JP2012190434A - Form defining device, form defining method, program and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a form defining device capable of saving labor in form definition as much as possible by properly performing automatic classification and grouping of processing areas.SOLUTION: A form defining device 1 for defining form information comprises: a processing area detection part 21 for automatically detecting a processing area in the form information from an inputted form image; a processing area grouping generation part 22 for performing classification so that contents of the same category are written in the same group, based on information about similarity of size of each of the detected processing areas, a connection situation of each of the processing areas in vertical and horizontal directions, and similarity of information about ruled lines surrounding each of the processing areas; a group selection I/F part 24 for selecting a processing area classified into the same group based on input information from an input part; and a group information defining part 25 for collectively defining setting information of the selected group.

Description

  The present invention relates to a form definition device, a form definition method, a program, and a recording medium for defining a form used in a fixed form processing system that automatically inputs form information to a fixed form.

There has been proposed a fixed form processing system that automatically inputs form information into a fixed form such as a purchase order. In such a fixed form processing system, there is a technology for setting a form information processing area on a form, grouping each processing area, and then defining and setting the form information for each group of processing areas. Already known. Furthermore, a method for automatically classifying a group of processing areas is already known.
For example, Patent Document 1 discloses that processing areas on a form are grouped in advance for the purpose of enabling a user to more easily define form information in a form processing system, and form information is collectively collected for each group. A configuration is disclosed in which character areas, table areas, figure areas, and the like are automatically classified and identified from form images when defining and grouping.

However, with the automatic classification method in the conventional fixed form processing system, automatic classification succeeds in a form in which the setting information of one category is an independent large frame, but a complicated form that is not so, specifically, Is a combination of various entries into a single large table, or a table with many checkmarks is not accurate enough to properly group the same type of frame There was a point.
If automatic grouping is appropriate, for example, although it is possible to save labor by batch setting in the subsequent processing, grouping cannot often be performed correctly. It will lead to things.
The present invention has been made in view of the problems as described above, and appropriately performs automatic classification and grouping of processing areas, and a form definition device, a form definition method capable of saving form definition as much as possible, It is an object to provide a program and a recording medium.

In order to achieve the above object, a first form of the present invention is a form definition device that defines form information, and a processing area detection unit that automatically detects a processing area of form information from the inputted form image And each processing area based on information on the similarity of the size of each detected processing area, the connection status of each processing area in the vertical and horizontal directions, and the similarity of the ruled line information surrounding each processing area. A processing area grouping creation unit for classifying areas into the same group in which the contents of the same category should be entered, and a group selection I / F for selecting the processing areas classified into the same group based on input information from the input unit And a group information definition unit that collectively defines setting information of the group selected in the group selection I / F unit.
According to the present invention, when grouping processing areas on a form, form information is collectively defined for each group by grouping different processing areas with the same processing area with high accuracy. Therefore, it is possible to save labor in setting up the form information.
In the second aspect of the present invention, the processing area grouping creation unit determines whether each processing area detected by the processing area detection unit is in the same group as another processing area that is closest to the top, bottom, left, or right. It is characterized by classifying all the processing areas into groups in which the contents of the same category should be entered.
According to the present invention, it is possible to shorten the processing time of the form information setting process.

According to a third aspect of the present invention, the processing area grouping creation unit is configured to, for each processing area detected by the processing area detection unit, size information of each processing area and surrounding ruled lines surrounding each processing area. Other processing areas between the respective processing areas and the other processing areas in a state similar to the information, and in a state where the respective processing areas are not directly connected to the other processing areas in the upper, lower, left, and right directions. When it is determined that there is no state, the other processing areas are classified into the same group.
According to the present invention, it is possible to classify even when processing regions are not connected.
According to a fourth aspect of the present invention, the processing area grouping creation unit predefines a combination of three pieces of information: size information of the processing area, ruled line information surrounding the processing area, and connection information of the processing area. A group creation rule storage unit for storing the converted information, and the grouping of the processing areas is determined according to the rules stored in the group creation rule storage unit.
According to the present invention, there is an advantage that it is easier to customize a user's form correction by storing the grouping determination as a rule.

According to a fifth aspect of the present invention, the processing region grouping creation unit includes a display unit that visually represents the processing regions classified into the same group.
According to the present invention, there is an advantage that even when an automatic grouping result is misidentified, the user can easily grasp the situation and easily correct it.
A sixth aspect of the present invention includes a region structure analysis unit that performs a structure analysis of a region detected by the processing region detection unit and generates a tree structure of the region, and the processing region grouping generation unit includes Based on the analysis result of the region structure analysis unit, region portions having the same region structure are determined to be the same group.
According to the present invention, it is possible to determine a region having the same content at a distant position that is not adjacent to one group by structurally searching for a matching portion of regions having the same structure and forming the same group. Easy to set up.
In the seventh aspect of the present invention, the processing area grouping creation unit matches the character information with the size information of the processing area in which the character information is entered, and the right of the processing area in which the character information is entered. Alternatively, when there are processing areas of the same size below, the processing areas of the same size are determined as the same group.
According to the present invention, by searching for the right or lower area of the area where the character information exists and making it the same group, it is possible to determine an area having the same content at a position not adjacent to each other as one group, Easy to set up.
According to an eighth aspect of the present invention, the group selection I / F unit is configured to be able to collectively select processing areas classified into the same group by an input operation of the input unit. .
According to the present invention, since the operation of the input unit can be reduced, a simple setting method can be provided to the user.

