JP2003308478A - Form recognizing device, storage medium and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a form recognizing device, a storage medium and a program capable of displaying an image and the character recognition contents of the image for comparison, and improving the efficiency in correcting the wrong recognition in the character recognition. <P>SOLUTION: An OCR server 130 is connected to an image controller 120 connected through a scanner 110 with a sheet feeder and an SCSI, a DB server 140, data correcting terminals 150, 151, a layout table editing terminal 160 connected to a scanner for registration 165, and an order receiving DB server 170, through a communication line 100. The OCR server 130 creates the contrast display data of the uniform layout regardless of the kind of forms, after performing the processing for recognizing the kind of the form read by the scanner 110 with the sheet feeder, the processing for cutting the image on the basis of the form item information of the searched layout, and the OCR processing of every cut part. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a form recognition device, a storage medium, and a program, and in particular, a form for displaying an image image and the contents obtained by character recognition of the image image in a contrasted manner, and performing correction work of erroneous recognition in character recognition The present invention relates to a recognition device, a storage medium, and a program.

[0002]

2. Description of the Related Art Images read by OCR and OC
A technique is known in which the contents recognized by R are compared and displayed, and the correction work of the contents erroneously recognized by OCR is made efficient. For example, as disclosed in Japanese Patent Laid-Open No. 11-167603, as shown in FIG. 12, a first technique for improving the efficiency of scrutiny by displaying the content of character recognition of the image information immediately below the image information cut out of the standard form. Is disclosed.

[0003] In addition, in the publication, regarding an atypical form, O
A second technique has also been disclosed in which the image image of the form read by the CR is displayed in the upper half of the screen, and the contents recognized by the OCR are displayed in the lower half of the screen to make the check more efficient.

Similarly, as shown in FIG. 13, a third technique is also widely used in which an image image of a form is displayed on the left half of the screen and the recognized character data is displayed on the right half of the screen.

[0005]

However, according to the first technique described above, it takes a lot of time and effort to create a contrasting screen, so that there is a problem that a great deal of time and labor is required when the number of forms increases. is there. Even if the data items are common, the above problem becomes particularly serious in a business in which many kinds of forms having different layouts of detailed parts are often used based on business needs.

Further, according to the second technique, although the comparison screen can be easily created, the image images of various forms are displayed on the screen as they are, and the contents of character recognition are displayed in the order of recognition, so that the layout is rough. However, there is a problem in that the work of comparing the image image of the form with the content of character recognition cannot be performed efficiently, the operator's burden is heavy, and erroneous recognition is easily overlooked.

Further, as a problem common to the first to third techniques, since the display position of the image image and the display position of the character-recognized content are different for each form, the comparison point is also different for each form. As a result, there is a problem in that the operator's burden of comparison work is heavy and it is easy to overlook erroneous recognition.

The present invention has been made in view of the above problems, and displays the image image and the contents obtained by character recognition of the image image in a contrasted manner to improve the efficiency of correction work of erroneous recognition in character recognition. An object is to provide a form recognition device, a storage medium, and a program that can be improved.

[0009]

In order to achieve the above-mentioned object, the form recognition processing apparatus according to claim 1 reads an image image of a form in which the contents of form items are not described, and converts the read image image into the read image image. Layout information generating means for generating layout information composed of at least one rectangular area based on the image information, and an image image of another form in which the content of the form item is already described, and the generation based on the read image image. Image image cutout means for cutting out an image image for each rectangular area of the layout information, character recognition means for character recognition of the content of the cutout image image, the cutout image image and the content of character recognition And a first display information generating means for generating display information for displaying the information in contrast. , Display information which the generated is characterized in that it is arranged with the same layout in a plurality of forms having different rectangular areas.

A form recognition processing apparatus according to a second aspect is the form recognition processing apparatus according to the first aspect, wherein the form item setting means sets a form item described in each rectangular area of the layout information, and the form item. And a storage unit that stores the contents of each of the above items, and determines whether or not the contents described in a predetermined rectangular area of the form and the characters recognized are the form item contents described in the rectangular region. When it is determined that the content is not the content, the first display information generation means changes the display attribute of the rectangular area.

