information in a predetermined direction, and belongs to the type based on the positional relationship between the first character information and the second character information. The arrangement rule of the character information group is determined. For example, it is determined that the character information group belonging to the above type is arranged at intervals of the first character information and the second character information in the column direction including the first character information and the second character information. .

  As a result, it is possible to determine the arrangement rule of the character information group having the same type from among a plurality of character information included in the document data.

[2] In the search for the first character information, the control unit sequentially uses character information having a high detection probability in the search, and sequentially uses a plurality of character information included in the first document data. The image processing apparatus according to [1], wherein the character information detected from the first character information is determined as the first character information.

  In the above invention, when the first character information is searched from among a plurality of character information included in the first document data, the character information (search keyword) having a high detection probability in the search is used in order. To search for the corresponding character information. When character information corresponding to the character information (search keyword) is detected, the character information is determined as first character information. Thus, the search efficiency of the 1st character information can be improved by using and searching in order from the character information with a high detection probability.

[3] The control unit determines, in a direction different from the first column direction, from third character information arranged in a first column including the first character information and the second character information. When the fourth character information belonging to the type can be specified by searching, the arrangement rule is determined in consideration of the position of the fourth character information. [1] or [2] The image processing apparatus described.

  In the above invention, the third character information is character information arranged in the first column including the first character information and the second character information. The third character information includes first character information, second character information, and other character information arranged in the first column. If the fourth character information belonging to the above type can be specified by searching from one or more of the third character information in a direction different from that of the first column, the character whose alignment rule is to be determined It can be estimated that the information group is arranged in a plurality of columns (two-dimensional arrangement), and the arrangement rule is determined in consideration of the position of the fourth character information. Thereby, the arrangement rule of a plurality of columns can also be determined.

[4] The control unit searches the fourth character information in a direction along the first column to identify fifth character information belonging to the type, and includes the fourth character information and the first character information. 5. The image processing according to [3], wherein the arrangement rule for the second column including the fourth character information and the fifth character information is determined based on a positional relationship with the character information of 5. apparatus.

  In the above invention, in the determination of the arrangement rule of the plurality of columns, the arrangement rule of the character information group in the second column (parallel to the first column) along the first column is different from the first column. Determined.

[5] The control unit is based on a positional relationship between the first character information and the second character information with respect to a second column including the fourth character information along the first column. The image processing apparatus according to [3], wherein the arrangement rule for the second column is determined by applying the arrangement rule determined in the above.

  In the above invention, in the determination of the arrangement rule of a plurality of columns, the arrangement rule of the character information group in the second column (parallel to the first column) along the first column is the same as that of the first column. It is determined that

[6] The control unit may determine the arrangement rule in a direction different from the first column direction based on a positional relationship between the third character information and the fourth character information. The image processing device according to [3].

  In the above invention, the arrangement rule of the character information group in a direction different from the first column direction is determined based on the positional relationship between the third character information and the fourth character information. For example, it is determined that the character information group is arranged at an interval between the third character information and the fourth character information in a direction different from the first column direction.

  Thereby, in the determination of the arrangement rule for the character information group of a plurality of columns, it is possible to determine the arrangement rule in a direction different from the first column direction.

[7] Any one of [1] to [6], wherein when the first document data is data of a plurality of pages, the control unit determines the arrangement rule for each page. The image processing apparatus described in one.

  In the above invention, for a plurality of pages of document data (first document data), the arrangement rule of the character information group is determined for each page.

[8] The image processing apparatus according to any one of [1] to [7], wherein the control unit determines the predetermined direction based on a character direction of the first character information.

  In the said invention, the predetermined direction (search direction) which searches 2nd character information from 1st character information is determined based on the character direction of 1st character information. This character direction is the direction in which characters are arranged, and refers to the direction of writing such as horizontal writing or vertical writing. For example, when the first character information is horizontal writing, there is a high possibility that the character information group for which the arrangement rule is to be determined is arranged in the vertical direction (vertical direction). In this case, since it is highly possible that the same character information group is arranged in the left-right direction (horizontal direction), a search is performed in the left-right direction. As a result, the second character information can be efficiently searched and specified.

[9] The control unit compares the format of the first character information with the format of character information adjacent to the first character information, and character information having a format different from that of the first character information. The image processing apparatus according to any one of [1] to [8], wherein the arranged direction is excluded from a search direction of the second character information.

  In the above invention, when the character information adjacent to the first character information is different in format (character size, typeface, etc.) from the first character information, the character information group for which the arrangement rule is determined is the first character information. Since there is a low possibility that the character information is arranged in the direction of character information having a different format, the second character information is not searched in that direction. As a result, the second character information can be efficiently searched and specified.

  When the above invention is combined with the invention for determining the predetermined direction (the search direction of the second character information) based on the character direction of the first character information described in [8], for example, the first character information Is a horizontal writing, so the predetermined direction is set to the vertical direction, but if the format of the character information arranged in one direction (for example, upward) in the vertical direction is different, the second character information is displayed in the one direction. It is possible not to search.

[10] The control unit extracts individual character information belonging to the type in accordance with the arrangement rule from a plurality of character information included in the first document data, and inputs the character information to the input The image processing apparatus according to any one of [1] to [9], wherein the image data is combined with the second document data input from the scanner unit and output.

  In the above invention, individual character information whose contents extracted from the first document data are of the same type is synthesized with the second document data and output.

[11] The control unit retrieves sixth character information corresponding to the character information from the second document data using the designated character information, and the position and format of the sixth character information. The image processing apparatus according to [10], wherein the position and format of character information to be combined with the second document data are determined based on the above.

  In the above invention, when the character information extracted from the first document data is combined with the designated character information (sixth character information) retrieved from the second document data, the position of the character information to be combined And the format (character size, character direction, etc.) are determined based on the position and format of the designated character information. As a result, it is possible to arrange the character information to be synthesized with the designated character information in a good-looking manner, or to make the format look good.

[12] The image processing apparatus according to any one of [1] to [11], wherein the control unit performs the specification with reference to a database in which character information belonging to the type is registered. .

  In the above invention, the character information (second to fifth character information) can be easily identified with reference to a database in which character information belonging to the type is registered.

[13] The control unit determines whether or not the extracted character information according to the arrangement rule belongs to the type by referring to a database in which character information belonging to the type is registered. The image processing device according to any one of [1] to [11].

  In the above invention, it is possible to easily determine (verify) whether or not the character information extracted in accordance with the arrangement rule belongs to the above type with reference to the database in which the character information belonging to the above type is registered. Since the first and second character information and the fourth and fifth character information have been found to belong to the above type, the determination may be omitted. As a result, if character information that does not belong to the above type is included in the extracted data, the character information can be excluded.

[14] In determining whether or not it belongs to the type, the control unit also refers to a database in which a condition indicating that it does not belong to the type is registered, and character information corresponding to the condition does not belong to the type [13] The image processing device according to [13],

  In the above invention, in determining whether or not the character information extracted according to the arrangement rule belongs to the above type, the determination is made by referring to the database in which the condition indicating that the character information does not belong to the above type is registered. Efficiency and accuracy can be improved.

[15] A display unit and an operation unit are provided,
The image processing according to [13] or [14], wherein the control unit displays the determination result on the display unit, and accepts a confirmation or correction operation for the determination result by the operation unit. apparatus.

  In the above invention, the result of automatically determining whether or not the character information extracted in accordance with the arrangement rule belongs to the above type is displayed, and the user can confirm or confirm or correct the determination result. Thereby, it is possible to suppress waste such that erroneously determined character information or character information that does not need to be output is combined with the second document data and output.

  According to the image processing apparatus of the present invention, it is possible to determine an arrangement rule of a desired character information group from a plurality of character information included in document data. Therefore, versatility and convenience are improved.

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

[First Embodiment]
FIG. 1 shows a schematic configuration of a multifunction peripheral 10 (MFP; Multi Function Peripheral / Multi Function Printer) as an image processing apparatus according to the first embodiment of the present invention.

  The multifunction machine 10 has a document copy function, a printer function, a scanner function, a facsimile function, and the like.

  The scanner function is provided with an OCR function for automatically recognizing character image data written on an optically read original and converting it into text data.

  In the copy function, in addition to the normal copy mode in which the read original is copied and printed under conditions according to default settings or user settings, desired character information (characters) included in the first original using the OCR function is used. Column) is extracted, and an automatic composition mode for synthesizing and printing on the second original is provided. In this automatic composition mode, for example, “address” and “destination” as desired character information are extracted from the “name list”, “address book”, etc. as the first original (extraction source original), and the second original. This is a mode that can reduce work such as address writing and address writing by automatically composing it to “Send Letter”, “Guidance Letter”, “Notification Letter”, etc. / Automatic address writing mode).

