JPH0785215A - Character recognizing device - Google Patents

Abstract

PURPOSE:To provide a character recognizing device capable of accurately reproducing the format image of a read original by segmenting the positions, sizes and ruled fines of characters together with the characters at the time of reading out an original and converting the segmented data into the one to be utilized by a character editing device or the like. CONSTITUTION:In this character recognizing device, an original is read out by an image scanner 1 and stored in an image memory 2 as image data. When an operator specifies an area, a CPU 3 executes layout analysis, character recognition and data conversion as recognizing processing. Ira the layout analysis, whether the specified area is a character field or a table field is judged. In the case of a table field table ruled line (coordinate) is segmented. In the case of a character field, a character string is extracted, the character field is segmented in each line and its inclination is corrected (similar processing is executed also when a character string exists in a table field). In the character recognition, recognition dictionary data are loaded from a storage device to a work memory 5 and characters in the specified area are recognized. In the data conversion processing, the result of the layout analysis is collated with that of the character recognition, processed data are converted into data to be edited by a document editing device and the converted data are stored in a recognized data file as recognized data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device.

[0002]

2. Description of the Related Art A conventional character recognition device is a dictionary for character recognition in which a character portion within a designated range of a print document or a handwritten document is cut out and read as a dot image of one character, and the read data is analyzed. Find the appropriate character from the, and automatically character code that character (text data)
Had been converted to.

[0003]

In the prior art as described above, the code-converted data is converted into a character editing device (or the code-converted character code (data) is converted into an editing function of the character recognition device (hereinafter referred to as a character). When referred to by a character editing device, etc.) for performing editing operations such as corrections and additions), only character data is displayed and the image of the read document (position and size of characters, ruled lines or table) is displayed. There was a problem in that since it could not be reproduced, it had to be corrected or edited again.

That is, in the conventional character recognition apparatus, only the character code of the document is converted, and the format layout cannot be reproduced (character size, character position, ruled line).
Therefore, conventionally, in order to reproduce the format layout, it has been necessary to edit the data (recognition data) after the character code conversion with a word processor or the like to reproduce the character size, line pitch, ruled line, and other formats.

The present invention has been made in view of the above problems, and when reading a document, the character position, size, ruled line, or table is cut out together with the character and can be used as data in a character editing device or the like. When it is converted to data (editing data by the text editing device) and called by a text editing device, the format image of the read original is reproduced to make corrections and edits again (character alignment, size change, ruled lines (table)). It is an object of the present invention to provide a character recognition device that does not require input).

[0006]

In order to achieve the above object, the character recognition device of the first invention reads a document including a table composed of written or printed characters and / or ruled lines and outputs it as image data. An image reader for storing image data, a processing device for executing recognition processing by the recognition processing means, a recognition dictionary data storage device for storing recognition dictionary data used for character recognition, and a recognition device obtained by the recognition processing. In a character recognition device having a storage device for storing recognition data and an input device for inputting operation information, a recognition processing means, an area designating means for designating a region of image data by the input device,
Layout analysis means for analyzing the layout of the designated area for each designated area designated by the area designating means to obtain a layout analysis result, and a dictionary for recognizing character strings of character fields in the designated area designated by the area designating means Character recognition means for recognizing characters based on data to obtain character recognition results, and data conversion means for synthesizing character recognition results and layout analysis results to add format information and creating recognition data as format documents It is characterized by

According to a second aspect of the present invention, in the character recognition apparatus according to the first aspect of the present invention, the layout analysis means determines the field type of the designated area, and the ruled line when the field type is a ruled line or a table field. A ruled line cutting-out means for cutting out a character string, a first character string cutting-out means for cutting out a character string from the table field when the field type is a character field, and a character field when the field type is a character field A second character string slicing means for slicing a character string.

