JP2011145725A - Character recognition unit and text processing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To shorten a time to be spent on the character recognition of a plurality of originals having a common section and individual sections. <P>SOLUTION: A character recognition unit is configured to perform the character recognition of a plurality of originals having the common section and individual sections, and is provided with: a first acquisition means for acquiring the respective images of the plurality of originals; a second acquisition means for acquiring specification information specifying the common section obtained on the basis of the first original which becomes the object of character recognition first among the plurality of originals; a setting means for setting the common section and individual sections as object regions to be subjected to the character recognition for the first original, and for setting the object region in regions other than the common section for the second and following originals; and a recognition means for performing character recognition only to the object regions. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a character recognition unit for recognizing characters of a document and a sentence processing apparatus for processing a result of character recognition.

  In recent years, a character recognition device that reads characters written on a manuscript and uses the read results has started to be used for reading business card information.

  In a character recognition device for a business card, for example, a paper image is first taken into the character recognition device using a scanner (image input), “company name”, “name”, “address”, “phone number”, etc. A character recognition target area in which there is a character group to be recognized for each item is set (target area setting).

  Next, it recognizes what each character contained in each target area is by using a dictionary according to the attribute of each item (character recognition), and uses language information such as words The read characters are corrected (post-processing).

  By the way, although the description position (layout) of each item in a business card is usually different for each company, for example, the same type of business card such as a business card of the same company often has almost the same layout.

  Therefore, in the character recognition device described in Patent Literature 1, region position information and item names are set and stored as a template in association with each target region for character recognition, and the same kind of business card character recognition processing is performed. At this time, by selecting and using a template corresponding to the business card to be processed, the efficiency of character recognition is improved by eliminating the trouble of setting a character recognition target area for each business card.

JP 2001-202475 A

  However, in the character recognition device of Patent Document 1, when performing character recognition of a plurality of business cards of the same type, common information such as “company name”, “address”, etc. that is common to the contents of each business card is used. As for items that have individual names, it is necessary to perform character recognition for each business card as in the case of items having individual information for each business card such as “name”.

  Character recognition is usually the process that requires the most processing cost from the viewpoint of the capability and processing time of the processing device among the above-described series of character reading processing steps. Depending on the character recognition device of Patent Document 1, There is a problem that it is difficult to shorten the processing time or to reduce the cost by adopting an apparatus having a low processing capability.

  The present invention has been made to solve these problems, and shortens the time required for character recognition by efficiently reducing the amount of data to be character-recognized when character recognition is performed on a plurality of similar business cards. It aims at providing the technology which can be done.

  In order to solve the above-described problems, the invention of claim 1 is a character recognition unit that performs character recognition of a plurality of documents having a common part and individual parts, and is a first unit that acquires each image of the plurality of documents. Acquisition means, second acquisition means for acquiring specific information for identifying the common part, obtained based on a first original that is first subjected to character recognition among the plurality of originals, Setting means for setting the common part and the individual part as target areas for character recognition for the first original and setting the target area in an area other than the common part for the second and subsequent originals And a recognition means for performing character recognition only on the target area.

  The invention according to claim 2 is the character recognition unit according to claim 1, further comprising storage means for storing the specific information.

  The invention of claim 3 is the character recognition unit according to claim 2, further comprising third acquisition means for acquiring the specific information from the storage means in response to a predetermined operation input, When the third acquisition unit acquires the specific information, (a) the character recognition unit disables the second acquisition unit, and (b) the setting unit acquires the acquired information. The target area is set for each of the plurality of documents based on specific information.

  According to a fourth aspect of the present invention, there is provided a text processing apparatus, the character recognition unit according to any one of the first to third aspects, and the common part in the first original. And a generation unit that generates character data of each document based on the result of character recognition and the result of character recognition for the target area in each of the second and subsequent documents.

  According to each of the first to fourth aspects of the present invention, specific information for specifying a common part common to each original is acquired based on the first original, and character recognition of the second and subsequent originals is performed. Since the target area for character recognition is set in an area other than the common part, character recognition for the common part is not required, and the amount of data subject to character recognition can be reduced efficiently. Can be shortened.

It is a figure which shows an example of the external appearance of the text processing apparatus which concerns on embodiment. It is a block diagram which shows an example of the main structures of the text processing apparatus which concerns on embodiment. It is a block diagram which shows one example of the main function structures of the text processing apparatus and character recognition unit which concern on embodiment. It is a figure which shows an example of a structure of a business card. It is a figure which shows an example of the selection screen of a common part. It is a figure which shows an example of the positional information on the common part in a business card. It is a figure which shows an example of a template. It is a figure which shows an example of a structure of a business card. It is a figure explaining the setting of the object area | region in the business card after the 2nd sheet. It is a figure explaining the setting of the object area | region in the business card after the 2nd sheet. It is a figure which shows an example of the selection screen of a template. It is a figure which shows an example of the operation | movement flow of the text processing apparatus and character recognition unit which concern on embodiment. It is a figure which shows an example of the operation | movement flow of the text processing apparatus and character recognition unit which concern on embodiment.

