JPH0934989A - Format parameter generating method for ocr - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate format parameter in a short time by a user by instructing the area for character description of a slip for OCR input displayed on the screen of a display while using a pointing device. SOLUTION: Image data fetched from a slip S for OCR input are outputted onto a display 2 so that the image of the slip S for OCR input (such as the outline of the slip or the frame line of the area for character description) can be displayed on the display 2. When the user instructs the position, where the area for character description is displayed, on the screen of that display 2 while using a pointing device 3, based on data showing that position, the format parameter concerning that slip S for OCR input is generated. Thus, since the work for generating the format parameter is performed through the screen of the display 2, the format parameter can be generated by easy work in a short time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the technical field of an information processing system for inputting data by reading characters on an OCR input form using an OCR (optical character reading) device. Is O
The present invention relates to a method for allowing a user to easily and highly accurately create a format parameter for a CR input form.

[0002]

2. Description of the Related Art Today, an OCR module (miniaturized OCR device) is used as a data input device in a small-scale information processing system centered on a personal computer.
The convenience and usefulness of is attracting attention. That is, for example, in a small-scale information processing system for managing the management of slips, it is not possible to manually enter the data such as the product number, the quantity, and the unit price written on many slips by using a keyboard. Being efficient,
It may cause an incorrect input due to a key operation error. On the other hand, the product number, quantity, unit price, etc. are entered in the OCR input form in which the character entry area (usually surrounded by a frame etc.) is printed, and the OCR module is used to By inputting these data by reading characters, the input process can be performed very efficiently and accurately.

By the way, in order for the OCR module to read the characters on the OCR input form, a format parameter as information on the format of the OCR input form is created and the format parameter is supplied to the OCR module. Must. The OCR module performs preprocessing such as character cutting and weighting and final character recognition on the image data captured from the OCR input form according to the format parameters.

The contents of such format parameters include the position, size and shape of the area (hereinafter referred to as a read field) in which the OCR module should cut out characters on the OCR input form.
The name of the reading field, the data format of the characters written in the reading field (such as unit price or quantity), and the logical relationship between the reading fields (for example, the first reading in the same line). The product of the number in the field and the number in the second reading field is 3
The relation such as the number in the second reading field)
And information subject to restrictions derived from the specifications of the OCR module (for example, the type of characters to be written in each reading field (such as only numbers or including katakana)) and the like.

Conventionally, among these format parameters, the format parameters relating to the position, size, shape, etc. of the reading field are such that the user sets the size and distance of the character entry area on the OCR input form on the form by the user. It was created by actually measuring with a keyboard and the like.

[0006]

However, it is a tedious work to measure the dimensions and distances on the OCR input form one by one, especially when the number of character entry areas is large.
It takes a considerable amount of time to complete the actual measurement for all the character writing areas. Further, since the measured values include errors more or less, the format parameters relating to the position, size, shape, etc. of the read field created in this manner are used for the position of the character entry area on the OCR input form. Often, it does not correspond well with the size and shape. Therefore, in practice, such a format parameter is supplied to the OCR module to read out the character, and if there is a misrecognition as a result, the actual measurement on the OCR input form is performed again. Was being done. Therefore, it takes more and more time to create the format parameter.

The present invention has been made in view of the above points, and enables a user to create format parameters in a short time in a small-scale information processing system employing an OCR device as a data input device (particularly, (EN) It is intended to provide a format parameter creating method capable of creating a format parameter relating to the position, size, shape, etc. of a reading field in a short time with high accuracy by a simple operation.

[0008]

A method of creating an OCR reading format parameter according to the present invention comprises a first step of capturing image data from an OCR input form and a second step of outputting the image data to a display. A third step of instructing a character entry area of the OCR input form displayed on the screen of the display using a pointing device, and data indicating a portion instructed by the pointing device,
And a fourth step of creating a format parameter for the OCR input form.

By outputting the image data captured from the OCR input form to the display,
An image of the input form (such as the outline of the form and the frame of the character entry area) is displayed on the display. When the user uses the pointing device to indicate the location where the character entry area is displayed on the screen of the display, a format parameter for the OCR input form is created based on the data indicating the location.

As described above, since the work for creating the format parameter is performed through the screen of the display, the format parameter can be created in a short time by a simple work. Further, since the position, size, shape, etc. of the reading field can be calculated from the data indicating the position designated by the pointing device (that is, the coordinate position on the screen of the position), the format parameter relating to them can be created with high accuracy. Like

It is preferable that the image data is captured in the first step using an OCR device of the same model as the OCR device used to read the characters on the OCR input form. By doing so, OCR
The resolution when importing image data from the input form, and O
Since the resolution when reading characters on the CR input form will be the same, the created format parameter will be OCR.
It is possible to prevent erroneous recognition from occurring due to a difference in resolution when the character is read on the OCR input form by being supplied to the device.

Further, in the third step, the character writing area is designated based on a plurality of points on the character writing area which are designated by the pointing device.
As a candidate for creating format parameters,
It is preferably displayed on a display. By doing so, it becomes possible to perform the work while seeing the display of such creation target candidates on the screen, so that the work can be performed more efficiently and comfortably. Further, in the case of displaying such creation target candidates, when the same pattern is repeatedly displayed on the display and a plurality of sets of character writing areas are displayed, only one of the sets is displayed as described above. The display as a creation target candidate is performed based on the designation of a plurality of locations using a pointing device, and the display as a creation target candidate for the above one set is set as a copy target for each of the remaining groups. However, it is preferable to display the display as a candidate for creation by copying the display on the screen of the display at an appropriate scale. By doing so, it is possible to reduce the number of points on the screen that the user has to use the pointing device to display the candidates to be created on the display. Can be done more efficiently.