The form recognition processing apparatus according to claim 3 is the form recognition processing apparatus according to claim 1 or 2, wherein the layout information generation means is a storage means for storing layout information of a plurality of forms, and the form items. The rectangular area recognizing means for recognizing a rectangular area in the image image read for the form whose contents are not described, and layout information consisting of a rectangular area similar to the rectangular area in the recognized image image. And a second display information generating unit for generating display information for superposing and displaying the searched layout information and the read image image. It is characterized by being provided.

In order to achieve the above object, the program according to claim 4 reads an image image of a form in which the contents of the form item are not described, and includes at least one rectangular area based on the read image image. A layout information generation step for generating layout information,
An image image cutting step of reading an image image of another form in which the contents of the form item have been described, and cutting out an image image for each rectangular area of the generated layout information based on the read image image; A character recognition step of recognizing the content of the cut-out image image by a character; and a first display information generation step of generating display information for displaying the cut-out image image and the content of the character recognition in contrast. It is a program to be executed by a computer, and the generated display information is arranged in the same layout in a plurality of forms having different rectangular areas.

In order to achieve the above object, a storage medium according to a fifth aspect of the present invention reads an image image of a form in which the contents of form items are not described, and reads at least one rectangular area based on the read image image. And a layout information generating step of generating layout information, reading an image image of another form in which the content of the form item is already described, and for each rectangular area of the generated layout information based on the read image image. An image image cutting step of cutting out an image image, a character recognition step of character-recognizing the content of the cut-out image image, and display information for displaying the cut-out image image and the character-recognized content in contrast. The first display information generating step for generating A storage medium readable by computer, display information which the generated are
A plurality of forms having different rectangular regions are arranged in the same layout.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION A form recognition processing apparatus according to an embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a configuration diagram of a form recognition system including a form recognition processing apparatus according to the first embodiment of the present invention.

In FIG. 1, a form recognition system 1 includes an image controller 120 connected to a scanner 110 with a sheet feeder via SCSI, and an OCR server 13.
0, the DB server 140, the data correction terminal 150,
151, a layout table editing terminal 160 connected to the registration scanner 165, and an order DB server 170, which are each connected to the communication line 100.

The image controller 120 corrects the inclination of the image of the form read by the scanner with sheet feeder 110 and sends the corrected image to the OCR server.

OCR server 130 (form recognition processing device)
Stores a layout table 135 for form recognition processing, and recognizes the form type and the data described in the form based on the layout table 135.

The DB server 140 includes a product master 145,
Recognition character work file 146, image file 14
7 and the customer master 148, and based on these, the data recognized by the OCR server 130 is registered and collated.

The data correction terminals 150 and 151 are devices composed of a PC or the like operated by an operator, and can display the recognized data and the data actually described in the form in a contrasted manner (FIG. 5). Possible display device and FIG.
And an input terminal capable of correcting the data that is erroneously recognized by referring to the screen of 1. to the data actually described in the form. This corrected data is registered as order data in the sales details 175 of the order DB server 170. The order data is data relating to the contents of all the form items ordered.

The layout table editing terminal 160 is
The image image of the new form read by the registration scanner 165 and the layout information stored in the layout table 165 of the OCR server 130 may be superimposed and displayed (FIG. 10), or the layout information of the new form and the form item list may be displayed. (FIG. 11) A display device capable of performing the layout information of a new form is created by referring to the screen of FIG. 10, and what rectangular area of the layout information is created by referring to the screen of FIG. It has an input terminal that can set whether to describe form items (hereinafter referred to as “form item information”).

The layout information is information for storing the size and position of each rectangular area in which predetermined form items are described for each form type.

The layout table editing terminal 16 is provided in the layout table 135 for each form type.
The layout information transmitted by 0 and the form item information indicating what kind of form item data is written in which part of the rectangular area stored as the layout information are stored.

Further, the OCR server 130 and the DB server 1
40 and the order DB server 170 are the OCR server 13
It has a common hardware configuration except that 0 has an OCR function for performing character recognition from an image image. Therefore, regarding the common hardware of these servers, O
The CR server 130 will be described as an example, and the DB server 140 will be described.
The description of the order DB server 170 will be omitted.

