JP6917688B2 - Form reader, form reading method, program, and form reading system - Google Patents

Description

An embodiment of the present invention relates to a form reading device for reading characters from a form image, a form reading method, a program, and a form reading system.

The form reader recognizes an image (digital photograph, etc.) of a form acquired by an optical reader using character recognition application software (hereinafter referred to as "OCR software").

When the form to be read is, for example, a contract document composed of a plurality of sheets, a unique item is generally arranged for each page. Therefore, the form reading device needs to have form definition information for each page. The definition information of the form is, for example, the recognition item name, the coordinates (x coordinate, y coordinate) of the reading position corresponding to the item name, the size, the character type, the number of characters, and the like.

However, in order to read a plurality of forms having different formats from each other, it is necessary to create definition information of a plurality of forms in advance, which requires time and labor. In addition, there are various formats for forms, and it is difficult to create definition information in advance for all forms to be read.

JP-A-2015-090623

An object of the present invention is to provide a form reading device, a form reading method, a program, and a form reading system for facilitating advance preparation for reading a form.

The form reading device of the embodiment includes a storage unit that stores an image of a form including an item name, data corresponding to the item name, a search unit that searches for a predetermined item name from the image of the form, and the above. It includes an input unit that receives information for selecting data on the image of the form, an association unit that associates the selected data with the searched item name, and a character recognition unit that recognizes the associated data as characters. ..

It is a schematic diagram which shows the form reading system 1 which concerns on embodiment. It is a block diagram which shows the internal structure of a mobile terminal 10. It is a block diagram which shows the internal structure of the data server 30a, the processing server 30b, the entry client 40a, the marking client 40b, and the data confirmation client 40c. It is a table which represents an example of the entry item stored in the entry item DB 32a. This is an example of a marking screen. This is an example of a marking screen. This is an example of a marking screen. This is an example of a data confirmation screen. It is a flow chart which shows an example of the operation procedure of the form reading system 1. It is a flow chart which shows an example of the procedure of highlighting an entry item. It is a flow chart which shows an example of the procedure for confirming the correspondence between an entry item and a coordinate range. This is an example of a marking screen. This is an example of a marking screen. This is an example of a marking screen.

The form reading system reads a plurality of various fixed and atypical forms as an image (image), and creates and holds entry information in which image data and character data are integrated.
A standard form is a form in which the definition information of the form is registered in advance in the form reading system, and a non-standard form is a form in which the definition information of the form is not registered in advance in the form reading system.

Examples of the form include documents for insurance business, such as a notification form, a handler report, and a health examination result notification form. Insurance assessments can be made using the clean data created from these forms.
The clean data is entry information after confirmation and confirmation by the operator and correction of an error in the character recognition result.

Of the documents for insurance business, the notification form and the handler report are standard documents with a fixed format by the insurance company, and the definition information of the form can be registered in advance, so it is easy to read. be.
On the other hand, the health examination result notification is, for example, an atypical document having a different format for each medical institution, and it is often necessary to read a plurality of documents. Therefore, it is difficult to register the form definition information in advance. That is, it is difficult to obtain the form of the form in advance. In addition, it takes a lot of time and effort to register a large amount of definition information in advance. In this way, it is not easy to create appropriate entry information from a plurality of atypical forms.

In the following, as an example of the form, the configuration and operation of the form reading system will be described using the health diagnosis result notification.

As shown in FIG. 1, the form reading system 1 includes a mobile terminal 10, an image reading device 20, a data server 30a, a processing server 30b, an entry client 40a, a marking client 40b, a data confirmation client 40c, and an external server 50. Have.

The mobile terminal 10 is carried by an operator (here, for example, an insurance salesperson), and an image of an atypical form (for example, a health examination result notification) is read by an insured person (customer). Used.

FIG. 2 is a block diagram showing an internal configuration of the mobile terminal 10.
The mobile terminal 10 is, for example, a tablet terminal and includes a camera unit 11 as an image acquisition unit, a touch panel 12, a form reading processing unit 13, a memory 14, a communication interface 15 (hereinafter referred to as “communication I / F15”), and the like. ..