Further, with respect to the character entry area designated in the third step, information of a predetermined type for creating a format parameter (for example, information other than information such as the position, size and shape of the reading field) , The name of the reading field, the data format of the characters to be written in the reading field, the logical relationship between the reading fields, information subject to restrictions derived from the specifications of the OCR module, etc.). It is preferable.
By doing so, format parameters relating to such information can be created in a short time by a simple operation through the screen of the display. Further, such set information is registered as part information used for setting such information about another character entry area, and by using this part information, such information about another character entry area is registered. It is preferable to make the setting. By doing so, the work for creating the format parameters for such information can be performed more efficiently.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing an example of the hardware configuration of a small-scale information processing system for implementing the format parameter creating method according to the present invention. This system has a personal computer 1 as a core, and has a display (CRT display or liquid crystal display) 2, mouse 3,
A keyboard 4, an OCR module 5, a disk device (magnetic disk device or optical disk device) 6 and a printer 7 are connected to the computer 1.

As is well known, the mouse 3 is a kind of pointing device and is provided on the upper surface of the main body after the main body is slid on the desk to move the cursor to a desired position on the screen 2a of the display 2. By clicking the button, the data indicating the coordinate position on the screen 2a at that position can be input to the computer 1. The OCR module 5 is a miniaturized OCR device as a device for inputting data to the personal computer 1, and reads the characters written on the OCR input form S printed using the printer 7. In addition, this OCR
The module 5 is a device of a model that performs reading by designating any one of a plurality of types of colors (for example, two types of colors, red and blue) as a dropout color.

FIG. 2 is a flow chart showing an example of the format parameter creating process executed by the computer 1. Hereinafter, this process will be described step by step. [Process in Step ST1] In step ST1,
The user uses the OCR module 5 to capture image data from the OCR input form printed using the printer 7.

If the color of the frame or the like of the character entry area on the OCR input form is either red or blue designated as a dropout color by the OCR module 5, that frame is selected. It is assumed that the color other than the color of the line or the like is designated as the dropout color by the OCR module 5, and then the image data should be captured.

Further, in the next step ST2, the OCR
In order to clearly display the outline of the input form on the display 2, as shown in FIG. 3, the OCR input form S1 is very different in brightness and darkness (for example, the background color of the form S1 is white. If it is, it is assumed that the image data should be captured with the (black) object M as the background.

[Processing in Step ST2] In step ST2, the user uses the keyboard 4 or the like to perform OCR.
In response to the command to display the image of the input form, the image data of the OCR input form captured in step ST1 is output to the display 2. As a result, the OC
An image of the R input form (such as the outline of the form and the frame of the character entry area) is displayed on the display 2.

[Processing in Step ST3] In step ST3, first, the OC displayed on the display 2 is displayed.
Based on the user designating a plurality of specific locations within the outline of the R input form using the mouse 3, the size of the OCR input form (horizontal direction, The length in the vertical direction) is calculated. Then, the calculated size is displayed on one screen of the screen 2a.

FIG. 4 is a diagram exemplifying some places on the screen 2a to be designated by the user in the size calculation process. In FIGS. 3A to 3C, the OCR module 5
The image of the OCR input form S1 in which the image data is captured without being inclined with respect to the scanning direction by the display 2
Is displayed in. In this state, if two specific places within the range of the outline of the form S1 are pointed by using the mouse 3, from the data indicating those places (specifically, the mouse 3 is adjacent on the screen 2a). With the minimum distance between the points that can be designated together as one point, the number of points in the horizontal direction (horizontal direction of the screen 2a) and vertical direction (vertical direction of the screen 2a) of the outline of the form which is found from those data is It is possible to calculate the horizontal and vertical lengths of the form S1 by multiplying the minimum distance (distance per point). As the two places, in FIG.
1 indicates that the upper left corner (the portion marked with an arrow a) and the lower right corner (the portion marked with an arrow b) of the outline of 1 should be designated. In the same figure (b), they are provided at the four corners of the form S1. It is indicated that the left edge edge mark E (the portion marked with an arrow a) and the edge mark E at the lower right corner (the portion marked with an arrow b) should be designated among the obtained edge marks E. Further, in the same figure (c), between the corner of the outline of the form S1 and the edge mark E, a portion for size calculation is displayed on the screen 2a by the symbol "+".
Indicates that the upper left symbol "+" (the portion marked with an arrow a) and the lower right symbol "+" (the portion marked with an arrow b) should be designated. However, contrary to what is shown in FIG. 9A, the lower left corner and the upper right corner of the outline of the form S1 are designated, and contrary to the one shown in FIG. Or the edge mark E is designated, or contrary to the one shown in FIG.
Of course, you may instruct or.

On the other hand, in FIGS. 4D to 4F, O
The image of the form S1 in which the image data is captured while being inclined with respect to the scanning direction by the CR module 5 is displayed on the display 2. In this state, FIG.
As shown in (c), the form S is created from only the data showing two locations.
Although it is not possible to calculate the size of 1, the size of the form S1 can be calculated from the data indicating the three parts by pointing to three parts within the outline of the form S1. As these three places, in addition to the two places shown in FIG. 7A, in addition to the two places shown in FIG. 7A, the upper right corner (the place marked with arrow c) or the lower left corner (arrow d) of the outline of the form S1
(E.g., a portion marked with a) indicates that the edge mark E (arrow c) in the upper right corner is added in addition to the two locations shown in FIG. Or the edge mark E at the lower left corner (the place with the arrow d) should be designated. In FIG. 6F, the two places shown in FIG. In addition to,
It indicates that either the upper right symbol “+” (the portion with the arrow c) or the lower left symbol “+” (the portion with the arrow d) should be designated.

FIG. 5 is a diagram showing an example of calculating the size of the OCR input form. In this example, the data regarding the distance per point of the mouse 3 is set to 0.127 mm in both the horizontal direction and the vertical direction, and the OCR found from the data indicating the position designated by the user as shown in FIG. The number of points in the horizontal direction and the vertical direction of the outline of the input form S is 165 respectively as shown in FIG.
4,2339. Therefore, the horizontal length X and the vertical length Y of the form S are 1654 × 0.127 = 210 mm and 2339 × 0.127 = 297 mm, respectively (rounded down from the second decimal place) Is displayed on one screen of the screen 2a as shown in FIG. When the size of the OCR input form is displayed in this way, step ST3 is ended and step ST4 is advanced in response to the user commanding the shift to the next step using the keyboard 4 or the like.