<1. Structure of sentence processing device:>
FIG. 1 is a diagram illustrating an example of an appearance of a text processing apparatus 1 according to an embodiment. FIG. 2 is a block diagram illustrating an example of a main configuration of the text processing apparatus 1.

  As shown in FIGS. 1 and 2, the text processing device 1 includes a document feeding unit 31, a document placing unit 32, a reading unit 41, an input / output unit 42, a recording unit 51, and a display unit 62. An operation unit 61 provided and a document detection unit 71 are mainly provided.

  The text processing apparatus 1 is configured to realize various functions such as a FAX function, a copy function, a scan function, a print function, a character recognition function, and an address book registration function by using each of these components under appropriate control. It has become. The address book registration function and the character recognition function realized by the character recognition unit 2 (FIG. 3) will be described later.

  The document feeding unit 31 feeds a document to the reading unit 41 by a so-called automatic document feeder (ADF) method. An image of a document fed by the document feeding unit 31 is acquired during feeding by the reading unit 41 at a predetermined position on the feeding path.

  As shown in FIG. 1, the document feeding unit 31 mainly includes a document table 33 on which a document to be read is placed, and a document discharge for storing a document that has been read by the reading unit 41 using the ADF scan method. And a table 34.

  Here, the document feeding unit 31 is configured to be able to feed various sizes of documents from the small business card size document to the large A3 size document to the reading unit 41, for example.

  As shown in FIG. 1, the document placement unit 32 is disposed below the document feeding unit 31, and allows the reading unit 41 to acquire an image of a document while the document is stationary. Bed) Used to scan images.

  For example, when the document feeding unit 31 is opened and closed with respect to the document placing unit 32, the document to be read is placed on the document placing unit 32.

  The reading detection light projected from the reading unit 41 is transmitted through the document placing unit 32 and reflected by the surface of the document, and the light transmitted through the document placing unit 32 is received by the reading unit 41 again. An image of the document is acquired by being read by the sensor.

  In the document placement unit 32, a business card holder that holds a plurality of business cards in a matrix along the main scanning direction and the sub-scanning direction related to image reading by the reading unit 41 is placed on the document placement unit 32. In this state, the plurality of business cards can be efficiently read at once by a flat bed scanning method.

  The reading unit 41 is a so-called scanner unit, and is provided below the document feeding unit 31 and the document placing unit 32 as shown in FIG. The reading unit 41 reads document information as image data using a scanner having a light projecting unit that projects detection light for image reading and a light receiving unit that photoelectrically converts the detection light reflected on the document surface. It is configured.

  Here, the reading unit 41 is configured to read a document as a flatbed scan type scanner unit and also to read a document as an ADF scan type scanner unit.

  The image data read by the reading unit 41 is compressed, for example, in the JPEG format and stored in the image storage unit 23 of the RAM 22.

  The input / output unit 42 is an interface unit for connecting the text processing device 1 to an external network. For example, the input / output unit 42 inputs image data supplied from an external computer to the text processing device 1 functioning as a printer. The image data acquired by the text processing apparatus 1 functioning as a scanner is output to an external computer or the like.

  The input / output unit 42 also includes an interface with a telephone line, and functions as an input / output interface when the text processing apparatus 1 functions as a FAX.

  The recording unit 51 is an image forming unit that records a toner image based on an electrostatic latent image on a recording sheet by an electrophotographic method.

  For example, the recording unit 51 forms a toner image on a photosensitive drum (not shown) based on the image data stored in the image storage unit 23, and supplies the toner image below the recording unit 51. The image read from the original is recorded on the recording paper by being transferred onto the recording paper supplied from the paper section.

  The operation unit 61 is a so-called operation panel. The operation unit 61 includes a display unit 62 that also functions as a touch panel, various operation buttons including a reading start button 63, and the like.

  The display unit 62 is configured by, for example, a liquid crystal display, and the operator of the text processing device 1 (hereinafter simply referred to as “operation”) according to the information displayed on the display unit 62 displays the operation status, operation contents, and the like of the text processing device 1. It functions as a display means to be presented to a person.

  Further, the display unit 62 has a function as a “touch panel” that can designate a position on the screen by touching the screen with a finger or a dedicated pen.