According to a third aspect of the present invention, in the character recognition apparatus of the first aspect, the layout analysis means detects the inclination of the character or the entire document, and the field type determination means which determines the field type of the designated area. When the field type is a table field, a ruled line cutting-out means for cutting out a ruled line, and when the field type is a table field including a character string, a character string is cut out from the table field and every one line is extracted from the character string. First character string cutting means for cutting out a character string, and second character string cutting out for cutting out a character string from the character field and cutting out a character string for each line from the character field when the field type is the character field And a tilt correction means for correcting the tilt for each line of the character string.

According to a fourth aspect of the present invention, in the character recognition device according to the first aspect of the present invention, the recognition data editable as a format document is a. Ruled line information which is a repetition of a set of ruled line data and at least a set of individual ruled line information including a ruled line type code, a start coordinate, and an end coordinate, and b. It is characterized in that it includes a code indicating a character data length, a character pitch, and a line interval, and at least character information including a repetition of a set of a control code and a character code.

A fifth aspect of the present invention is the character recognition apparatus according to any of the first to fourth aspects, further comprising a recognition data storage means for storing the recognition data in a storage device.

A sixth aspect of the present invention is the character recognition device according to any one of the first to fourth aspects, further comprising a recognition data storage means for storing the recognition data in a storage device and an image data storage means for storing the image data in the storage device. It is characterized by having.

[0012]

With the above construction, the character recognition device of the first invention is
The recognition processing means specifies the area of the image data with the input device by the area specifying means, analyzes the layout of the specified area for each specified area specified by the area specifying means by the layout analyzing means, and obtains a layout analysis result. The character recognition unit obtains a character recognition result by performing character recognition on the character string of the character field in the designated area designated by the area designation unit based on the recognition dictionary data, and the character conversion result and layout analysis by the data conversion unit. The results are combined and format information is added to create recognition data as a format document.

According to a second aspect of the present invention, in the character recognition apparatus according to the first aspect, the layout analyzing means discriminates the field type of the designated area by the field type discriminating means and cuts out a data string based on the field type. In this case, when the field type is a table field, the ruled line cutting means cuts out a ruled line, and when the field type includes a character string, the first character string cutting means cuts out a character string from the table field. In this case, the character string is cut out from the character field by the second character string cutting means.

According to a third aspect of the present invention, in the character recognition apparatus of the first aspect, the layout analysis means detects the inclination of the character or the entire document by the inclination detection means, and the field type determination means determines the field type of the designated area. To do.
Then, when the field type is a table field, the ruled line is cut out by the cutting out means, and when the field type includes a character string, the character string is cut out from the table field by the first character string cutting out means, and then the character string A character string for each line from the character field, and in the case of a character field, the character string is cut out from the character field by the second character string cutting means, and then 1 from the character field.
Cut out a character string for each line. Further, the inclination correction means corrects the inclination of the character for each line of the character string.

A fourth aspect of the present invention is the character recognition device according to the first aspect of the present invention, wherein a. Ruled line information which is a repetition of a set of ruled line data and at least a set of individual ruled line information including a ruled line type code, a start coordinate, and an end coordinate, and b. A recognition data editable as a format document is created, which includes a code indicating a character data length, a character pitch, and a line spacing, and at least character information including a repetition of a set of a control code and a character code.

A fifth aspect of the present invention is the character recognition device according to any one of the first to fourth aspects, wherein the recognition data storage means further stores the recognition data in the storage device.

According to a sixth aspect of the present invention, in the character recognition device according to the first to fourth aspects, the recognition data storage means stores the recognition data, and the image data storage means stores the image data in the storage device.

[0018]

1 is a block diagram of an embodiment of the character recognition device of the present invention, and FIG. 2 is a block diagram of another embodiment of the character recognition device of the present invention. The apparatus of FIG. 2 continuously reads character (and format) originals and stores recognition data in a storage device.
After the reading process is completed, it has a function of collectively performing a patch process for correcting with a character editing device or the like (not shown).