[Processing in Step ST4] In step ST4, the OCR input form in which the image data is captured in step ST1 is a form in which the actual size is known from the beginning (for example, from a size of paper conforming to the JIS standard). If the size of the form displayed on the display 2 in step ST3 is different from the actual size, the user selects the keyboard 4
You should enter a number that represents the actual size using. (Thus, otherwise, step S
Since it is not necessary to perform the processing of T4, the user should immediately use the keyboard 4 or the like to order the shift to the next step. ) Then, the numerical value input by the user is displayed again as a size of the form on one screen of the screen 2a, and the data regarding the distance per point of the mouse 3 is corrected according to the numerical value input by the user.

FIG. 6 is a diagram showing an example of this correction processing. In this example, the OCR input form S in which the image data is captured in step ST1 is a JIS standard A4 size (horizontal 2
It is a form consisting of 10 mm, 297 mm long) paper, and the data on the distance per point of the mouse 3 is 0.127 in both the horizontal and vertical directions.
It is set to millimeters. Therefore, normally, as shown in FIG. 5, the horizontal direction of the outline of the form,
The number of points in the vertical direction is 1654 and 2339, respectively.
Therefore, the horizontal and vertical lengths should be accurately calculated and displayed.

However, due to the fact that the operation of the OCR module 5 by the user at the time of taking in the image data in step ST1 was inaccurate, the horizontal outline of the form S as shown in FIG. If the number of points in the vertical and vertical directions is 1654 and 2400 respectively,
The horizontal length X and the vertical length Y of the form S are calculated as 1654 × 0.127 = 210 millimeters 2400 × 0.127 = 304.8 millimeters (rounded down from the second decimal place). Sisters,
As shown in FIG. 6B, a value different from the actual length for the length in the vertical direction is displayed on one screen of the screen 2a. In this case, the user should use the keyboard 4 or the like to input the numerical value "297" representing the actual length in the vertical direction, and display the size according to this numerical value as shown in FIG.
In addition, the data on the distance per point in the vertical direction of the mouse 3 is corrected to 297 ÷ 2400≈0.124 millimeters.

When this correction is completed, step ST4 is ended and step ST5 is completed in response to the user commanding the shift to the next step using the keyboard 4 or the like.
Proceed to.

In this way, by correcting the data concerning the distance per point of the mouse 3 according to the actual size of the OCR input form, in the following steps, the position, size and shape of the reading field are calculated. The processing can be performed with higher accuracy.

[Processing in Step ST5] In step ST5, based on the user designating a plurality of locations on the character entry area of the OCR input form displayed on the display 2 using the mouse 3, the character entry is performed. The display area is displayed on the display 2 as a candidate for which a format parameter is created (that is, as a candidate for a reading field).

7 to 10 are diagrams exemplifying this display processing. FIG. 7A shows that the character entry area displayed on the display 2 has a configuration in which a plurality of character frames fr for one character are arranged in the horizontal direction at equal intervals. In this case, first, the user on the screen 2a
As shown in the figure, the upper left corner (the portion marked with an arrow a) of the frame line of the leftmost character frame fr and the lower right corner of the frame line of the rightmost character frame fr of the character frames fr forming the character writing area. (Location marked with arrow b) should be designated using the mouse 3. Then, from the data indicating these two locations designated by the user, the entire range of the reading field candidate F is displayed on the display 2 by a line 11 that surrounds the entire character writing area, as shown in FIG. To do. (For convenience of illustration, the line 11 is made a broken line, and the broken line is shifted from the frame line of the character frame fr so that the line 11 can be identified from the frame line. However, in practice, the line 11 is made a solid line. In addition, the solid line is displayed so as to overlap with the frame line of the character frame fr, and in order to distinguish the line l1 from the frame line of the character frame fr on the screen 2a, as will be described later, both Are color-coded and displayed on the display 2. The same applies to the line displaying the reading field candidate F in FIGS. 7 (c), 8, 9, and 10.)

Next, as shown in FIG. 7 (c), the user selects the upper left corner (arrow a) of the frame line of each character frame fr on the screen 2a.
It is assumed that the mouse 3 is used to indicate the position marked with and the lower right corner (position marked with arrow b). Then, from the data indicating these points designated by the user, the range of one character of the reading field candidate F is displayed on the display 2 by a line 12 overlapping the frame line of each character frame fr as shown in FIG. To do. (In FIG. 7C, for convenience,
The arrows a and b and the line 12 are displayed only in the leftmost character frame fr. )

FIG. 8A shows that the character entry area displayed on the display 2 is composed of a horizontally long character frame fr for 6 characters (the numeral "1" in the character frame fr). "-" 6 "are illustrated for the sake of convenience to show that the character frame is for 6 characters.) In this case, first, the user, as shown in FIG. The upper left corner (the portion marked with an arrow a) and the lower right corner (the portion marked with an arrow b) should be designated using the mouse 3. Then, from the data indicating these two points designated by the user, as shown in FIG. 8B, a line (here, a line overlapping the frame line of the character frame fr) surrounding the entire character writing area l
3, the entire range of the reading field candidate F is displayed on the display 2. Next, the user should input "6", which is the number of characters written in the character writing area, using the keyboard 4 or the like. Then, based on this numerical value input by the user, as shown in FIG.
Five lines l that equally divide the inside of the line l3 into six areas
4, the range of one character of the reading field candidate F is displayed on the display 2.

FIG. 9A shows that the character entry area displayed on the display 2 is composed of a character frame fr of 7 rows and 4 columns (more specifically, a cell of a summary table). Shows. In this case, first, as shown in the figure, the user selects the upper left corner of the top row (the portion marked with arrow a) and the lower right corner of the bottom row (arrow b) on the screen 2a. (The part marked with) should be designated using the mouse 3. Then, from the data indicating these two locations designated by the user, the entire range of the reading field candidate F is displayed on the display 2 by a line 15 surrounding the entire character writing area as shown in FIG. 9B. To do. Next, the user:
It is assumed that "7", which is the number of rows, and "4", which is the number of columns, should be input using the keyboard 4 or the like. Then, based on this numerical value input by the user, as shown in FIG. 9C, the six lines l6 that divide the inside of the line l5 into seven areas evenly in the horizontal direction and the inside of the line l5 in the vertical direction. With the three lines 17 equally dividing into four ranges, the range for one row and the range for one column of the reading field candidate F are displayed on the display 2.