FIG. 2 shows the OCR server 130 shown in FIG.
It is a figure which shows the hardware constitutions of.

In FIG. 2, the OCR server 130 is a C
PU21, RAM22, ROM23, LAN adapter 2
4, video adapter 25, keyboard 26, mouse 2
7, a hard disk 28, and a CD-ROM drive 29, which are connected to each other via a system bus 20. The system bus 20 is, for example, PCI.
It means a bus, an AGP bus, a memory bus, or the like. Further, a connection chip between each bus, a keyboard interface, and an input / output interface such as so-called SCSI and ATAPI are omitted.

The CPU 21 performs various operations such as four arithmetic operations and comparison operations and controls hardware based on a program of an operation system (hereinafter referred to as "OS") and an application program. The RAM 22 stores an OS program and application programs read from a storage medium such as a CD-ROM or a CD-R mounted on the hard disk 28 or a CD-ROM drive 29, and these are stored under the control of the CPU 21. To be executed.

The ROM 23 stores a so-called BIOS and the like that controls input and output to and from the hard disk 28 and the like in cooperation with the OS. The LAN adapter 24 cooperates with the communication program of the OS controlled by the CPU 21 to communicate with the outside via the network. The video adapter 25 generates an image signal to be output to the display device 30, and the keyboard 26 and mouse 27 are used to input control instructions of the server 100.

The hard disk 28 stores an OS, application programs for controlling the information input system, input information input by the keyboard 26 and mouse 27, and the like.

The CD-ROM drive 29 is a CD-RO.
It is used to install a storage medium such as M, CD-R or CD-R / W and install the application program on the hard disk 28. Needless to say, a CD-R drive, a CD-R / W drive, an MO drive or the like may be used instead of the CD-ROM drive 29.

FIG. 3 is a flowchart of the form recognition processing executed by the form recognition processing system 1 of FIG.

In FIG. 3, first, the image image of the form in which the order details are described is the scanner 1 with sheet feed.
Image controller 120 after being read by 10
(Step S301).

The image of the transmitted image is rotated by the image controller 120 so as to have a predetermined orientation (step S302).

Further, when the image controller 120 determines that the image image after the image rotation has the inclination (YES in step S303), the OCR server 1 performs the inclination correction (step S304).
Sent to 30. For this determination, for example, a tilt identification mark such as "" is printed at a specific position on the form,
This is performed based on the shape of the tilt identifying mark in the image image after the image rotation.

The processing from steps S301 to S304 may be performed by the OCR server 130.

Next, the type of the form read in step S301 is recognized based on the transmitted image image (step S305). Specifically, the layout information of the form is generated from the image image of the form read in step S301, and the same layout information as the generated layout information is stored in the layout table 135 of the OCR server 130. The type of the form read in step S301 is recognized by searching from.

After that, the image image of the portion corresponding to each rectangular area of the retrieved layout information is cut out (step S306), and character recognition is performed by OCR processing for each cut out portion (step S307). The content of character recognition for each area is transmitted to the DB server 140. At this time, what form items are written in each rectangular area is set in advance, and thus the content of characters recognized for each rectangular area is recognized as the content of each form item.

Next, the contents recognized in each rectangular area are sorted in the order of form items (step S
308), and is registered as recognition information for each form item in the recognition character work file 146 (step S30).
9). Note that this processing may be performed by the OCR server 130.

Each image image cut out in step S307 is also registered in the image file 147 of the DB server 140 as an image image for each form item (step S309 ').

Thereafter, of the recognition information registered in step S309, the information whose form item relates to the product information is collated with the information of the product master 145 (step S310), and similarly, the form item relates to the customer information. The information that is the one is collated with the information of the customer master 148 (step S311). As a result of the collation in steps S310 and S312, when the recognition information is inconsistent (step S
(NO in 312), after turning on the flag for changing the data display attribute of the recognition information, the recognition character work file 1
The recognition information of 46 is updated (step S313).