In FIGS. 1 and 2, reference numeral 1 is an image scanner (image reading device) for reading characters of a printed original or handwritten original, 2 is an image memory for temporarily storing (storing) image data read by the image scanner, Reference numeral 3 denotes a CPU (central processing unit) that controls the operation of the character recognition device and executes recognition processing by the recognition processing means described later.
4 is a memory for storing the recognition data recognized by the recognition process and the image data from the image memory 2.
A device 5 is a work memory (internal memory) for temporarily storing each processing means (program), data, dictionary data, or set values or parameters required for recognition processing by the CPU 3, and 6 is an image reader 1. A storage device that stores a recognition dictionary file composed of recognition dictionary data for recognizing the read characters, 7 is a display device that displays operation information necessary for operating the character recognition device, operation information such as menus or icons, and 8 is The input device has a keyboard or a pointing device such as a mouse for inputting operation information necessary for operating the character recognition device.

Further, in FIG. 2, 9 is an output device (printer or magnetic disk for storing recognition data). The storage devices 4 and 6 may be physically the same device. Further, in the embodiment, the recognition processing means such as the layout analysis means, the character recognition means, the data conversion means, etc. are stored in the storage device 4 as the program means, and are read out when the character recognition device is activated and the devices are driven into the work memory 5. Held for
It is executed by the CPU 3 in a timely manner.

FIG. 3 is an operation block diagram of the character reading device, and FIG. 11 is an example of a read document. The operation of the character reading device of the present invention will be described below with reference to FIGS. In addition, in FIG. 11, for the sake of explanation, a sentence (character group) 1
11 and Table 112 are shown above and below the document, but Table 112
The present invention can be applied to the case where a character is written inside.

The setting block 30 sets recognition conditions. The recognition conditions include, for example, the type of document (for example, word processing printing, book (printed characters), etc.), character direction determination conditions (automatic determination, manual operation, etc.), whether or not format can be reproduced, whether ruled lines can be reproduced, or in the table. Whether or not characters can be reproduced, etc., and these conditions are displayed on the display device 7 at the time of setting, and are specified and input by the operator with the mouse. Each input setting condition is CPU3
It is stored in a predetermined address of the work memory 5 after being checked in. In the reading block 31, the image data is drawn in the image memory 2 while reading the original 110 (see FIG. 11) set in the image scanner 1 in units of a reading buffer (54 Kbytes), and at the same time, the entire image of the original is displayed on the display device 7. indicate.

Next, in the area designating block 32, the position of the starting point and the ending point of the rectangle displayed on the screen by the area designating means are designated by the mouse of the input device 8 and input. FIG. 12 shows FIG.
This is an example of image display of one document 110 and area designation. In FIG. 12, two areas F1 are designated areas.
(FIX1, FIY1), F2 (F2X1, F2Y1)
It is shown.

In this case, the areas are recognized in the order in which the areas are designated. When the area is not designated, the entire original is designated as one area. In the recognition block 33, processes such as layout analysis, character recognition, and data conversion are executed based on the flowchart of FIG. Note that FIG.
4 is a detailed flowchart of the layout analysis (step S3) of FIG. 4, FIG. 6 is a detailed flowchart of the character recognition (step S13) of FIG. 4, and FIG. 7 is a detailed flowchart of the data conversion (step S15) of FIG.

In FIGS. 4A and 4B, steps S1 and S2 are the operation of securing the area by the above-mentioned area designating block, and the description thereof is omitted. In step S3, the CPU 3 executes the layout analysis unit shown in FIG. 5 to perform layout (based on the setting conditions set in the setting block 30 and stored in the work memory 5 and the coordinates of the area specified in the area setting block 32). la
y-out) Perform analysis. In the layout analysis, as shown in the flowchart of FIG. 5, first, in the preprocessing of step S30, a region where black dots are gathered in unknown data is searched for and enclosed by a rectangle (internal processing).