FIG. 10A shows that, similarly to FIG. 9A, the character entry area displayed on the display 2 is composed of a character frame for 7 lines. However, FIG.
Unlike in the case of (a), there is a gap at each interval between the character frames. In this case, first, as shown in the same figure, the user first selects the upper left corner of the top row (the portion marked with arrow a) and the lower right corner of the bottom row (arrow b) on the screen 2a.
(The part marked with) should be designated using the mouse 3. Then, from the data indicating these two points designated by the user, the entire range of the reading field candidates is displayed on the display 2 by a line 15 surrounding the entire character writing area as shown in FIG. 10B. . Next, the user, as shown in FIG. 10 (b), changes the lower right corner (the part with the arrow c) of the one character range at the left end of the top row to
It should be indicated using mouse 3. From the data indicating the portion indicated by the arrow c and the data indicating the portion indicated by the arrow a, the length in the vertical direction of one character (that is, in the vertical direction in the range of one line of the reading field candidate). (Length) and the horizontal length of the range of one character are obtained. Therefore, as shown in FIG. 10C, the range of one line of the reading field candidates is displayed on the display 2 by the line 16 and The range of one character in each line of the reading field candidates is displayed on the display 2 by the line 17. still,
Regarding the character writing area of FIG. 9A, it is needless to say that the reading field candidate display processing may be performed by the same method as that for the character writing area of FIG. 10A instead of the method described above. Is.

In this way, the screen 2 of the display 2
When a character entry area to be a reading field candidate among the character entry areas displayed in a is displayed as a reading field candidate, the user uses the keyboard 4 or the like to instruct the shift to the next step. Accordingly, step ST5 is ended and the next step ST6 is executed.
Proceed to.

When a plurality of sets of character writing areas are displayed by repeating the same pattern on the screen 2a of the display 2, the display as a reading field candidate in the process of step ST5 is one of the sets. Should be done only for. Then, the remaining sets should be displayed as reading field candidates by the copying process in the next step ST6.

[Processing in Step ST6] In step ST6, when the same pattern is repeated on the display 2 to display a plurality of character writing areas,
The display as a reading field candidate in step ST5 for one of them is set as a copy target, and this display is copied at an appropriate scale on the screen 2a, so that each of the remaining groups is set as a reading field candidate. Is displayed. (Therefore, in other cases, there is no need to perform the processing of step ST6, so the user should immediately use the keyboard 4 or the like to order the transition to the next step.)

FIG. 11 is a diagram showing an example of this copying process. In FIG. 1A, the same pattern is repeated on the display 2 (that is, with the same shape) to form six sets of character writing areas C1 to C6 (one set of character writing areas has a horizontal character frame for one character). A character writing area in which five character writing areas are arranged at equal intervals in the direction and a character writing area in which three character frames for one character are horizontally arranged at equal intervals are displayed). Only one set C1 among them is displayed as the reading field candidate F1 and the reading field F2 by the process of step ST5.

In this state, the user first touches the screen 2a.
In the above, the upper left end of the set C1 (here, it matches the upper left end of the reading field candidate F1) and the lower right end (here, it matches the lower right end of the reading field candidate F2) are to be designated using the mouse 3. To do. Then, the display of the reading field candidates F1 and F2, which are the reading field candidates existing in the set C1, is set as one copy target group from the data indicating these two locations designated by the user. It should be noted that the range in which the group to be copied is set is displayed differently from the other areas on the screen 2a so that the user can clearly understand the range (for example, the display is reversed with light and dark, or shaded display). Preferably).

Next, the user displays the screen shown in FIG.
As shown in (b), the upper left corner of the upper left character frame of any one of the sets C2 to C6 should be designated by using the mouse 3 as a base point for copying. Then, from the data indicating the base point designated by the user, the display of the copy target group is copied on the screen 2a by aligning the upper left end of the display of the copy target group with the base point. As a result, the same set of display field candidates as the set C1 is displayed on the set. Next, for the rest of the sets C2-C6, FIG.
By performing the same processing as shown in (b), display as a read field candidate is performed.

If there is a set C2 to C6 which has the same shape as the set C1 but a different size (that is, a similar shape), the user can select the set as shown in FIG. In addition to designating the base point of copying as in (b), the lower right corner of the character frame at the lower right corner of the set should be designated using the mouse 3 as the end point of copying, as shown in FIG. 11 (c). And Then, based on the data indicating the base point and the end point designated by the user, the upper left end of the display of the copy target group is aligned with the base point and the lower right end of the display of the copy target group is aligned with the end point to display the copy target group. Copy on the screen 2a (that is, enlarge or reduce the display of the group to be copied).

When the copying process as described above is finished, step ST6 is ended and step ST7 is advanced in response to the user commanding the shift to the next step using the keyboard 4 or the like. By performing such a copying process, it is possible to reduce the number of places on the screen 2a that the user has to point using the mouse 3 to display the screen 2a of the display 2 as a candidate for a read field. , The work for creating the format parameter can be performed more efficiently.

In the above steps ST2, ST5 and ST6, in order to make it possible to clearly distinguish the display as the reading field candidate and the display of the character frame of the character writing area on the screen 2a, The lines for displaying the candidates (the lines corresponding to the lines 11 to 16 in FIGS. 7 to 9) and the frame lines of the character frame are displayed on the display 2 in different colors. FIG. 12 is a diagram showing an example of such color-coded display. In this example, the frame of the character frame is displayed in black, and a line that displays the entire range of the reading field candidates (lines corresponding to lines l1, l3, and l5 in FIGS. 7 to 10).
Is displayed in blue, and a line for displaying the range of one character of the reading field candidate (lines l2, l4, l in FIGS. 7 to 10) is displayed.
The lines corresponding to 6 and 17 are displayed in red. In addition, the part where the frame line of the character frame and the line for displaying the entire range of the reading field candidates overlap is displayed in yellow, and the frame line of the character frame and the line for displaying the range of one character of the reading field candidate are displayed. The overlapping area is displayed in sky blue, and the overlapping area of the line showing the entire range of the reading field candidate and the line showing the range of one character of the reading field candidate is displayed in green. . In addition, the portion where the frame line of the character frame and the line displaying the entire range of the reading field candidate and the line displaying the range of one character of the reading field candidate all overlap are displayed in pink. There is.