  Accordingly, the operator gives an instruction using the “touch panel” function of the display unit 62 based on the content displayed on the display unit 62, thereby causing the text processing apparatus 1 to perform predetermined processing (for example, an image on a recording sheet). Recording process).

  That is, the display unit 62 also functions as an input unit that receives various operations of the operator for the text processing device 1 via the touch panel and various buttons.

  As described above, the operator acquires the state of the text processing device 1 from the display unit 62 on the operation unit 61 and inputs settings and operations to the text processing device 1 using the display unit 62 and various operation buttons. can do.

  The document detection unit 71 includes, for example, a contact type or non-contact type sensor provided in the document feeding unit 31, and detects the presence of a document placed on the document table 33.

  An MPU (Micro Processing Unit) 11 is a control processing device that controls each functional unit of the text processing device 1, and executes control and processing according to a program stored in the ROM 21.

  The MPU 11 also functions as a recognition unit 12, a setting unit 13, a third acquisition unit 14, a second acquisition unit 15, an initial draft detection unit 16, and a generation unit 17, as will be described later.

  Further, the MPU 11, the ROM 21, the RAM 22, the storage device 26, the reading unit 41, the recording unit 51 and the like are electrically connected via the signal line 19. Therefore, the MPU 11 can execute, for example, a recording process by the recording unit 51 at a predetermined timing.

  A RAM (Random Access Memory) 22 is a readable / writable volatile memory, and temporarily stores processing information of the MPU 11 and the image storage unit 23 that stores an image read by the reading unit 41 and an image input from the input / output unit 42. It functions as a work memory that stores data.

  A ROM (Read Only Memory) 21 is a read only memory and stores a program for operating the MPU 11 and the like. A readable / writable nonvolatile memory (for example, a flash memory) may be used instead of the ROM 13.

  The storage device 26 is configured by a readable / writable nonvolatile memory such as a flash memory, for example, and permanently records various information such as setting information for the text processing device 1.

  The storage device 26 also includes a template storage unit 27 and an address storage unit 28, which store a template and an address book, which will be described later.

<2. Structure of character recognition unit:>
◎ About text processing device 1 related to the address book registration function:
FIG. 3 is a block diagram illustrating an example of main functional configurations of the text processing device 1 and the character recognition unit 2 according to the embodiment.

  In FIG. 3, among the functional components of the text processing apparatus 1, an image of a document such as a business card is read and character recognition is performed on the image, and necessary information of the character recognition result is stored in the address book. A functional section related to a function to be registered (hereinafter simply referred to as “address book registration function”) is shown.

  As shown in FIG. 3, the sentence processing apparatus 1 exhibits an address book registration function by mainly operating each function unit such as the character recognition unit 2, the generation unit 17, and the address storage unit 28.

  The generation unit 17 is a functional unit realized by the MPU 11, and generates registration data for registration in the address book based on character recognition results for various documents such as business cards generated by the character recognition unit 2.

  The address storage unit 28 is a functional unit realized by the storage device 26, and stores the registration data generated by the generation unit 17 as an address book.

◎ About character recognition unit 2:
The character recognition unit 2 shown in FIG. 3 is a functional unit that performs character recognition that acquires character data as a result of character recognition by performing character recognition on images of various documents.

  The character recognition unit 2 includes a first acquisition unit 43, an image storage unit 23, a display unit 62, a recognition unit 12, a setting unit 13, a third acquisition unit 14, a second acquisition unit 15, an initial draft detection unit 16, and a template storage unit. 27 as a main functional unit.

  The character recognition unit 2 has a common part common to each original and individual parts for each original such as a business card of the same company and the same type of form by appropriately operating these functional units. In character recognition for a plurality of documents, specific information for identifying a common part common to each document is acquired based on the first document that is first subjected to character recognition among the plurality of documents.

  The character recognition unit 2 sets a character recognition target area in an area other than the common part when performing character recognition of the second and subsequent originals.

  Accordingly, in the second and subsequent originals, it is not necessary to recognize the character again for the common portion, and the time required for character recognition can be shortened by efficiently reducing the amount of data to be subjected to character recognition.

  As described above, in the configuration example illustrated in FIG. 2, the functional units of the recognition unit 12, the setting unit 13, the third acquisition unit 14, the second acquisition unit 15, and the initial draft detection unit 16 are performed by the MPU 11. Although implemented, each of these functional units may be implemented by a dedicated hardware circuit, for example.

  The first acquisition unit 43 is a functional unit realized by, for example, the reading unit 41 or the input / output unit 42 and acquires images of a plurality of documents that are processing targets of the text processing device 1.

  The acquired image data is stored in the image storage unit 23 of the RAM 22 by the MPU 11 and displayed on the display unit 62 as an image or character data that has undergone character recognition by the recognition unit 12.