The camera unit 11 is, for example, a CCD camera or the like, and captures, that is, optically reads the form to be read, for example, the form image (form image) of the form A.

The touch panel 12 has two functions, an information display function and an input function, and displays a form image captured by the camera unit 11 and inputs information by touching it (touch operation).
The form reading processing unit 13 controls the form reading processing by the camera unit 11, determines the suitability of the form image read by the camera unit 11, and displays a warning on the touch panel 12 when the image is inappropriate. ..
After this warning, the form reading processing unit 13 confirms whether or not the improper state of the image is resolved. When the inappropriate state of the image is resolved, the image can be taken automatically without waiting for the operator's operation.

Inappropriate images are, for example, overexposure, poor size, tilt, and camera shake.
Overexposure means that at least a part of the area of the image exceeds the upper limit of the gradation of the camera unit 11. The gradation (upper limit) is uniform in a part of the area, and even if there are characters and images in the area, later processing such as character recognition cannot be performed.

Poor size means that the size of the form occupied within the imaging range of the camera unit 11 is not appropriate. If the form is too small or too large for the imaging range (only a part of the form can be imaged), subsequent processing such as character recognition becomes difficult.

The tilt means that the form is tilted with respect to the imaging range of the camera unit 11. If the form is tilted with respect to the imaging range, subsequent processing such as character recognition becomes difficult. If the inclination is small to some extent, the form reading processing unit 13 can correct the inclination of the image of the form by processing the image data (tilt correction processing). However, if the inclination of the form with respect to the imaging range becomes too large, it becomes difficult to correct the inclination.

The camera shake means that the camera unit 11 vibrates with respect to the form. The image of the form becomes unclear, and it becomes difficult to perform processing such as character recognition later.

The form reading processing unit 13 functions as a detection unit that detects at least one of overexposure, size defect, inclination, and camera shake of the image of the form to be read. That is, the form reading processing unit 13 detects overexposure, size defect, inclination, and camera shake as follows.
Overexposure can be detected by whether or not there is a continuous region of the upper limit of gradation on the image.
The size defect can be detected by calculating the ratio of the area of the form to the entire image and checking whether or not this ratio is out of the appropriate range. For example, the area of the form can be obtained by extracting the outline of the form from the image.
The inclination can be detected by calculating the inclination of the form with respect to the entire image and checking whether or not this value is out of the appropriate range. For example, the slope of the form can be obtained by extracting the outline of the form from the image. The form reading processing unit 13 can issue a warning to the operator when the inclination correction processing is difficult.
The camera shake can be detected by whether or not the image of the form captured by the camera unit 11 changes with time.

The memory 14 stores the read form image and also functions as a work area for reading the form image.
The communication I / F15 transmits the image of the form to the data server 30a through the network NET.

The image reading device 20 is, for example, a scanner device used for reading an image of both a fixed form and an atypical form. The image reading device 20 scans the surface of the form with, for example, a CCD or the like to acquire (generate) image information (image data).

The data server 30a, the processing server 30b, the entry client 40a, the marking client 40b, and the data confirmation client 40c function as a character recognition system, and use the image of the form read by the mobile terminal 10 or the image reading device 20. , Generates entry information that integrates character data and image data.
The external server 50 receives the entry information from the data server 30a and uses it for, for example, an insurance assessment.

FIG. 3 is a block diagram showing the internal configurations of the data server 30a, the processing server 30b, the entry client 40a, the marking client 40b, and the data confirmation client 40c.

The data server 30a is a server computer, and is a form image database 31a (hereinafter referred to as "form image DB 31a"), an entry item database 32a "(entry item DB 32a), a character database 33a (character DB 33a), and an entry information database 34a (. It has entry information DB34a) and communication I / F35a.

The form image 71 acquired by the camera unit 11 is stored in the form image DB 31a (details will be described later). The form image DB 31a functions as a "storage unit for storing an image of a form including a plurality of sets of an item name and data corresponding to the item name".