Next, in step S31, the inclination of the character is checked. If ± 5 ° or more, read again.
As shown in FIG. 8, there are two types of inclination of characters: the case where the position of the original is normal and the characters themselves are inclined, and the case where the original is not inclined as shown in FIG. 9 and a combination of both. is there.

When the document is tilted, the upper left coordinates (x1, y1) and the lower left coordinates (x2, y2) as shown in FIG. 9A.
Since it is difficult to determine the inclination only by itself, the lower left coordinates (x1, y4) and the upper right coordinates (x2, y3) as shown in FIG. 9B.
Is used to determine the inclination, and whether it is falling to the right or to the left is determined as a whole. If the tilt check is OK, it is determined in step S32 of FIG. 5 whether the designated area of the read document is a character field or a front field. If the table field is discriminated, the process proceeds to steps S33 to S35 and then moves to step S36. If it is not the table field, the process proceeds to step S36. Whether the field is a table field is determined by extracting ruled lines when the line segment is larger (longer) than a predetermined size, and when the number exceeds a predetermined number, it is determined as a table. In addition, ruled lines are extracted by rungless (black dot contact length).

If there is a character string in the table in step S33, the character string is extracted from the marginal distribution in step S34,
The character field is cut line by line and the inclination is corrected. If there is no character string, the process proceeds to step S35.

FIG. 8 shows a case of correction (licking correction) when a character is inclined, and FIG. 8A shows a case of downward sloping. In the case of downward sloping, the coordinates are rotationally moved (coordinate conversion) about the origin 0 in the positive direction (the direction of the arrow in the drawing) to correct the x and y coordinates. In the case of downward sloping, the coordinates are rotationally moved in the negative direction (the direction of the arrow in the figure) about the origin 0 to correct the x, y coordinates. Even if the document is tilted, it is determined whether the document is on the right or left as a whole, and then, for example, the coordinate (x1, y1) is rotated as the origin to convert the x, y coordinates.

In step S35, only the table ruled lines (coordinates) are cut out, and in step S36 it is determined whether or not the ruled line is the flat distribution of the peripheral distribution and the aspect ratio of the circumscribed rectangle. If the determination result is a ruled line, the ruled line is cut out in step S37, and the process proceeds to step S38. If the result of determination is other than a ruled line, the process immediately proceeds to step S38.

In step S38, it is determined whether the character is a character based on the size of the rectangle. In the case of a character, a character string is extracted from the peripheral distribution in step S39, the character field is cut out line by line, and the inclination is corrected, and the process proceeds to step S40.
The case where there is no character is the case of a figure or a photograph, the layout analysis is ended, and the process proceeds to step S4 (FIG. 4).

In step S40, it is checked whether or not there is a designated area for which layout analysis has not been completed, and if there is a designated area for which layout analysis has not been completed, steps S32 to S39 are repeated to process the next designated area. If the layout analysis of the designated area has been completed, the layout analysis is terminated and the process proceeds to step S4 (FIG. 4). The layout analysis result is stored in a predetermined area of the work memory 5 in a format as shown in FIG. The coordinates of the row field that is the result of the layout analysis are the row coordinates that have not been licked (the row coordinates that are faithful to the original image).
Is.

Here, as a concrete example of steps S32 to S39, taking the case of the document 110 of FIG. 11 as an example, it is determined in step S32 whether the area F1 is a front field. Since the area F1 is not a table, it is determined in step S36 whether it is a ruled line. Further, since the area F1 is not a ruled line, it is determined in step S38 whether it is a character string. Area F
Since there is only a character string in No. 1, the character field is cut out line by line and inclination is corrected line by line in step S39. Note that FIGS. 16 and 17 show layout analysis results and the like using the document 110 of FIG. 11 as an example. Here, FIG. 17 shows layout analysis information (image data of the original 110), and FIG. 16 shows layout analysis results (row coordinates and ruled line coordinates of the original 110).