  When the reading unit 41 operates as the first acquisition unit 43, the first acquisition unit 43 performs an image reading process such as an ADF scan method on a business card to be processed to acquire an image of a document.

  When the input / output unit 42 operates as the first acquisition unit 43, the first acquisition unit 43 inputs an image of a document already acquired by an external computer or the like to the text processing apparatus 1. To obtain an image of the original.

  Thus, the reading unit 41 and the input / output unit 42 function as the first acquisition unit 43 of the character recognition unit 2.

  The initial draft detection unit 16 is a functional unit that detects a first original that is a target of character recognition among a plurality of originals having a common part and an individual part. The second acquisition unit 15 acquires specific information for identifying a common part common to a plurality of originals obtained based on the first original detected by the initial draft detection unit 16.

  The common part specification information acquired by the second acquisition unit 15 is a template (hereinafter simply referred to as a “template”) containing specific information for specifying a common part in a plurality of documents having a common part and individual parts. It is stored in the template storage unit 27 and acquired from the template storage unit 27 by the third acquisition unit 14.

  The third acquisition unit 14 is a functional unit that acquires, from the template storage unit 27, specific information for specifying a common part common to documents such as a plurality of business cards as character recognition targets in response to a predetermined operation input.

  Specifically, for example, the operator selects a template corresponding to a document that is a character recognition target from among the templates stored in the address storage unit 28, and performs an input operation for confirming the selection from a touch panel or the like. Accordingly, the third acquisition unit 14 acquires specific information stored in the selected template in response to the input operation.

  Also, for example, first, the third acquisition unit 14 automatically specifies a desired template by comparing the image of the business card to be processed with the common part area stored in each template, The template is displayed on the display unit 62, and then the operator performs input for confirming that the template is an appropriate template corresponding to the business card to be processed via the touch panel or the like, and responds to the input operation. Thus, a configuration in which the third acquisition unit 14 acquires specific information from the specified template may be employed.

  The setting unit 13 is a functional unit that sets a target area for character recognition with respect to an image of a document, and the recognizing unit 12 performs character recognition only on the target area for character recognition set by the setting unit 13. It is a functional part to perform.

  Next, the operation of each functional unit shown in FIG. 3 will be described in detail with reference to the flowcharts shown in FIGS. 12 and 13 and other drawings.

<3. Explanation of operations of the text processing device 1 and the character recognition unit 2:>
FIGS. 12 and 13 show a plurality of business cards (hereinafter simply referred to as “a plurality of business cards”) to be processed for address book creation when a plurality of documents having a common part and individual parts are a plurality of business cards. It is a figure which shows an example of the operation | movement flow of the text processing apparatus 1 and the character recognition unit 2 which are taken as.

◎ About the first card-based character recognition process and address book creation:
In the following, templates (hereinafter simply referred to as “templates”), which are identification information for specifying common parts in a plurality of originals having common parts and individual parts, are still templates for these plural business cards and other originals. Operations of the text processing device 1 and the character recognition unit 2 when not stored in the storage unit 27 will be described with reference to flowcharts shown in FIGS. 12 and 13.

  First, in order to register the information on the business card in the address book, the operator places a plurality of business cards on the document table 33 of the text processing device 1, and the placement is detected by the document detection unit 71. A detection signal is supplied to the MPU 11.

  Then, a screen that prompts the user to select an operation mode of the text processing apparatus 1 is displayed on the display unit 62 under the control of the MPU 11 supplied with the detection signal.

  The operator selects the “address book registration from a plurality of documents of the same type” mode (hereinafter simply referred to as “address book registration mode”) on the screen displayed on the display unit 62.

  By selecting the mode, the sentence processing apparatus 1 starts “address book creation processing from a plurality of similar business cards” shown in FIG.

  Next, when the operator presses the reading start button 63 in the address book registration mode, a signal of this operation is supplied to the MPU 11, and the first draft detection unit 16 uses the first business card as the business card to be processed. It is determined that there is, and a first sheet detection flag indicating the first business card is set in the RAM 22.

  The MPU 11 determines whether or not the business card to be processed is the first sheet according to the setting state of the first sheet detection flag (step S10).

  Here, since the first sheet detection flag is set, the MPU 11 moves the process to step S20 and determines whether one or more existing templates are stored in the template storage unit 27.

  Whether or not there is a template that identifies the common part by presenting to the operator that the first business card is the current character recognition target when character recognition is performed on a plurality of business cards of the same type. Prompting you to select, you can reduce the occurrence of time-consuming situations by creating unnecessary templates, even if there are existing templates, and speeding up character recognition for multiple similar business cards Can be

  Also, if there is no appropriate template, according to the method of creating a template based on the first business card, after performing character recognition for all items for the first several business cards, the subsequent business cards Compared with the case where a common part is set on the basis of, the processing time required for character recognition of the entire plurality of business cards can be shortened.