According to a ninth aspect of the present invention, there is provided a form defining method for defining form information, the step of automatically detecting a process area of form information from the input form image, and the detection of each of the detected process areas. Based on information on the similarity in size, the connection status of each processing area in the vertical and horizontal directions, and the similarity in the ruled line information surrounding each processing area, the contents of each processing area are filled in with the same category. A step of classifying the same group, a step of selecting processing regions classified into the same group based on input information from the input unit, and setting information of the group selected in the group selection I / F unit And defining all at once.
According to the present invention, when grouping processing areas on a form, form information is collectively defined for each group by grouping different processing areas with the same processing area with high accuracy. Therefore, it is possible to save labor in setting up the form information.
A tenth aspect of the present invention is a program for causing a computer to execute the form defining method described in the ninth aspect.
An eleventh aspect of the present invention is a computer-readable recording medium on which the program according to the tenth aspect is recorded.

  According to the present invention, when grouping processing areas on a form, form information is collectively defined for each group by grouping different processing areas with the same processing area with high accuracy. Therefore, it is possible to save labor in setting up the form information.

It is a block diagram of the form input processing system provided with the form definition apparatus which concerns on embodiment of this invention. It is the figure which showed the hardware constitutions of the form definition apparatus which concerns on this embodiment. It is a functional block diagram of the form definition part of the form definition apparatus of this embodiment. It is explanatory drawing of a grouping process. It is the figure which showed the example in which automatic grouping fails. It is the figure which showed the classification result of the information which considered size and connectivity. It is the flowchart which showed an example of the table frame extraction process in the process area | region detection part 21 of this embodiment. It is the flowchart which showed an example of the grouping process in the process area grouping preparation part. It is the figure which showed the process of the grouping process. It is the figure which showed the process and result of grouping processing. It is the figure which showed an example of the setting screen which inputs an attribute into definition information. It is the figure which showed an example of the money amount column which fills in the money amount. It is the figure which showed an example of the continuous rectangular frame. It is the figure which showed an example of the column of the postal code which fills in a postal code. It is the figure which showed the case where the entry column was comprised with the independent frame. It is the figure which showed the other example of the table frame. It is the flowchart which showed an example of the grouping process in case each entry area | region is an independent frame. It is another functional block diagram of the form definition part of the form definition apparatus of this embodiment. It is the figure which showed an example of the area | region divided | segmented frame. It is the figure which showed an example at the time of showing the frame shown in FIG. 19 by the tree structure. It is the figure which showed the process of provision of the cell ID in the frame shown in FIG. It is the figure which showed the other example of the form. FIG. 23 is a diagram showing a process of assigning a cell ID to the form frame shown in FIG. 22. It is the figure which showed an example at the time of showing the frame shown in FIG. 23 by the tree structure. It is a functional block diagram of the form definition part of 2nd Embodiment. It is the figure which showed an example of the process which produces | generates the tree structure of a frame. It is the flowchart which showed an example of the grouping process in the process area grouping preparation part 22 of 2nd Embodiment. It is the flowchart which showed the other example of the grouping process in the process area grouping preparation part 22 of 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described.
FIG. 1 is a configuration diagram of a form input processing system including a form definition device according to an embodiment of the present invention. A form input processing system 100 shown in FIG. 1 includes a form definition apparatus 1 and a form input processing apparatus 50 according to this embodiment.
The form definition device 1 is a processing device necessary as advance preparation for operating the form input processing device 50, and includes a master image input unit 10, a form definition unit 20, a form definition storage unit (form definition database (DB)) 30. Consists of.
The master image input unit 10 performs an input process of an image for defining form information (hereinafter referred to as “master image”). The form definition unit 20 performs a form definition process based on the master image information from the master image input unit 10. The form definition DB 30 stores form definition information defined in the form definition unit 20.

On the other hand, the form input processing device 50 includes a processing image input unit 51, a recognition processing unit 52, a recognition result unit 53, and a business related DB 54.
The processed image input unit 51 performs input processing of a form image to be processed. The recognition processing unit 52 performs recognition processing of the form image information from the processing image input unit 51 based on the form definition information stored in the form definition DB 30. The recognition result unit 53 displays the recognition result of the recognition process performed by the recognition processing unit 52. The business relation DB 54 stores the recognition result information of the recognition result unit 53.

FIG. 2 is a block diagram showing a hardware configuration of the form definition device of the present embodiment.
The form definition device 1 shown in this figure is constructed so that it can be executed by digital processing such as a microprocessor. In this figure, a scanner 101 optically reads a document image original such as a master image and converts it into image data that is an electric signal. The CPU 102 reads out the program from the program storage ROM / RAM 103 in which the processing program according to the present embodiment is stored or from the CD-ROM / FD recording medium 104 by the CD-ROM / FD drive 105, and in the work area RAM 106. Division processing and image processing are executed on the document image data. The processed document image data is stored in the memory 107 in accordance with an instruction from the CPU 102, and is displayed on the display 108 or output from the printing device 109 as necessary.

[First Embodiment]
The form definition unit according to the first embodiment of the present invention will be described below.
FIG. 3 is a functional block diagram of the form definition unit of the form definition apparatus of this embodiment.
In the form definition device 1, processing is started by inputting the master image 2 from the master image input unit 10.
The form definition unit 20 includes a processing region detection unit 21, a processing region grouping creation unit 22, a display unit 23, a group selection I / F unit 24, and a group information definition unit 25.
The processing area detection unit 21 extracts a processing area in which the user fills in the form from the input master image (form image), and obtains processing area information including position coordinates and surrounding ruled line information. Output. In addition, the automatic detection of the processing region in the processing region detection unit 21 uses a known method disclosed in Japanese Patent Application Laid-Open Nos. 2001-125010, 2001-266068, 2002-042143, and the like. be able to.
The processing area grouping creation unit 22 performs grouping of which areas of the processing areas extracted by the processing area detection unit 21 and which areas are in the same category. The group having the same category means a state in which the same setting information is recorded as a form definition by the user.