Resulting area F2 of the presence judgment of the designated area in S40
Is an unprocessed area, the process returns to step S32 and the area F2
Is a table field. Since the area F2 is a table, it is determined in step S33 whether there is a character string in the table. In this example, it is determined that there are no characters in the area F2,
In step S35, ruled line coordinates are cut out. Area F2
The ruled line coordinates of are as follows.

Ruled Line 1 K1 (K1X, K1Y), K2 (K2X, K2
Y) Ruled line 2 K3 (K3X, K3Y), K4 (K2X, K4
Y) Ruled line 3 K1 (K1X, K1Y), K3 (K3X, K3
Y) Ruled line 4 K2 (K1X, K1Y), K4 (K4X, K4
Y) In step S4 of FIG. 4, the CPU 3 determines whether or not there is a layout analysis error, and if there is a layout analysis error, releases the designated area (interface area) and terminates the process (the operator displays a block 32 due to an error display). Start again from the area setting). If there is no layout analysis error, the CPU 3 displays on the screen of the display device 7 in step S5.
For example, the character string is enclosed in a rectangle and the layout analysis result is displayed (FIG. 13).

Next, in steps S6 to S14, character recognition processing is executed. In step S6, the CPU 3 displays a message asking whether to execute the character recognition process on the display device 7 and prompts the operator to input the support. The operator gives an input instruction with a mouse. If it cannot be executed, the designated area (interface area) is released and the processing is terminated. If it is executable, the CPU 3 loads the recognition dictionary data from the recognition dictionary file stored in the storage device 6 into the work memory 5 in step S7 (block 34), checks whether or not there is a load error in step 8, and then executes step S9. Then, the layout analysis screen of the display device 7 is switched to the recognition screen and the process proceeds to step S10. For example, row K of area F1 in FIG.
The part 1 is highlighted.

In step S10, the CPU 3 determines whether or not there is an unrecognized area. If there is an unrecognized area, the process proceeds to step 11, and if there is no unrecognized region, the process proceeds to step S17 to release the recognition dictionary data in the work area and release the designated area in step S18 to move to the next block. Move on to.

In step S11, the CPU 3 determines whether or not to continue the recognition process. If not, the process proceeds to step S17, the recognition dictionary data in the work area is released, and the designated area is released in step S18. Move on to the next block. In step S12, the CPU 3 displays the image line to be recognized on the display device 7. In step S13, the CPU executes character recognition means based on the layout result based on the flowchart of FIG. 6 to perform character recognition. As a specific example, the character recognition of the portion of the row K1 of the area F1 of FIG. 13 will be described with reference to FIG.

In step S131, the CPU cuts out characters from the line L1 (1 in this example, the CPU cuts out characters from the line L1 in step S131 (10 in this example; in FIG. 13, the fourth and subsequent characters are specific characters). Is displayed for the sake of explanation (space is displayed), the character coordinate C1
(C1X1, C1Y1), C2 (C2X1, C2Y
1), C2 (C2X1, C2Y1), ... C10 (C1
0X1, C10Y1) and the character shape (full-width; 10 characters) are stored in the work memory 5.

Next, in step S132, preprocessing (normalization of size,
Position normalization, noise removal, and dot loss correction) are performed in step S133 to recognize one character (extract a feature amount from preprocessed image data and compare it with a recognition dictionary) to obtain a character code and recognition accuracy.

In step S134, if the recognition result is outside the predetermined threshold, the recognition result is rejected and step S131
Then, the process proceeds to the recognition process of the next line (L2). The rejected character coordinates are coded, for example, as “black square characters”.

If there is an unrecognized character in the original line (L1) in step S132, the process returns to step S132 to repeat the recognition process (preprocessing, character recognition) of the next character. When the character recognition of FIG. 6 is completed for one line of the character field (line L1 in the specific example), the CPU 3 determines in step S14 of FIG. 4 whether there is a recognition error.
15, the data conversion processing is performed by executing the data conversion means as shown in the flowchart of FIG. Note that FIG.
Is an example of a recognition result (a character code and a character coordinate taking the document 110 as an example). The recognition result is stored in the work memory 5 in the format shown in FIG.

