JPS58192180A - Character reader - Google Patents

Abstract

PURPOSE:To correct easily a rejected or erroneous character, by displaying a similar character set corresponding to a recognition candidate character and inputting the set selectively. CONSTITUTION:A character on a slip 1 is photoelectric-converted, a character pattern obtained with the scanning at a recognition means 3 and a reference pattern in a recognition dictionary 4 are compared to obtain the degree of similarity for the character recognition. The result of recognition is displayed on a display means 5 and a correct character is inputted from an input means 6 selectively. If no correct character is included in a recognition candidate character, the input means 6 is used and one character in a displayed recognition candidate character is inputted selectively and the similar character set in a storage means 8 corresponding to the recognition candidate character is displayed on the display means 5. When the character to be inputted is displayed, the correct character is inputted selectively by using the input means 6 and given to an edition means 9.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a character reading device that reads many types of characters, and more particularly to a character reading device w that can easily correct reject characters and error characters.

According to “Kanji in Modern Newspapers” published by the National Institute for Japanese Language and Linguistics, the frequency distribution of kanji used in Japanese texts is too.
s19% for O-shaped woodcutter, 9s for 2000 characters. It accounts for 6%. Therefore, more character types &l! 9, it is necessary to have an open dictionary for character types that rarely occur.
Since the equipment is large-scale, IIl! ! ! It is necessary to limit the types of characters that can be written. For this reason, the conventional device was constructed as shown in Figure 1. Character patterns obtained by photoelectrically converting the characters on the form (11) by the scanning means (2) and scanning them by the recognition means (3). .

The character is recognized by comparing it with a reference pattern in the first recognition dictionary (4) to determine the degree of similarity. The recognition result is displayed on the display means (5), and if the g-stain result is different from the correct character, the recognition result is displayed on the input means (6).
) to send the correct characters to the editing means (7).

- The east means (7) controls the display means (5) and the input means (6) and edits the recognition results. However, when there are many types of characters to be recognized, it is nearly impossible for the person writing the information to remember whether or not the character is a recognition target character, and non-recognized characters are often entered.

Often rejects or errors.

Therefore, some form of external character input means is required, and conventionally the sound M
I input foreign characters by converting kana to kanji using i11 input or by directly inputting the 9-character code.

However, checking the character code takes a lot of time,
Even with the method of inputting kana, it is always a painful task for amateurs and has the drawback of requiring a dedicated operator.

In order to eliminate these drawbacks, this invention has a similar character set corresponding to the character to be recognized, displays a similar character set corresponding to the character that becomes the first candidate character as a result of recognition, and selects a character set from among the displayed characters. It is characterized by the selection of necessary characters, and its purpose is to easily correct and input errors and reject characters mainly caused by non-standard characters.

This will be explained in detail below using the drawings.

FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention. The characters on the form (1) are photoelectrically converted by the scanning means (2), and the character pattern obtained by scanning by the &I precision means (3) is compared with the reference pattern in the 2g knowledge dictionary (4) to find similarities. Recognize characters by searching for degrees. The recognition result is displayed on the display means (5), and the correct character is selected and inputted from the input means (6). Specifically, if the recognition result is different from the correct character, the input means (6) is used to instruct the display means (5) to display the recognition candidate character, and if the correct character is included in the recognition candidate characters, Select and input the correct character using the input means (6).

Also, if the correct character is not included in the recognition candidate characters,
Using the input means (6), select and input one character from the displayed recognition candidate characters, and display the similar character set in the storage means (8) corresponding to the recognition candidate character on the display means (5). . If the characters to be entered are displayed, use the input method (
Select and input the correct characters using 6) and enter the 9m collection means (8)
send to If the character to be input is not displayed, one character from among the recognition candidate characters is selected and inputted again, and the similar character set is displayed on the display means (5), and the process is repeated.

FIG. 3 is a diagram showing examples of input character patterns, V recognition results, and recognition candidate characters. The recognition results for the input character pattern ``''01 shown in Figure 3 (-f) are shown in Figure 3 (B).
The recognition candidate characters are -H, '1m, On, 'f4, as shown in the same figure (c).
It will be J04.

Figure 4 is a conceptual diagram showing an example of a similar character set corresponding to each of the nine characters. The similar character set for (1, kidney (14)) is -, '哄■. Therefore, the above recognition candidate characters 0 and α
If you select 1- from 4, the display means (5) will display the similar character set 1-, 'f14. (Since 1!9 is displayed, select and input the character using the input means (6).

As described above, according to the present invention, a similar character set corresponding to each character is stored, and a similar character set corresponding to a recognition candidate character is displayed and selectively input. You can select and input characters from the similar character set, and regarding errors and rejections of internal characters.

This method has the advantage that it is possible to manually select characters from the similar character set of recognition candidate characters, and it is possible to easily input corrected nine characters.

Although the case where one character from among the displayed recognition candidate characters is selected and input has been described above, the present invention is not limited to this, and a plurality of characters may be selected and input.

In addition, although we have described a case in which all characters of a set of similar characters are repeatedly displayed until the input character is displayed, this invention is not limited to this, but instead of displaying similar characters that have already been displayed, the number of characters to be input by one selection is reduced. The more you reduce the number, the easier the selection input will be.

As described above, according to the present invention, a similar character set corresponding to a recognition candidate character is displayed and selectively input, so that reject or error characters can be easily corrected.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a conventional device. FIG. 2 is a block diagram showing the configuration of the device according to the present invention;
FIGS. 3A, 3B, and 3C are diagrams showing examples of input character patterns, μ iron results, and candidate characters, and FIG. 4 is a conceptual diagram showing similar character sets. +11 is a form, (2) is a scanning means, (3) is a recognition means,
(4) is a recognition dictionary, (5) is a display means, (6) is an input means, +71. +91 is an editing means, (8) is a similar character set storage means, Ql is an input character example, (lυ is an authentication result example α ri ~ α 荀 is a candidate character example 09 ~ 0 η is a similar character set example. Identical or equivalent parts in the figures are indicated by the same reference numerals. Agent Shin Kuzuno - Figure 1 Figure 2 smi [lft practice -

Claims (1)

[Scope of Claims] A scanning means for scanning and charging nine characters in a character reading device that recognizes and reads characters recorded on paper or the like;
a recognition dictionary storing reference patterns of characters to be recognized, a storage means storing a set of characters having similar glyph shapes for each character to be recognized, and a recognition means for comparing input character patterns and the reference pattern to recognize characters; , Shosen results and gwII
t display means for displaying candidate characters; and input means for selecting and inputting a character from among the characters displayed on the two display means. When the recognition result is different from the correct character and the correct character is not included in the displayed recognition candidate characters. When a part or all of the twill candidate characters are selected and input using the input means, the corresponding character set in the storage means is displayed on the display means, and 2 characters are selected from the displayed character set. A character reading device characterized by selectively inputting.