The display unit 23 displays the grouping result created by the processing area grouping creation unit 22.
The group selection I / F unit 24 is an I / F that enables the processing area displayed on the display unit 23 to be selected by an input unit such as a pointing device. The group selection I / F unit 24 is used by the user to correct a mistake and select a necessary group range when an error occurs in the grouping process in the automatic processing area grouping creation unit 22.
When the group form definition unit 25 performs the setting corresponding to the processing area after the group is selected, the group form definition unit 25 defines the processing area information common to the group as the form definition.
The form definition information defined in the group form definition unit 25 is stored in the form definition DB 30. Note that the form definition information stored in the form definition DB 30 can be output in either a memory format or a file format.

Next, the grouping process in the form definition unit 20 will be described.
FIG. 4 is an explanatory diagram of the grouping process. For example, when there is a postal code entry frame 41 as shown in FIG. 4 as a form, a numerical value indicating the postal code is entered by hand for each frame 41a. There are many cases. At that time, for example, “recognition: handwritten OCR, character type: number, ruled line frame: excluding and recognizing” is set for each frame. Less effort. And in the case where one group exists independently as shown in FIG. 4, grouping is automatically possible using area identification technology, and batch setting is also possible.

  A method of automatic area identification is disclosed in, for example, Japanese Patent Application Laid-Open No. 2002-42143. According to this method, black pixel rectangle extracting means for extracting a circumscribed rectangle of a connected component of black pixels from input image data, and the black pixel rectangle extracted by the black pixel rectangle extracting means, the area information of the black pixel rectangle is obtained. A primary determination means for discriminating the black pixel rectangle into a frame candidate or a non-frame candidate, and extracting a connected component rectangle of a white pixel with respect to the inside of the black image rectangle discriminated from the frame candidate; Sort into frame and non-frame. By this method, a block of frames can be automatically extracted. When the mass of the frame is automatically extracted by this method, the frame lump and the frame inside the lump are extracted. A technique for performing grouping on the frame is disclosed in Patent Document 1 as a prior art. In this invention, it is assumed that the position of the block of frames is designated by the user as the processing area setting means, but an application for processing the automatically extracted area is possible. Since the block of frames and the individual frames belonging to the block are already known, grouping can be easily performed, and it is also easy to set the grouped frames at once. Various means for extracting the frame region have been proposed in addition to the method described here, and the method is not limited to the method described above. For example, for the automatic detection of the processing area, several known methods can be applied as described above in Japanese Patent Application Laid-Open Nos. 2001-122010, 2001-266068, and 2002-042143.

However, in Patent Document 1, in the case of a form 42 as shown in FIG. 5, frames 42d having almost the same size are arranged and there is no space independently for each group. Could not be grouped.
Therefore, for example, the processing area detection unit 21 of the form definition unit 20 acquires the size of the processing area (frame), the ruled line information of the processing area, and the connectivity information of the processing area, and first classifies the processing area according to the size of the processing area. However, in the form 42 shown in FIG. 5, the size of the entry area is the same, and even if it is an image after printing and scanning, there is no significant difference. The classification in consideration of the size and connectivity as shown in FIG. 6, for example, the classification results classified into five frames 42a to 42e.

  Therefore, in the present embodiment, the processing region detection unit 21 automatically performs table frame extraction and subsequent individual frame extraction. The processing area grouping creation unit 22 is characterized in that grouping processing is performed using ruled line information of the processing area. That is, the grouping process is performed using the rule that relatively thick ruled lines are at the grouping breaks.

  The technology for automatically extracting the table frame can be realized by the above-described region extraction technology, but as another example, according to the invention of Japanese Patent No. 3215163, a circumscribed rectangle of a black connected component from a document image (or a reduced image thereof) is used. To extract. The circumscribed rectangle is extracted from the circumscribed rectangle using only the connected components of black pixels whose threshold value is abnormally long in a specific direction, and the number of vertical and horizontal circumscribed rectangles obtained there and the table area are determined. Therefore, the table area can be extracted by such a method. In the method shown in Japanese Patent No. 2851089, the contents of the table area thus obtained are extracted from the ruled line rectangle in the main scanning direction and the sub-scanning direction, and the frame surrounded by the ruled line is placed outside the ruled line rectangle. It is considered to be one frame using the coordinates of. The frame used for the template is basically the one that the user has not filled in. Basically, it is possible to obtain the blank area in one frame or the inner coordinate value constituting the frame. Each frame area can be obtained.

FIG. 7 is a flowchart showing an example of a table frame extraction process in the processing area detection unit 21 of the present embodiment.
In this case, the processing area detection unit 21 aims to detect all items / frames that the user who fills in the form fills in and can be read by the system.
Therefore, the processing area detection unit 21 divides the image input (entire surface) from the master image input unit 10 into areas (S1), and extracts a table-like area (S2).
Next, ruled lines constituting the table are extracted from the extracted table (S3), and a frame composed of the ruled lines is extracted (S4) and output.
The table frame extraction processing in the processing region detection unit 21 described above is disclosed in, for example, Japanese Patent No. 3140079 as an existing technique.
Note that the frame extraction process described above is merely an example, and the frame extraction process in the present invention is not limited to the frame extraction process described above, and other frame extraction processes can be applied.

