JPH0362280A - Character reading device - Google Patents

Abstract

PURPOSE:To correct the misrecognized areas of a document with high efficiency by deciding these misrecognized areas based on the sum total of recognizing results of characters included in a prescribed range of the document written into a business form as well as the numerical information written into another corresponding business form and displaying these areas. CONSTITUTION:A check sum check part 12 adds the rows and the columns of a document respectively via a calculation part 13 based on the recognizing result (character code) received from a reading part 11. When the sum of recognizing results (character codes) of rows and columns is calculated, the part 13 transfers the recognizing results of the check sums corresponding to the rows and columns respectively as well as the calculation data to a deciding part 14. The part 14 decides the mixture of a misrecognized character if no coincidence is obtained between th two lower rank digits of the calculation data and the recognizing result of the check sum. Then a display part 15 display the reading result as a corrected screen. Thus a read document can be easily corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION Object of the Invention 1 (Field of Industrial Application) The present invention relates to a character reading device for reading documents written on forms.

(Prior Art) Generally, a character reading device (OCR) is used to read general documents (characters) written on a form. When reading general documents using a character reading device, the positions of figures and characters written on a form are manually determined. That is, the line where characters are written in the form and each character are automatically detected, and recognition processing is performed for each detected character.

This recognition processing is combined with recognition technology for many character types, including Kanji, to realize a general-purpose Japanese document reader.

(Problem to be solved by the invention' There is.

Specifically, corrections were confirmed by comparing the display of the results read by the document reader with the document written on the form that was actually read. For this reason, there is a problem in that the efficiency of correcting the read result is low and a lot of time is required.

The present invention has been made in view of the above-mentioned points, and an object of the present invention is to provide a character reading device that can efficiently correct the result of reading a document.

[Structure of the Invention] (Means for Solving the Problems) The present invention provides information on the document and the numerical information from a document and a form in which numerical information corresponding to a group of characters included in a predetermined range of the document is entered. a reading means for reading, a calculating means for calculating the sum of character codes of characters included in the predetermined range of the document read by the reading means, a sum of the character codes calculated by the calculating means and the reading means; a determining means for determining whether or not the document read by the reading means has been correctly read based on the numerical information read by the reading means; In this case, the display means is provided with a display means for displaying a display indicating this fact.

The present invention also provides a reading means for reading the document and the numerical information from a document and a form in which numerical information corresponding to character groups included in each row and each column of the document is written; a calculation means for calculating the sum of character codes of characters included in each row and each column of the document, based on the sum of character codes calculated by the calculation means and numerical information read by the reading means; a determining means for determining whether or not the document read by the reading means has been correctly read for each row and each column; and one row and column determined by the determining means to have not been correctly read. and display means for displaying characters that have not been correctly read based on the position of the characters.

(Function) In the character reading device configured in this way, since the correction location is automatically displayed, it becomes possible to easily correct the read document.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a character reading device according to the same embodiment. In the figure, a reading unit 11 detects character information such as a document written on a form. A checksum check section 12 is connected to the reading section 11 to detect errors in the document read by the reading section 11. Check sum check section 12
The system is comprised of a calculation unit 13 that calculates the sum of character codes of read characters, and a determination unit 14 that determines the location of an error based on the sum value and checksum obtained in the calculation unit 18. Further, a display section 15 is connected to the checksum check section 12 for displaying the read result according to the checksum check.

FIG. 2 is a diagram showing an example of a method of filling in a form to be processed in the same embodiment. In the example shown in FIG. 2, a checksum (numerical information) corresponding to a general document is written in each row and each column of the same document. In the figure, the checksum corresponding to each row is indicated as "xx", and the checksum corresponding to each column is indicated as rYYJ. Also, the checksum (XX, YY) is for each row and column (
This indicates the last two digits of the total of the character codes of characters included in the predetermined range.

Next, the operation of this embodiment will be explained.

First, regarding a form as shown in FIG.
performs character reading processing. The reading unit 11 is a general-purpose Japanese document reader, and the document size is not limited. The reading unit 11 detects a form image including an image of characters written on the form by optically scanning the form. The reading unit 11 detects the position of each line of the document from the detected form image, and further detects each character included in each line. Then, a character pattern is cut out for each character. The character patterns thus cut out are each subjected to recognition processing. Here, general documents (characters) and checksums (numbers) are recognized in the same way.