  Note that the first business card may be detected by using, for example, a counter that counts business cards to be processed instead of the first sheet detection flag.

  Here, since there is no existing template as described above, the MPU 11 shifts the processing to step S50 in order to create a template based on the current business target business card.

  In step S50, the first acquisition unit 43 (here, the reading unit 41) acquires the image data of the first business card. The acquired image data is stored by the MPU 11 in the image storage unit 23 (FIG. 3).

  Here, FIG. 4 is a diagram showing an example of the configuration of the first business card 6a. As shown in FIG. 4, a business card generally includes information about items having common information such as “company name”, “address”, etc., whose contents are common among business cards. There are a portion and an individual portion in which information about items having individual information for each business card such as “name” is described.

  In the example shown in FIG. 4, the common portions 8a, 8b, and 8c are common portions in which “company name”, “address”, and “telephone number” are described, and the individual portions 9a, 9b, 9c, and 9d. Are individual parts in which “affiliation”, “position”, “name”, and “e-mail address” are respectively described.

  As shown in FIG. 4, each individual part and each common part has an area in the entire area of the business card 6 a.

  The shared part is a part that does not need to be recognized again once character recognition is performed and the result of character recognition is stored, and the individual part is a part that needs to be recognized for each original. It is.

  Next, returning to FIG. 12, the setting unit 13 analyzes the entire image data of the acquired business card, and common parts and individual parts such as “company name”, “name”, “address”, “phone number”, etc. A character recognition target area in which a character group to be recognized for each item is set (step S60).

  That is, the setting unit 13 sets the common part and the individual part as target areas to be recognized for the first original. At this stage, character recognition processing is usually not performed, and the target area is set by a known image processing technique.

  When the target area for character recognition is set, the recognition unit 12 extracts an image of each character from the set target area and stores it in the ROM 21 or the storage device 26. Character recognition is performed to recognize what each character is by collating with the contents of the dictionary (step S70).

  Next, the MPU 11 estimates the item name of each target area based on the result of character recognition for each target area performed by the recognition unit 12.

  For the estimation, for example, when a character group such as “prefecture” is included in the character group of the target area, a method of estimating that the item of the target area is “address” or the like is used.

  Here, FIG. 5 is a diagram showing an example of the selection screen for the common part. When the item name of each target area can be estimated, the MPU 11 includes the estimated item name, the contents of each item, that is, the character data for the character group included in each target area, and the items for the common part. An interface ("Y" / "N" button in FIG. 5) for enabling the operator to select whether or not there is displayed on the display unit 62 line by line for each item as shown in FIG.

  In the image, an item name and its contents are also provided with an interface for shifting to a correction mode that can be corrected by an operator instructing on the display unit 62. Is not valid, the character recognition unit 2 is shifted to the correction mode to correct the character recognition result (step S80).

  In the correction mode, for example, correction of characters that have been mistakenly recognized, correction of recognition of item names such that the e-mail address is erroneously recognized as a telephone number, correction to supplement missing information, In the manuscript, when only one of the addresses written on two lines is recognized as an address and the second line is recognized as an item other than an address, the item name on the second line is changed. Etc. are performed.

  For correction, automatic correction by referring to a word dictionary stored in the ROM 21 or the like may be employed.

  In the example shown in FIG. 5, the position information of the common part is automatically corrected in conjunction with the correction of the result of character recognition. For example, the acquired business card image is displayed around the target area. A configuration in which the operator can correct the common part region on the image using the touch panel in a state where the region is displayed on the display unit 62 so as to be identified from the region may be adopted. .

  As described above, since the common part can be specified based on the image regardless of the character recognition result, the usefulness of the present invention is not impaired. However, as illustrated in FIG. 5, the character recognition of the first original is performed. According to the method of acquiring the identification information of the common part identified based on the result, a smaller display unit can be adopted as compared with the case where the common part is identified by displaying an image. Even in a small text processing apparatus equipped with the above, it is possible to perform character recognition processing on a plurality of originals having common parts and individual parts at high speed by appropriately specifying the common parts.

  When the operator completes the necessary correction for the character recognition result, the operator determines whether each item displayed on the display unit 62 corresponds to a common part or an item corresponding to an individual part, The result of character recognition corresponding to the common part is selected by instructing the “Y” button of the “Y” / “N” button on the touch panel (step S90).