Next, the grouping process in the processing area grouping creation unit 22 of this embodiment will be described.
To classify frames into groups, information on the height and width of the frame, information on the connectivity of the frame, and information on the ruled lines around the frame (thickness, type of ruled lines such as solid and dotted lines), etc. Is used.
This is because the majority of the series of frames that should be in the same group are basically the same size, and if the series of frames is not formed of independent frames, This is because the majority is connected somewhere. This is also because most of the ruled line information constituting a series of frames is similar.
However, there may be an exception in which only the ruled lines constituting the outermost frame are exceptionally solid lined lines and the inner ruled lines are dotted lines. In such a case, the ruled line information that forms the frame is different between the leftmost frame and the adjacent frame, but the ruled line information is used by treating the outermost circumference as an exception. it can.

FIG. 8 is a flowchart showing an example of grouping processing in the processing area grouping creation unit 22.
In this case, the processing area grouping creation unit 22 extracts the ruled line information of the frame from the input (image, frame area coordinates) (S11).
This ruled line information includes ruled line type (dotted line / solid line) and ruled line thickness information. Further, the ruled line information extraction includes a process of acquiring the ruled line thickness. The thickness of the ruled line can be roughly calculated by calculating the average run length in the thickness direction of the ruled line, or by dividing the number of black pixels in the ruled line area by the value of the length of the ruled line area (number of pixels). Number) can be estimated.
If the average run length is a vertical ruled line, the black pixel connected component and run length run length obtained when the ruled line is manipulated in the horizontal direction are obtained and averaged. For example, it is an image in which a rod is cut into circles and the average of the diameters is obtained.

Next, grouping based on the thickness of the ruled line of the frame is performed (S12). For example, it is roughly classified using the thickness information of the ruled line of the frame, and is divided into chunks 61a as shown in FIG.
Next, grouping is performed within the same group with the area size (vertical / horizontal) (S13). For example, when a small frame exists in a thick frame, it is determined as a different group by “unit of thick frame”. In the example shown in FIG. 9, the same vertical and horizontal size frames are arranged in a thick frame, so that the frames 62 a to 62 g are formed when grouped as shown in FIG. 9B (individual types of ruled lines: solid lines). , Dotted lines are not used).
Next, when there is only one frame in the group (in the thick frame), it is determined whether it is the same as the group having only one other frame (S14). Of the frames 62a to 62g shown in FIG. 9B, both the frame 62b and the frame 62g are one group and one frame. Thereafter, if different character strings existing in the frame are determined to be different groups, they can be divided into frames 62a to 62f and 63a to 63e as shown in FIG.
Next, when there is one frame in the group, this corresponds to the frames 63 a to 63 f in which bibliographic items such as “zip code” are written. Comparing the sizes of the group frames, the “zip code” and the “name” can be grouped in the same size because the size of the frame is within a certain range. Similarly, since “name reading” and “address” have the same size, they can be grouped together.
Thereafter, the grouping attribute of the frame is output. The grouping attribute may be a serial number such as group numbers 1 and 2 or a group A, B, or C, and may be an ID indicating which group in the table.

  In the present embodiment, the processing area grouping creation unit 22 extracts frame ruled line information, but this is merely an example, and in the table frame extraction process of the processing area detection unit 21 shown in FIG. In step S3, the ruled line information of the frame can be extracted, and the processing order can be changed as appropriate.

FIG. 10B is a diagram showing a result of grouping using the ruled line information of each processing region (frame) in the processing region grouping creation unit 22.
As described above, when the inside of one group is composed of several frames, the frames are grouped by size. In this example, a frame constituting one group, that is, a postal code entry box 60d, a name entry box 60f, a name reading box 60g, an address entry box 60h, and an address reading box 60i, The result is that all are in the same group.
On the other hand, when the group is one frame, this corresponds to a place where bibliographic items such as “zip code” are described. Comparing the sizes of the group frames, the “zip code” and the “name” can be grouped in the same size because the size of the frame is within a certain range. Similarly, “name reading” and “address” have the same size. Accordingly, the frames 60a and 60a printed with “zip code” and “name” are in the same group, and the frames 60b and 60b printed with “name full name” and “address” are in the same group. A frame 60c printed with “address reading” is a single frame.
This is a result of classification into groups of the same type because there is no relative difference in one frame because the ruled line type is the same thick ruled line in all four directions. If there is a frame detection device that can acquire the information of the thickness of the ruled line with high accuracy, such classification can be grouped in detail.

  Note that the preprinted areas (frames), such as the frames 60a to 60c, often do not serve as entry frames for the user to enter, and therefore, special processing may not be performed on the assumption that they do not interfere with batch setting. .

  As described above, in the present embodiment, the processing area grouping creation unit 22 can group form information for each group by grouping the same processing area and different processing areas with high accuracy. It is possible to save labor for setting up the form information.

For example, in a template creation tool that is used as a prototype, if you click one frame while holding down the Ctrl button, all the frames belonging to the same group are selected in the same way. The process is as follows. Of course, a shift button may be used in place of the Ctrl button, and any operation may be used as long as the entire group is selected when a part of the group is touched. This configuration has an advantage that it can be easily performed compared to operations such as surrounding a range to be selected by mouse dragging or clicking a necessary region one by one while pressing the shift button.
When form information is defined collectively for each group, for example, an attribute may be input to the definition information on a setting screen 70 as shown in FIG.