Next, when the DB server 140 receives a data contrast display request from any of the data correction terminals 150 and 151 (step S320), it generates the contrast display data (step S314).

The comparison display data is, as shown in FIG.
Image data that is displayed so that the image and the recognition information about the image can be compared. At this time, in step S313, for the rectangular area in which the data display attribute flag is ON, data in which the display attributes such as the color of the display screen are changed is created. As a result, it is possible to surely correct the case where the contents to be described in the form are described in the wrong rectangular area.

The comparison display data is displayed in the same layout regardless of the type of form. This allows
Data correction terminals 150, 1
As a result of displaying the display information of the same layout on 51, it is possible to more reliably improve the efficiency in correcting the content when the character is erroneously recognized.

The contrast display data generated by the DB server 140 is transmitted to one of the data correction terminals 150 and 151 that has transmitted this data contrast display request, and is displayed on the display device for comparison. If the operator confirms the displayed content in contrast and finds the recognition information that is erroneously recognized, he / she makes an input to correct the recognition information using the input terminal. (Step S321).

At this time, the comparison display data may be transmitted to both the data correction terminals 150 and 151. As a result, a plurality of operators can simultaneously perform confirmation work on a plurality of different parts, and as a result, the confirmation work time per sheet can be greatly reduced. Furthermore, by processing the partial data by a plurality of persons in charge, it is possible to prevent privacy infringement and confidentiality leakage even when the form is information relating to personal privacy, transaction information, or the like.

Next, the recognition information confirmed and corrected by the operator is transmitted to the order DB server 170 and registered in the sales details 175 (step S315), and this processing ends.

According to the processing of FIG. 3, the OCR server 130
The recognition processing of the type of the form read by the scanner 110 with the sheet feeder (step S305), the image image cutting process based on the form item information of the searched layout (step S306), and the OCR of each cut part After performing the process (step S307), contrast display data having the same layout is generated regardless of the form type (step S314). Therefore, even if the form type is different, the data correction terminal 15 operated by the operator 15
Since the same layout information is displayed at 0 and 151, the efficiency of correction work for erroneous recognition can be greatly improved.

FIG. 4 is a flowchart of the form registration process executed by the form recognition processing system 1 of FIG.

In the following processing, the registration scanner 16
5 performs the processing of step S401, and the OCR server 1
30 performs the processes of steps S402 to S404 and S411, and the layout table editing terminal 160 performs the processes of steps S405 to S410.

In FIG. 4, first, the registration scanner 16
5, the image of the form for which the layout information and the form item information are newly registered in the layout table 135 is read (step S401).

The image image of the form to be registered read in step S401 is transmitted to the OCR server 130 via the layout table editing terminal 160, and the ruled line portion in the image image of the form to be registered is recognized (step S402). ).

FIG. 6 is an explanatory view of the recognition processing of the ruled line portion in the image image of the form to be registered, and FIG. 6A is a diagram showing the image image of the form and the recognized ruled line portion.
(B) shows a pixel value N (X
It is a graph which shows the relationship of i).

The range of X coordinates that can be processed by the OCR server 130 is Xmin to Xmax from the left end, and the range of Y coordinates is Ymin to Ymax from the bottom end.

Value n (X, Y) of each pixel of the image image of the form
Is 0 for a color recognized as white and 1 for a color other than white, the total value N (Xi) of the pixel values whose X coordinate is Xi is given by the following formula (1). Become.

[0055] When the value of Xi is changed from Xmin to Xmax,
As shown in (b), a peak of N (Xi) appears at the position of the X coordinate of the ruled line in the form. The X coordinate values at which the N (Xi) peak appears are X1, X2 ,. ． ． , Xn, the ruled line portion for the X coordinate is recognized.

Similarly, the Y-coordinate values at which the peak of the total value N (Yi) of the pixel values whose Y-coordinate is Yi appear are Y1, Y in order from the bottom.
2 ,. ． ． , Yn, the ruled line portion for the Y coordinate is recognized.

With respect to the portion recognized as the ruled line portion in this manner, the OCR server 130 will be described later with reference to FIGS.
After performing the rectangular area recognition process of (step S403),
The layout table 135 is searched for layout information including a rectangular area similar to the recognized rectangular area (step S404), and the searched similar layout information is transmitted to the layout table editing terminal 160.

