JPH02220186A - Character reader - Google Patents

Abstract

PURPOSE:To reduce the number of times of correction by means of an operator when similarly deformed characters cannot be read for many times by correcting the similar illegible characters in a lump. CONSTITUTION:A correcting part 31 displays a character image described in a character frame 2 as the character image of the illegible character on a CRT 61, inputs a corrected character code '8' from an operator, compares characteristics data of the illegible characters stored and displayed on a memory 42 with the characteristic data of the characters in character frames 6, 9 and 11 as the other illegible characters in the same slip, decides whether or not they are the similar characters, and obtains a result that the all characters described in the character frames 6, 9 and 11 are similar to the character described in the character frame 2. The corrected part 31 collectively replaces the read result of the characters described in the character frames 2, 6, 9 and 11 stored into the memory 42, obtains the correction result, outputs the correction result to a floppy disk driver 81, and then outputs it to a floppy disk. Thus when many similar deformed characters are generated, the number of times of the correction by the operator can be reduced.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a character reading device, and more particularly to a method for correcting unreadable characters. [Prior Art] In conventional character reading devices, when correcting unreadable characters, an operator has to input character codes for all unreadable characters one by one. [Problems to be Solved by the Invention] In the conventional character reading device described above, when correcting unreadable characters, the operator had to input character codes for all unreadable characters one by one. There is a drawback that when a large number of similar deformed characters are misread, the number of corrections made by the operator increases accordingly. For example, when the form shown in Figure 2 is read and the judgment result shown in Table 1 is obtained, the operator selects the character frames ■ and ■. Similar correction operations were required four times for the four illegible characters with similar glyph shapes written in ■ and ■. Table 1 [Means for solving the problem] The character reading device of the present invention has an image memory and a document.
means for scanning each sheet and storing image data in the image memory; a recognition result storage memory; and a means for reading out the image data stored in the image memory, cutting out and recognizing characters one by one, and outputting recognized characters as character recognition results. A means for storing a code or a predefined unreadable character code and characteristic data of unreadable characters obtained upon recognition in a recognition result storage memory, and a modified character code for characters that are unreadable by an operator. a correction means for replacing an unreadable character code in the recognition result with the character code input by the correction character code input means; and correction means for referring to the recognition result memory and correcting it by the correction means means for searching for similar unreadable characters having similar characteristics to the unreadable characters, means for collectively replacing unreadable character codes of the similar unreadable characters with the specified character codes, and means for replacing the unreadable characters by the collective replacing means. and a cup means for outputting the recognized recognition results. [Function] Since similar illegible characters are corrected all at once, the number of corrections by the operator is reduced when a large number of similar deformed characters are illegible. [Example] Next, an example of the present invention will be described with reference to the drawings. FIG. 1 is a schematic diagram of a character reading device according to an embodiment of the present invention. The image input unit 11 operates the image data of the form for each form and inputs the image data to the memory 41 via the image data bus 51.
Store in. The recognition unit 21 reads the image data stored in the memory 41 via the bus 51, cuts out and recognizes characters one by one. At this time, if one candidate character is determined as a character recognition result, the corresponding character code is used, and if the other character is unreadable, the predefined unreadable character code and the corresponding unreadable character code obtained along with the recognition. The characteristic data of the characters and the character images of the illegible characters are stored in the memory 42 via the bus 52. The correction section 31 has a keyboard 71 and a CRT 61 for inputting the code of the character to be corrected by the operator, and a floppy disk driver 81 for outputting the reading result.
is provided. The correction unit 31 displays the image of the unreadable characters stored in the memory 42 on the CRT 61, and inputs a correction character code from the operator. At this time, the feature data of the displayed unreadable character stored in the memory 42 is searched and compared with the feature data of other unreadable characters in the form, and an unreadable character similar to the displayed unreadable character is found. Determine whether or not. If there are no similar unreadable characters, operate the recognition results stored in the menu 42 only for the displayed unreadable characters, or if similar unreadable characters exist, operate the recognition results stored in the menu 42 for the displayed unreadable characters and all similar unreadable characters. Replace with the character code input by the user. The correction unit 31 operates every time the recognition unit 21 completes processing one form, and outputs the recognition result of the corrected word for each form to the floppy disk driver 81, thereby outputting it to a floppy disk. Next, the operation of the apparatus of this embodiment will be described in the case where the form shown in FIG. 2 is input. The image input section 11 scans the form shown in FIG. 2, inputs image data, and stores it in the memory 41.
Store in. The recognition unit 21 sequentially extracts and recognizes the character images written in the character frames ■ to ■ from the image data of the form stored in the memory 41, and as a result, Table 1 is obtained.
Obtain the reading result. At this time, the character frame ■, ■, ■, ■. For the characters written in ■, ■, [phase], the character code corresponding to the reading result, and the characteristic data of the character written in the character frame ■, ■, ■, ■ whose recognition result was illegible. , a character image, and a predefined unreadable character code are stored in the memory 42 . The correction unit 31 displays the character image written in the character frame ■ as a character image of the unreadable character on the CRT 61, inputs the correction character code "8" from the operator, and at the same time inputs the character image of the displayed unreadable character stored in the memory 42. Same as the characteristic data - Compare the characteristic data of the characters in the character boxes ■, ■, ■ as other unreadable characters in the form to determine whether they are similar characters, and then compare the characteristic data of the characters written in the character frames ■, ■, ■. It is concluded that all the characters are similar to the characters written in the character box ■. Therefore, the correction unit 31 edits the character frames ■ and ■ stored in the memory 42. The reading results of the characters written in (1) and (2) are replaced all at once to obtain the correction results shown in Table 2, and the correction results are output to the floppy disk driver 81 to be output to the floppy disk. Table 2 As described above, in the case of this embodiment, the operator can correct four unreadable characters in one correction operation and output the result. In this embodiment, the similar character search range of the correction unit 31 is within the same form, but the search range may be expanded between multiple forms in the case of a device used in an operation where the same person fills in multiple forms. It is clear that this is possible.

【Effect of the invention】

As explained above, by collectively correcting similar illegible characters, the present invention has the effect of reducing the number of corrections made by the operator when a large number of similar deformed characters are illegible. In cases where it is not possible to provide sufficient training to operators on character entry, for example, when reading forms that have been filled out by an unspecified number of people and have many curly or distorted characters, the number of times operators must correct unreadable characters may be significantly reduced. There are effects that can be reduced.

[Brief explanation of the drawing]

FIG. 1 is a schematic block diagram of a character reading device showing an embodiment of the present invention, and FIG. 2 is a diagram showing an example of a form read by the character reading device. 11... Image input section, 21...
...Recognition unit, 31...Modification section, 41.42--Memory, 51, 52...Bus, 61... ...CRT. 71...Keyboard, 81...Floppy disk driver. Figure 1

Claims (1)

[Claims] 1. An image memory, means for scanning a form one by one and storing image data in the image memory, a recognition result storage memory, and reading the image data stored in the image memory. , means for cutting out and recognizing characters one by one and storing, as a character recognition result, a recognized character code or a predefined unreadable character code and feature data of the unreadable character obtained along with the recognition in a recognition result storage memory; A corrected character code input means for inputting a corrected character code for an illegible character by an operator; and a correction for replacing the unreadable character code in the recognition result with the character code input by the corrected character code input means. means for referring to the recognition result memory and searching for similar unreadable characters having similar characteristics to the unreadable character corrected by the correcting means; and converting the unreadable character code of the similar unreadable character into the specified character. A character reading device comprising: means for collectively replacing a code; and means for outputting a recognition result replaced by the batch replacing means.