[Processing in Step ST7] In step ST7, predetermined field information set by the user is stored in the reading field candidates displayed on the display 2 by the processing in step ST5 or step ST6. Then, this field information is registered as, so to speak, component information used for setting field information for another read field candidate. The field information that the user should set here is the position of the read field candidate among the format parameters,
Items other than the size and shape, that is, the name of the reading field candidate, the data format of characters to be written in the reading field candidate, the reading field candidate and other reading field candidates, as described above. Of the OCR module 5 and information subject to restrictions derived from the model of the OCR module 5.

FIG. 13 is a diagram showing an example of the field information setting process. FIG. 3A shows a state in which a plurality of reading field candidates F are displayed on the display 2. In this state, the user should first specify the reading field candidate F for which the field information is to be set on the screen 2a by using the mouse 3. Then, for the read field candidate F designated by the user, a field information setting box as illustrated in FIG. 13B is displayed on one screen of the screen 2a. In this example, in the box, there is a field for setting the name (field name) of the reading field candidate F, a field for setting the data format of the characters written in the reading field candidate F, and the reading field candidate F. And the logical relationship between other reading field candidates F (referred to as an operator in the figure)
And a field for setting information (such as the number and type of characters to be written in the reading field candidate F) subject to restrictions derived from the model of the OCR module 5. Also, below these fields, there are "Cancel" command, "Modify" command, "Register" command, and "OK".
Commands are provided.

Next, the user should set the field information of the reading field candidate by inputting a desired or predetermined numerical value or symbol using the keyboard 4 in each field in the field information setting box. To do. Then, the field information set by the user is stored. Next, the user should select the "register" command using the mouse 3 or the keyboard 4.
Then, in response to the user selecting this command, this field information is registered as component information. In this way, for each read field candidate displayed on the display 2, when the field information is set and registered as the component information, the user can enter the keyboard 4
In response to commanding the shift to the next step by using, etc., step ST7 is ended and the process proceeds to the next step ST8.

If a plurality of reading field candidates having the same size and shape are displayed on the display 2 (this is the case in FIG. 13), step ST7.
The setting of the field information by the process of 1 should be performed only for one of the read field candidates. Then, with respect to the remaining read field candidates, the field information should be set using the component information by the process of step ST10 described later.

[Processing in Step ST8] In step ST8, the read field candidates are canceled or the size and shape of the read field candidates are corrected based on the selection by the user. (Therefore, when it is not necessary, the user should immediately use the keyboard 4 or the like to order the shift to the next step.)

The cancel processing is, for example, as shown in FIG.
This is done by selecting the "Cancel" command in the field information setting box shown in. In other words, if there is a character entry area displayed on the display 2 as a reading field candidate for which it is desired to cancel the instruction as the reading field candidate at step ST5 (that is, to be excluded from the target for creating the format parameter), First, the user uses the mouse 3 to specify the reading field candidate F to be canceled on the screen 2a. Then, for the read field candidate designated by the user, the field information setting box shown in FIG. 13B is displayed on one screen of the screen 2a. Next, the user should select the "Cancel" command using the mouse 3 or the keyboard 4. And in response to the user selecting this command,
The display of the read field candidate is erased from the screen 2a, and the field information stored in step ST7 for the read field candidate is cleared. As a result, the character entry area is excluded from the target for creating the format parameter.

The correction process is performed based on, for example, selecting the "correction" command in the field information setting box shown in FIG. 13B. That is, in the character entry area displayed on the display 2 as a reading field candidate, since the location designated by the user using the mouse 3 in step ST5 is incorrect, the size and shape of the reading field candidate display and the character entry are displayed. If the size or shape of the use area does not match on the screen 2a, the user sets the read field candidate field information setting box on the screen 2a in the same manner as in the canceling process. Mouse 3 after displaying
Alternatively, the "correct" command should be selected using the keyboard 4.

Next, based on the user designating a location on the screen 2a with the mouse 3, the character entry area is re-displayed on the display 2 as a candidate for a reading field. FIG. 14 is a diagram showing an example of this redisplay processing. In the figure (a), the display of the reading field candidate F on the screen 2a shows that the shape matches the character writing area but the size does not match the character writing area (that is, the character writing area). It is similar to the case). In this case, on the screen 2a, the user re-displays the upper left corner of the character frame at the upper left corner and the lower right corner of the lower right corner of the character frame on the screen 2a, respectively, as the base points for redisplay. And mouse 3 should be used as the end point. Then, from the data indicating the base point and the end point designated by the user, the upper left end of the display of the reading field candidate F is aligned with the base point and the lower right end of the display of the reading field candidate F is aligned with the end point, and the reading field candidate F is displayed. Is redisplayed on the screen 2a (that is, the read field candidate F is enlarged or reduced on the screen 2a and redisplayed). When the display size and shape of the read field candidate F matches the size and shape of the character entry area by this re-display, the user uses the mouse 3 or the keyboard 4 to select "OK" in the field information setting box. Command should be selected. This completes the correction of the display of the reading field candidate F.

On the other hand, as shown in FIG. 14B, when not only the size of the display of the reading field candidate F but also its shape does not match the character writing area, the user performs the step ST5. Similarly to the above, it is assumed that a plurality of locations on the character writing area should be redesignated using the mouse 3 (see FIGS. 7 to 10). Then, similarly to step ST5, the data indicating the location designated by the user is displayed again as a reading field candidate of the character writing area. Next, the user should select the "OK" command of the field information setting box using the mouse 3 or the keyboard 4. This completes the correction of the display of the reading field candidate F. When the reading field candidates are canceled or the size and shape of the reading field candidates are corrected, in response to the user's command to move to the next step using the keyboard 4 or the like,
The step ST8 is ended and the process proceeds to the next step ST9.