The layout table editing terminal 160 is
When this similar layout information is received, as shown in FIG. 10, a screen superimposed on the image image of the form in which this similar layout information is registered is displayed on the display device (step S405).

A form layout list may be provided on a screen different from the superimposed display screen so that the operator can select different layout information instead of the similar layout information and display it in an overlapping manner.

Next, it is determined whether the layout information of the form to be registered is the same as the layout information retrieved in step S404, based on the result of the operator's visual judgment on the superimposed screen (step S404). S40
6).

As a result of the determination in step S406, when the layout information is the same as the retrieved layout information, it is determined that the layout information of the form to be registered is the layout information displayed in the superimposed manner, and the process directly proceeds to step S409.

On the other hand, as a result of the determination in step S406, when the layout information is different from the retrieved layout information, it is determined that the layout information of the form to be registered needs to be partially corrected. The layout information displayed in the superimposed manner is corrected and input so as to be the same as the image of the new form (step S407).
This correction input is performed by dragging the rectangular area of the layout information displayed in a superimposed manner to a predetermined position using a mouse or the like, or by changing the vertical and horizontal sizes to a predetermined size (FIG. 10). The OCR server 1 is used as the layout information of the form registered by the corrected layout information.
The layout table 135 of 30 is additionally updated (step S408), and the process proceeds to step S409.

In step S409, all the form items of the sales details 175 of the order DB server 170 are read and displayed as a form item list. Specifically, FIG.
As shown in, the layout information of the form to be registered is displayed on the left side of the screen, and the form item list is displayed on the right side of the screen.

The operator sets the form item information of each rectangular area of the layout information displayed on the left side of FIG. 11 (step S410). Specifically, this processing is performed by the operator selecting a predetermined rectangular area and the form items described in the rectangular area by clicking with the mouse.

When the setting of the form item information is completed for all rectangular areas, the registration button is clicked to send the layout information and the form item information to the OCR server 130.

The OCR server 130 registers the layout information and the form item information of the transmitted form by updating the layout table 135 (step S411).

According to the processing of FIG. 4, the registration scanner 16
When layout information including a rectangular area similar to the rectangular area of the image image of the form to be read that is read in step 5 is searched from the layout table 135 (step S4).
04), the layout information of the form to be registered is created by superimposing it with the image of the form to be registered (step S405) (step S406-).
(S408), layout information of the form to be registered can be easily created.

Further, when the layout information of the new form and all the form items in the sales details 175 of the order DB server 170 are displayed on the layout table editing terminal 160 (step S409), the form item information of the form to be registered. Is set (step S410), it is possible to easily set the form item information of the form to be registered.

Next, the rectangle recognition processing in step S404 of FIG. 4 described above will be described.

7 and 8 are flowcharts of the rectangle recognition processing executed by the layout table editing terminal 160.

First, the coordinates P (N) obtained by sorting the X and Y coordinates of the part recognized as the ruled line part in step S403 of FIG. 4 are stored (step S700). This sort is performed so that the X coordinate has priority over the Y coordinate and is in ascending order.

Next, after setting N = 1 (step S70)
1), it is determined whether or not the coordinate P (N) exists (step S702). As a result of this determination, if the coordinate P (N) exists, the coordinate P (N) is set to the coordinate candidate S (X1, Y1) at the lower left corner of the rectangular area as shown in FIG. 9 (step S70).
3), it is determined whether or not there is an upward vertical line that passes through S (X1, Y1) (step S704). Specifically, the value n (X1, Y1) of the pixel whose Y coordinate is one pixel above the coordinate of S (X1, Y1)
It is determined whether or not the value of (1 + 1) is 1.

As a result of the determination in step S704, S (X1, Y1)
When there is no upward vertical line passing through, it is determined that there is no rectangular area including the set S (X1, Y1), N = N + 1 is set (step S705), and the process of step S702 is performed again.
On the other hand, when there is an upward vertical line passing through S (X1, Y1), P (N + 1) on that vertical line is set as the coordinate candidate S (X2, Y2) in the upper left corner, as shown in FIG. (Step S706), S (X
It is determined whether or not there is a rightward horizontal line passing through (2, Y2) (step S707).

