JPS6095688A - Character recognizing device - Google Patents

Abstract

PURPOSE:To improve the character recognition factor by reading the corresponding 2nd character code out of a replacing character memory to use it as the final result of recognition in case the result of recognition is equal to the 1st character code. CONSTITUTION:The pattern information is obtained by scnanning characters with photoelectric conversion and then separated every character to be supplied to a character recognizing part 2. A control circuit 5 delivers a recognition start command to the part 2 and a dictionary pattern memory 3. The part 2 calculates the similarity between each standard character pattern in the memory 3 and segmented unknown pattern and delivers the patterns having the highest similarity to the circuit 5 as a result of recognition. The circuit 5 searches a replacing character memory 7 and discriminates whether the codes of characters to be replaced are registered to a corresponding memory area. Then the character code if registered is used as the final result of recognition.

Description

DETAILED DESCRIPTION OF THE INVENTION [Branch Field of the Invention] The present invention relates to a character recognition device such as an original character reading device.

[Technical background of the invention and its problems]

In recent years, the range of character types that character recognition devices can target has been expanding more and more. It has been expanded to include kana, symbols, hiragana, and kanji.

However, as the target character types expand, the recognition rate inevitably decreases. These are "Ten" and "Husband".

Characters with similar fonts such as ``大'' and ``尤'' (for example, ``bi'')
One of the reasons for this was the increase in the number of ``bi'' characters and the increase in characters that resembled other characters due to errors in writing or printing, making it extremely difficult to improve the recognition rate.

[Purpose of the invention]

The present invention has been made in consideration of the above circumstances, and an object of the present invention is to provide a character recognition device capable of increasing the recognition rate.

[Summary of the invention]

In order to achieve this object, the character recognition device according to the present invention includes a replacement character memory storing a second character code corresponding to the first character code, and a character recognition device in which the recognition result is based on the first character code. The present invention is characterized by comprising a processing means for making the second character code read from the lost character memory when the character code is lost as the final recognition result.

[Embodiments of the invention]

FIG. 1 shows the main parts of a character recognition device according to an embodiment of the present invention. The character recognition section 2 inputs an unknown pattern 7 cut out character by character from the pattern information obtained by scanning and photoelectrically converting a document in which characters to be read are recorded and output from a photoelectric conversion section (not shown).

It is recognized by comparing it with the character pattern in the dictionary pattern memory 3 in which standard character patterns are registered in advance.

Conventionally, the recognition result was used as the final recognition result, but in this embodiment, the ``transformation'' character is written as ``1.''. A replacement character memory 7 is provided. The replacement character memory 7 has memory areas corresponding to all the characters that can be read by the character reading device of this embodiment, and this memory area contains the characters to be replaced with the characters. The character code of is memorized.

If there are no characters to replace, all 10" is memorized. For example, if you want the final recognition result to be "large" even if it is recognized as "Yu", use the character code of Correlate and memorize the character code for "large". The replacement character memory 7, the character recognition unit 2, and the dictionary pattern memory 3 are controlled by a control circuit 5.

Assume that an unknown pattern as shown in FIG. 2(a) is input to this character recognition device. 11 and 12 are field numbers of the character line, which are assigned to each field of one character line to perform format control. Figure 2 (
The pattern information obtained by scanning and photoelectrically converting the characters shown in a) is cut out character by character by a circuit (not shown) and input to the character recognition section 2. Thereafter, the control circuit 5 issues a recognition start command to the character recognition section 2 and dictionary pattern memory 3. Then, the character recognition unit 2 performs a similarity change N between each standard character pattern in the dictionary pattern memory 3 and the cut out unknown pattern, and outputs the one with the highest degree of similarity as the missing recognition result to the control circuit 5. The control circuit 5 searches the replacement character memory 7 and determines whether the character code of the character to be replaced is registered in the memory area corresponding to the character code resulting from this recognition.

If the character code is registered, the character code is used as the final recognition result, and if it is not registered, the character code of the recognition result is used as the final recognition result.

Now, suppose that in the replacement character memo IJ7, there are 5 characters "dai" which should be replaced with the character "尤" and a character "bi" which should be replaced with the character "bi". As shown in Figure 2 (b), even if the recognition result from 2 shows "Peel bottle" and "Yukii Lake", in the final recognition result, "Pi" is replaced with "BijVc'尤'Ka1dai". "
The correct recognition results are obtained as ``peel bottle'' and ``large lake.''

Registration in the replacement character memory 7 may be performed in any way, but the user may freely register the data, such as inputting format control data of a form at the same time. A method that allows this is preferable.

[Effects of the Invention] As described above, according to the present invention, the final recognition result can be changed by the user registering in the replacement character memory. Therefore, if there are similar characters that appear less frequently, the recognition rate will improve if the characters VC that appear more frequently are replaced with the characters VC that appear more frequently. Also, if you want to recognize a new character other than the target character, you can register a similar character among the target characters to be replaced with a new character.
It is also possible to recognize characters other than those to be recognized. This is effective when the user wants a certain character to be read as a completely different character.

[Brief explanation of drawings]

FIG. 1 shows the block ratio of the main parts of a character recognition device 4 according to an embodiment of the present invention, and FIG. 2 shows the block ratio of the main part of the character recognition device 4 according to an embodiment of the present invention.
FIG. 2...Character recognition section, 3...Dictionary pattern memory. 5... Control circuit, 7... Replacement character memory. Applicant's agent Kiyoshi Inomata

Claims (1)

[Claims] In a character recognition device that outputs a recognition result as a character code, a replacement character memory stores a second character code corresponding to the first character code, and the recognition result is the first character code. A character recognition device characterized in that the second character code χ read from the replacement character memory is further provided with processing means for producing a final recognition result.