  As used herein, “character information” refers to names or words having some meaning (information) made up of one or a plurality of characters, such as “character strings” that are generally used. It is synonymous. This character information can be classified by contents (information contents) such as personal names (last names, first names), company names, place names (in the case of Japan, prefectures, municipalities), and postal codes. Hereinafter, “character information” may be simply referred to as “character”.

  A name (name) such as a person name or a company name described in a name list or the like is entered in the address column of a document (document) such as a cover sheet, thereby becoming the address in the document. However, in the present specification, for convenience, the name list will be referred to as an address.

  The multifunction machine 10 includes a CPU (Central Processing Unit) 11 serving as a control unit, a ROM (Read Only Memory) 13, a RAM (Random Access Memory) 14, a nonvolatile memory 15, and a hard disk device (Hard) via a bus 12. Disk Drive (HDD) 16, operation unit 17, display unit 18, facsimile communication unit 19, network communication unit 20, document feeder control unit 21, scanner unit 22, image processing unit 23, printer The unit 24 is connected.

  The CPU 11 controls the operation of the multifunction machine 10 based on a program stored in the ROM 13. In the automatic composition mode, character (character information) is extracted from the original data and synthesized. The RAM 14 is used as a work memory for temporarily storing various data when the CPU 11 executes a program, and is also used for an image memory for temporarily storing image data.

  The nonvolatile memory 15 is a memory that retains memory even when the power is turned off, and stores device-specific information and various setting information. The hard disk device 16 stores various saved data and the like, and also saves image data for scanning and copying originals, image data (print data) for printing requested from a client terminal, and the like.

  The operation unit 17 includes various buttons such as a mode selection button, a start button, a stop button, and a numeric keypad, and fulfills a function of accepting various operations performed on the multifunction machine 10 by a user or the like. The display unit 18 is composed of a liquid crystal display or the like, and has a function of displaying various screens such as an operation screen, a setting screen, and a guidance screen.

  The facsimile communication unit 19 performs a function of communicating with an external device having a facsimile function through a public line. The network communication unit 20 functions to communicate with external devices such as client terminals and servers through a network such as a LAN (Local Area Network).

  The document feeder controller 21 has a function of controlling the operation of an automatic document feeder (ADF) (not shown). The automatic document feeder has a function of automatically feeding a document to a reading position by the scanner unit 22. For example, a document tray on which documents are stacked, a conveyance path for guiding and discharging the documents stacked on the document tray to the reading position of the scanner unit 22, and a document stacked on the document tray one by one in the conveyance path. A plurality of transport rollers for feeding and passing through the reading position of the scanner unit 22 and discharging from the transport path, and a discharge tray for discharging the document read by the scanner unit 22 from the transport path, and the like.

  The scanner unit 22 performs a function of optically reading a document and acquiring image data (document data). For example, a light source that irradiates light on a document, an image sensor that receives reflected light from the document and reads the document for one line in the width direction (main scanning direction), and a reading position in units of lines by the image sensor is the length of the document. A scanning mechanism that moves in the vertical direction (sub-scanning direction) and an optical path (optical system) including a lens and a mirror that guides reflected light from the document to the image sensor to form an image.

  The image processing unit 23 functions to perform various types of image processing such as image correction, rotation, enlargement / reduction, compression / expansion on the image data. Also, it fulfills the function of performing rasterization processing on print data (vector format image data) in printing requested from the client terminal.

  The printer unit 24 functions to form and output an image corresponding to the input image data on a sheet. For example, it is equipped with a paper transport device, photosensitive drum, charging device, laser unit, developing device, transfer separation device, cleaning device, fixing device, etc., and forms an image on the paper by an electrophotographic process Configured as a laser printer.

  FIG. 2 shows a functional block diagram of the CPU 11 in the automatic synthesis mode. In addition, this figure also shows the main configuration of the above-described multifunction peripheral 10 related to the automatic synthesis mode.

  In the automatic synthesizing mode, for example, the document 50 in which the address is not entered and the name list 60 in which a plurality of names (names, first names), company names, etc. as addresses to be entered in the document 50 are read and characters by the OCR function are read. Automatic recognition is performed. After character recognition by this OCR processing, search for titles such as “sama”, “dono”, “middle” described in the document 50, and an array of a plurality of addresses (character information group) described in the name list 60 A rule is determined (array prediction) and a determination is made as to whether the characters arranged in the rule are addresses. Then, it is possible to extract all addresses from the name list 60 and print the number of documents 50 corresponding to the number of addresses obtained by combining the addresses according to the titles.

  The data input unit 22 has a function of inputting the image data of the document 50 in which the address is not entered as in the above example, and the image data of the name list 60 in which the address to be entered in the document 50 is described. Here, the data input unit 22 includes the scanner unit 22 described above. The data input unit 22 (scanner unit 22) optically reads a document 50 and a name list 60 that are set in an automatic document feeder and automatically sent, and acquires each image data (original data) to input data. I do.

  Further, in order to identify the image data of the document 50 and the name list 60 by the data input by reading the original as described above, it can be realized by a configuration in which the type of original to be read is designated by the user and read individually.

  For example, the operation unit 17 is provided with a composition destination document designation button for designating a composition destination document (document 50) and an extraction source document designation button for designating an extraction source document (name list 60). The user sets one type of document on the multifunction peripheral 10, operates the document designation button corresponding to the type, and then operates the reading start button (start button). After the reading of the original is completed, the other type of original is set in the multi-function peripheral 10, and the original designation button corresponding to that type is operated, and then the reading start button is operated. With this configuration, it is possible to identify that the input image data is data corresponding to the type of document specified by the user.

  Alternatively, the input image data may be subjected to OCR processing to analyze the contents of the document and be automatically identified.

  Hereinafter, image data of a document is abbreviated as “document data”, and image data of a name list is abbreviated as “name list data”. Further, when collectively referring to them, it is referred to as “original data”.

  The data storage unit 14 functions to store document data and name list data input from the data input unit 22. Here, the data storage unit 14 is configured by the RAM 14 described above or a memory such as a DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory) mounted in the multi-function device 10.

  The CPU 11 includes a document information output unit 31, a character collation unit 32, a character coordinate detection unit 33, an address array prediction unit 34, an address reading unit 35, an address determination unit 36, and an address pasting unit 37. These functional units extract addresses from name list data and synthesize them into document data.

  The data output unit 24 functions to output document data created by the CPU 11 and combined with an address. Here, the data output unit 24 includes the printer unit 24 described above. The data output unit 24 (printer unit 24) performs data output by forming an image corresponding to the document data on a sheet and outputting it.

  A manuscript information output unit 31 in the CPU 11 creates information (manuscript information) related to a manuscript necessary to determine (calculate) the position and size of desired characters (address, honorific, etc.) included in the manuscript data. And play the role of output. In this example, the position and size of the character are represented by the number of lines (number of dots).

  Specifically, when document data is input from the data input unit 22, the document information output unit 31 creates a document information table 41 that stores document information related to the document data. The document information is the document size, the reading resolution, and the number of lines in the document width direction (X direction) and length direction (Y direction) calculated from the information. The document information output unit 31 outputs the document information table 41 storing the document information to the character coordinate detection unit 33 and the address array prediction unit 34.

  For example, if the original is the above-described document 50 and the name list 60, and the original size and the reading resolution are the same, one (common) original information table storing the same information is created. If at least one of the document size and the reading resolution is different and the number of lines is different, information about each document is stored individually in one document information table, or a document information table is created for each document. May be.

  The character collating unit 32 plays a role of presenting a search keyword used for searching for a desired character included in the document data and performing collation in specifying and determining a desired character. Specifically, a character that becomes a preset search keyword is presented to the character coordinate detection unit 33, or a character requested to be collated by the address array prediction unit 34 or the address determination unit 36 is collated.

  The characters used for the above presentation and verification are registered in advance in the database 38 stored in the hard disk device 16 or the like. In the case of this example, the name of the person as the address (the last name used in Japan) and “sama”, “dono”, “san”, etc. as the titles are registered in the database 38 (address list / name list) ). Further, if necessary, information on a destination (destination list) such as a company name as an address, “Gion” as a title, a place name, a postal code, and a telephone number may be registered.

  The character collation unit 32 accesses the database 38 to acquire characters (honorific search keyword / address search keyword) to be presented as a search keyword and collate the character requested for collation. The search keyword can also be specified (set) by the user through the operation unit 17 or the like.