The entry item DB 32a stores data for the search unit 31b, which will be described later, to search for the name of the entry item to be entered (input).
FIG. 4 shows an example of an entry item stored in the entry item DB 32a. Here, height, weight, maximum blood pressure, minimum blood pressure, etc. are shown as entry items for the health examination.

The item index represents the superordinate concept of the entry item.
The target character string is used by the search unit 31b to search for an entry item. That is, it can be determined that the character string that matches any of the target character strings corresponds to the entry item.

The necessity of OCR indicates the necessity of OCR (character recognition) in each entry item. For some entry items, for example, chest X-rays, it is assumed that the image of the X-ray photograph is used as data, and character recognition is unnecessary. That is, the data of the entry item that does not require OCR is not a character but an image such as an X-ray photograph.
The necessity of this OCR can be changed as appropriate by setting.

The character DB 33a is a dictionary of character types, and IDs, types, and rules for identifying a plurality of character types are stored in association with each other. The ID is, for example, a numerical value such as "1" or "2", the type is "kana" or "English", and the rule is "aiueo ..." for "kana" or "English" for "English". "Abcdef ..." and the like are set.

Entry information is stored in the entry information DB 34a. The entry information integrates image data and character data, and includes the item name and the recognition result or image combination of the image on the form corresponding to the item name.
The communication I / F35a communicates with the processing server 30b.

The processing server 30b includes a search unit 31b, an association unit 32b, a character recognition unit 33b, an entry information creation unit 34b, a display control unit 35b, and a communication I / F 36b.

The search unit 31b searches for a target character string (predetermined item name) corresponding to the entry item stored in the entry item DB 32a from the image of the form. Simple character recognition is performed on the image of the form, it is searched whether or not the character-recognized character string matches the target character string, and the coordinates are specified.

The association unit 32b associates a character image (data represented by a character string) and a data image (data represented by the image itself such as an X-ray photograph) on a form with an entry item.

(1) Character image association The character image association will be described first.
When the reading target corresponding to the entry item is specified by the input to the form screen (data selection), the association unit 32b selects the item name (target character string) searched by the search unit 31b. Associate with. That is, the association unit 32b cuts an image of the area corresponding to the selected data, specifies the coordinates thereof, and stores the cut image, the coordinates, and the corresponding entry item in a memory or the like as associated information.

If necessary, the association unit 32b confirms that the selected data and the searched item name (target character string) correspond to each other before the association. This correspondence can be confirmed by comparing the item name (target character string) with the coordinates of the selected data. For example, when the left side of the coordinates of the selected data is the coordinates of the searched item name (target character string), it is determined that there is a correspondence between them, and the association is performed.

Here, a plurality of data may be selected (for example, as described later, selection of a plurality of data (character strings) existing on a line connecting two points on the image of the form). In this case, the association unit 32b associates the plurality of data (character strings) corresponding to the entry items.

The association unit 32b sends an image of the associated data (cut image) to the character recognition unit 33b for character recognition, and the character recognition result is also included with the cut image, its coordinates, and the corresponding entry item. Store it in a memory or the like as associated information.

(2) Association of data image In the case of a data image, an association is generally made by selecting a data image (data) and an entry item.
That is, the association unit 32b cuts out the image of the selected range based on the item name and the information for selecting the range of the image of the form, associates it with the selected item name, and stores it in the memory.

The character recognition unit 33b character-recognizes an image (read target) of the data associated with the association unit 32b.

The entry information creation unit 34b creates entry information in which a plurality of entry items, character data corresponding to the items, and image data are associated with each other.

The display control unit 35b displays the marking screen on the touch panel 41b of the marking client 40b.
5 to 7 show an example of a marking screen.
The marking screen has a column for displaying the form image 71 and an entry item designation column 72. The entry item specification field 72 is used to specify an entry item.

In the form image 71, some character strings (target character strings) are highlighted. That is, the display control unit 35b functions as a display control unit that highlights the searched item name and displays the image of the form.

The highlighted character string is the target character string (entry item) searched by the search unit 31b. In FIG. 5, the items of "height" and "weight" are highlighted, in FIG. 6, the items of "HDL cholesterol" to "occult blood" are highlighted, and in FIG. 7, the items of "electrocardiogram" and "chest" are highlighted.

