JPH0475557B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an optical character reading device that can perform character recognition processing for a group of similar characters.

[Technical Background of the Invention and Problems Therewith] In recent years, optical character reading devices (OCR) are recognizing a wide range of character types, including alphanumeric characters, katakana characters, hiragana characters, and kanji characters. With such OCR, it is difficult to recognize similar characters, such as the number "0" and the alphabet "O" or the number "5" and the alphabet "S".

Conventionally, when writing similar characters like the ones above, for example, a line was added above the letter "O", or a serif was added at the bottom left of the letter "S".
Things are being done to add . This can improve the recognition rate for similar characters, but if kanji are included in the characters to be recognized,
The number of similar characters increases dramatically. For this reason, when writing on a form, it is actually difficult to distinguish between similar characters. For this reason, it has been extremely difficult for conventional OCR to reliably recognize similar characters including kanji.

[Object of the Invention] The present invention has been made in view of the above points, and an object thereof is to provide an optical character reading device that can reliably recognize similar characters including Chinese characters.

[Summary of the Invention] The present invention includes similar character determining means for determining whether the recognition result of the character recognizing means is a similar character type. If the recognition result belongs to a similar character type, the character type determining means determines whether the recognition result is the same as the character type of the characters before and after it. If the character types are the same, output means is provided for outputting the character code of the recognition result corresponding to the character type of the preceding and succeeding characters as the final recognition result. Also, if the character types are not the same,
Determination means is provided for determining the character type of the recognition result using a character type determination table prepared in advance.

With such a configuration, character recognition processing for similar characters can be reliably executed.

[Embodiment of the Invention] An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an OCR according to an embodiment. In FIG. 1, a character recognition unit 10 executes character recognition processing on a photoelectrically converted character pattern, and outputs a recognition result R to a buffer memory 11. The buffer memory 11 normally stores recognition results R for one line. The similar character determination unit 12 determines whether the recognition result R output from the buffer memory 11 belongs to a similar character type. If the recognition result R does not belong to the similar character type, the similar character discrimination section 12 outputs the recognition result R to the answer output section 17.

When receiving the recognition result R belonging to the similar character type, the preceding and following character discrimination unit 13 extracts the recognition results of the characters existing before and after the recognition result R from the buffer memory 11.
FB is read and it is determined whether the recognition result R is the same as the character type of the recognition result FB. When the discrimination results of the preceding and following character discriminating section 13 are the same, the final code determination section 14 determines the character code of the recognition result R corresponding to the character type of the recognition result FB as the final code, and outputs it to the answer output section 17. . If the discrimination results of the preceding and following character discrimination sections 13 are not the same, the final code determination section 15 determines the character type of the recognition result R using a character type determination table stored in the table memory 16 in advance, and determines the character type corresponding to the character type. The code is output to the answer output section 17.

In the OCR having such a configuration, the operation according to one embodiment will be explained. First, characters recorded on paper are subjected to photoelectric conversion and then provided to the character recognition section 10. The character recognition unit 10 performs character recognition processing on character patterns. The recognition result R obtained thereby is stored in the buffer memory 11 as a character code. At this time, if the character to be recognized is a similar character type, the character code corresponding to the Chinese character or the like is temporarily stored in the buffer memory 11. For example, if the recognition result R is "-", a character code indicating a long sound is stored in the buffer memory 11. The similar character determination unit 12 determines whether the recognition result R from the buffer memory 11 is a similar character type. Here, similar character discriminator 1
2 stores in advance a table consisting of a group of similar characters (including first level kanji) as shown in FIG. 2, for example. If the recognition result R does not belong to the similar character group shown in FIG. 2, the recognition result R is output to the answer output section 17 with the same character code.

On the other hand, if the recognition result R belongs to a similar character group,
The preceding and following character determination unit 13 determines the character types of the characters before and after the corresponding character in the recognition result R. In this case, the preceding and following character discrimination section 13 reads out the recognition results FB corresponding to the preceding and following characters from the buffer memory 11, and executes character type discrimination. As a result of this discrimination, if the characters are of the same type, the character code of the recognition result R corresponding to the character type is output to the final code determining section 14. For example, when the recognition result R is "Ha",
If the preceding and succeeding recognition results FB are katakana characters, the character code corresponding to the katakana character "ha" is output to the final code determination unit 14. Final code determination section 1
In step 4, the character code output from the preceding and succeeding character discriminating section 13 is determined as the final character code, and the character code is output to the answer output section 17.

Further, when the character type of the recognition result R is different from that of the preceding and succeeding characters, the character code of the recognition result R is determined in the final code determining section 15. The final code determination section 15 executes character code determination using, for example, a character type determination table as shown in FIG. Here, in Figure 3, long is a long sound, da is datsushiyu, - is a Chinese numeral, ma is a minus, ? indicates that it cannot be determined. For example, when the recognition result R is an unknown "-", it is assumed that the preceding and succeeding characters are respectively Chinese characters based on the discrimination results of the preceding and following character discriminating unit 13. In this case, the recognition result R is determined to be the Chinese numeral "-" according to the table shown in FIG. Then, the final code determination section 15
The character code determined in is outputted to the answer output section 17 as the final code of the recognition result R. In addition, the final code determination unit 15 determines that the determination result is "?"
In this case, it is determined that there are two or more character type candidates or the frequency is rare, and the character type candidate is rejected as a judged gait performance.

In this way, when a character to be recognized belongs to a similar character group, the character type is determined based on the character types of the characters before and after the character. Based on this determination result, the character code corresponding to the determined character type is output as the final character recognition result. Therefore,
There is no need to write in advance on paper to distinguish between similar characters, and similar characters including kanji can be reliably recognized.

[Effects of the Invention] As described in detail above, according to the present invention, similar characters including Chinese characters can be reliably recognized with a simple configuration. Therefore, the recognition rate for characters to be recognized belonging to a wide range of character types can be greatly improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an OCR according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a table stored in the similar character discriminator shown in FIG. 1, and FIG. 3 is a diagram similar to the one shown in FIG. FIG. 2 is a diagram showing an example of a character type determination table stored in a table memory of FIG. 10...Character recognition unit, 12...Similar character discrimination unit, 13...Next and subsequent character discrimination unit, 14...Final code determination unit, 15...Final code determination unit, 16...
table memory.

Claims (1)

[Scope of Claims] 1. A character recognition means that performs character recognition processing on a photoelectrically converted character pattern, and a similar character group table prepared in advance to determine whether or not a recognition result by this character recognition means belongs to a similar character group. and a similar character discriminating means that discriminates by referring to the similar character discriminating means, and when the recognition result belongs to a similar character group according to the discrimination result of the similar character discriminating means, the character type of the recognition result and each character existing before and after the recognition result. a preceding and following character discriminating means for discriminating whether or not the preceding and following character types are the same; and when the recognition result of the preceding and following character discriminating means is the same character type as the character type of each character, the recognition corresponding to the character type; a final code determination means for outputting the resulting character code as a final character recognition result; and a character type prepared in advance when the character type of the recognition result is different from the character type of each character according to the determination result of the preceding and following character determining means. The present invention is characterized by comprising a final code determination means for referring to a determination table, determining a character code of the recognition result based on the character type of each character, and outputting this character code as a final character recognition result. Optical character reader.