The data conversion process is step S151 in FIG.
Then, the layout analysis result of step S3 and step S13
Results of character recognition are compared, and steps S152, S
At 153, the data is converted into data (document editing device edit data, see FIG. 18) that can be edited by the document editing device (not shown).
The converted document editing device edit data is stored in the recognition data file as recognition data and the display device 7 displays the recognition data.

Specifically, ruled line information is created for ruled line coordinates, line fields are calculated for line coordinates, line feed codes are added, and character control codes (character points) and character pitches are calculated for character coordinates. Also, the pitch code and space code are added. The calculation of the character points is performed by calculating the size from the character coordinates.

As shown in FIG. 18, the recognition data basically has a simple structure composed of text information (see below) of document data and graphic ruled line information (see below). (1) Text information Character data length Character pitch Line field Control code, character code ... (Repeated as a set) (2) Ruled line information Ruled line data length Ruled line type, page, ruled line code, start coordinate (X, Y), end coordinate ( (X, Y) ... (repeat as a group)

When the recognition character is displayed in step S15, the line uses the y coordinate of the recognition line coordinate and the digit uses the X coordinate of the recognition character coordinate. As shown in the lower part of FIG. 10A, the recognized character coordinates are obtained by performing character recognition by licking and correcting the coordinates of the close rectangle of the layout result of the characters, and the recognized character coordinates (see FIG. 10A) are not changed. If used as the display coordinates of the recognition result, the Y coordinate will be different for each character, and the lines may be disjointed. Therefore, the CPU 3 calculates a rectangle in which all the recognized character coordinates are included in each line,
Let it be the recognition line coordinate.

When the data conversion for one line, the storage of the recognition data, and the display are completed, the CPU 3 repeats steps S11 to S15 to process the next line in step S16. In the specific example, when the data conversion of the character area F1 is completed, the process proceeds to the table (ruled line) area F2. In this case, since there is no character string in the table, only the layout result is obtained without performing character recognition (S13, S14). Based on the above, ruled line information is created in step S15 and stored as recognition data (document editing device edit data: see FIG. 18) in the recognition data file. When the processing of all the designated areas of the read document is completed, the recognition dictionary is released in step S17, the designated area is released in step S18, and the processing of the next block is performed.

After the recognition process of block 33 in FIG. 3 is completed, the recognition data (document editing device edit data) is sequentially displayed on the display device 7 (see the display example of FIG. 14), and the character recognition device of FIG. Subsequently, correction processing (change or deletion of rejected characters, addition of characters, etc.) is performed, but the character recognition apparatus of FIG. 2 returns to block 31 to read the next document and repeats the processing of blocks 31 to 33 ( The correction can be collectively performed separately after reading all the originals and storing the recognition data). Further, in the case of batch processing (batch processing) of FIG. 2, the image data read in block 31 is stored in the storage device.

As described above, according to the character recognition device of the first invention, it is possible to obtain the recognition data converted into the recognition data (document editing device editing data) which can be edited by the document editing device such as a word processor. Therefore, it is possible to reproduce the layout in the read format (reproduction of ruled lines, character size, character pitch, etc.). Further, according to the character recognition device of the second invention, the correction of the recognition data can be separately corrected by the document editing device such as a word processor.

[0050]

As described above, according to the character recognition device of the first invention, it is possible to obtain the recognition data converted into the data editable by the document editing device such as a word processor (document editing device editing data). As a result, it is possible to reproduce the layout in the read format (reproduction of ruled lines, character size, character pitch, etc.). Therefore, it is not necessary to add ruled lines and edit the size of characters again as in the conventional case. Further, according to the character recognition device of the second invention, since the correction of the recognition data can be separately corrected by the document editing device such as a word processor, it is necessary to perform the correction work in the case of making a database of a large number of documents (for example, books). It can be distributed to multiple people and can contribute to shortening the work period.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of a character recognition device of the present invention.