This highlight display is for making the part visually conspicuous compared to other parts, and various methods such as addition of colors and frames, temporal change (blinking) of light and dark, etc. are used. be able to.
Multiple data can be efficiently selected using the marking screen. The details will be described later.

The display control unit 35b causes the display unit 41c of the data confirmation client 40c to display the data confirmation screen having the entry item field 81, the character image field 82, the recognition character field 83, and the data image field 84. That is, the display control unit 35b displays a second display having a plurality of predetermined item name fields, the selected data field, and a character recognition result field of the selected data. Functions as a control unit.

FIG. 8 shows an example of a data confirmation screen.
The data confirmation screen has an entry item field 81, a character image field 82, and a recognition character field 83.
The name of the entry item is displayed in the entry item field 81.
The character image field 82 is a field representing an image cut out from the image of the form, and the recognition character field 83 is a field representing characters as a result of character recognition of the image.
By comparing the character image field 82 and the recognition character field 83, it is possible to easily find an error in character recognition. When there is an error in character recognition, the error can be easily corrected by designating the recognition character field 83 and inputting characters from a keyboard or the like.
The data image column 84 is a column representing an image such as an X-ray photograph cut out from the image of the form.

The communication I / F 36b communicates with the mobile terminal 10, the image reading device 20, the data server 30a, the entry client 40a, the marking client 40b, the data confirmation client 40c, and the external server 50.

The entry client 40a is a client computer for inputting data of a standard form, and has a display unit 41a, an input unit 42a, and a communication I / F 43a.

The marking client 40b and the data confirmation client 40c are client computers for data input and data confirmation of atypical forms (for example, health examination result notification), respectively.

The marking client 40b has a touch panel 41b, an input unit 42b, and a communication I / F 43b, and mainly displays and inputs for marking and character recognition. The marking means the designation of the range on the form image such as character recognition, and is performed by the input from the touch panel 41b.

The touch panel 41b has two functions, an information display function and an input function, and displays a form image captured by the camera unit 11 and inputs information by touching (touch operation).

It is convenient to use an input pen (input pen) for this touch operation. By bringing the tip of the input pen into contact with the touch panel 41b, moving it on the line, and releasing it, two points on the line can be specified.
In addition, different inputs can be used depending on the direction of the moving line.
For example, a plurality of data (a plurality of character strings) can be selected by sliding the screen in a predetermined reference direction (for example, a vertical direction).
Further, the range of 1 data (1 character string) can be specified by sliding the screen in a direction perpendicular to a predetermined reference direction (for example, a horizontal direction).
Further, the range of the image can be specified by sliding the screen in a predetermined reference direction and an oblique direction. In this case, a rectangular area whose diagonal lines are the start point and the end point of the diagonal line is selected.

The above shows the case where the character string or the image is selected using the touch panel 41b, but the selection can be canceled using the touch panel 41b. For example, the tip of the pen is relatively hard and pointed, and the other end is relatively soft and rounded. In this way, if the contact areas of the tip and the other end of the pen with respect to the touch panel 41b are different, each can be used for different inputs. For example, the tip of the pen can be used for selection and the other end can be used for deselection.
Not limited to the shape of the tip of the pen, for example, a character string or an image can be appropriately selected or deselected by input for specifying a selection mode or a deselection mode.

As described above, the touch panel 41b is a touch input mechanism capable of continuously inputting coordinates on the display unit, and functions as an input unit for receiving information for selecting data on the image of the form and information for deselecting the data. ..

The data confirmation client 40c has a display unit 41c, an input unit 42c, and a communication I / F 43c.
The input unit 42c functions as a second input unit that receives information for correcting the result field.

Hereinafter, the operation of the form reading system will be described with reference to FIGS. 9 to 11.
FIG. 9 shows the overall operation (outline operation) of the form reading system 1.

A. Image acquisition of form (step S11)
When the operator operates the mobile terminal 10, the camera unit 11 takes a picture of the form and stores it in the memory 14.

The form is, for example, a health examination result notification, and basically, images of a plurality of forms are acquired. The operator puts the form to be processed in the screen of the touch panel 12 and takes a picture. As a result, one or more form images are taken into the memory 14.