As a result of the determination in step S707, S (X2, Y2)
When there is no horizontal line to the right passing through, set S (X1, Y1), S
It is determined that there is no rectangular area including (X2, Y2), N = N + 2 is set (step S706), and the process of step S702 is performed again. On the other hand, when there is a horizontal line to the right passing through S (X2, Y2), it is the same as the Y coordinate of S (X2, Y2) among the coordinates sorted in step S700 on that horizontal line, and S (X2, Y2) A coordinate that is larger than the X coordinate but is closest is searched and set as the coordinate candidate S (X3, Y3) in the upper right corner of the rectangular area as shown in FIG. 9 (step S709).

Next, it is judged whether or not there is a horizontal line to the right passing through S (X1, Y1) (step S710). Step S7
As a result of the determination in 10, when there is no horizontal line in the right direction passing through S (X1, Y1), it is recognized that the rectangular area does not include the set S (X1, Y1), S (X2, Y2), and step S708 is performed. Go to processing. on the other hand,
When there is a horizontal line to the right passing through S (X1, Y1), it is the same as the Y coordinate of P (N) among the coordinates sorted in step S700 on that horizontal line, and from the X coordinate of S (X1, Y1) The largest but closest coordinate is retrieved and set to the coordinate candidate S (X4, Y4) at the lower right corner of the rectangular area, as shown in FIG. 9 (step S711).
Then, it is determined whether or not there is an upward vertical line passing through this S (X4, Y4) (step S712).

As a result of the determination in step S712, S (X4, Y4)
When there is no upward vertical line passing through, S (X1, Y1), S (X2, Y2),
It is determined that there is no rectangular area including S (X4, Y4), and step S7
The process proceeds to 08. On the other hand, when there is an upward vertical line passing through S (X4, Y4), among the coordinates sorted in step S700 on the vertical line, S (X4, Y4) has the same X coordinate and S (X4, Y4). Y4)
A coordinate which is larger than the Y coordinate but is closest is searched for and set to the coordinate candidate S (X5, Y5) in the upper right corner of the rectangular area as shown in FIG.

It is determined whether or not the coordinates S (X3, Y3) and the coordinates S (X5, Y5) match (step S714). If they do not match, S (X1, Y1), S (X2, Y2) ), S (X3, Y3), and S (X4, Y4), it is determined that there is no rectangular area, and the process proceeds to step S706. On the other hand, when they match, S (X1, Y1), S (X2, Y2), S (X
The area cut out by (3, Y3), S (X4, Y4) is recognized as a rectangular area, and the process proceeds to step S715.

In step S715, X4-X1 (= X3-X
It is determined whether the value of 2) and the value of Y2-Y1 (= Y3-Y4) are smaller than a certain value. As a result of this determination, when these values are smaller than the fixed values, it is determined that the recognized rectangular area is not a rectangular area having a predetermined height and width, and step S7
From the coordinates sorted by 00, S (X1, Y1), S (X2, Y2), S (X
The four coordinates of (3, Y3) and S (X4, Y4) are deleted, and it is determined that the area is not a rectangular area (step S71).
6) Then, the process from step S700 is performed again. on the other hand,
When these values are larger than a certain value, the following nested deletion processing is performed.

Specifically, the nest deletion processing is as follows.

First, among the coordinates sorted in step S700, except for the four coordinates S (X1, Y1), S (X2, Y2), S (X3, Y3), and S (X4, Y4), the following formula See the relationship between (2) and (3) 4
It is determined whether or not one coordinate T (X6, Y6), T (X7, Y7), T (X8, Y8), T (X9, Y9) exists (step S717).

X1 (= X2) ≦ X6, X7, X8, X9 ≦ X4 (= X3) ・ ・ ・ (2) Y1 (= Y4) ≦ Y6, Y7, Y8, Y9 ≦ Y2 (= Y3) ・ ・ ・(3) As a result of this discrimination, four coordinates T (X6, Y6), T (X7, Y7), T (X
8, Y8) and T (X9, Y9) exist, it is determined that the recognized rectangular area is a rectangle outside the nest, and step S716 is performed.
Go to processing.