[Processing in Step ST9] In step ST9, the user selects, on the screen 2a, a position to be a reference point for position calculation within the outline of the OCR input form and a plurality of positions on the reading field candidates. The position and the size of the reading field candidate on the OCR input form are calculated based on the instruction using the mouse 3 at. Then, the calculation result is stored and displayed on the display 2 according to the user's instruction.

FIG. 15 is a diagram showing an example of this calculation processing. In the same figure, (a) shows a state in which three character entry areas in which character frames fr for one character are arranged horizontally at equal intervals are displayed on the display 2 as a pick field candidate F. In this state, first, as shown in the figure, the user sets the upper left corner of the outline of the OCR input form S (the part with the arrow a) on the screen 2a to the form base point that is the reference point for position calculation. The mouse 3 should be used as an instruction. Next, the user displays the screen 2a as shown in FIG.
In the above, it is assumed that the upper left end of the reading field candidate F (the portion marked with an arrow b) should be designated by using the mouse 3 as the field candidate base point. Then, from the data indicating the form base point and the data indicating the field candidate base point, in the vertical direction of the reading field candidate F with respect to the form base point,
Calculate the horizontal position (in the figure). And
The calculation results are displayed together with the field name (set in step ST7) as the position information of the reading field candidate F on one screen of the screen 2a as shown in FIG.

Next, the user, as shown in FIG.
On the screen 2a, the lower right corner of the reading field candidate F (the portion marked with an arrow d) and the lower right corner of the left one character range of the reading field candidate F (that is, the leftmost character frame fr) (marked with an arrow c) And the place to be used) should be designated using the mouse 3. Then, this reading field candidate F
The horizontal length (field width) (in the figure) of the read field candidate F is calculated from the data indicating the lower right corner of the field and the data indicating the field candidate base point.
From the data indicating the lower right corner of the leftmost character frame fr and the data indicating the field candidate base point, the horizontal length (character width) (in the figure) of the range of one character of the reading field candidate F and the reading field The vertical length (character height) of the candidate F (in the figure) is calculated. Then, those calculation results are stored as position information. The field width data is also registered as a part of the component information about the read field candidate F registered in step ST7.

When the user uses the mouse 3 to designate a position in the read field candidate F in which the position information is stored, the position information of the read field candidate F is correspondingly changed as shown in FIG. 15 (b). , And the field name are displayed on one screen of the screen 2a. Needless to say, it is also possible to give such an instruction for position calculation together with the display processing of the read field candidates in step ST5 described above.

In FIGS. 16A and 16B, three character writing areas of the same pattern are arranged in the vertical direction at the same intervals (the “condition” in FIG. 16 indicates that). When displayed as a candidate on the display 2 (however, the display as a reading field candidate is omitted in the figure), the points to be designated for position calculation are indicated by arrows a, b, c, d, and FIG. 6 is a diagram showing a method of calculating the position and size of each reading field candidate based on the data indicating the designated portion, and the display state of the position information. The details thereof will be easily understood from the description of FIGS. 15A and 15B, and thus the description thereof will be omitted.

When such a calculation process is completed for each reading field candidate displayed on the display 2, the step ST9 is ended and the next step is completed in response to the user commanding the transition to the next step using the keyboard 4 or the like. To step ST10. When a plurality of read field candidates having the same size and shape are displayed on the display 2, the size of the read field candidates is calculated by the process of step ST9 (field width and character height in FIG. 15). Should be performed only on one of the read field candidates. Then, for the remaining read field candidates, size information should be set by the processing of the next step ST10.

[Processing in Step ST10] In Step ST10, when a plurality of reading field candidates having the same size and shape are displayed on the display 2, one of the reading field candidates is displayed in Steps ST7 and ST9. The user retrieves and uses the registered component information as library information to set the information for each of the remaining sets. (Therefore, in other cases, it is not necessary to perform the process of step ST10, so the user should immediately use the keyboard 4 or the like to order the shift to the next step.)

FIG. 17 is a diagram showing an example of an information setting process using this component information. In FIG. 3A, the reading field candidates F1 to F5 having the same size and shape are displayed on the display 2. Of these reading field candidates F1, the processing of steps ST7 and ST9 has been completed, but for the other reading field candidates F2 to F5, the size of the processing of step ST7 and the processing of step ST9 is the same. It is in a state where calculation processing is not performed.

In this state, the user first selects the screen 2a.
Above, it is assumed that any one of the reading field candidates F2 to F5 should be designated using the mouse 3. And
For the reading field candidates designated by the user, a parts information search box as illustrated in FIG. 17B is displayed on one screen of the screen 2a. In this example, one part information at a time is displayed in the box together with the shape of the reading field candidate in which the part information is registered. As a column for displaying the part information, a column for displaying the part name (serial number in the library) of the reading field candidate, a field width and a character height that are the size of the reading field candidate (calculated in step ST9) No.) (see FIG. 15) and a column for displaying the data format, the number of characters, the type of characters, etc. (the one set in step ST7) which is the field information of the reading field candidate. There is. At the bottom of the box, an icon for displaying the next part information in the box and a field information and size information setting using the part information currently displayed in the box. A "set" command is provided.

Next, when the component information currently displayed in the component information search box is not the component information for the reading field candidate F1, the user uses the mouse 3 to point to the icon to read it. It is assumed that the part information about the field candidate F1 should be displayed in the box. On the other hand, when the part information about the reading field candidate F1 is already displayed in the box, this work is unnecessary.

Next, the displayed reading field candidate F is displayed.
If the part information for 1 is directly available for the user-specified reading field candidate, then the user should use the mouse 3 to select the "set" command. On the other hand, if the displayed part information about the read field candidate F1 cannot be used unless a part is changed (for example, the read field candidate designated by the user and the read field candidate F1 have different data formats). If you want to change the information, enter the new numerical value or symbol by using the keyboard 4 in the field where the information that needs to be changed is displayed, and then select the “Set” command. To do. Then, in response to the user selecting this command, the component information is stored as field information and size information regarding the read field candidate designated by the user.