At the time of this shooting, the form reading processing unit 13 determines the suitability of the image of the form read by the camera unit 11, and displays a warning on the touch panel 12 when the image is unsuitable.
After that, the form image stored in the memory 14 is sent from the communication I / F 15 to the data server 30a via the network NET, and is stored in the form image DB 31a.

Although the image of the form was acquired by the mobile terminal 10 here, the image of the form may be acquired by using the image reading device 20.

B. Marking / OCR (character recognition) (steps S12 to S17)
The marking and character recognition are basically executed by the processing server 30b by the input from the marking client 40b.

(1) Form selection (step S12)
Upon input from the marking client 40b, the display control unit 35b of the processing server 30b reads one or more form images 71 stored in the form image DB 31a and displays them on the touch panel 41b.
Generally, thumbnails of the form image 71 are displayed, and a desired thumbnail is selected from the thumbnails.

(2) Display of marking screen (highlight display of entry item, step S13)
A marking screen including the form image 71 corresponding to the designated thumbnail is displayed on the touch panel 41b (see FIGS. 5 to 7).
The marking screen has a form image 71 and an entry item designation field 72, and character strings (target character strings, “height” and “weight” in FIG. 5) corresponding to the entry items on the form image 71 are highlighted. Will be done.

The search unit 31b searches for the target character string corresponding to the entry item on the image of the form, specifies the coordinates thereof (step S21), and the searched target character string is highlighted (step S22). ..
At this time, the target character string is searched and the coordinates are specified for the entire area of the related series of forms.

(3) Designation of coordinate range of a plurality of character strings (step S14)
Coordinate ranges (plurality of OCR points) of a plurality of character strings are specified on the form image 71 (marking).
For example, a plurality of character strings from "this time result" to "30.7" in FIG. 5 are vertically traced with an input pen. As a result, a coordinate range starting from the "result of this time" and ending at "30.7" is specified.

(4) Confirmation of correspondence between entry items and coordinate range (step S15)
It is checked whether the specified coordinate range corresponds to the entry item. This confirmation result is also displayed on the touch panel 41b.
For example, as shown in FIG. 12, the character strings (here, "173.0", "93.2") corresponding to the highlighted target character strings (here, "height", "weight"). ) Is highlighted, and a line segment indicates that the two are related.
It should be noted that this association display is not limited to the line segment, and can be appropriately used as long as it is a display associated with the association between the two.

FIG. 11 is a flow chart showing the details of this association.
a) The association unit 32b determines whether or not there is an entry item (target character string) corresponding to the specified coordinate range (step S31).
It is determined whether or not there is a highlighted target character string at a position corresponding to the specified coordinate range (for example, to the left of the specified coordinate range).
Since some of the entry items (chest X-ray, electrocardiogram) do not require OCR as shown in FIG. 4, even if the coordinate range corresponding to these is specified, they are excluded from this determination target.

b) When there is an entry item (target character string) corresponding to the specified coordinate range, the association unit 32b cuts out a character string (image) around the marking and specifies the coordinates (step S32).

c) The association unit 32b associates this coordinate value with the entry item corresponding thereto and stores it in a memory or the like (step S33).
The above is repeated until the entire specified coordinate range is processed (until the remaining character strings are exhausted) (step S34), and the combination of the entry item and the corresponding OCR target is determined.

As shown in FIG. 12, even when the coordinate range of a plurality of character strings from "this result" to "30.7" is specified, the "current result" does not correspond to the entry item (target character string). , "30.7" are not highlighted, and only "173.0" and "93.2" corresponding to the entry item are highlighted.

(5) OCR processing of the corresponding character string (step S16)
The character recognition unit 33b recognizes a character image cut out using the character DB 33a. The association unit 32b stores the result of character recognition in a memory or the like in association with the entry item.

(6) Designation of entry items and input images (step S17)
Some entry items require only images and do not require character recognition. An image on the form image 71 is designated for such an object, and is stored in association with an entry item.
For example, the range of the image is specified by running the input pen diagonally on the form image 71.
By specifying the entry items on the entry item specification field 72 before and after this, the input image, the entry items can be specified, and the association can be performed.