Also, it is often the case that the type of ruled line is changed mainly by the three-digit break in the amount column. When designing a form for a form, in addition to improving the appearance, there are many cases where the user does not confuse the entry frame with the aim of reducing erroneous entry by the user.
Therefore, in the present embodiment, form design know-how for reducing user entry errors is incorporated into the automatic classification method.
Here, the rules stored in the group creation rule storage unit will be described as the form design know-how.
FIG. 12 is a diagram showing an example of an amount field for entering an amount. In FIG. 12, two different types of money amount columns 71a and 71b are shown.
In the amount fields 71a and 71b shown in FIG. 12, when “three-digit break in the amount field” is ruled, the following is an example.
In this case, the target is a block in which the frames 72a or 72b of the amount column 71a or 71b are connected.
(1) The ruled lines constituting the outer peripheries of the connected frames 72a and 72b have the same thickness.
(2) When viewed from the outer peripheral frames 72a and 72b, the sizes (vertical and horizontal sizes) of the child frames 73a and 73b inside are the same. The same size means that the threshold value is within a certain value, the variance is within a certain value, and the like.
(3) Every three frames, one or more of the type or thickness / ruled line length of the ruled lines 74a and 74b changes.
When all of the conditions (1) to (3) are satisfied, it is determined not to group in units of three, but to determine that the outer frames are the same group in the amount column. If comprised in this way, the automatic classification precision of a cell frame can be raised.

  Further, in the present embodiment, the processing area grouping creation unit 22 performs grouping on the size of each processing area in the grouping process shown in FIG. 8, and there is connectivity among the groups determined to be the same size. It is also possible to adopt a method in which only things are further grouped and finally made into the same group.

Whether or not the areas are actually connected can be determined by a method of whether the two processing areas share a ruled line.
Furthermore, in the present embodiment, in the processing area grouping creation unit 22, there are four rectangles (in some cases, four or more in some cases) that are located on the top, bottom, left, and right of a processing area (frame). In contrast, a method of determining whether the sizes are the same may be used. In other words, in this case, the processing region grouping creation unit 22 has one processing region (frame) in each processing region detected by the processing region detection unit 21 having four or less other than itself closest to the top, bottom, left, and right. It is determined whether or not it is the same group as the other processing areas, and this is repeated for the detected processing areas, thereby classifying all the processing areas into groups in which the contents of the same category should be entered.

  In this way, connectivity can be ascertained by examining whether one processing frame is in contact with another processing region that is closest to the top, bottom, left, and right, so there is an entry frame 43 as shown in FIG. In this case, it is possible to eliminate the waste of determining the connectivity and size of the entry frame 43a and the entry frame 43c by checking the state of the entry frame 43a and the entry frame 43c.

  Further, in the determination of connectivity in this embodiment, it is sufficient if there is a rule of grouping basically connected items into the same group. However, as a typical example, the frame is independent. Even if they exist, add rules to group them into the same group.

FIG. 14 is a diagram showing an example of a postal code column for entering a postal code. FIG. 14 shows two different types of postal code columns 81a and 81b.
As a rule in this case,
(1) The ruled lines 82a and 82b constituting the outer periphery of the connected frames have the same thickness.
(2) Frames 84a and 84b having three child frames 83a and 83b are adjacent to frames 85a and 85b having four child frames 83a and 83b.
Although the order of the frames 84a and 84b and the frames 85a and 85b is related, the adjacent rules do not have to be in contact with each other. In addition, the presence or absence of a symbol such as “−” between the frames 84a and 84b and the frames 85a and 85b is ignored. As an ignoring method, it is conceivable to create exception handling in advance.
When these conditions (1) and (2) are satisfied, it is determined that the three frames 84a and 84b and the four frames 85a and 85b are not in different groups but are in the same group as a postal code column. Do.

  The above rule can be applied to, for example, a product number. When the number of digits of a model number such as “R11-A082034” is determined, whether or not grouping is performed depending on the adjacent state and thickness, and a child frame is provided inside. Information on how many items are necessary is made into rules, and by setting conditional statements with those rules, it is possible to acquire parts that appear to be product numbers as the same group.

FIG. 15 is a diagram showing the structure of the entry field in an independent frame. As shown in FIG. 15, each entry area 44a is an independent frame (not sharing ruled lines with other processing areas). There are many cases. Only the examples described so far cannot deal with such a form, so two unprocessed processing areas (between one processing area and another processing area) are examined.
(1) There is no other processing area between the two areas.
(2) The sizes of the two areas are the same (or a difference equal to or less than a threshold value),
(3) There is no difference in the correspondence of ruled line information (ruled line state such as dotted line, solid line, broken line) constituting the above two regions (frames),
If the three conditions are satisfied, the classification accuracy can be improved by considering the same group.
More specifically, when the corresponding ruled lines are two frames A and B, the upper and lower ruled lines of the frame A are solid lines and the left and right ruled lines are dotted lines. The ruled lines above and below the frame B are also solid lines, and the left and right are dotted lines.
However, as an exception, a table frame 90 of a different type such as a dotted line as shown in FIG. 16 is normally present, and therefore processing for treating the ruled lines constituting the outline as exceptions when a table lump is added. Is desirable.

FIG. 17 is a flowchart showing an example of grouping processing when each entry area is an independent frame.
In this case, the processing area grouping creation unit 22 performs grouping by size from the input image / extraction frame (area) (S21).
Next, horizontal grouping is performed (S22), and it is determined whether there is pre-printing on the left of the grouped area (S23). If it is determined that there is pre-printing on the left of the grouped area, it is determined as one group (S24).
Next, it is determined whether or not all checks of the grouped areas have been completed (S25). If it is determined that all checks have not been completed, the process returns to step S23 and continues. On the other hand, if it is determined in step S23 that there is no pre-printing to the left of the grouped area, the process proceeds to step S25 without determining that it is one group.
If it is determined in step S25 that all of the grouped areas have been checked, grouping is performed in the vertical direction for the area where the group has not been determined (S26). Then, it is determined whether or not there is pre-printing on the grouped area (S27). If it is determined that there is pre-printing on the grouped area, it is determined as one group (S28).
Next, it is determined whether or not all the checks of the grouped areas have been completed (S29). If it is determined that all the checks have not been completed, the process returns to step S23 to continue the processing. On the other hand, if it is determined in step S27 that there is no pre-printing to the left of the grouped area, the process proceeds to step S29 without determining that it is one group. If it is determined in step S29 that all checks for the grouped areas have been completed, the process ends.

By the way, there are few cases where the type of form changes dramatically, and there are many cases where small-scale correction of the format is performed on the operation.
Therefore, in the present embodiment, the processing region grouping creation unit 22 stores information in which a combination of three pieces of information, that is, processing region size information, surrounding ruled line information surrounding the processing region, and processing region connection information, is ruled in advance. It is also possible to provide a group creation rule storage unit and to determine the grouping of the processing areas according to the rules stored in the group creation rule storage unit. With this configuration, it becomes easier to customize the correction of the form of the user who has installed the system.

FIG. 18 is another functional block diagram of the form definition unit of the form definition apparatus of this embodiment. Note that the same blocks as those in FIG.
The form definition unit 20 shown in FIG. 18 includes a grouping rule storage DB 26 as a group creation rule storage unit. Then, the processing area grouping creating unit 22 determines the grouping of the processing areas in accordance with the rules stored in the grouping rule storage DB 26.
The grouping rule storage DB 26 can be edited by a grouping rule editing editor 27.

As a rule stored by the grouping rule storage DB 26,
(1) If the tolerance of the ruled line thickness of the child frame is different, NG, the outer periphery must be the same and each frame must be the same. (2) The child frame size tolerance is (threshold is set) threshold The above is NG
(3) Concerning the continuity of frames, allow space, allow characters, but NG if not adjacent
(4) Consecutive number of child frames “3 ・ 3 ・ 3” “3 ・ 4”
It is described in the form.

In the example of FIG.
(1) The outer circumferences must be the same. (2) Thresh1 (Thresh1 is a smaller value because we want to make it approximately the same size.)
(3) NG if not adjacent
(4) “3 ・ 3 ・ 3”
It becomes.
In the example of FIG.
(1) The outer periphery must be the same (2) Thresh2 (Thresh2 is a smaller value because we want to make it approximately the same size)
(3) Allow characters (4) "3.4"
It becomes.

  In the determination process for determining whether or not the above rule is satisfied, it may be determined whether or not all of the conditions (1) to (4) are satisfied. At this time, necessary parameters are obtained from the grouping rule storage DB 26 and used for determination.

  The grouping rules are basically stored on the system side in advance, but can be changed by the user. With such a configuration, the rule can be changed by the rule editing editor 27 when it is desired to add or modify the rule in the state stored in advance. That is, grouping of processing areas can be executed according to a new rule. Note that template creation and grouping rule editing are not particularly synchronized. Therefore, the rule is changed when it is determined that it is more efficient to edit the rule, and the new rule is applied when a new template is subsequently created.

A system that can update the setting file and its setting rule by a learning function or the like can be considered, and the setting can be further facilitated.
In the present embodiment, the processing area grouping creation unit 22 displays the frames of the processing areas classified into the same group (processing areas belonging to the same group) with the same color, and visually shows the difference between the different groups (visually). ), The processing region grouping result waits for the user to be selected in the subsequent group selection I / F unit 24, so that the state can be expressed visually. Therefore, even if there is a misidentification of the automatic grouping result, there is an advantage that the user can easily grasp the situation and can easily correct the situation.

Further, the group selection I / F unit 24 of the present embodiment is configured such that the processing areas belonging to the same group can be easily selected in a batch by operating an input unit such as a pointing device. For example, it is assumed that one of the processing areas classified into the same group can be selected at once by Shift-Key + left double-clicking on the mouse. is doing. By doing this, it is possible to provide a simple setting method to the user by reducing operations such as selecting the range of the processing area by dragging the mouse or clicking one by one and performing batch setting.
In the form definition device 1 of the present embodiment, the processing of the frame area has been described as an example. However, the processing area is not necessarily limited to the frame area. For example, Japanese Patent Application Laid-Open No. 2001-12610 As in the publication, it is also possible to automatically extract the processing area using pre-printed characters. Further, for example, methods such as Japanese Patent Application Laid-Open Nos. 2001-266068 and 2002-042143 by the present applicant can be used.

[Second Embodiment]
Next, a form definition unit according to the second embodiment of the present invention will be described.
FIG. 19 is a diagram showing an example of a frame that is divided into regions.
When there is a block of frames 201 having the configuration shown in this figure, a frame ID as shown in FIG. 19 can be assigned to each of the frames 201a to 201c. That is, the frame ID of the frame 201a is (1, 1), the frame ID of the frame 201b is (1, 2), and the frame ID of the frame 201c is (2, 1). At this time, the frame 201a is interpreted as having the frame 201b and the frame 201c as children.
FIG. 20 shows an example in which the frame shown in FIG. 19 is shown in a tree structure.
The method of attaching the frame ID shown in FIGS. 19 and 20 is described in paragraph [0016] of Japanese Patent No. 3157534.