  FIG. 6 is a diagram illustrating position information of the common part 8a in the business card to be processed. As shown in FIG. 6, the position information of the area on the business card is specified by, for example, the coordinates of the point P1 at the upper left end and the point P2 at the lower right end of each area in the coordinate system with the upper left corner of the business card as the origin P0. Is done.

  When the selection of the common part is completed, the operator terminates the selection of the common part by instructing a “confirm” button on the screen.

  Returning to FIG. 12, when the selection of the common part is completed by the instruction operation, the selection result for the common part is supplied to the MPU 11.

  The second acquisition unit 15 acquires specific information for identifying each common part such as an item name, item content, and each common part based on the selection result of the operator, creates a template, Store in the storage unit 27 (step S100).

  In addition, when correction is performed in the correction mode described above, specific information reflecting the correction content is acquired.

  FIG. 7 is a diagram showing an example of a template, and the template corresponds to the business card shown in FIG.

  In the template shown in FIG. 7, for example, the item name, the content, the coordinates of the upper left corner and the coordinates of the lower right corner of each common portion are associated as specific information of each common portion. In general, one template is set for a group of business cards of the same type.

  Each template includes, for example, a header section that records the name of the template, data size, the number of common parts included, the item name for each common part, and the number of characters of each content, and each common part included. Are stored in the template storage unit 27 based on a data structure having a data part that sequentially stores item names, contents, upper left coordinates of the area, lower right coordinates of the area, and the like.

  The specific information of each common part in the template can be referred to in any order by referring to the information of the header part.

  The template storage unit 27 also stores, for example, a template library management table that describes the name and storage destination address of each template, and the templates are referred to in any order by referring to the management table. Can be done.

  In the above description, the second acquisition unit 15 acquires the common part specific information determined by the operation of the operator, but the second acquisition unit 15 is an item of common information stored in advance in the ROM 21 or the like. A process of automatically specifying a common part based on information such as a name and setting and recording a template may be performed.

  That is, the second acquisition unit 15 acquires specific information for specifying the common part in response to an operation of the operator based on the first document or automatically based on the first document. I do.

  In the template shown in FIG. 7, the coordinates of the upper left end point and the lower right end point of the area range of the common part are adopted as the specific information about the position information of each common part. Even if position information such as coordinate information indicating a region range is employed, the usefulness of the present invention is not impaired.

  In the example of FIG. 6, each common area is shown as a rectangle that encloses a character string made up of all the characters that form one item. Alternatively, an area range surrounding the character may be designated for each character forming the character string of one item.

  Next, returning to FIG. 12, the generation unit 17 determines a predetermined item to be registered in the address book from among the character recognition results for the common part and the individual part that have been subjected to the necessary correction in step S80. Registration data in the address book is generated by arranging information in a predetermined order (step S110).

  When the registration data to the address book is generated by the generation unit 17, the MPU 11 stores the registration data generated in the address storage unit 28, thereby registering the acquired registration data in the address book (Step S1). S120).

  Upon completion of registration in the address book, the MPU 11 clears the first sheet detection flag indicating that it is the first business card.

  Next, the MPU 11 checks whether or not the address book creation process for all business cards has been completed (step S130). The confirmation of the completion is performed, for example, based on whether or not the business card is detected by the document detection unit 71 in the case of reading by the ADF scan method.

  In addition, when a plurality of business cards held in the dedicated holder described above are read by the flatbed scanning method, the confirmation of the completion is whether the processing for the business cards held at the predetermined end position in the holder is completed. It is done by confirming.

  Further, when creating an address book while repeating a series of operations of inserting a business card into the text processing device 1, character recognition and address book registration, and removing a business card from the text processing device 1 for each business card, for example, Each time the address book registration process for each business card is completed, a screen having an interface that allows selection of “continue” or “end” is displayed on the display unit 62 to prompt the operator to select. A method of determining whether or not the address book creation process for all business cards is completed based on the result of the selection according to the above can be adopted.

  If the address book creation processing for all business cards has been completed, the text processing apparatus 1 ends the address book creation processing, but here the address book creation processing for all business cards has not been completed. The process returns to step S10.

  In step S10, since the second business card is the processing target and the first sheet detection flag is cleared here, the MPU 11 moves the process to step S170 (FIG. 13). Similarly, business card image data is acquired.

  FIG. 8 is a diagram showing an example of the configuration of the second business card 6b. In the business card 6b, the positions and description contents of the common portions 8a to 8c are the same as those of the business card 6a.

  Further, the individual portions 9e to 9h of the business card 6b correspond to the respective items of the individual portions 9a to 9d of the business card 6a, but the description contents thereof are different from those of the business card 6a.

  Further, with the variation in the number of characters, the position information also differs from the position information of the individual portions 9a to 9d.