Next, the field information and size information are set for the remaining read field candidates F2 to F5 by repeating the same processing as described above. Then, in response to the user commanding the shift to the next step using the keyboard 4 or the like, step ST10 is ended and the process proceeds to the next step ST11.

In the example of FIG. 17, the part information registered for the reading field candidate F1 is used for setting the information of the reading field candidates F2 to F5 on the same OCR input form. However, even if there is no reading field candidate having the same size and shape as the reading field candidate F1 on the same OCR input form, the format parameter creating process is performed later for another OCR input form. At this time, the reading field candidate F is added to the OCR input form as illustrated in FIG.
Reading field candidate F6 having the same size and shape as 1
It is needless to say that, if the items F to F11 exist, the component information about the reading field candidate F1 can be used for setting the information of the reading field candidates F6 to F11 in step ST10. By performing the information setting process using such component information, the work for creating the format parameter can be performed more efficiently.

[Processing in Step ST11] In step ST11, step ST7 (or step ST7)
The field information of each read field candidate stored in 10) and the position and size information of each read field candidate stored in step ST9 (or step ST9 and step ST10) are converted into a format that the OCR module 5 can interpret. By converting, a format parameter for the OCR input form is created. Then, the created format parameter is recorded in the disk device 6. This completes the format parameter creation processing.

Next, of the above format parameter creation processing, the image data acquisition processing (step ST
A modification example of the process 1) will be described. In the above format parameter creation processing, in step ST1, OC
Image data was captured from the OCR input form using the R module 5, but if the OCR input form is a multi-colored form, the image data can be imported using an image scanner that can capture the image data in color. May be performed. In such an image scanner, the frame lines of all the colors on the OCR input form are once captured as image data, and then any one of a plurality of types of colors is designated as a dropout color. Color processing. Therefore, even when the same color as any of the plurality of types of colors exists on the OCR input form, it is not necessary to specify the dropout color before the image data is captured as in step ST1 described above. Instead, after the image data is captured, by the time the image data is output to the display 2 in step ST2, a color other than the color existing on the OCR input form may be designated as the dropout color. .

Next, a modified example of the reading field candidate display processing (the processing of steps ST5 and ST6) in the above-described format parameter generation processing will be described. FIG. 19 is a diagram showing a modification of this display process. FIG. 11A shows a state in which the image of the OCR input form S is displayed on the display 2 by the processing in step ST2 of the above format parameter creation processing. In this state, when the processing of step ST3 and step ST4 of the above-described format parameter creation processing is completed, first, the frame structure (that is, the character frame) of each one is automatically generated from the image displayed on the display 2. To detect. The detection of such a frame structure can be performed by a known technique. Next, an area in which a plurality of these frame structures are regularly arranged (at the same interval) is automatically detected. As a result, in this figure, six sets of regions each having three character frames arranged at the same interval are detected. And
As shown in FIG. 7B, all the detected areas are displayed on the display 2 as the reading field candidates F1 to F6 in the same manner as shown in FIGS.

Next, the user selects the reading field candidate F
If there is an area that is not adopted as a reading field candidate (for example, an area for printing comment characters for the user, not an area for writing characters) among the areas displayed as 1 to F6, the screen is displayed. It is assumed that the area on 2a should be designated by using the mouse 3. Then, with respect to the area designated by the user, the display as the reading field candidate is deleted from the screen 2a. FIG. 19 (c)
Are read field candidates F2, F4 and
The display of F6 is erased, and reading field candidates F1 and F
Only the display of F3 and F5 remains. As a result, only the character entry area is displayed on the display 2 as a reading field candidate. When this display process ends, the process proceeds to step ST7 of the above format parameter creating process.

In the above embodiment, the mouse is used to indicate the location on the screen 2a of the display.
A point on the screen 2a of the display may be designated using a pointing device other than the mouse (eg, a digitizer). In the above embodiment,
Although the present invention is applied to a small-scale information processing system having a personal computer as a core, the present invention may be applied to other information processing systems.

[0071]

As described above, according to the OCR format parameter creating method of the present invention, the work for creating the format parameter is performed through the screen of the display, so that the format parameter can be easily created. It has an excellent effect that it can be created in a short time. Further, since the position, size, shape, etc. of the reading field can be calculated from the data indicating the position on the screen of the display designated by the pointing device, the excellent effect that the format parameters relating to them can be created with high accuracy Play.

When the image data is taken in from the OCR input form by using the OCR device of the same model as the OCR device used to read the characters on the OCR input form, Since the resolution when capturing image data from the input form and the resolution when reading characters on the OCR input form are equal, the created format parameters are supplied to the OCR device.
When the characters on the input form are read, it is possible to prevent an erroneous recognition from occurring due to a difference in resolution, which is an excellent effect.

[Brief description of drawings]

FIG. 1 is a block diagram showing the overall configuration of an information processing system for implementing the present invention.

FIG. 2 is a flowchart showing an example of processing executed by a personal computer.

FIG. 3 is a diagram showing an example of a process of capturing image data.

FIG. 4 is a diagram showing an example of a display on a display.

FIG. 5 is a diagram showing an example of a calculation process of a size of an OCR input form.

FIG. 6 is a diagram showing an example of a calculation process of a size of an OCR input form.

FIG. 7 is a diagram showing an example of display processing of a read field.

FIG. 8 is a diagram showing an example of read field display processing.

FIG. 9 is a diagram showing an example of read field display processing.

FIG. 10 is a diagram showing an example of display processing of a read field.

FIG. 11 is a diagram showing an example of a copy process for displaying a read field.

FIG. 12 is a diagram showing an example of color-coded display on a display.

FIG. 13 is a diagram showing an example of field information setting processing.

FIG. 14 is a diagram showing an example of correction processing of display of a read field.

FIG. 15 is a diagram showing an example of a calculation process of position information of a read field.