As described above, the entry information associated with the entry item, the character data (character recognition result), and the image data is created and stored in the entry information DB 34a.

C. Data confirmation The created entry information can be confirmed and confirmed on the data confirmation client 40c.
(1) Display of data confirmation screen (step S18)
The data confirmation screen is displayed on the display unit 41c of the data confirmation client 40c.
(2) Confirmation of OCR result (step S19)
The operator can easily find an error in character recognition by comparing the character image field 82 and the recognition character field 83.
When there is an error in character recognition, the error can be easily corrected by designating the recognition character field 83 and inputting it from a keyboard or the like.

In the above, the operator specifies the character string of the corresponding data while the target character string corresponding to the entry item is highlighted. On the other hand, when the positional relationship between the target character string and the character string of the corresponding data is known, both the target character string and the character string of the corresponding data are highlighted as shown in FIG. , The operator may confirm the propriety of this correspondence.

In FIG. 13, the “previous” inspection result is selected (highlighted) instead of the “current” inspection result.
In this way, if the character string of the corresponding data is selected incorrectly, it can be corrected as follows.

(1) Deselecting an erroneous selection Deselecting a character string of data. For example, the other end (deselection side) of the pen traces the column of the "previous" inspection result of the touch panel 41b in the vertical direction. As a result, only the target character string as shown in FIG. 6 is highlighted.

Also, if the item name (target character string) is erroneously searched and highlighted, its selection is deselected. For example, by touching the part of the target character string with the other end of the pen, the selection is canceled and the highlight of the part disappears.

(2) Input of correct selection Specify the correct range of character strings. For example, the “previous” inspection result column in FIG. 6 is traced vertically with the tip of a pen.
As a result, as shown in FIG. 14, both the target character string and the character string of the corresponding data are highlighted, and the entry information can be created by OCR processing.

In the above, a plurality of data have been input at once, but the data may be input individually.
For example, by sliding the pen in the horizontal direction on the image, the range of 1 data (1 character string) can be specified. Then, by touching the entry item designation field 72, the entry item can be associated with the character string, OCR processing can be performed, and entry information can be created.

In the above embodiment, as compared with the case where the operator selects the entry item and the corresponding data, the entry item and the data are efficiently associated with each other, the data is recognized as characters, and the confirmed entry information (clean data) is obtained in a short time. Can be created.
That is, since the target character string corresponding to the entry item is automatically searched on the form image, it is not necessary to visually search for and select the entry item. Then, just by specifying the coordinate range, the character image within that range is cut out and automatically associated with the entry item.

In the above embodiment, a character recognition system having a mobile terminal 10, a data server 30a, a processing server 30b, an entry client 40a, a marking client 40b, and a data confirmation client 40c is used. On the other hand, these functions may be realized by one device (computer).

Further, a program for realizing these functions in a computer may be stored in a computer-readable electronic medium: electronic media, and may be appropriately read by the computer. Examples of the electronic medium include a recording medium such as a CD-ROM and a DVD-ROM, a flash memory, a removable media: and the like.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other embodiments, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are also included in the scope of the invention described in the claims and the equivalent scope thereof.

1 Form reading system 10 Mobile terminal 11 Camera unit 12 Touch panel 13 Form reading processing unit 14 Memory 15 Communication I / F
20 Image reader 30a Data server 31a Form image DB
32a entry item DB
33a character DB
34a Entry information DB
35a communication I / F
30b Processing server 31b Search unit 32b Association unit 33b Character recognition unit 34b Entry information creation unit 35b Display control unit 36b Communication I / F
40a Entry client 41a Display 42a Input 43a Communication I / F
40b Marking client 41b Touch panel 42b Input unit 43b Communication I / F
40c Data confirmation client 41c Display 42c Input 43c Communication I / F
50 External server 71 Form image 72 Entry item specification field 81 Entry item field 82 Character image field 83 Recognition character field 84 Data image field

Claims (18)