On the other hand, four coordinates T (X6, Y6), T (X7, Y7), T
When (X8, Y8) and T (X9, Y9) do not exist, a new rectangular area other than the recognized rectangular area is recognized, and the process proceeds to step S716.

According to the above rectangle recognition processing, among those recognized as rectangular areas, those which do not have a predetermined height and width or those which are rectangles outside the nest are judged not to be rectangular areas ( It is possible to reliably search for similar layout information in steps S718 and S720) and step S404.

Further, an object of the present invention is to supply a storage medium (or storage medium) recording a program code of software for realizing the functions of the above-described embodiments to a system or apparatus, and to supply a computer of the system or apparatus ( Alternatively, by the CPU or MPU) reading and executing the program code stored in the storage medium,
It goes without saying that it will be achieved.

In this case, the program code itself read from the storage medium realizes the function of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. Further, by executing the program code read by the computer, not only the functions of the above-described embodiment are realized, but also an operating system (OS) running on the computer is executed based on the instruction of the program code. It goes without saying that a case where some or all of the actual processing is performed and the functions of the above-described embodiments are realized by the processing is also included.

Furthermore, after the program code read from the storage medium is written in the storage medium provided in the function expansion card inserted into the computer or the function expansion unit connected to the computer, the program code is instructed. It goes without saying that a case where the CPU provided in the function expansion card or the function expansion unit performs a part or all of the actual processing based on this and the processing realizes the functions of the above-described embodiments is also included.

Further, an object of the present invention is to distribute the program code of software for realizing the functions of the above-mentioned embodiment via a network, thereby storing the hard disk or memory of the system or apparatus or the CD-RW. Or a computer (or CPU or M) of the system or apparatus stored in a storage medium such as a CD-R.
It is needless to say that it is also achieved by the PU) reading and executing the program code stored in the storage means or the storage medium.

[0088]

As described in detail above, according to the form recognition device of claim 1, the program of claim 4, and the storage medium of claim 5, the layout information is cut out for each rectangular area. The display information for comparing and displaying the image image and the character-recognized content of the cut-out image image are arranged in the same layout in a plurality of forms having different rectangular areas, so that the type of form is Even if they are different, as a result of the display information of the same layout being displayed on the computer on which the operator works, it is possible to greatly improve the efficiency in correcting the erroneous recognition in the character recognition. Further, when this display information is transmitted to a plurality of computers, a plurality of operators can simultaneously perform a confirmation work on a plurality of different parts, and as a result, the confirmation work time for one form can be significantly shortened. Further, by causing the plurality of persons in charge to process the partial data, it is possible to prevent privacy infringement and confidentiality leak even when the form is information regarding personal privacy, transaction information, or the like.

According to the form recognition apparatus of the second aspect, when the contents recognized in the predetermined rectangular area of the form and subjected to the character recognition are not the contents of the form item described in the rectangular area, it is generated. Since the display information is obtained by changing the display attribute of the rectangular area, it is possible to surely correct even when the contents to be described in the form are described in the wrong rectangular area.

According to the form recognition apparatus described in claim 3, when the layout information is generated for the form in which the contents of the form item are not described, a rectangle similar to the one recognized as the rectangular area in the image image of the form. The computer on which the operator works, because the layout information consisting of areas is searched from the stored layout information of a plurality of forms and the display information for superimposing the searched layout information and the image image of the form is generated. As a result of being able to display this display information on, the layout information of the form to be newly registered can be easily created.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a form recognition system including a form recognition processing device according to a first embodiment of the present invention.

2 is a diagram showing a hardware configuration of an OCR server 130 in FIG.

FIG. 3 is a flowchart of a form recognition process executed by the form recognition processing system 1 of FIG.

FIG. 4 is a flowchart of a form registration process executed by the form recognition processing system 1 of FIG.

FIG. 5 is an explanatory diagram of comparison display.