However, the recognition results (character codes) are classified into documents and checksums. The reading section 11 transfers the recognition result to the checksum checking section 12.

In the checksum/check section 12, the calculation section 13 performs addition calculations on the recognition results (character codes) from the reading section 11 for each row and each column of the document.

When the sum of the recognition results (character codes) for each row and each column is calculated, the calculation unit 13 uses this calculation data and the recognition result for the checksum corresponding to each row and each column to the discrimination unit 14.
Transfer to.

The determination unit 14 performs a determination process to determine whether or not the read document has been correctly read based on each piece of information from the calculation unit 13 . That is, the last two digits of the calculated data for each row and each column are compared with the recognition results for the corresponding checksums. Here, if the last two digits of the calculated data and the recognition result of the checksum match (match), it is assumed that the document has been read correctly; if they do not match (inconsistency), the erroneously recognized characters are It is determined that it is included. When the discrimination processing for each row and each column of the read document is completed, the discrimination section 14 notifies the display section 15 of the discrimination results.

The display unit 15 displays the reading result of the document written in the form according to the determination result, and uses this as a correction screen. In this case, the display unit 15 displays the row and column that is determined to include the erroneously recognized character (inconsistency) using double-brightness display or the like. Also, row 1 that is inconsistent
The character at the intersection of the columns is treated as a reject and is displayed, for example, as shown in FIG. 3, replacing it with a symbol "?" indicating a reject. Further, on this correction screen, control is performed so that the cursor moves only to the reject symbol.

In this way, the location to be corrected can be easily determined from the display screen. In addition, when correcting rejected characters by inputting correct character data via a human power section (not shown) (consisting of a keyboard, etc.), it is possible to edit the rejected characters efficiently. Become. Further, when there is one character that is misrecognized, only that character is pointed out, but even when there are multiple characters, it is possible to point out multiple candidates for the misrecognized character.

This makes it possible to significantly improve the efficiency of corrections, compared to performing corrections while displaying the read results and comparing them with the document.

Such a method can be applied to all forms in which a checksum can be written on the form read by the reading section 11.

Further, as a specific usage method, if it is applied to a test sheet when calculating the recognition rate of a character reading device, it is possible to reduce the effort required to create master data.

In the above embodiment, a case was described in which a Japanese document was entered in a form, but it is of course applicable to numbers, alphabets, kana, etc.

Further, although the checksum (numerical information) corresponding to each row and each column is entered in the form in advance, the present invention is not limited to this.

Furthermore, various methods can be considered for calculating the last two digits of the sum of character codes in each row and each column.

[Effects of the Invention] As described above, according to the present invention, the total sum of recognition results (character codes) of characters included in a predetermined range (row, column, etc.) of a document entered in a form and the corresponding Erroneous recognition points are identified based on the numerical information (checksum) also entered in the form, and the points to be corrected are displayed on the correction screen so that the points to be corrected can be easily seen, making it possible to efficiently correct the reading results of the document. It is possible to do so.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a character reading device according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a method of filling in a form to be processed in the same embodiment, and FIG. FIG. 3 is a diagram for explaining an example of a display screen in an embodiment. 11...Reading section (reading means) 12...Checksum check section, 13...Calculation section (calculation means)
, 14... Discrimination section (discrimination means) 15... Display section (display means).

Claims (2)

[Claims] (1) A reading means for reading the document and the numerical information from a document and a form in which numerical information corresponding to a group of characters included in a predetermined range of the document is written; and a document read by the reading means. a calculating means for calculating the sum of character codes of characters included in the predetermined range of the character code; a discriminating means for discriminating whether or not the document read by the discriminating means has been read correctly; and a display means for displaying an indication to that effect when the discriminating means determines that the document has not been read correctly; A character reading device characterized by comprising: (2) a reading means for reading the document and the numerical information from a document and a form in which numerical information corresponding to each character group included in each row and each column of the document is written; and a reading means for reading the document and the numerical information; a calculating means for calculating the sum of character codes of characters included in each row and each column of the document, based on the sum of character codes calculated by the calculating means and numerical information read by the reading means, determining means for determining whether or not the document read by the reading means has been correctly read for each row and each column; and the position of the row or column determined by the determining means to have not been correctly read. A character reading device comprising: display means for displaying characters that have not been correctly read based on the following.