A storage unit that stores an image of a form including an item name, data corresponding to this item name, and
A search unit that searches for a predetermined item name from the image of the form, and
A display control unit that displays the image of the form and the name of the searched item,
An input unit that receives information for selecting the coordinate range of the data on the displayed image of the form, and
An association unit that associates the coordinate range of the selected data with the searched item name, and
A character recognition unit that recognizes the coordinate range of the associated data and
A form reading device comprising. The form reading device according to claim 1, wherein the input unit receives information for selecting a coordinate range of a plurality of data. The selection of the coordinate range of the plurality of data is an input along a predetermined reference direction of the image.
The form reading device according to claim 2. The information for selecting the coordinate range of multiple data is
The form reading device according to claim 2 or 3, which is information for designating at least two points on the image of the form. The association unit cuts out an image of the selected data and associates it with the searched item name.
The form reading device according to claim 1, wherein the character recognition unit recognizes the cut image as characters. The first aspect of claim 1 is that the associating unit confirms that the coordinate range of the selected data and the searched item name correspond to each other, and associates the coordinate range of the confirmed data with the item name. Form reader. The form reading device according to claim 6, wherein the association unit confirms the correspondence based on the coordinate range of the selected data and the coordinates on the image of each of the searched item names. The input unit receives information for selecting the item name and the range of the image of the form, and receives the information.
The form reading device according to claim 1, wherein the associating unit cuts out an image of the selected range and associates it with the selected item name. The selection of the range of the image is an input along a direction oblique to the predetermined reference direction.
The form reading device according to claim 8. Wherein the display control unit, the search item name document reading apparatus of the highlight to Ru <br/> claim 1, wherein an image is displayed in the form a. The form reading device according to claim 10, wherein the display control unit highlights the associated item name and data to display an image of the form. The input unit receives the information for deselecting the item name on the image of the form, and receives the information.
The form reading device according to claim 10 or 11, wherein the display control unit cancels the highlighting of the item name. A display unit that is controlled by the display control unit to display an image is further provided.
The form reading device according to claim 1, wherein the input unit is a touch input mechanism capable of continuously inputting coordinates on the display unit. A second display control unit that displays an image having the plurality of predetermined item name fields, the selected data field, and the character recognition result field of the selected data.
A second input unit for receiving information for correcting the result column is further provided.
The form reading device according to claim 1, wherein the second display control unit displays the corrected information in the result column. A reading unit that optically reads the image of the form,
A detection unit that detects at least one of overexposure, size defect, tilt, and camera shake of the image of the form to be read.
The form reading device according to claim 1, further comprising. A step to store an image of a form including an item name, data corresponding to this item name, and
A step of searching for a predetermined item name from the image of the form, and
A step of displaying the image of the form and the searched item name, and
A step of receiving information for selecting a coordinate range of data on the displayed image of the form, and
A step of associating the coordinate range of the selected data with the searched item name,
The step of character recognition of the coordinate range of the associated data and
A form reading method comprising. Computer,
A storage unit that stores an image of a form including an item name, data corresponding to this item name, and
A search unit that searches for a predetermined item name from the image of the form, and
A display control unit that displays the image of the form and the name of the searched item,
An input unit that receives information for selecting the coordinate range of the data on the displayed image of the form, and
An association unit that associates the coordinate range of the selected data with the searched item name, and
A character recognition unit that recognizes the coordinate range of the associated data and
A program that functions as. A form reading system including a terminal device and a character recognition device.
The terminal device is
A reading unit that optically reads the image of the form including the item name and the data corresponding to this item name.
A detection unit that detects at least one of overexposure, size defect, tilt, and camera shake of the read form image.
A transmission unit that transmits the image of the read form to the character recognition device is provided.
The character recognition device is
A receiving unit that receives the image of the read form and
A search unit that searches for a predetermined item name from the image of the form, and
A display control unit that displays the image of the form and the name of the searched item,
An input unit that receives information for selecting the coordinate range of the data on the displayed image of the form, and
An association unit that associates the coordinate range of the selected data with the searched item name, and
A form reading system including a character recognition unit that recognizes the coordinate range of the associated data.

Images