FIG. 2 is a configuration diagram of another embodiment of the character recognition device of the present invention.

FIG. 3 is an operation block diagram of the character reading device.

FIG. 4 is a flowchart showing a recognition operation of the character recognition device of the present invention.

5 is a detailed flowchart of the layout analysis of FIG.

6 is a detailed flowchart of the character recognition of FIG.

FIG. 7 is a detailed flowchart of the data conversion of FIG.

FIG. 8 is an explanatory diagram of the inclination correction of a character.

FIG. 9 is an explanatory diagram of document skew correction.

FIG. 10 is an explanatory diagram of recognized character coordinates and recognized line coordinates.

FIG. 11 is an example of a document to be read.

FIG. 12 is an example of area designation.

FIG. 13 is a display example of a layout analysis result.

FIG. 14 is a display example of edit data of a document editing device.

FIG. 15 is an example of layout analysis information.

FIG. 16 is an example of a layout analysis result.

FIG. 17 is an example of a recognition result.

FIG. 18 is an example of the data edited by the document editing apparatus.

[Explanation of symbols]

 1 Image Scanner 2 Image Memory 3 CPU 4 Storage Device 5 Work Memory 6 Recognition Dictionary File 7 Display Device 8 Input Device

Claims (6)

[Claims] 1. An image reader for reading a document including a table of written or printed characters and / or ruled lines and outputting it as image data, an image memory for storing the image data, and a recognition process by a recognition processing means. Character recognition having a processing device, a recognition dictionary data storage device that stores recognition dictionary data used for character recognition, a storage device that stores recognition data obtained by the recognition process, and an input device that inputs operation information. In the apparatus, the recognition processing means analyzes the layout of the designated area for each designated area designated by the area designating means, the area designating means designating the area of the image data by the input device, and a layout analysis result. A layout analysis means for obtaining the specified area specified by the area specifying means. Character recognition means for recognizing a character string in a character field based on the recognition dictionary data to obtain a character recognition result, and combining the character recognition result and the layout analysis result to add format information and recognizing as a format document. A character recognition device comprising: a data conversion unit that creates data. 2. The character recognition device according to claim 1, wherein
When the layout analysis means determines the field type of the designated area, and the field type is a ruled line or a table field,
A ruled line cutting-out means for cutting out a ruled line; a first character string cutting-out means for cutting out a character string from the table field when the field type is a table field containing a character string; and a character field when the field type is a character field A second character string cutout unit for cutting out a character string from the character recognition device. 3. The character recognition device according to claim 1, wherein
The layout analysis unit detects a character or the inclination of the entire document, a field type determination unit that determines the field type of the designated area, and a ruled line cutting unit that cuts out a ruled line when the field type is a table field. When the field type is a table field including a character string, a first character string cutout unit that cuts out a character string from the table field and cuts out a character string for each line from the character string, and the field type is a character field At this time, a second character string cutout unit that cuts out a character string from the character field and a character string for each line from the character field, and a tilt correction unit that corrects the tilt for each line of the character string are provided. A character recognition device having. 4. The character recognition device according to claim 1,
The recognition data that can be edited as a format document includes a. Ruled line information which is a repetition of a set of ruled line data and at least a set of individual ruled line information including a ruled line type code, a start coordinate, and an end coordinate, and b. A character recognition device comprising: a character data length, a character pitch, a code indicating a line spacing, and at least character information including a repetition of a set of a control code and a character code. 5. The character recognition device according to claim 1, further comprising a recognition data storage means for storing the recognition data in a storage device. 6. The character recognition device according to claim 1, further comprising a recognition data storage means for storing recognition data in a storage device, and an image data storage means for storing image data in the storage device. A character recognition device comprising: