JPH0554178A - Character recognizing device and slip for correction - Google Patents

Abstract

PURPOSE:To provide the character recognizing device, which can easily execute correction on a slip even without providing any terminal facility for correction equipped with a display device, and the slip for correction which can efficiently execute the correction. CONSTITUTION:A slip identification part 102 judges whether slip image data inputted from an image input part 101 are image data of a slip for character recognition or image data of the slip for correction, when the input slip is the slip for character recognition, the character recognition is executed by a character recognizing part 103, and the result is stored in a result storage part 104. A slip output part 105 for correction prepares the slip for correction based on the recognized result stored in the result storage part 104. Concerning the slip for correction described confirmation marks or the like and inputted from the image input part 101, correcting contents are read by a correcting slip reading part 106 and the slip is corrected by a correction part 107.

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 and a correction form that can be corrected efficiently.

[0002]

2. Description of the Related Art As a method for confirming and correcting an input character recognized by a conventional character recognition device (hereinafter referred to as correction), for example, a method disclosed in Japanese Patent Application No. 63-049134 "Character Recognition Device" is disclosed. There is a character pattern of the input character to be recognized, a candidate character, or information such as the degree of similarity is displayed on a display device such as a display, the operator visually checks the information, and further operates an input device such as a keyboard. A method of performing corrective work such as inputting correct characters or selecting correct characters from displayed candidate characters has been widely used.

[0003]

However, in the above-mentioned conventional apparatus, since the contents of the displayed information and the correction procedure must be understood, the correction work requires the skill of an operator, and a terminal connected by a line. There was a problem that the correction work could not be performed except around the device.

Particularly when a large amount of forms are recognized by batch processing or the like, it is necessary to perform a large amount of correction work. In the conventional device, however, terminal equipment equipped with a display device and an input device is used for the correction work. Since it is necessary, there is a problem that correction work can not be done by multiple people unless the terminal equipment is expanded.When the terminal equipment is expanded, the system becomes expensive and large-scale, and character recognition is also required. There is a problem that the processing speed of the device is reduced, and it is practically impossible to quickly recognize, correct and input a large amount of data.

The present invention requires the operator to be skilled in the correction work described above, the correction work cannot be performed except in the vicinity of the terminal device, and the large amount of data can be efficiently stored because a large amount of correction work cannot be performed at the same time. In order to eliminate the problem of not being able to enter, it is possible to make corrections on the form without any terminal equipment for correction equipped with a display device, so that anyone can easily make corrections, and at the same time, any person can make corrections. An object of the present invention is to provide a character recognition device capable of quickly and efficiently inputting a large amount of data, and a correction form that enables efficient correction.

[0006]

In order to achieve the above object, a character recognition apparatus according to a first aspect of the present invention includes an image input section for obtaining form image data, and correction of whether the form image data is image data for character recognition form. A form identification section for identifying whether it is image data of a form, a character recognition section for recognizing character patterns on the image data of the character recognition form, and a correction form printed based on the result of character recognition by the character recognition section. The present invention is characterized by comprising a correction form output unit for outputting, a correction form reading unit for reading correction contents on the form from image data of the correction form, and a correction unit for making corrections based on the correction contents.

Further, the correction form of the second invention displays the character pattern for a character for which a single candidate character is obtained as a result of character recognition by the character recognition device of the first invention, and confirms it in the vicinity thereof. A mark entry frame for entering the mark is provided, and for characters for which two or more candidate characters have been obtained, the candidate character pattern is displayed and a mark for selecting one character is entered near each candidate character pattern. It is characterized in that a mark entry frame is provided, and for a rejected character without a candidate character, a character entry frame for re-entering the character is provided near the display symbol indicating that the character is rejected. ..

[0008]

According to the character recognition device of the first aspect of the invention, the form identification unit is provided, and when the input form is the form for character recognition, the character recognition unit performs character recognition, and the correction form output unit performs the result of character recognition. Create a correction form based on. By inputting this correction form, the form identification unit discriminates it as a correction form, and the correction contents are read by the correction form reading unit and corrected by the correction unit. Therefore, even if there is no correction terminal equipment including a display device, it is possible to easily make a correction by writing a mark or the like on the correction form while looking at the recognition result output on the form.

According to the correction form of the second aspect of the invention, a mark entry frame is provided in the vicinity of the candidate character pattern based on the result of the character recognition, in order to determine whether the candidate character is acceptable or not, or to select the candidate character. As for the rejected characters, a character entry frame for re-inputting is provided, so even if there is no terminal device for correction equipped with a display device, you can easily and efficiently see the recognition result output on the correction form. Can be modified.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the character recognition device of the present invention will be described below with reference to FIGS. FIG. 1 is a block diagram showing an embodiment of the character recognition device of the present invention. Character recognition device 1
00 is an image input unit 101, a form identification unit 102, a character recognition unit 103, a result storage unit 104, and a correction form output unit 10
5, correction form reading unit 106, correction unit 107, control unit 1
08 and an output terminal 109.

FIG. 2 is a diagram for explaining the character recognition unit 103 of this embodiment. The character recognition unit 103 uses the line cutout unit 2
01, character cutout unit 202, sub-pattern extraction unit 20
3, a feature extraction unit 204, and a matching unit 205.

FIG. 3 is a diagram showing an example of a character recognition form used in this embodiment. Reference numeral 300 is a character recognition form, 301 is a character entry frame, and 302 is a form identification mark. The form identification mark 302 is the character identification form identification mark 30.
3 and the result correction form identification mark 304, each mark is painted black to indicate that the form is a character recognition form or a result correction form. The form shown in FIG. 3 is a form identification mark 30 for character recognition.
3 is painted in black, indicating that it is a character recognition form.

FIG. 4 is a diagram for explaining the character recognition result.
FIG. 4A is a result candidate character string when the form shown in FIG. 3 is recognized. 401 is a candidate character, 402 is a first-ranked candidate character, 403 is a second-ranked candidate character, and 404 is a third-ranked character. It is a candidate character. 4B is a diagram for explaining the data format when the result candidate character string shown in FIG. 4A is stored in the result storage unit 104, 405 is the result candidate character string data, and the number of candidates is , Each candidate character is stored in a table format as shown in the figure.

FIG. 5 is a diagram showing an example of a result correction form. 500 is a result correction form, 501 is a result file identification number, 502 is a form identification mark, 505 is a recognition result character, and 506 is a recognition result. A candidate character, 507 is a candidate character selection mark entry frame, 508 is a recognition result confirmation mark entry frame, and 509 is a reject character re-input character entry frame. The form identification mark 502 includes a character recognition form identification mark 503 and a result correction form identification mark 504.
Each mark is painted black to indicate that the form is a character recognition form or a result correction form. In the form shown in FIG. 5, the result correction form identification mark 504 is painted in black, which indicates that the form is a result correction form.

The operation of the character recognition apparatus of this embodiment will be described in detail below with reference to FIGS. Image input unit 101
Photoelectrically converts an optical signal S from a form in which characters, figures, symbols, etc. (hereinafter, referred to as characters) are described, and for example, a black and white binary value in which a character line part is represented by black pixels and a background part is represented by white pixels. A quantized electric signal (hereinafter referred to as “form image data”) is generated and output to the form identification unit 102.

The form identification unit 102 determines the type of form from the form image data input from the image input unit 101, and based on the result of the determination, the form image data is recognized by the character recognition unit 1.
03 or to the corrected form reading unit 106.

In the present embodiment, the type of the form is judged as follows.
Form identification mark (302, 5
02) is carried out as follows. That is, from the form image data, the form identification mark for character recognition (303, 50
3) and the form identification mark for correcting results (304, 504)
The number of black pixels in each area is counted and the number of black pixels in the character recognition form identification mark area (Bm) is calculated as the number of black pixels in the result correction form identification mark area (Bs). If it is above (Bm ≧ Bs), it is determined that the form is a character recognition form, and the number of black pixels (Bm) in the character recognition form identification mark area is within the result correction form identification mark area. When the number is smaller than the number of black pixels (Bs) (Bm <Bs), it is determined that the form is a result correction form.

If the form is determined to be a character recognition form, the form image data is output to the character recognition unit 103, and if the form is determined to be a correction form, the form image data is modified form. It is output to the reading unit 106.

The character recognition section 103 cuts out character pattern data for each character from the character recognition form image data input from the form identification section 102, and performs recognition processing on the cut out character pattern data. 1 for each character
The above candidate character names are obtained, and the candidate character names are stored in the result storage unit 10.
4 and can be performed by any of various conventionally well-known methods. In the present embodiment, the character recognition unit configured as shown in FIG. 2 performs as follows. ..

That is, first, in the line cutout unit 201, character line data is extracted from the character recognition form image data input from the form identification unit 102, and the character cutout unit 20 is extracted.
2 is output sequentially.

The character slicing unit 202 extracts character patterns for each character from the character line data input from the line slicing unit 201 and stores the character patterns in a pattern register (not shown).

The sub-pattern extraction unit 203 scans the character patterns in the pattern register in a plurality of directions,
A predetermined specific value h on each scanning line (in the present embodiment, h =
5) The continuous black pixel row is detected, and the continuous black pixel row is extracted as the black pixel component of the sub-pattern, thereby extracting the sub-pattern for each scanning direction from the character pattern and sequentially characterizing the sub-pattern. It is output to the extraction unit 204.

In the present embodiment, the scanning direction is a direction (vertical direction) perpendicular to the character line direction (hereinafter, X-axis direction), a parallel direction (horizontal direction), and 45 ° counterclockwise from the X-axis. The direction (the left diagonal direction) and the direction of 45 ° clockwise (the left diagonal direction) are set, and character patterns are scanned in each of these directions to extract four sub patterns for each direction.

For example, in the extraction of a sub pattern in the vertical direction, the vertical direction is the main scanning direction, continuous black pixels (black runs) are detected on the vertical scanning line of the pattern register, and 1 ≧ h
The vertical sub-pattern is extracted by extracting a black run having a length of 1 as the black pixel portion of the vertical sub-pattern. Similar to the vertical sub-pattern extraction, the sub-pattern extraction is performed when the remaining other direction is the main scanning direction. The extracted sub patterns are sequentially extracted by the feature extraction unit 20.
4 is output.

In the feature extraction unit 204, a rectangular region corresponding to the character circumscribing frame of the character pattern is set on the extracted sub-patterns in each direction, and N × M rectangular regions (N and M are arbitrary). It is divided into small regions of a suitable natural number), the feature amount representing the length of the character line of each sub-pattern included in each small region is extracted, and the feature amount is normalized by the size of the character circumscribing frame and normalized. Feature matrix F consisting of the feature quantities f _i
Is generated and the feature matrix F is output to the matching unit 205.

In this embodiment, the number of divisions N and M = 8.
Further, the normalization of the feature amount with the size of the character circumscribing frame is performed with a value of (dX + dY) / 2.
However, dX is the horizontal length of the character circumscribing frame, and dY is the vertical length of the character circumscribing frame. The feature quantity f _i is
The number i (i = 1, 2, ..., 1) from 1 to N × M is assigned to each small area.
(N × M) are sequentially added to represent small areas, the number i
Represents the feature amount of the small area and is an element value of the feature matrix F.

The collating unit 205 uses the feature matrix F as a feature dictionary matrix G of a standard pattern prepared in advance (not shown).
The similarity R represented by the following equation (1) is obtained, the character names of the dictionary matrix whose similarity R is equal to or more than a predetermined value P are used as candidate character names, and the first similarity is calculated in descending order. The rank candidate character and the second rank candidate character are ranked, and the candidate character string is output as a recognition result character string.

[0028]

[Equation 1]

However, g _i represents an element of the dictionary matrix. When the recognition process as described above is performed on the character recognition form as shown in FIG. 3, a candidate character string as shown in FIG. 4A is obtained.

The result storage unit 104 is a means for storing the result candidate character string input from the character recognition unit 103, and can be easily realized by various recording media such as an IC memory, a floppy disk and a hard disk. The result candidate character string as shown in FIG. 4 (A) is stored in a table format as shown in FIG. 4 (B). The information stored in the result storage unit is not limited to the example shown in FIG. 4B, and information such as similarity may be added. In addition, the result storage unit 1
An identification number 04 is given to each recognition result in order to match the stored recognition result with the correction information that may be input from the correction unit later, and the recognition result is stored together with the identification number. ..

The correction form output unit 105 includes a result storage unit 1
The recognition result candidate character string stored in 04 is printed out on the correction form, and the format of the printed correction form is output to a format storage unit (not shown) in the correction form reading unit 106 described later. .. The output format of the recognition result may be any suitable format, but for example, it is assumed that the output is performed in a format similar to a form for character recognition, and the character pattern is selected for a character that can be determined as one candidate character as a result of recognition. Display and put a confirmation mark in the vicinity of the candidate character. If there are two or more candidate characters, display the candidate character pattern and put a mark for selecting one character in the vicinity of each candidate character. It is preferable that the characters are rejected because there is no candidate character, and the characters are re-entered in the vicinity of the display symbol indicating that the character is rejected. Therefore, in this embodiment, when the character recognition form as shown in FIG. 3 is input and the recognition result as shown in FIG. 4 is obtained, the correction form is output in the format as shown in FIG. 5, for example. ..

In FIG. 5, reference numeral 501 denotes the result storage unit 104.
Is an identification number of the recognition result of the form stored in the table, and 502, 503, and 504 are identification marks indicating that the form is a result correction form as described above. 5
05 is the character of the recognition result, "8", "child", "high",
Since "tail" and the like have only one candidate character, the candidate character is printed as it is. 506 has three candidate characters of "king", "ball", and "sat", so the candidate character and the mark entry frame 507 for selection are printed, and the correct character can be selected by entering the mark. It is a thing. Also,
Since all the characters in the field are recognized only by the first candidate character, reference numeral 508 is a mark entry frame for confirming whether or not these are finally correct. Furthermore, 50
Reference numeral 9 is a character entry frame for recognizing again because it is rejected as unrecognizable because there is no candidate character for the character "8" indicated by "●".

As described above, the form as shown in FIG. 5 is prepared as the correction form, but it is not necessary to use all the various correction methods described above, and for example, only the confirmation mark entry frame is used to confirm the recognition result. There is no problem even if it is a form for confirmation, a form for only selecting candidates, a form only for re-inputting reject characters, or a form combining them.

In the correction form reading unit 106, the format of the correction form stored in the format storage unit in the correction form reading unit 106 (not shown) is determined from the correction form image data input from the form identification unit 102. With reference to this, the correction content for the recognition result stored in the result storage unit 104 is read, and the read result is corrected by the correction unit 107.
Output to. In this case, using the recognition result number (501 in the example of FIG. 5) printed by the correction form output unit 105, the input correction form is associated with the format information and the recognition result of the form. .. Further, when the character pattern is re-input for the rejected character, the character name of the recognition result is obtained after performing the recognition processing in the character recognition unit in the correction form reading unit 106 (not shown) in this embodiment. Is output to the correction unit 107 as the correction content. In addition,
The mark detecting method and the character recognizing method in the modified form reading unit 106 can be realized by the same methods as the mark detecting method in the form identifying unit 102 and the character recognizing method in the character recognizing unit 103, respectively.

