JPS6039276A - Optical character reader - Google Patents

Abstract

PURPOSE:To attain correction by the method executed for a document or the like by reading a line segment overlapped on a described character, recognizing it as a deletion symbol for the described character and processing a character above or under the column as a character to be corrected. CONSTITUTION:When the character position detecting section 9 of a character recognizing section 4 detects the description of a character, a deletion discriminating section 10 detects the presence or absence of a horizontal line segment representing deletion and when the line exists, the section 10 outputs the presence of information relating to the deletion to a deletion buffer 11 to store the information. Then the address calculation required for recognizing the character for correction described at the outside of a column of a character frame upper part is conducted by an address calculation section 12 for detection of described character to be corrected and the information is returned to a pre-processing control section 8, where the pre-processing and positioning processing is conducted with the corrected character. Then the recognition is processed by a characteristic extracting section 13, a discriminating section 14 and a dictionary section 19 or the like, the information from the buffer 11 is referenced at the edition of output at an output edition section 15 and the result of recognition of the corrected input character is outputted to an output buffer 20 to store it.

Description

Detailed Description of the Invention (a) Technical Field of the Invention The present invention relates to an optical character reading device that optically reads characters written within a character frame of a form, and particularly relates to an optical character reading device that optically reads characters written within a character frame of a form. The cancellation of 1'+1 is cut off.

This invention relates to an optical character reading device that reads corrected characters written above or below a character frame.

(b) Technical background When reading handwritten characters optically, numbers, numbers,
The reading of alphabetic and kana characters is now commonly put into practical use. Furthermore, the difficulty of recognizing handwritten characters varies greatly depending on how much the degree of deformation of the characters is restricted. Currently, the commonly used method is to write characters in a square box according to certain rules.

However, both the method of entering handwritten characters and the method of correcting already written characters must be corrected in a manner that follows one rule that is easy for machines to process.

There is a sense of incompatibility with the method of correcting entries on general forms that is already accepted in society, and there is a strong desire from users to use the method of correcting entries on forms that are already accepted as much as possible. be.

(c) Prior Art and Problems As a method for correcting and IFing characters entered in a form with a character frame, an example of a method that has been put into practical use in the past will be explained with reference to the drawings.

FIG. 1 shows a correction entry side that has been generally put into practical use.

In the figure, 1 is a frame for writing handwritten characters, 2 is a correction frame, and 3 is a frame for writing handwritten characters.
indicates the correction character entry frame.

Conventionally, when you want to correct a character written in character entry frame 1, enter a predetermined symbol (× in this example) in correction frame 2, and then enter the corrected character in correction character entry frame 3 located at the bottom. When reading, read the characters in handwritten character entry frame 1, then read the correction frame 2, and if there is a correction symbol (× in this example), the characters in correction character entry frame 3 are corrected. Read it as text.

” - In the method described, it is necessary to change not only the characters you want to correct but also all the letters you have entered, and it is not necessary to change the characters you want to correct with two lines and write the corrected characters in the column above. There is a problem with deceiving that a certain method cannot be implemented.

(d) Purpose of the Invention The present invention is a novel optical character NFT which eliminates the drawbacks mentioned above.
The purpose of this system is to provide a device that can read and recognize characters that have been corrected using the method commonly used for forms, in which characters to be corrected are crossed out with two lines and 81 regular characters are entered in the field above. The object of the present invention is to realize an optical character reading device that can shift the characters.

(e) Structure of the Invention The present invention is an optical character reading device that optically reads characters written in a character frame of a form, and recognizes line segments that overlap the characters written in the character frame of the form. and a means for determining that a written character whose line segments overlap is a cancellation, and a means for reading a character written above or below the character frame of the canceled character as a corrected character, and a corrected final entry. A means to recognize and output characters is provided and is implemented for general forms. This can be achieved by an optical character reading device that is characterized in that it can read and recognize characters that have been corrected by erasing the characters to be corrected, for example, with two lines.

(f) Examples of the invention The present invention will be explained below with reference to the drawings.

FIG. 2 shows an embodiment of the optical character reading device according to the present invention, and FIG. 3 shows an example of the form taken by 7'e in FIG.

(A) indicates the input form, and (+3) indicates the output result.

In the figure, 4 is a character recognition unit, 5 is a Blε cutting frame information storage unit, 6 is a character recognition control unit, 7 is a video extraction address calculation unit, 8 is a preprocessing control unit, 9 is a character position detection unit, 10 is a cancellation determination unit, 116J is a cancellation information storage temporary buffer, 12 is an address calculation unit for detecting corrected characters, 1
3 is a feature extraction unit, 14 is a determination unit, 15 is an output w unit, 1
6 is the image buffer (1), 17 is the preprocessing section, 18 is the image 1 image buffer (2), 19 is the dictionary section, 20 is the output buffer, 21 is the entry frame (1), 22 is the entry frame (2) ), ■
indicates output (1) and ■ indicates output (2), respectively.

An example of the optical character reading device according to the present invention is a video recorder with 10 copies (
A character recognition unit 4 recognizes and identifies entered characters using the following devices (devices with symbols 5 to 20) from the image information transferred from a field (generally, characters that can be read on a form are entered). i'43 is called a field) Address specification 1 Character frame height specification etc. Reading format information storage section 5 which stores information regarding the format of the readable eye tag, reading format Ili 9
1j The character recognition control unit 6 controls the rest of the character recognition unit 4 according to the information from the storage unit 5, and the position information of two characters among the information from the reading format information storage unit 4 (field 1 address designation 5 character frame iRr) (1) Give the address information necessary to extract the video to be recognized from the video information stored in 16 to the preprocessing control unit 8. Video extraction address calculation unit 7
, a preprocessing control unit 8 that controls the preprocessing unit 17 using the address information from the video extraction address calculation unit 7; positioning (move the reading position a little downward, left and right to ensure proper positioning, to avoid misrecognizing written characters due to deviations from the predetermined position or protrusion from the character frame, etc.).

A character position detection unit 9 detects the assumed position of a character frame that is not stored as information and whether or not a character has been entered, and detects if there is a horizontal line indicating cancellation in the entered character part detected by the character position detection unit 9. Cancellation determination unit 10, Cancellation determination unit 10
A cancellation information storage buffer 11 stores the information upon receiving an output signal indicating that the written character portion determined in step 9 has been canceled; (Address calculation unit 12 for detecting correction entry characters that calculates the address necessary to recognize the entry characters for correction expected outside the cabinet) at the top of the character frame;
Video buffer (2) Input characters output from 18 and 5
A feature extraction unit 13 extracts the features of the written characters based on the information from the cancellation determination unit 10, and features of the written characters output from the '11r feature extraction unit 13.

Registered in the dictionary section 19 1. □□□The judgment unit 14, which recognizes the entered characters by comparing it with the information in
< 14, the ratio/correction is judged and the final output information (output result ■ or ■), the output editing section 15 edits the video record tQ section (
(not shown) Iff, I+'7 report as sentence -j・-
Receipt record 1. Needle J-ru video buffer (1
) 16, Video buffer (1) Cut out the part to be recognized from the video information stored in 16, and video buffer (2) Part 53 that seems to be noise or dirt on the video information cut out to obtain it in 18. (1) A video buffer that stores the information obtained by preprocessing the part of the information in 16 that is to be recognized by the preprocessing unit 17. (2) Information input to the determination unit 14 (18) It consists of a dictionary section 19 that stores and registers information for comparing and determining the characteristics of characters, and an output buffer 20 that stores recognized character information for sending it to the next circuit.

The operation of this embodiment will be explained using FIGS. 2 and 3.

The manual form characters in entry frame (2) 22 are scanned by a scanning mechanism section (not shown) and the information is stored in the video buffer (1) 16 via the video recorder 10 section (not shown).

On the other hand, it is possible to read the information stored in RIX
Under the control of the character recognition control section 6, which controls the entire character recognition processing process for form-1- in the two-character recognition section 4 according to the information from the tJd address calculation section 5, Referring to the information from
71 pulse information necessary for cutting out a video to be recognized from among the videos stored in the video buffer (1) 16 is calculated and sent to the preprocessing control unit 8.

The preprocessing section 17 is under the control of the preprocessing control section 8.

Using the address information necessary to cut out the video that is the recognition target, cut out the video that is the recognition target from the video information stored in the video buffer (2). ) to output to 1B.

Unnecessary parts such as dust and noise on the cut out video information (obtained) are removed, and the cut out video information as a recognition target is stored in the video buffer (2) 18.

Next, the character position detection unit 9 performs positioning for recognizing the -C character for the video information written in the video buffer (2) 1B, and determines the expected position of one character frame and whether or not a character is written in the character frame. The information is detected and output to the cancellation determination unit 10.

When the character position detection unit 9 detects that a character has been written, the cancellation determination unit IO detects the presence or absence of a horizontal line segment indicating cancellation, and selects the corresponding number (number 3 in entry frame (2) 22 in FIG. 3). ,4... Make me remember.

Next, enter the characters (number 9) for corrections in the margin at the top of the character frame.
, 8.7) is performed by the address calculation unit 12 for detecting corrected characters, and the 1η+U
is returned to the preprocessing control unit 8, and the corrected characters are preprocessed.

After performing the positioning process, the feature extraction unit 13, '1'
+1 constant part 14. The dictionary section 19 etc. performs recognition processing, and the output editing section 15 outputs the cancellation information (1, 0/\Nov) during output editing.
Referring to the information from 11, the recognition result of the corrected input character (output (2) ■) is output to the output/nofa 20 and delivered to each station.

(g) Effects of the invention As described above, the character to be corrected spans the left and right ends of the character frame, and the characters are read with multiple horizontal line segments, which are cities, and are recognized as the cancellation symbol for the written characters. By correcting the character (m) above as a corrected character and canceling it, it is possible to read the corrected character using the method used for general forms and correct the 8J character. This has the effect of providing a reading device.

[Brief explanation of the drawing]

FIG. 1 shows an example of correction entry that has been generally put into practical use in the past, and FIG. 2 shows an example of an optical character reading device according to the present invention.
FIG. 3 shows examples of forms that can be read in FIG. 2. In the figure, 1 is a hand 71 character entry frame, 2 is a correction frame, 3 is a correction sentence large frame, 4 is a character recognition unit, 5 is a readable human information storage unit, 6 is a character recognition control unit, and 7 is for video clipping. Address calculation section, 8 is preprocessing control section, 9 is character position detection section, 10 is cancellation determination section, 11 is cancellation information storage 1g buffer, 12 is address calculation n15 for Genshoki character detection
．． 13 is a feature extraction section, 14 is a foot section, 15 is an output editing section, 16 is a video buffer (IL, 17 is a preprocessing section, 18 is a video buffer (2), 19 is a dictionary section, and 20 is an output section. Ka＼Sofa, 21 is entry space (1) 12 is entry space (2)
, ■ indicates the output (1), and ■ indicates the output (2), respectively. (company) “ )

Claims (1)

[Scope of Claims] An optical character reading device that optically reads characters written in a character frame of a form, and recognizes line segments that overlap the characters written in the character frame of the form. i: means for determining that a written character overlapping the i-marked line segment is a cancellation, and reading a character written above or below the character frame of the cancellation character as a correction character;
An optical character reading device having a 94 feature of being provided with means for recognizing and outputting corrected final written characters.