In this case, it is assumed that there is actually no ruled line, a cell frame is attached, and the frame using the virtual ruled line is finally absorbed by the one with the smaller frame number. To take the idea of
Therefore, in the case of the frame 201 shown in FIG. 19, a cell ID is given on the matrix as shown in FIG. Then, (2, 1) is expressed as shown in FIG. 19 in a form integrated with (1, 1). This notation method is known.

FIG. 22 is a diagram showing another example of a form.
The form 210 shown in FIG. 22 has a common form frame at the top and bottom.
Such a form structure can be given a frame ID as shown in FIG. 23 by the above-described frame ID assigning method.
Furthermore, when the parent-child relationship of the frame IDs shown in FIG. 23 has a tree structure, it can be shown in FIG. In this case, in FIG. 24, the two tree structures are identical from the fact that the configuration of the two existing trees is the same and the corresponding frames: for example, the sizes of (1, 3) and (7.3) do not change. Then it can be determined. It is a feature of the second embodiment that it is determined from the determination result that the corresponding frames such as (1, 3) and (7, 3) are the same group.

FIG. 25 is a functional block diagram of a form definition unit according to the second embodiment. Note that the same blocks as those in FIG.
The form definition unit 20 shown in FIG. 25 includes a frame structure analysis unit 28 between the processing region detection unit 21 and the processing region grouping creation unit 22.
The frame structure analysis unit 28 analyzes the structure of the region detected by the processing region detection unit 21 and understands the tree structure of the region. For example, the structure of the frame is detected and a tree structure from a large lump to a small part is generated.
FIG. 26 is a diagram illustrating an example of processing for generating a tree structure of a frame.
Note that the processing of steps S31 to S34 is processing executed by the processing region detection unit 21, and the processing of steps S35 to S38 is processing executed by the frame structure analysis unit 28.
In this case, the processing area detection unit 21 divides the input image (entire surface) into areas (S31), and extracts an area of the “table” type (S32).
Next, ruled lines constituting the table are extracted from the extracted table (S33), and a frame constituted by the ruled lines is extracted.
Next, the frame structure analysis unit 28 assigns a frame ID to the extracted frame (S35), generates a tree structure from the dependency relationship of the frame ID (S36), and outputs it.

  Thus, according to the present embodiment, based on the analysis result of the frame structure analysis unit 28, the matching portion of the frames (regions) having the same structure is structurally searched and set as the same group. It becomes possible to determine a frame with the same content at a distant position as a single group, and the setting of the user becomes easy.

Next, in the form definition device according to the second embodiment, a grouping process performed by the processing area grouping creation unit using information on preprinted characters will be described.
In this embodiment, a case where character recognition (OCR) is used and a case where determination is made based on the degree of coincidence of images in a pre-printed part without using character recognition will be described as examples.
FIG. 27 is a flowchart illustrating an example of grouping processing in the processing area grouping creation unit 22 according to the second embodiment. This process is an example using character recognition (OCR).
In this case, the processing area grouping creation unit 22 first extracts a frame area and a character string from the input image (S41, S42). Next, the characters are recognized (S43), the recognized characters in the frame with the pre-printed characters are compared (S44), and if the recognized characters in the frame with the pre-printed characters do not match, the process returns to step S44. , Search for the next character.
On the other hand, if the recognized characters in the frame with the pre-printed character match, it is next determined whether or not the frame sizes of the regions with the pre-printed character are equal (S46).
If it is determined in step S46 that the frame sizes of the areas with the pre-printed characters are equal, it is next determined whether or not there is an entry frame of the same size to the right or below the pre-printed character position. Perform (S47).
If it is determined in step S47 that there are entry frames of the same size, the corresponding two frames are set to the same group (S48).
If it is determined in step S46 that the frame sizes of the areas with the pre-printed characters are not equal, or if it is determined in step S47 that the size is not the same, the process returns to step S44, and in step 6, If it is determined that the frame sizes of the areas with the pre-printed characters are equal, if the recognized characters in the frame with the pre-printed characters do not match, the process returns to step S44 to search for the next character.

FIG. 28 is a flowchart illustrating another example of the grouping process in the processing area grouping creation unit 22 according to the second embodiment. This process is an example in which determination is made based on the degree of coincidence of images in the preprinted portion without using character recognition.
In this case, the processing area grouping creation unit 22 first extracts a frame area and a character string from the input image (S51, S52). Next, character recognition is performed (S53), the sizes of the frames of the areas with the pre-printed characters are compared (S54), and if the frame sizes are not equal, the process returns to step S54 to search for the next character. I do.
On the other hand, if the size of the frame between the areas where the pre-printed characters are equal, image matching of the pre-printed characters is performed next (S56), and if it is determined in step S57 that the image difference is small, It is determined whether there is an entry box of the same size on the right or below the pre-print character position (S58).
If it is determined in step S58 that there are entry frames of the same size, the corresponding two frames are set to the same group (S59).
If it is determined in step S57 that the image differences do not match greatly, or if it is determined in step S58 that the size of the entry boxes is not equal, the process returns to step S54 to search for the next character.

  If the above-described grouping process is executed, for example, in FIG. 22, there are a plurality of corresponding “frigana” items, but the region sizes coincide with each other when (1, 3) (5, 3 ) (7, 3) (11, 3), which are the same group. In addition, (3, 3) and (9, 3) are the same group.