6A and 6B are explanatory diagrams of recognition processing of a ruled line portion in an image image of a form, FIG. 6A is a diagram showing an image image of the form and a recognized ruled line portion, and FIG. It is a graph which shows the relationship of pixel value N (Xi) about coordinate value Xi.

FIG. 7 is a flowchart of rectangle recognition processing executed by a layout table editing terminal 160.

FIG. 8 is a flowchart of rectangle recognition processing executed by the layout table editing terminal 160.

FIG. 9 is an explanatory diagram of rectangle recognition processing.

FIG. 10 is an explanatory diagram of superimposed display.

FIG. 11 is an explanatory diagram of a form item list display.

FIG. 12 is an explanatory diagram of conventional misrecognition correction.

FIG. 13 is an explanatory diagram of conventional erroneous recognition correction.

[Explanation of symbols]

1 Form recognition system 20 system bus 21 CPU 22 RAM 23 ROM 24 LAN adapter 25 video adapter 26 keyboard 27 mice 28 hard disk 29 CD-ROM drive 100 communication lines 110 Sheet Feeder Scanner 120 image controller 130 OCR server 135 layout table 140 DB server 150 Data confirmation and correction terminal 151 Data confirmation and correction terminal 160 Layout table editing terminal 165 Registration Scanner 170 Order DB server

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5B029 BB03 CC26 EE04                 5B064 AA01 EA10 FA03 FA05 FA07                       FA13                 5L096 AA07 BA17 EA27 FA03 FA05                       FA16 FA36 FA64 FA69

Claims (5)

[Claims] 1. A layout information generation unit for reading an image image of a form in which the contents of the form item are not described, and generating layout information including at least one rectangular area based on the read image image, and the form item. And an image image cutout unit that cuts out an image image of each of the rectangular areas of the layout information generated based on the read image image of the other document whose contents have been described, and the cutout. Form recognition including character recognition means for recognizing the content of the image image by character, and first display information generation means for generating display information for displaying the cut-out image image and the content for which the character is recognized in a contrasting manner. In the processing device, the generated display information is the same in a plurality of forms having different rectangular areas. Form recognizing processing apparatus characterized by being arranged in a layout. 2. A predetermined rectangular area of the form, further comprising: a form item setting unit for setting a form item described in each rectangular region of the layout information; and a storage unit for storing each content of the form item. And a means for determining whether or not the character-recognized content is the content of a form item described in the rectangular area, and the first display information generation means when it is determined that the content is not the content. The form recognition processing apparatus according to claim 1, wherein the display attribute of the rectangular area is changed. 3. The layout information generating means includes a storage means for storing layout information of a plurality of forms, and a rectangular area for recognizing a rectangular area in an image image read for a form in which the contents of the form items are not described. Recognizing means, searching means for searching layout information of rectangular areas similar to the rectangular area in the recognized image image from the stored layout information of the plurality of forms, the searched layout information, and the 3. The form recognition processing apparatus according to claim 1, further comprising a second display information generation unit that generates display information for superimposing the read image image. 4. A layout information generating step of reading an image image of a form in which the contents of the form item are not described, and generating layout information including at least one rectangular area based on the read image image, the form item The image image cutting step of reading the image image of the other form in which the contents of the above are described, and cutting out the image image for each rectangular area of the generated layout information based on the read image image; The computer executes a character recognition step of recognizing the content of the image image by a character and a first display information generation step of generating display information for displaying the cut-out image image and the content of the character recognition in contrast. The display information to be generated is a different rectangular area. A program characterized by being arranged with the same layout in a plurality of forms having. 5. A layout information generating step of reading an image image of a form in which the contents of the form item are not described, and generating layout information including at least one rectangular area based on the read image image, the form item The image image cutting step of reading the image image of the other form in which the contents of the above are described, and cutting out the image image for each rectangular area of the generated layout information based on the read image image; A character recognition step of recognizing the content of the image image by a character; a first display information generation step of generating display information for displaying the cut out image image and the content of the character recognized in contrast; A computer-readable storage medium that stores the program The storage medium is characterized in that the generated display information is arranged in the same layout in a plurality of forms having different rectangular areas.