  The character coordinate detection unit 33 plays a role of detecting a desired character included in the document data and calculating its position (coordinate position) and size (character size).

  More specifically, the title search keyword presented by the character collating unit 32 is searched from the document data, and the position and size of the searched title are searched for using the number of XY lines of the document stored in the document information table 41. Calculate as the number of lines. Furthermore, the character direction (horizontal writing / vertical writing) of the retrieved title is determined. A title coordinate table 42 storing the retrieved title, the number of XY lines indicating the position, the number of XY lines indicating the character size, and the character direction is created and output to the address pasting unit 37.

  Further, the address search keyword presented by the character collating unit 32 is searched from the name list data, and the position and size of the searched address are determined by using the number of XY lines of the document stored in the document information table 41. Calculate as Further, the character direction (horizontal writing / vertical writing) of the searched address is determined. An address coordinate table 43 storing the retrieved address, the number of XY lines indicating the position, the number of XY lines indicating the character size, and the character direction is created and output to the address array prediction unit 34.

  The address arrangement prediction unit 34 plays a role of predicting the arrangement of the address included in the name list data (determining an arrangement rule) and outputting the result. Specifically, based on the document information table 41 and the address coordinate table 43, an array of addresses included in the name list data is predicted (calculated), and a predicted arrangement position is determined for each address (for each character that can be estimated as an address). An address arrangement table 44 in which the number of XY lines indicated and the number of XY lines indicating the character size are stored in association with each other is created and output to the address reading unit 35.

  The address reading unit 35 plays a role of reading and outputting the characters arranged at the position predicted by the address array prediction unit 34 from the name list data. Specifically, based on information stored in the address arrangement table 44 (predicted arrangement position and corresponding character size), characters are read from the name list data, and the characters are stored in the address arrangement table 44 in association with the original information. And output to the address determination unit 36.

  The address determination unit 36 plays a role of determining whether or not the character read from the name list data is an address and outputting the result. Specifically, the character collation unit 32 is requested to collate the characters stored in the address arrangement table 44, and address determination information indicating the result is stored in the address arrangement table 44 in association with the characters. A new address determination table 45 is created from the array table 44 and output to the address pasting unit 37.

  The address pasting unit 37 plays a role of pasting characters determined to be an address to document data and outputting them. Specifically, the character addressed based on the address determination information is read from the characters stored in the address determination table 45, pasted according to the title of the document data based on the title coordinate table 42, and the data Output to the output unit 24.

  In addition, when the CPU 11 determines the arrangement rule of the address (character information group) included in the name list data by executing the program, it corresponds to the character using the character (address search keyword) belonging to the address. The first character is searched, the second character is searched from the first character in a predetermined direction, the second character belonging to the address is specified, and the address array is based on the positional relationship between the first character and the second character. Serves to determine rules. In the present embodiment, the predetermined direction is the vertical and horizontal directions in the name list data (name list 60). The direction from the first character to search is the character direction of the first character (horizontal writing / vertical writing). ) And so on.

  Next, the operation of the multifunction machine 10 will be described.

  FIG. 3 shows a flow of processing in the automatic synthesis mode by the multifunction machine 10. The flowchart shown in FIG. 3 describes the outline of the processing, and details will be described later.

  In the multifunction machine 10, when the automatic synthesis mode (automatic address writing mode) is selected by the user operating the mode selection button, the CPU 11 starts this processing (Start). Subsequently, the above-described document 50 and name list 60 are read by the user's operation, and manuscript data (document data / name list data) is input (step S101).

  In inputting the manuscript data, for example, the user sets the document 50 on the automatic manuscript feeder, designates the type of manuscript as a document (composition destination manuscript), and then presses the start button. The CPU 11 sends the document 50 by the automatic document feeder via the document feeder controller 21. The document 50 is read by the scanner unit 22, and the image data input from the scanner unit 22 is recognized as image data (document data) of the document 50 and stored in the data storage unit 14.

  When the reading of the document 50 is completed, the user sets the name list 60 on the automatic document feeder, designates the type of document as the name list (extraction source document), and then presses the start button. The CPU 11 sends the name list 60 by the automatic document feeder via the document feeder controller 21. The name list 60 is read by the scanner unit 22, and the image data input from the scanner unit 22 is recognized as image data (name list data) of the name list 60 and stored in the data storage unit 14.

  The CPU 11 creates a document information table 41 storing document information necessary for determining (calculating) the title and address location and size included in the data from the document data and the name list data ( Step S102).

  Subsequently, the title described in the document 50 is searched by OCR processing on the document data and a search using the title search keyword. Also, a predetermined number (at least one) of addresses is searched from among a plurality of addresses described in the name list 60 by OCR processing for name list data and a search using address search keywords (step S103). Other addresses arranged around this address are specified with reference to the database 38, and the arrangement positions of all addresses are predicted based on the positional relationship of these addresses (step S104).

  The CPU 11 reads out individual characters existing at the predicted arrangement position from the list data as address candidates (step S105), and individually determines whether the address is an address with reference to the database 38 (step S106). The character determined to be the address is pasted in the vicinity of the title in the document data (step S107). The document data is input to the printer unit 24, the document with the combined address is printed out (step S108), and the process ends (End).

  Next, details of the automatic synthesis mode will be described with reference to FIGS. 4 to 16 in accordance with the procedure shown in the flowchart of FIG.

  In the present embodiment, a case will be described in which the name of the parents is individually entered in “notification material” distributed to the parents of the students by the school teacher. For example, since it takes time to hand-write the address one by one on the notification material, the document 50 as the notification material as shown in FIG. 4 and the name and address of the student are described using the automatic composition mode of the multifunction machine 10. By reading the name list 60 of the class (○ year ○ set), the address is automatically written on the notification material.

  The document 50 has a horizontal format for the entire document. An address entry field is provided on the right shoulder (upper right part) of the document, and a title 51 indicating “Like” is arranged. The title 51 has a size of 12 pt. The address is written horizontally on the front side (left side in the figure) of the title 51, the address itself is not filled in, and the front side of the title 51 is blank.

  In the name list 60, the format is written horizontally, and the names 61 and addresses 62 of a plurality of students are described in a list format. Specifically, the name 61 and address 62 for one student are arranged in the horizontal direction (left and right direction), and arranged in the vertical direction (up and down direction) in the order of the 50 sounds of the name 61. The character size is 8 pt.

  In this example, the document 50 and the name list 60 are both in the A4 size portrait orientation and will be described in the case of reading with a resolution of 600 dpi (dot per inch). Further, a case will be described in which the width direction (left-right direction) of the document is set to the X direction, the length direction (vertical direction) is set to the Y direction, and the origin position in the XY coordinates is set to the lower left corner shown in the drawing.

(1) Document data (document, name list) input In step S101 of FIG. 3, as shown in FIG. 4, the image data (document data 70) of the document 50 and the image data (name of the name list 60) input from the data input unit 22 are displayed. The name list data 80) is temporarily stored in the data storage unit 14.

(2) Document Information Creation In step S102 of FIG. 3, as shown in FIG. 4, the document information output unit 31 determines the line in the XY direction of the document from the document size and reading resolution of the input document data 70 and name list data 80. Calculate the number (number of dots).

  In the case of this example in which both the document 50 and the name list 60 are read in A4 size portrait orientation and resolution of 600 dpi, the number of lines in the XY direction is calculated as [X: Y = 4960 Line: 7015 Line], respectively. The document information output unit 31 creates a common document information table 41 that stores the document size, the resolution, and the number of XY lines. When at least one of the document size and the reading resolution is different between the document 50 and the name list 60 and the number of XY lines is different, each document information is individually stored in the document information table 41 or the document information Create tables individually.

(3) Title / Address Search Here, a method for searching for designated characters (name / address) from the document data 70 and the name list data 80 in step S103 of FIG. 3 will be described.

(3-1) Method for Retrieving Designated Titles from Document Data As shown in FIG. 5, document data 70 stored in the data storage unit 14 is input to the character coordinate detection unit 33. Along with this data input, the document information output unit 31 outputs the document information table 41 to the character coordinate detection unit 33. The character matching unit 32 acquires a preset title search keyword 90 from the database 38 and outputs it to the character coordinate detection unit 33. In this example, the title search keyword 90 is “sama”.

  The character coordinate detection unit 33 performs OCR processing on the document data 70 and searches for whether the character “sama” of the title search keyword 90 exists in the document data 70. If the character “sama” does not exist in the document data 70 and cannot be detected, the character coordinate detection unit 33 requests the character collation unit 32 to output the next title search keyword, and newly Search again using the title search keyword entered.

  As an example of the title search keyword, in addition to “sama”, “dono”, “san”, and the like can be cited. In addition, when the address is a company name or the like, “Gochu” or the like may be mentioned. Each time the character collation unit 32 receives an output request for a title search keyword from the character coordinate detection unit 33, the character collation unit 32 outputs the title search keyword to the character coordinate detection unit 33 in a preset order.

  In the case of this example, since the character “sama” exists in the document data 70 (honor title 51), the character coordinate detection unit 33 detects the character (search designated character). Then, the coordinate position and character size of the detected character (“sama”) are calculated.

  The coordinate position is the center position of the character for a single character, and the center position of the first character for a plurality of characters. The character size is the size of only that character in the case of one character, and the size of the entire plurality of characters in the case of a plurality of characters.

  In this example, the coordinate position is X: Y = 4600 Line: 6800 Line (equivalent to 180 mm: 267 mm), and the character size is X: Y = 105 Line: 105 Line (roughly converted value from 12 pt).

  The character coordinate detection unit 33 further detects a character direction as information necessary for pasting the address. If there are multiple search specified characters, the character direction is detected as horizontal writing (“horizontal”) if the character sequence is horizontal, and the character direction is written vertically (“vertical”) if the character sequence is vertical. ) To detect.

  In the case of a single character like “sama” in this example, since the character direction cannot be detected only by that character, the character coordinate detection unit 33 searches for another plurality of characters again to detect the character direction, The user may designate horizontal writing / vertical writing of the document 50, or the like. In the re-search of a plurality of characters, for example, the direction of the character can be detected by searching for a character that is always present in the sentence, such as “is.”, “Is.”, “Is”.

  The document 50 in this example is horizontal writing, and the above character direction is detected as horizontal writing. The character coordinate detection unit 33 stores the detected character (name 51) in the title field, and creates a name coordinate table 42 that stores the coordinate position and character size calculation result of the character, and the character direction detection result. .

(3-2) Designated Address Retrieval Method from Name List Data As shown in FIG. 6, name list data 80 stored in the data storage unit 14 is input to the character coordinate detection unit 33. Along with this data input, the document information output unit 31 outputs the document information table 41 to the character coordinate detection unit 33. The character matching unit 32 acquires a preset address search keyword 91 from the database 38 and outputs it to the character coordinate detection unit 33. In this example, the address search keyword 91 is “Sato”.

  When the document information table 41 is common to the document data 70 (document 50) and the name list data 80 (name list 60) as in this example, the character coordinates are read from the document information output unit 31 when the name list data 80 is input. The output of the document information table 41 to the detection unit 33 is omitted, and the character coordinate detection unit 33 proceeds with processing using the document information table 41 input from the document information output unit 31 when the document data 70 is input. Also good.

  The character coordinate detection unit 33 performs OCR processing on the name list data 80 and searches the name list data 80 for the character “Sato” as the address search keyword 91. If the character “Sato” does not exist in the list data 80 and cannot be detected, the character coordinate detection unit 33 requests the character collation unit 32 to output the next address search keyword, and newly Search again using the entered address search keyword.

  For example, the last name used in Japan is registered in the database 38 in advance, and the address search keyword is frequently used every time the character collation unit 32 receives an output request for the address search keyword from the character coordinate detection unit 33. It is possible to output from the last name (high detection probability) as the address search keyword 91 in order, or output the last name set by the user as the address search keyword 91.

  In the case of the former, for example, as shown in FIG. 7A, “Sato”, “Suzuki”, “Takahashi”, “ "Tanaka" etc. are output as address search keywords. In the latter case, for example, as shown in FIG. 7B, the last name “Suzuki” set by the user through the operation unit 17 is output as the address search keyword.

  In the case of this example, since the character “Sato” does not exist in the name list data 80, in the configuration illustrated in FIG. 7A, the character coordinate detection unit 33 makes the next address search keyword to the character matching unit 32. Request output. The character collation unit 32 outputs “Suzuki” to the character coordinate detection unit 33 as the next address search keyword 92 (see FIG. 6).

  Since the character “Suzuki” is present in the name list data 80 of this example (character 63), the character coordinate detection unit 33 detects the character (search designated character) and stops the search here. Then, the coordinate position and character size of the detected character 63 (“Suzuki”) are calculated. Note that the calculation method of the coordinate position and the character size according to whether it is a single character or a plurality of characters is the same as in the case of titles.

  In the case of this example, the coordinate position (center position of the first character “bell”) is X: Y = 800 Line: 4300 Line, and the character size (total size of two characters “Suzuki”) is X: Y = 140 Line: 70 Line (approximately converted value from 8 pt).

  The character coordinate detection unit 33 further detects a character direction as information necessary for pasting the address. Note that the method of detecting the direction of the character according to whether it is a single character or a plurality of characters is the same as in the case of titles. If the arrangement of two characters is “left and right” like “Suzuki” in this example, the character direction is detected as horizontal writing (“horizontal”). The character coordinate detection unit 33 stores the detected character (character 63) in the address field, and creates an address coordinate table 43A that stores the calculation result of the character coordinate position and character size, and the detection result of the character direction. .

  FIG. 8 is a flowchart showing a flow of the above mentioned title / address search process, and is a subroutine showing in detail step S103 of FIG.

  When the title / address search process is started, the character collation unit 32 outputs a title search keyword to the character coordinate detection unit 33 (step S111). The character coordinate detection unit 33 searches the document data 70 for a character (name) that matches (corresponds) to the designated title search keyword (step S112). When the search designated character cannot be detected (step S113; No), the character collating unit 32 is requested to output the next title search keyword (repetition after step S111).

  When the search designated character is detected (step S113; Yes), the arrangement position of the character (honor title) is calculated (step S114). Details are the above-described calculation of the coordinate position and character size, and detection of the character direction. Then, the character coordinate detection unit 33 stores the result in the title coordinate table 42.

  Subsequently, the character collation unit 32 outputs the address search keyword to the character coordinate detection unit 33 (step S115). The character coordinate detection unit 33 searches the name list data 80 for a character (address (last name)) that matches (corresponds) the designated address search keyword (step S116). When the search designated character cannot be detected (step S117; No), the character collating unit 32 is requested to output the next address search keyword (repetition after step S115 (in the case of the configuration illustrated in FIG. 7A)). .

  When the search designated character is detected (step S117; Yes), the arrangement position of the character (address) is calculated (step S118). Details are the above-described calculation of the coordinate position and character size, and detection of the character direction. Then, the character coordinate detection unit 33 stores the result in the address coordinate table 43 and ends this processing (Return).

(4) Address Array Position Prediction Here, a method for predicting the array position of the address (name described in the name list 60) included in the name list data 80 (determining the order rule) in step S104 of FIG. Will be described. In (3-2) described above, one address “Suzuki” is detected, and based on the detection result, how addresses other than “Suzuki” are arranged is predicted.

  As shown in FIG. 9, in general, when the name written in the name list is written horizontally, it is highly likely that the name is arranged vertically, and when the name is written vertically, the name is arranged horizontally. There is a high possibility. The address coordinate table 43A described above stores the character direction (horizontal writing / vertical writing) of the detected address. Based on this character direction, the next address is detected from the detected address ("Suzuki" in this example). Determine the direction to search.

  As shown in FIG. 10, the OCR-processed name list data 80 and the address coordinate table 43 are input from the character coordinate detection unit 33 to the address array prediction unit 34.

  When the detected address ("Suzuki") is written horizontally as in this example, it can be predicted that the other addresses are arranged vertically. The address array predicting unit 34 grasps the position of the first address (the character 63 of “Suzuki”) from the OCR-processed name list data 80 based on the coordinate position of the address coordinate table 43A, and the address coordinate table 43A. When the character direction of the character is “horizontal”, the character 64 of “Saito” adjacent to the upper side of the address (character 63 of “Suzuki”) and the character of “Tanaka” adjacent to the lower side are included in the list data 80. 65 is detected, and the character collation unit 32 is requested to collate whether or not the address is addressed.

  Further, when the character direction of the address coordinate table 43A is “vertical”, the character adjacent to the left side and the character adjacent to the right side of the first address is detected from the name list data 80, and character verification is performed. Requests the unit 32 to check whether the address is addressed.

  If there is a character that is adjacent to only one side of the top, bottom, left, or right, that one character may be detected, and the character collation unit 32 may be asked to collate whether the address is an address.

  The character collation unit 32 collates the input characters 64 and 65 with the address list of the database 38 and outputs the result to the address arrangement prediction unit 34 with conformity or nonconformity. In the case of this example, the match is output for both the characters 64 and 65.

  In response to this result, the address arrangement predicting unit 34 specifies (determines) that the character 64 “Saito” and the character 65 “Tanaka” are addresses (surnames), and the addresses are placed above and below the character 63 “Suzuki”. Predict that they are arranged. Subsequently, the address arrangement prediction unit 34 designates the character 64 (“Saito”) and the character 65 (“Tanaka”) adjacent to the character coordinate detection unit 33 in the upper and lower sides of the character 63 (“Suzuki”) and arranges them. Request position calculation.

  The character coordinate detection unit 33 detects the designated character 64 (“Saito”) and the character 65 (“Tanaka”) from the OCR-processed name list data 80 held therein, and the character 63 (“Suzuki”) described above. An address coordinate table is created in the same procedure as in the case, and is output to the address array prediction unit 34 as a calculation result of the arrangement position.

  For the character 64 (“Saito”), the coordinate position (center position of the first character “Sai”) is X: Y = 800 Line: 4370 Line, and the character size (the overall size of the two characters “Saito”) is X: Y = 140 Line: 70 Line, and the character direction is written horizontally. The character coordinate detection unit 33 stores “Saito” of the character 64 in the address column and creates an address coordinate table 43B storing the above-described information (see FIG. 12), and outputs it to the address array prediction unit 34.

  For the character 65 (“Tanaka”), the coordinate position (center position of the first character “Ta”) is X: Y = 800 Line: 4230 Line, and the character size (the overall size of the two characters “Tanaka”) is X: Y = 140 Line: 70 Line, and the character direction is written horizontally. The character coordinate detection unit 33 stores “Tanaka” of the character 65 in the address field and creates an address coordinate table 43C storing the above-described information (see FIG. 12), and outputs it to the address array prediction unit 34.

  Accordingly, the manuscript information output unit 31 outputs the manuscript information table 41 storing the manuscript information used for calculating the address arrangement to the address arrangement prediction unit 34.

  As shown in FIG. 12, the address array prediction unit 34 uses the information stored in the stored address coordinate table 43A, the input address coordinate tables 43B and 43C, and the manuscript information table 41. Calculate the address array. More specifically, the number of XY lines indicating the coordinate positions of the character 63 (“Suzuki”), the character 64 (“Saito”), and the character 65 (“Tanaka”) already known to be addressed and name list data 80 Based on the number of XY lines, the arrangement position of the address and the maximum arrangement number are calculated.

  As for the arrangement position of the address, in the case of this example, since the positions in the X direction of the above three characters are all 800 lines, the individual addresses are arranged in the vertical direction along X = 800 Line in the name list data 80. Is determined.

  Further, the address arrangement interval is calculated from the positions of the three characters in the Y direction. In the case of this example, if the distance (interval) in the Y direction between “Suzuki” and “Saito” adjacent to the upper side is calculated, 4370Line-4300Line = 70Line. When the distance (interval) in the Y direction between “Suzuki” and “Tanaka” adjacent to the lower side is calculated, 4300 Line−4230 Line = 70 Line. Thus, since the interval between adjacent characters is 70 lines, it is determined that the address arrangement interval is 70 lines.

  By combining these, it is determined that each address is arranged (arranged) at intervals of 70 lines in the vertical direction along X = 800 lines in the name list data 80 as the address arrangement position.

  The maximum number of addresses is calculated based on the position of any of the characters “Suzuki”, “Saito”, and “Tanaka”. For example, when “Suzuki” is used as a reference, the number of addresses arranged above and the number of addresses arranged below are calculated by the following equations. In addition, if a fraction occurs in the division result of the following formula, it is rounded down.

Number of upper arrangements with respect to the reference character: A = {number of Y lines of the original (7015 Line) −number of Y lines of the reference character (4300 Line)} ÷ address arrangement interval (70 Line)
Number of lower arrangements with respect to the reference character: B = number of Y lines of the reference character (4300 lines) ÷ address arrangement interval (70 lines)

  As shown in the following formula, the maximum number of addresses is a value obtained by adding the above A and B and “1” for the reference character.

Maximum number of addresses to be arranged: N _max = A + B + 1

  In this example, the calculation of the address arrangement position and the maximum number of addresses is described in the case where the address arrangement direction is the vertical direction. However, even when the address arrangement direction is the horizontal direction, “X” and “Y ”Is replaced, and“ upper and lower ”is replaced with“ left and right ”, and the calculation method itself is the same.

  Upon completion of the above calculation, the address array predicting unit 34 has a matrix-like configuration in which an address field, a coordinate position field (number of XY lines), and a character size field (number of XY lines) are provided in correspondence with the number of addresses. An address array table 44 is created, and the coordinate position and character size for each address are stored. In addition, “Suzuki”, “Saito”, and “Tanaka”, which are already known as addresses, are stored in the corresponding address fields.

  FIG. 13 is a flowchart showing a flow of the above-described address array position prediction process, which is a subroutine showing in detail step S104 of FIG.

  When the address array position prediction process is started, the character coordinate detection unit 33 outputs the OCR-processed name list data 80 and the address coordinate table 43A to the address array prediction unit 34. Based on the character direction of the address (“Suzuki”) stored in the address coordinate table 43A, the address array predicting unit 34 predicts the address array direction in the name list data 80 based on the address (step S121).

  Subsequently, the characters (“Saito”, “Tanaka”) arranged in the predicted arrangement direction with respect to the above address (“Suzuki”) are detected from the name list data 80 as address candidates (step S122), and the character matching unit 32 is requested to verify whether it is the address. The character collation unit 32 performs collation with the address list in the database 38 and outputs the result to the address array prediction unit 34 (step S123).

  If not collated (step S124; No), the next character arranged in the predicted arrangement direction is detected as a candidate address and a similar collation is requested to the character collation unit 32 (repetition after step S122).

  When collation is performed (step S124; Yes), the character coordinate detection unit 33 calculates an arrangement position of the character (address) (step S125). Details are the calculation of the coordinate position and character size, and the detection of the character direction. The character coordinate detection unit 33 stores the result in the address coordinate tables 43B and 43C, and inputs the address coordinate tables 43B and 43C to the address array prediction unit 34. The document information output unit 31 inputs the document information table 41 to the address arrangement prediction unit 34.

  The address array predicting unit 34 determines the character (address) based on the number of XY lines of each address stored in the address coordinate tables 43A, 43B, and 43C and the number of XY lines of the name list 60 stored in the document information table 41. ) (Address arrangement position and maximum number of arrangements) are calculated, and the result is created as the address arrangement table 44 (step S126), and this process is terminated (Return).

(5) Address Reading In step S105 in FIG. 3, the address is read from the name list data 80 based on the above-described address array prediction (calculation result). As shown in FIG. 14, the name list data 80 that has been subjected to the OCR processing and the address arrangement table 44 are input from the address arrangement prediction unit 34 to the address reading unit 35. The address reading unit 35 is arranged at each coordinate position (number of XY lines) of the address array table 44 from the name list data 80 except for “Suzuki”, “Saito”, and “Tanaka”, which are already known to be addressed. Is detected and stored in the corresponding address field. The detected and stored results are as shown in the figure. Note that “blank” indicates a blank space and no characters.

(6) Address determination Since various characters other than the address are described in the name list 60, it is necessary to determine whether or not the character read from the name list data 80 is the address. This determination is performed in step S106 of FIG. As shown in FIG. 15, the address arrangement table 44 is input from the address reading unit 35 to the address determination unit 36. The address determination unit 36 inputs the characters stored in the address column of the address array table 44 in order to the character verification unit 32 and requests verification of whether the address is the address.

  The character collation unit 32 collates the input characters with the address list in the database 38 and outputs the result to the address determination unit 36 with conformity (◯) or non-conformity (×).

  For example, the “(blank)” portion without characters is not registered in the address list, and is not suitable. Since characters such as “Year ○ Class” and “Name” are not registered in the address list, they are not suitable. On the other hand, “Ando”, “Saito”, “Suzuki”, “Tanaka”, “Watanabe”, etc. are registered in the address list and are relevant.

  The address determination unit 36 creates an address determination table 45 in which an address determination column is added to the address arrangement table 44, and matches (◯) or non-conformity (×) from the character matching unit 32 to the address determination column of the corresponding character. The address determination table 45 is stored and completed. The information stored in the address determination column indicates that “o” is the address (address conformance) and “x” indicates that the address is not addressed (address incompatibility).

  For “Suzuki”, “Saito”, and “Tanaka” that have already been identified as addresses, an address determination column is provided in advance when the address array table 44 is created, and conformance (○) is stored. Thus, the collation request to the character collation unit 32 may be omitted.

  In addition, it is estimated that there are more than tens of thousands of last names used in Japan, and it is said that there are more than several thousand even general ones. To increase the accuracy of address (last name) judgment, the number of registered databases is increased. If it is increased, there is a concern that it takes time for verification. In addition, when the company name of a business partner is registered and collated as an address, there is a similar concern when the number of registrations increases to several hundred companies. In any case, when the number of addresses in the name list is large and the number of characters to be verified (address determination number) is large, there is a concern that the time required for verification of all characters may become long.

  Therefore, although it is clear that it is not an address, a character frequently used in a name list or the like, for example, a word indicating the name of the name list or the type of description (information) is registered in the database 38 as a non-address list, and the target character May be searched first in the non-address list.

  For example, in addition to the above “Year ○ Class” and “Name”, the characters to be registered in the non-address list are “name list”, “name”, “male”, “female”, “school” (elementary / junior high school) / Vocational school), "high school", "university", "class" (class list), "contact network" (emergency contact network), "company", "executive" (director list), "employee" (employee list) , “Group” (group list), “circle” (circle list), “resident's association” (local community list), and the like. Furthermore, “address”, “postal code”, “phone number”, “date of birth”, “remarks”, and the like may be added.

  By searching for the target character first in such a non-address list and making it non-conforming (x) if applicable, the efficiency and accuracy of address determination can be improved. Furthermore, this non-address list may be used for collation of characters (in this example, “Saito” and “Tanaka”) when predicting the above-described address arrangement.

  Further, the result of the address determination may be displayed on the display unit 18 so that the user can check whether there is an error, or can be confirmed or corrected through the operation unit 17. For example, the characters detected as address candidates and the address determination result (“◯” / “×”) are displayed. If there is no error, it will be confirmed, and “x” added to the addressed character by incorrect determination will be corrected to “○”, and “○” added to the non-addressed character will be corrected to “×”. You may do it.

  Further, depending on the contents of the document, not all addresses listed in the name list may be targeted, but only specific addresses may be targeted. In such a case, the user can intentionally change from “◯” to “X” for characters that are addressed but do not need to be output.

(7) Address Pasting In step S107 of FIG. 3, as shown in FIG. 16, the character coordinate detection unit 33 outputs the honorific coordinate table 42 to the address pasting unit 37, and the address determination unit 36 sends the address determination table 45 to the address. Output to the pasting unit 37.

  The address pasting unit 37 calculates the pasting position of the address based on the input information of each table. In this example, the following three conditions are set in advance as conditions for the pasting position.

・ Condition A: Align the address and the title.
⇒If the text direction is horizontal writing, align the Y direction.
If the character direction is vertical writing, align the position in the X direction.
Condition B: The size of the address and title characters (font size) are aligned.
Condition C: A space for one character is provided between the address and the title.

  According to the title coordinate table 42, the coordinate position of the title 51 (“sama”) is X: Y = 4600 Line: 6800 Line, the character size is X: Y = 105 Line × 105 Line, and the character direction is horizontal writing.

  For example, when the address 52 “Suzuki” is pasted, the position in the Y direction is 6800 lines from the condition A. From condition B, the character size is expanded 1.5 times from 8 pt (70 Line × 70 Line) to 12 pt (105 Line × 105 Line) per character, and the character size of the entire two characters “Suzuki” is X: Y. = 140 Line × 70 Line is changed to X: Y = 210 Line × 105 Line. From condition C, a space of X: Y = 105 × 105 lines corresponding to one character is provided between “Suzuki” and “Sama”.

  As a result, the position where “Suzuki” is pasted (the center position of the first character “bell”) is X: Y = 4285 Line: 6800 Line. It should be noted that the pasting position (center position) of “tree” is X: Y = 4390 Line: 6800 Line.

  Further, the address pasting unit 37 counts the number of “○” of address determination in the address determination table 45, obtains the number of finally confirmed addresses, and requests output of document data 70 for the number of addresses. . As a result, the document data 70 stored in the data storage unit 14 is sequentially input to the address pasting unit 37.

  The address pasting unit 37 sequentially reads addresses whose address determination is “◯” from the address column of the address determination table 45, changes the character size, and pastes it at the above position in the document data 70. The output of the document data 75 with the address pasted to the printer unit 24 is repeated until the last address is completed.

(8) Addressed Document Output In step S108 of FIG. 3, as shown in FIG. 16, the printer unit 24 forms an image corresponding to the document data 75 input from the address pasting unit 37 on a sheet, and the addressed document It outputs as 55.

  As described above, in the MFP 10 according to the present embodiment, the character information (address search keyword) belonging to the address for which the arrangement rule is to be determined from among the plurality of character information included in the input name list data 80. ) To search for the first character information corresponding to the character information (“Suzuki” in this example). The second character information (in this example, “Saito” and “Tanaka”) belonging to the address is specified by searching from the first character information in a predetermined direction (up / down / left / right direction), and the first character information and the second character information The arrangement rule of the address is determined based on the positional relationship with the character information. For example, as described above, it is determined that the address is arranged at an interval between the first character information and the second character information in the column direction including the first character information and the second character information. Thereby, the arrangement rule of the address included in the name list data 80 can be determined.

  Further, when searching for the first character information from the name list data 80, it is used in order from the address search keyword having the highest detection probability in the search for the address (surname) (in this example, “Sato” → “Suzuki” → “ The corresponding character information is searched, and when the character information corresponding to the address search keyword is detected, the character information is determined as the first character information. Thus, the search efficiency of the first character information can be improved by performing the search using the address search keywords having the highest detection probability in order.

  Further, when searching for the second character information from the first character information, if the first character information is horizontal writing, the search is performed in the vertical direction by predicting that the address is arranged in the vertical direction. In the case of writing, it is predicted that the address is arranged in the left-right direction and the search is performed in the left-right direction. Thus, by determining the search direction of the second character information based on the character direction (horizontal writing / vertical writing) of the first character information, the second character information can be efficiently searched and specified. it can.

  When the address extracted from the name list data 80 (name list 60) is combined with the document data 70 (document 50), the first manuscript data for the title (designated character information) searched from the document data 70 is used. When combining the character information extracted from the above, the position and format (address size, character direction, etc.) of the address to be combined are determined based on the position and format of the title. As a result, it is possible to arrange the addresses to be synthesized with respect to the titles in a good-looking manner or to make the format look good.

  In addition, when it is configured such that the result of address determination for characters extracted from the list data 80 is displayed on the display unit 18 and can be confirmed and corrected by the user through the operation unit 17, it is determined erroneously by automatic determination. It is possible to suppress waste such as a mailing address or a mailing address that does not need to be output is synthesized with a document.

[Second Embodiment]
In the first embodiment, the case where the address is described in one column in the name list has been described. In the second embodiment, the case where the address is described in a plurality of columns will be described.

  Some names have names (names) listed in multiple columns. In this case, as shown in FIG. 17, it is possible that other addresses are arranged on the top, bottom, left and right of the first detected address (for example, “Suzuki”) regardless of whether the address is written horizontally or vertically. High nature.

  As an example of such a name list, as shown in FIG. Using this name list 200 as an example, a method for determining an array rule when addresses are arranged in a plurality of columns (two-dimensional array) will be described.

  First, the image data (document data 70) of the document 50 and the image data (name list data 210) of the name list 200 are input from the data input unit 22 by the document data input described in the first embodiment. It is stored in the storage unit 14. Subsequently, (2) document information creation and (3-1) designated title retrieval method from document data are similarly performed.

  Subsequently, the character coordinate detection unit 33 performs OCR processing on the name list data 210, and (3-2) the designated address search method from the name list data described in the first embodiment (see FIG. 6) is used as the address search keyword. Search for the corresponding character.

  Since the first address search keyword “Sato” does not exist in the name list data 210 of this example, the second address search keyword “Suzuki” is searched. There are two “Suzuki” characters 201 arranged in the male column (first column) and characters 202 arranged in the female column (second column). The character coordinate detection unit 33 stops the search when any one of these is detected. Here, the case where the character 201 in the first column is detected will be described.

  Next, (4) address arrangement position prediction described in the first embodiment (see FIGS. 10 to 13), the OCR-processed name list data 210 is input from the character coordinate detection unit 33 to the address arrangement prediction unit 34. Then, the address arrangement prediction unit 34 detects each character arranged above and below the character 201, and creates an address arrangement table 220 as shown in FIG. In the address column of the address arrangement table 220, the characters arranged in the first column of the name list data 210 are stored by (5) address reading described in the first embodiment (see FIG. 14). .

  Subsequently, the character coordinate detection unit 33 performs address search in the left-right direction (second and subsequent columns). The address search in the left-right direction starts, for example, with reference to characters (including blanks) of address candidates arranged immediately near the origin position (lower left corner) in the XY coordinates of the name list data 210. In the case of this example, as shown in FIG. 19, the address search in the horizontal direction is sequentially performed in the order of “(blank)” → “Watanabe”.

  As shown in FIG. 20, first, the characters arranged in the left-right direction of the blank 203 are detected. In the case of this example, since there are no characters on the left and right of the blank 203, it cannot be detected. In such a case, the process proceeds to the next character (“Watanabe”).

  In the detection of the character arranged in the left-right direction of the second “Watanabe” character 204, the “Watanabe” character 205 arranged on the right side is detected in this example. The address arrangement prediction unit 34 requests the character collation unit 32 to collate whether or not this character is the address.

  In this example, “Watanabe” of the character 205 is identified (determined) as an address (surname), and the address array predicting unit 34 predicts that the address is arranged above the character 205. Subsequently, a character adjacent to the upper side of the character 205 is detected, and the character collation unit 32 is requested to collate whether or not the address is the address. When it is specified that this character is the address, the address array predicting unit 34 designates the character 205 (“Watanabe”) and the character located above it to the character coordinate detecting unit 33, and sets the positions of those characters. Request a calculation.

  The character coordinate detection unit 33 outputs the calculation result of the arrangement position for the designated character to the address arrangement prediction unit 34 as an address coordinate table. Based on the address coordinate table, the address array prediction unit 34 calculates the address position and the maximum number of addresses for the second column in the same procedure as the first column. Subsequently, an address array table is created by adding an address field in the second column, a coordinate position field (number of XY lines), and a character size field (number of XY lines) to the address array table 220 in the first column. (See FIG. 12), and output to the address reading unit 35 (see (4) Address array position prediction).

  The address reading unit 35 detects characters arranged at each coordinate position in the second column in the address array table from the name list data 210 and stores them in the corresponding address field. As a result, an address arrangement table 221 as shown in FIG. 21 is created (see (5) Address reading).

  Thereafter, (6) address determination to (8) addressed document output is performed with the contents obtained by replacing the address arrangement table 44 described in the first embodiment with the address arrangement table 221 described above. As a result, as many documents as the number of addresses in which the individual addresses described in two columns in the name list 200 are entered are printed out from the multifunction device 10.

  As described above, according to the present embodiment, it is possible to determine an address arrangement rule for a name list in which addresses are described in a plurality of columns. In addition, here, the case where the first column of characters 201 in the name list 200 (name list data 210) is detected has been described. However, even when the second column of characters 202 is detected, the arrangement rule for a plurality of columns is similar in the above-described procedure. Can be determined.

[Third Embodiment]
In the first and second embodiments, the case where the name list is one page (when the address is described on one page) has been described. However, in the third embodiment, the name list is composed of a plurality of pages (the number of addresses is plural). (If it is listed on the page).

  Some names have names (names) written over a plurality of pages. As an example of such a name list, as shown in FIG. 22, there is a case where names are described on different pages for each gender. Taking the name list 300 as an example, a method for determining an arrangement rule when addresses are described on a plurality of pages will be described.

  The name list 300 of this example has a total of two pages. A male name is described on the first page 301 and a female name is described on the second page 302.

  First, as described in the first embodiment (1), by inputting document data, the image data of the document 50 (document data 70) from the data input unit 22 and the image data of the first page 301 in the name list 300 (name list one page) Eye data 311) and image data of the second page 302 (name list second page data 312) are input and stored in the data storage unit 14.

  The multifunction device 10 acquires the number of pages of the name list in this data input. The number of pages can be acquired by, for example, counting the number of pages by reading a designated name list in a configuration in which the user can input the number of pages in the name list through the operation unit 17 or specify the type of document to be read as described above. It can be done by.

  In the subsequent (2) document information creation, the document information output unit 31 adds the number of pages in the name list (two pages in this example) as name list data in addition to the document information table 41 described in the first embodiment. The prepared document information table 320 is created.

  (3) In the title / address search, in (3-2) the designated address search method from the name list data, and (4) address array position prediction, the name list is determined from the number of pages stored in the name list data of the document information table 320. The number of 300 pages is grasped, and the above (3-2) and (4) are sequentially performed on the name list first page data 311 and name list second page data 312. And the address arrangement | sequence prediction part 34 produces the address arrangement | sequence table 330 as shown in FIG. 23 as a result of address arrangement position prediction.

  In the address arrangement table 330 of this example, a page field (“page”) corresponding to the number of addresses is additionally provided in the storage field of the coordinate position (the number of XY lines), and the number of pages of the name list from which the address is read is the page. Stored in the column.

  (5) In the address reading, the address reading unit 35 reads the address from the name list first page data 311 and name list second page data 312 using the address arrangement table 330. FIG. 23 shows the result of reading the address from the two-page name list 300 shown in FIG.

  Thereafter, (6) address determination to (8) addressed document output is performed with the contents obtained by replacing the address arrangement table 44 described in the first embodiment with the address arrangement table 330 described above. As a result, documents corresponding to the number of addresses in which individual addresses described in the two-page name list 300 are entered are printed out from the multifunction device 10.

  As described above, according to the present embodiment, it is possible to determine an address arrangement rule for each page with respect to a name list in which the address is described over a plurality of pages.

  The embodiment of the present invention has been described with reference to the drawings. However, the specific configuration is not limited to that shown in the embodiment, and there are changes and additions within the scope of the present invention. Are also included in the present invention.

  For example, in the above-described embodiment, the address (name) described in the name list has been described as an example of the type of character information group for determining the arrangement rule. However, the type of character information group to be targeted is not limited to this. Absent. An address other than the personal name (name) or a destination may be used.

  Examples of the former include various group names such as company names, store names, institution names, organization names, and corporate names. Examples of the latter include an address, a postal code, a telephone number, and an e-mail address.

  In addition, an arrangement rule of a plurality of types of character information groups included in the first manuscript data is determined and extracted and combined, such as extracting an address and an address from a name list and combining them into a document. The second original data may be combined and output.

  The second document data is not limited to document data as described in the embodiment. For example, it is also possible to extract character information described in a name list of a specific format and combine it with a name list of another format. In this case, it is different from name list data as the first document data. The list data becomes the second document data.

  Specifically, only a certain type of text information (for example, only name and phone number) is extracted from a list that contains multiple types of text information (for example, name, address, postal code, telephone number). Editing to transfer to another name list, format change (format change) to change from a vertically written name list to a horizontally written name list, and the like.

  (4) In the address arrangement position prediction, the search direction of the next address is determined based on the character direction of the address detected first, but the second and subsequent addresses are specified by searching in the determined direction. If not, the address may be specified by searching in a direction different from that direction. For example, if the letter direction of the first detected address is horizontal writing, and if the second and subsequent addresses cannot be specified by first searching up and down from that address, the next address will be searched in the left and right direction. And so on.

  Further, the search direction of the second and subsequent addresses may be determined based on conditions other than the above character direction.

  Specifically, the search direction specified by the user is determined, or the first detected address is compared with the character format adjacent to the surroundings (up, down, left and right), and the search direction is determined based on the comparison result. It is possible to do. As the format in the latter case, a character size or a typeface can be used.

  For example, it can be predicted that the address is unlikely to be arranged in the direction in which characters with different formats are arranged, so that direction can be excluded from the search direction and the remaining direction can be determined as the search direction It is. Further, the search direction of the second and subsequent addresses may be determined comprehensively based on the condition based on the character format and the condition based on the character direction.

  (7) The conditions of the pasting position described in the address pasting are not limited to the above three conditions. Various conditions can be set for the pasting position according to the format of the document. Moreover, you may comprise so that an administrator, a user, etc. can set and change conditions.

  In the determination of the arrangement rule for a plurality of columns described in the second embodiment, address search (address search in the left-right direction) after the second column performed after the address sequence prediction of the first column is performed immediately after the origin position of the document. The addresses are arranged in order from the addresses arranged in the above, but the order is not limited to this.

  In addition, although the arrangement rule of the address is performed for each column, for example, the arrangement rule determined in the first column may be applied to other columns. In this case, the arrangement rule determination processing can be reduced.