  As described above, in the present embodiment, the processing area grouping creation unit 22 searches the right or lower area of the area with the same pre-printed character to make the same group, and thus the same at a position that is not adjacent to each other. It becomes possible to determine the content area as one group, and the setting of the user becomes easy.

Note that when character recognition is used, it can be combined with more intelligent processing from the information of the recognized character string. However, character recognition is a technology that is difficult to achieve 100%, and in the present application, good accuracy can be obtained not by character string comparison but also by determination of whether the difference between preprinted images is large or small.
In the flowcharts shown in FIG. 26 and FIG. 28, when entering the double loop, the frame size is determined first or later, but this depends on the target and the required speed. It can be changed accordingly (tuning range).

Further, the present invention is not limited to the above-described embodiment, but the functions constituting the form definition device 1 of the above-described embodiment are programmed, and the program is executed by the CPU 102 shown in FIG. The object of the present invention can be achieved.
As recording media, semiconductor media (eg, ROM, nonvolatile memory card, etc.), optical media (eg, DVD, MO, MD, CD-R, etc.), magnetic media (eg, magnetic tape, flexible disk, etc.) Either may be sufficient.
Further, the case where the operating system or the like performs part or all of the actual processing based on the instruction of the loaded program and the functions of the above-described embodiments are realized by the processing is also included. Furthermore, the case where the above-described program is loaded into a memory provided in a function expansion board or a function expansion unit and the functions of the above-described embodiments are realized by executing the program is included.
Further, when the above-described program is stored in a storage device of a server computer and downloaded from a user's computer connected via a network and distributed, or when distributed and distributed from a server computer, A storage device is also included in the recording medium of the present invention.
As described above, by recording and distributing the program having the function of the present invention on a recording medium, cost, portability, and versatility can be improved.

  DESCRIPTION OF SYMBOLS 1 ... Form definition apparatus, 10 ... Master image input part, 20 ... Form definition part, 21 ... Processing area detection part, 22 ... Processing area grouping preparation part, 23 ... Display part, 24 ... Group selection I / F part, 25 ... Group information definition unit, 26 ... Grouping rule storage DB unit, 27 ... Grouping rule editing editor, 30 ... Form definition database (DB), 50 ... Form input processing device, 51 ... Processed image input unit, 52 ... Recognition processing unit, 53 ... Recognition result part, 54 ... Business-related DB, 54 ... Setting screen, 100 ... Form input processing system, 101 ... Scanner, 102 ... CPU, 103 ... Program storage ROM / RAM, 104 ... Recording medium, 105 ... CD-ROM / FD drive, 106 ... work area RAM, 107 ... memory, 108 ... display, 109 ... printing device

Japanese Patent Laid-Open No. 11-66232

Claims (11)

A form definition device for defining form information,
A processing area detection unit for automatically detecting a processing area of the form information from the input form image;
Based on the information on the similarity of the size of each detected processing area, the connection status of each processing area in the vertical and horizontal directions, and the similarity of the surrounding ruled line information surrounding each processing area, A processing area grouping creation unit that classifies the same category contents into the same group to be filled in;
A group selection I / F unit that selects processing regions classified into the same group based on input information from the input unit;
A group information definition unit that collectively defines setting information of the groups selected in the group selection I / F unit;
A form definition device characterized by comprising: The processing area grouping creation unit
By determining whether or not each processing region detected by the processing region detection unit is in the same group as another processing region that is closest to the top, bottom, left, and right, all processing regions are classified into the same category. The form definition device according to claim 1, wherein the form is classified into a group to be filled in. The processing area grouping creation unit
There is similarity between the size information of each processing area detected by the processing area detection unit and the ruled line information surrounding each processing area, and each processing area is directly connected to other processing areas vertically and horizontally. If it is determined that there is no further processing area between each processing area and the other processing area, the other processing areas are classified into the same group. The form definition device according to claim 1 or 2. The processing area grouping creation unit
A group creation rule storage unit for storing information in which a combination of three pieces of information of size information of the processing region, surrounding ruled line information surrounding the processing region, and connection information of the processing region is previously ruled; The form definition device according to any one of claims 1 to 3, wherein the grouping of the processing areas is determined according to a rule stored in a rule storage unit.   5. The form definition device according to claim 1, further comprising a display unit that visually expresses the processing regions classified into the same group in the processing region grouping creation unit. An area structure analysis unit that performs a structure analysis of the area detected by the processing area detection unit and generates a tree structure of the area,
The form definition device according to claim 1, wherein the processing area grouping generation unit determines that the area portions having the same area structure are the same group based on the analysis result of the area structure analysis section. The processing area grouping creation unit
When the character information and the size information of the processing area in which the character information is entered match and there is a processing area of the same size on the right or below the processing area in which the character information is entered, the processing of the same size The form defining apparatus according to claim 1, wherein the areas are determined to be the same group.   The group selection I / F unit is configured to be able to collectively select processing regions classified into the same group by an input operation of the input unit. Form definition device described in the section. A form definition method for defining form information,
Automatically detecting the processing area of the form information from the input form image;
Based on the information on the similarity of the size of each detected processing area, the connection status of each processing area in the vertical and horizontal directions, and the similarity of the surrounding ruled line information surrounding each processing area, Classifying the same category content into the same group to be filled in;
Selecting processing regions classified into the same group based on input information from the input unit;
Defining collectively the setting information of the groups selected in the group selection I / F unit;
A form definition method characterized by including:   A program for causing a computer to execute the form defining method according to claim 9.   The computer-readable recording medium which recorded the program of Claim 10.

Images