FIG. 16 is a diagram showing an example of a calculation process of position information of a read field.

FIG. 17 is a diagram showing an example of information setting processing using part information.

FIG. 18 is a diagram showing an example of information setting processing using part information.

FIG. 19 is a diagram showing a modified example of the reading field display processing.

[Explanation of symbols]

 1 Personal Computer 2 Display 3 Mouse 4 Keyboard 5 OCR Module 6 Disk Unit 7 Printer S OCR Input Form

Claims (21)

[Claims] 1. A method of creating a format parameter as information defining a format of an OCR input form, the first step of capturing image data from the OCR input form, and outputting the image data to a display. A second step; a third step of instructing a character entry area of the OCR input form displayed on the screen of the display using a pointing device; and data indicating a portion instructed by the pointing device And a fourth step of creating a format parameter for the OCR input form based on the above. 2. The OCR reading according to claim 1, wherein the OCR device of the same model as the OCR device used to read the characters on the OCR input form is used to capture the image data in the first step. Format parameter creation method for. 3. An apparatus for capturing image data in the first step specifies and reads any one of a plurality of types of colors as a dropout color which is a color invisible to the apparatus. And the same color as any one of the plurality of types of colors is OCR.
The OC according to claim 1 or 2, wherein, when the data exists on the input form, the color other than the same color is used as a dropout color to output the image data in the second step.
Method for creating R read format parameter. 4. The OCR according to claim 3, wherein a color other than the same color is designated as a dropout color before the image data is captured in the first step.
How to create read format parameters. 5. The OCR input form is captured in the second step by capturing image data of the form with an object whose brightness is different from that of the OCR input form in the background. 5. The OCR read format parameter creating method according to claim 1, wherein the outline of the above can be clearly displayed on the display screen. 6. The plurality of OCR input forms displayed on the screen of the display in the second step are designated by using a pointing device to specify a plurality of specific positions within the outline of the OCR input form. 6. The OC according to claim 1, further comprising a step of calculating a size of the OCR input form from data indicating a location.
Method for creating R read format parameter. 7. The size of the calculated OCR input form is
The method further comprising the step of correcting data relating to a minimum distance between points which the pointing device can point to next to each other on the screen when the size is different from the actual size of the OCR input form known in advance. O described in 6
Method of creating format parameter for CR reading. 8. In the third step, based on instructing a plurality of locations on a character writing area with a pointing device, the character writing area is set as a candidate for creating a format parameter. The OCR read format parameter creating method according to any one of claims 1 to 7, wherein the OCR read format parameter is displayed on a screen of a display. 9. The whole of the range of the candidates based on the designation of the positions on both sides of the frame for character writing areas having a configuration in which a plurality of frames for writing one character are arranged. 9. The method for creating a format parameter for OCR reading according to claim 8, wherein is displayed. 10. Regarding a character writing area having a configuration in which a plurality of frames for writing one character are arranged, a plurality of portions on each of the frames are designated, and one character of the candidate The range of each is displayed, respectively.
A method of creating a format parameter for OCR reading according to [4]. 11. With respect to a character entry area consisting of one frame for entering a predetermined number of characters, the entire range of the candidate is displayed based on indicating a plurality of locations on the frame. 9. The range of one character of the candidate is displayed by dividing the entire range into the predetermined number based on subsequently inputting data indicating the predetermined number.
A method of creating a format parameter for OCR reading according to [4]. 12. With respect to a character entry area composed of a frame for entering a predetermined number of lines of characters, the entire range of the candidate is displayed based on a plurality of points on the frame, 9. The OCR read format parameter according to claim 8, wherein the range for one line of the candidate is displayed by dividing the entire range into the predetermined number based on inputting data indicating the predetermined number. How to make. 13. In the third step, when a plurality of sets of character writing areas are displayed by repeating the same pattern on the screen of the display, the character writing is performed only on one of the sets. The display as the candidate is performed on the basis of pointing a plurality of locations on the use area using a pointing device, and the display as the candidate for the one set is copied for each of the remaining sets. 13. The OCR read format parameter creating method according to claim 8, wherein the display is made as the candidate by setting the display as a candidate and copying the display at an appropriate scale on the screen of the display. . 14. The OCR reading format according to claim 8, wherein the display of the character writing area and the display of the character writing area as the candidate are performed in different colors on the screen of the display. How to create parameters. 15. The method according to claim 1, wherein a predetermined type of information for creating a format parameter is set in the character entry area designated in the third step. OCR read format parameter creation method. 16. The predetermined type of information is registered as part information used to set the predetermined type of information for another character entry area, and the other character entry area is registered as part information. 16. The OCR read format parameter creating method according to claim 15, wherein the predetermined type of information is set by using the component information. 17. The information of the predetermined type includes a name of a character entry area, a data format of a character entered in the character entry area, a character entered in the character entry area, and other characters. 17. The method for creating an OCR reading format parameter according to claim 15 or 16, which includes a logical relationship with a character to be written in the writing area and information that is subject to a constraint derived from the model of the OCR device. 18. The position and size of the character entry area on the OCR input form are calculated from the data indicating the location on the character entry area designated in the third step. Item 18. A method for creating an OCR reading format parameter according to any one of Items 1 to 17. 19. A step of detecting a frame structure displayed on the display screen by the second step, and displaying all regions including the detected frame structure on the display screen as candidates for creating format parameters. A step of selecting whether or not to adopt the area displayed as the candidate as the candidate, and a step of deleting the display of the area that is not adopted as the candidate as the candidate from the display screen. 7. The method according to claim 1, further comprising: wherein the processing in the fourth step is performed on the area adopted as the candidate based on the instruction in the third step.
Alternatively, the OCR read format parameter creating method described in any one of 14 to 18. 20. The OCR reading device according to claim 1, wherein the character entry area designated in the third step is adapted to selectively cancel or modify the designation. How to create format parameters. 21. In the fourth step, the format parameters are created by converting the predetermined type of information and the information about the position and size of the read field into a format that can be interpreted by an OCR device. The OCR read format parameter creating method according to claim 18, which is performed.