  In addition, the arrangement rule can be determined even in three or more columns. In this case, for example, based on the positional relationship (interval) between the first column address and the second column address, the direction (vertical direction in the embodiment) is different from the direction (vertical direction in the embodiment). The direction rule may be determined.

  In the determination of the arrangement rule for a plurality of pages described in the third embodiment, the arrangement rule is determined for each page. However, the arrangement rule determined for a specific page is applied to other pages. Also good. The specific page in this case is not limited to the first page. For example, a page including character information corresponding to the first search keyword is a specific page.

  Further, although the database 38 used for presentation and collation of character information is provided inside the multi-function device 10, such a database may be provided outside and accessed by the multi-function device 10.

  Data input to the multifunction machine 10 is not limited to input by reading a document described in the embodiment. For example, print data may be input from a client terminal connected via a network, or file data may be input from an information storage device capable of external connection such as a USB (Universal Serial Bus) memory.

  Data output from the multifunction machine 10 is not limited to printing output described in the embodiment. For example, output by data transfer to an external device such as a client terminal or server connected via a network, or output of file data to the USB memory or the like may be used.

  The printer unit 24 in the multifunction machine 10 may be another type of printer such as an LED printer that scans and exposes a photosensitive drum with an LED (Light Emitting Diode) instead of a laser beam.

  The present invention is not limited to the multi-function device described in the embodiment, but can be applied to other image processing apparatuses such as a copying machine and a printer.

1 is a block diagram illustrating a schematic configuration of a multifunction peripheral as an image processing apparatus according to a first embodiment of the present invention. FIG. 3 is a functional block diagram illustrating a functional block configuration of a CPU in an automatic synthesis mode of the multifunction peripheral according to the first embodiment of the present invention. 3 is a flowchart showing processing in an automatic synthesis mode by the multifunction peripheral according to the first embodiment of the present invention. FIG. 4 is an explanatory diagram for explaining from document data input to document information creation in FIG. 3. It is explanatory drawing explaining the designation title search method of title / address search in FIG. It is explanatory drawing explaining the designation | designated address search method of honorific name / address search in FIG. It is explanatory drawing explaining the structure which outputs the address search keyword in the designated address search of FIG. FIG. 4 is a flowchart showing a title / address search process (subroutine) in FIG. 3. FIG. It is explanatory drawing explaining the relationship between the character direction of an address and the search direction of an address. It is explanatory drawing explaining the address arrangement position prediction (address collation) in FIG. It is explanatory drawing explaining the address arrangement | positioning position prediction (address arrangement position calculation) in FIG. It is explanatory drawing explaining the address arrangement | sequence table produced from the address coordinate table and the original information table by the address arrangement | positioning position prediction in FIG. It is a flowchart which shows the process (subroutine) of address arrangement position prediction in FIG. It is explanatory drawing explaining the address reading in FIG. It is explanatory drawing explaining the address determination in FIG. FIG. 4 is an explanatory diagram for explaining from address pasting to addressed document output in FIG. 3. It is explanatory drawing explaining the example in which the address of horizontal writing / vertical writing is arranged in multiple columns in the determination of the arrangement rule for the multiple columns of addresses according to the second embodiment of the present invention. It is a figure which shows the name list with which the address which concerns on the 2nd Embodiment of this invention was described in multiple columns. It is a figure which shows the address arrangement | sequence table in which the address of the 1st column in the name list | wrist of FIG. 18 was stored. It is explanatory drawing explaining the address search and address arrangement position prediction of the 2nd column in the name list of FIG. It is a figure which shows an example of the address arrangement | sequence table produced with respect to the name list | wrist with which the address is described in several columns, and the state in which the 1st column and the 2nd column address in the name list | wrist of FIG. 18 were stored. It is a figure which shows the manuscript information table produced by the name list by which the address which concerns on the 3rd Embodiment of this invention was described in several pages, and the input of the name list data. The figure which shows an example of the address arrangement | sequence table produced with respect to the name list with which the address was described in several pages, and the state which shows the state where the address described in the 1st page and the 2nd page in the name list of FIG. 22 was stored. is there.

Explanation of symbols

10 ... MFP (MFP)
11 ... CPU
12 ... Bus 13 ... ROM
DESCRIPTION OF SYMBOLS 14 ... RAM / data preservation | save part 15 ... Non-volatile memory 16 ... Hard disk apparatus 17 ... Operation part 18 ... Display part 19 ... Facsimile communication part 20 ... Network communication part 21 ... Document feeder control part 22 ... Scanner part / data input part 23 ... Image processing unit 24 ... Printer unit / data output unit 31 ... Document information output unit 32 ... Character collation unit 33 ... Character coordinate detection unit 34 ... Address array prediction unit 35 ... Address reading unit 36 ... Address determination unit 37 ... Address pasting unit 38 ... Database 41 ... Document information table 42 ... Title coordinate table 43 ... Address coordinate table 43A, 43B, 43C ... Address coordinate table 44 ... Address array table 45 ... Address determination table 50 ... Document 55 ... Document 51 ... Title 52 ... Address 60 ... Name list 61 ... Name 62 ... Address 63, 64, 65 ... Character 70 Document data 75 ... Document data 80 ... Name list data 90 ... Title search keywords 91 and 92 ... Address search keywords 200 ... Name lists 201 and 202 ... Characters 203 ... Blanks 204 and 205 ... Characters 210 ... Name list data 220 and 221 ... Address array table 300 Name list 301 ... First page 302 ... Second page 311 ... Name list first page data 312 ... Name list second page data 320 ... Original information table 330 ... Address array table

Claims (15)

An input section for inputting document data;
The character information belonging to the type is used when determining the arrangement rule of the character information group of the same type from among the plurality of character information included in the first document data input from the input unit The first character information corresponding to the character information is searched, the second character information belonging to the type is specified by searching from the first character information in a predetermined direction, and the first character information and A control unit that determines the arrangement rule based on a positional relationship with the second character information;
An image processing apparatus comprising: In the search for the first character information, the control unit detects the first character information from a plurality of character information included in the first document data in order from character information having a high detection probability in the search. The image processing apparatus according to claim 1, wherein the character information is determined as the first character information. The control unit searches from the third character information arranged in the first column including the first character information and the second character information in a direction different from the first column direction. 3. The image processing apparatus according to claim 1, wherein when the fourth character information belonging to the type can be specified, the arrangement rule is determined in consideration of a position of the fourth character information. . The control unit searches the fourth character information in a direction along the first column to identify fifth character information belonging to the type, and includes the fourth character information and the fifth character. The image processing apparatus according to claim 3, wherein the arrangement rule for the second column including the fourth character information and the fifth character information is determined based on a positional relationship with the information. The control unit determines a second column including the fourth character information along the first column based on a positional relationship between the first character information and the second character information. The image processing apparatus according to claim 3, wherein the arrangement rule for the second column is determined by applying the arrangement rule. The control unit determines the arrangement rule in a direction different from the first column direction based on a positional relationship between the third character information and the fourth character information. An image processing apparatus according to 1. 7. The image according to claim 1, wherein when the first document data is data of a plurality of pages, the control unit determines the arrangement rule for each page. Processing equipment. The image processing apparatus according to claim 1, wherein the control unit determines the predetermined direction based on a character direction of the first character information. The control unit compares the format of the first character information with the format of character information adjacent to the first character information, and character information having a different format is arranged with respect to the first character information. The image processing apparatus according to claim 1, wherein the current direction is excluded from a direction in which the second character information is searched. The control unit extracts individual character information belonging to the type according to the arrangement rule from a plurality of character information included in the first document data, and inputs the character information from the input unit. The image processing apparatus according to any one of claims 1 to 9, wherein the image processing apparatus synthesizes and outputs the synthesized second original data. The control unit retrieves sixth character information corresponding to the character information from the second document data using the designated character information, and based on the position and format of the sixth character information. The image processing apparatus according to claim 10, wherein a position and a format of character information to be combined with the second document data are determined. The image processing apparatus according to claim 1, wherein the control unit performs the identification with reference to a database in which character information belonging to the type is registered. The control unit determines whether or not the extracted character information according to the arrangement rule belongs to the type by referring to a database in which character information belonging to the type is registered. Item 12. The image processing device according to any one of Items 1 to 11. In determining whether or not it belongs to the type, the control unit also refers to a database in which a condition indicating that the type does not belong is registered, and determines that character information corresponding to the condition does not belong to the type The image processing apparatus according to claim 13. A display unit and an operation unit;
The image processing apparatus according to claim 13, wherein the control unit displays the result of the determination on the display unit, and accepts a confirmation or correction operation for the determination result by the operation unit.

Images