  In this case, since the specific information for specifying the common part based on the first business card 6a has already been acquired as a template, when the image data for the second business card 6b is acquired, the setting unit 13 Based on the position information of each common part set as a template, the character recognition target area for the business card 6b is set to an area other than each common part (step S180).

  FIG. 9 and FIG. 10 are diagrams for explaining setting of target areas in the second and subsequent business cards, taking the second business card 6b as an example.

  FIG. 9 shows a non-common part 7a which is an area other than the common parts 8a to 8c in the business card 6b, and FIG. 10 shows individual parts 9e to 9h which are target areas for character recognition in the business card 6b. Yes.

  The setting unit 13 is an area other than the common part for the second and subsequent originals as shown in FIGS. 9 and 10 by using a template in which common part specifying information for the business card to be processed is stored. A target area for character recognition is set in the non-common part 7a.

  Next, returning to FIG. 13, when the target area is set by the setting unit 13, the recognition unit 12 performs character recognition only for the target area in the same manner as in step S70 (step S190).

  When the character recognition is completed, the result of character recognition is displayed on the display unit 62. The operator refers to the displayed content and performs the necessary correction of the character recognition result as in step S80 (step S200).

  When the necessary correction for the character recognition result is completed, the generation unit 17 generates the character recognition result for the second and subsequent business cards and the first business card stored in the template. Then, registration data for the address book is generated by arranging the result of the character recognition of the common part and information on the predetermined item to be registered in the address book in a predetermined order (step S210).

  When the registration data for the address book is generated by the generation unit 17, the MPU 11 registers the registration data for the second business card in the address book in the same manner as the first business card (step S120).

  Next, the MPU 11 confirms whether or not the address book creation processing for all business cards has been completed in the processing of step S130 in the same manner as the processing for the first business card.

  If the address book creation processing for all business cards has not been completed, the MPU 11 returns the processing to step S10, and the functional units of the text processing device 1 and the character recognition unit 2 perform 2 for all unprocessed business cards. The same processing as that for the first business card is repeated.

  If the address book creation process for all business cards has been completed, the MPU 11 ends the address book creation process.

  By the processing procedure described above, an address book is created for a plurality of documents having common parts and individual parts from a state in which no template, which is identification information for identifying common parts, is stored in the template storage unit 27. An address book creation process is performed when the process is started.

  As described above, the character recognition unit 2 acquires specific information for specifying a common portion common to each document based on the first document to be character-recognized first among a plurality of documents. When character recognition is performed on the first and subsequent documents, character recognition target areas are set in areas other than the common part, so that it is not necessary to repeat character recognition for the common part, and the character recognition is efficiently performed. By reducing the amount of data, the time required for character recognition can be shortened.

◎ About template-based character recognition processing and address book creation:
Next, address book creation processing for a plurality of documents having common parts and individual parts from a state in which one or more templates created based on the first document are already stored in the template storage unit 27 A description will be given of address book creation processing in the case where is started.

  In FIG. 12, when the address book creation process is started, it is determined that the business card to be processed is the first one in the determination process in step S10, and then one or more in the determination process in step S20. It is determined that the template already exists, and the process proceeds to step S30.

  At this timing, the MPU 11 disables the processing function of the second acquisition unit 15 that acquires the specific information for specifying the common part based on the first business card.

  Also, at this timing, the third acquisition unit 14 of the MPU 11 creates a new template based on the first business card and applies it to the second and subsequent business cards in the address book creation process for a plurality of business cards to be processed. To prompt the operator to select whether to apply the first card-based character recognition process or to apply an existing template to all the first and lower business cards (template-based character recognition process) A screen having an interface is displayed on the display unit 62.

  The operator operates the touch panel of the screen displayed on the display unit 62 to determine whether the process to be performed is “character recognition process based on the first business card” or “character recognition process based on the template”. Is selected (step S30).

  If “first business card reference character recognition processing” is selected in step S30, character recognition is performed according to the processing flow described above in the description of “first business card reference character recognition processing”. If the address book is created and “template-based character recognition process” is selected, the process proceeds to step S140 (step S40).

  When the process moves to step S140 based on the operator's selection, the MPU 11 displays each template stored in the template storage unit 27 on the display unit 62, for example, to select a template to be used for the operator. Encourage selection.

  FIG. 11 is a diagram showing an example of a template selection screen displayed on the display unit 62 in step S140.

  In the example shown in FIG. 11, the name of each template is displayed on the screen based on the template library management table stored in the template storage unit 27 together with each template.

  As described above, the template library management table stores the name and storage address of each template.

  The template selected by the operator is referred to by the MPU 11 referring to the template library management table.