From the result correction form as shown in FIG. 5, a candidate selection mark 507, a recognition result confirmation mark 508,
The re-input character 509 is extracted, the re-input character pattern is subjected to recognition processing to obtain the character name of the recognition result, and the extracted mark and recognition result character name are collated with the correction form format information, thereby the mark and character. The pattern is associated with the recognition result in the recognition result storage unit, and the correction content is output to the correction unit 107.

The correction unit 107 is a correction form reading unit 106.
The result storage unit 104 is based on the correction content input from
Make corrections to the results stored in. For example, FIG. 4 (B)
In the recognition result shown in, the second character "king",
“Oh” is selected from the candidate characters “ball” and “Sat” based on the candidate character selection mark 507 of the correction form in FIG. 5,
As for the rejected “8” character, the recognition result of the newly input character pattern entered in the character entry box 509 of FIG. 5 is stored. Further, when there is a recognition result to which the recognition result confirmation mark is not added, the correction form output unit 105 creates a correction form again for the recognition result and prompts the input of a new character pattern. In this way, the process of setting the number of candidates in the recognition result storage unit to 0 is performed.

The control section 108 controls each section of the character recognition device 100, controls the operation of each section and data synchronization, and controls the interface with the outside through a control signal line (not shown).

The output terminal 109 is a data output terminal for outputting the recognition result to the outside, and is connected to another system, a medium for recording the recognition result, a communication network, another information processing system or the like. ..

The present invention is not limited to the above-mentioned embodiment, and the operation of each component, the processing method, the flow of input / output signals, the number of arranged, the position, the shape, the number and other conditions are set. It can be changed arbitrarily.

[0040]

As described above in detail, according to the character recognition device of the first invention, a correction form is created based on the result of character recognition, and a mark or the like is written on the correction form. The correction contents are read by and the correction is performed. Therefore, even if there is no terminal device for correction equipped with a display device, it is possible to easily make corrections on the correction form, and at the same time, it is possible for a large number of people to make corrections and a large amount of data can be input. The device can be realized.

According to the correction form of the second aspect of the invention, a mark entry frame is provided in the vicinity of the candidate character pattern based on the result of the character recognition, and a mark entry frame for determining whether the candidate character is acceptable or not, or for selecting the candidate character. With regard to rejected characters, a character entry frame for re-inputting is provided, so even if there is no correction terminal equipment equipped with a display device, it is easy to enter marks while watching the recognition result output on the form. The work allows anyone to easily and efficiently make corrections.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a configuration of a character recognition unit of the present invention.

FIG. 3 is a diagram showing an example of a character recognition form.

FIG. 4 is an explanatory diagram of a character recognition result.

FIG. 5 is a diagram showing an example of a correction form.

[Explanation of symbols]

 101 image input unit 102 form identification unit 103 character recognition unit 104 result storage unit 105 correction form output unit 106 correction form reading unit 107 correction unit 108 control unit 109 output terminal

Claims (2)

[Claims] 1. An image input unit for obtaining form image data, a form identifying unit for identifying whether the form image data is image data for character recognition form or image data for correction form, and image data for character recognition form. A character recognition unit for recognizing the above character pattern, a correction form output unit for printing and outputting a correction form based on the character recognition result in the character recognition unit, and correction on the form from the image data of the correction form. A character recognition device comprising: a correction form reading unit for reading the contents, and a correction unit for making a correction based on the correction contents. 2. A mark for displaying a character pattern for a character for which a single candidate character is obtained as a result of character recognition by the character recognition device according to claim 1 and for writing a confirmation mark in the vicinity thereof. A mark box is provided for characters for which two or more candidate characters have been obtained, and a mark entry frame is provided in the vicinity of each candidate character pattern to mark a character for selecting one character. For a rejected character without a character, a correction form is provided in which a character entry frame for re-entering the character is provided near a display symbol indicating that the character has been rejected.