  In the example of FIG. 11, templates for business cards of each company from Company A to Company D are named “A Company Business Card”, “B Company Business Card”, “C Company Business Card”, and “D Company Business Card”, respectively. 27, and a template corresponding to a business card of “XX Co., Ltd.” which is a plurality of business cards to be processed this time is stored in the template storage unit 27 under the name “XX Co., Ltd. business card”. The case is shown as an example.

  Returning to FIG. 13, the operator selects the template by designating on the touch panel the name of the template “xxx business card” shown in FIG. 11.

  By this selection operation, the display of the name is highlighted from the black character on the white background to the white character on the black background, so that the operator confirms that the selection has been properly made and then presses the “confirm” button. The template is pressed to confirm the selection, and the template selection process is terminated (step S150).

  The third acquisition unit 14 detects, for example, a push signal from the “confirm” button pushed by the operator as a trigger, and thus the third acquisition unit 14 identifies specific information of the common part based on the template selected by the operator, that is, the common part. Item names, character recognition results, position information, and the like are acquired (step S160).

  When the specific information about the common part is acquired, each process for the second and subsequent business cards described in the explanation section of “First business card standard character recognition process” is performed in the “template standard character recognition”. This is applied to each of the first and subsequent business cards in “Process”, and the address book creation process for each business card is completed.

  From the state in which one or more templates storing the specific information for specifying the common part are already stored in the template storage unit 27 according to the processing procedure described above, a plurality of documents having the common part and the individual part are targeted. The address book creation process is performed when the address book creation process is started.

  In character recognition for a plurality of same-type documents having individual parts and a common part, a template created based on the first document is saved, and the plurality of same-type business cards are again targeted. If the saved template is applied to the first and subsequent originals when character recognition is performed, in the second character recognition process, specific information for specifying the common portion again based on the first original Therefore, it is possible to shorten the character recognition processing time for the plurality of originals again.

<4. Variation:>
Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and various modifications can be made.

  (1) In the present embodiment, it has been described that each piece of specific information about the common part is stored in one template. However, if the common part and each piece of specific information are appropriately associated, for example, each piece of specific information May be individually stored for each specific information.

  (2) In the present embodiment, it has been described that the template is permanently stored in the template storage unit 27 of the storage device 26. However, even if the template is temporarily stored in the RAM 22, a plurality of similar types Since the business card is often divided into a plurality of business card groups and character recognition processing is performed, the usefulness of the present invention is not impaired.

  (3) In the present embodiment, the first acquisition unit 43 also acquires an image for the entire area of the business card even for the business card when the common part is specified. The image acquisition area may be set based on the common part so that at least all the images of the non-common part 7a in the area can be acquired.

  According to this configuration, for example, when an image is acquired by placing a business card on a flat bed so that the vertical direction of the business card is determined along the sub-scanning direction of the reading unit 41, it is not common to only the upper half of the business card. If the portion 7a exists, only the upper half can be targeted for image acquisition, so the image acquisition time for the second and subsequent business cards can be reduced, and characters for a plurality of business cards can be reduced. The total time for recognition can be shortened.

DESCRIPTION OF SYMBOLS 1 Text processing apparatus 2 Character recognition unit 31 Document feeding part 32 Document placing part 33 Document stand 34 Document discharge stand 41 Reading part 42 Input / output part 43 First acquisition part 51 Recording part 61 Operation part 62 Display part 63 Reading start button 71 Document detection unit 6a, 6b Business card 7a Non-common part 8a, 8b, 8c Common part 9a, 9b, 9c, 9d, 9e, 9f, 9g, 9h Individual part

Claims (4)

A character recognition unit that performs character recognition of a plurality of originals having a common part and individual parts
First acquisition means for acquiring each image of the plurality of documents;
Second acquisition means for acquiring specific information for specifying the common part, obtained based on a first original to be character-recognized first among the plurality of originals;
For the first original, the common part and the individual part are set as target areas for character recognition, and for the second and subsequent originals, the target area is set in an area other than the common part. Means,
Recognition means for performing character recognition only on the target area;
A character recognition unit comprising: A character recognition unit according to claim 1,
A character recognition unit, further comprising storage means for storing the specific information. A character recognition unit according to claim 2,
Further comprising third acquisition means for acquiring the specific information from the storage means in response to a predetermined operation input;
When the third acquisition unit acquires the specific information,
(a) the character recognition unit disables the second acquisition means;
(b) The character recognition unit characterized in that the setting means sets the target area in each of the plurality of documents based on the acquired specific information. A character recognition unit according to any one of claims 1 to 3,
Generation for generating character data of each original based on the result of character recognition for the common part in the first original and the result of character recognition for the target area in each of the second and subsequent originals And
A sentence processing apparatus comprising:

Images