JP3111521B2 - Recognition character correction method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for correcting a character in which a character pattern is read by a character recognizing device and an error relating to the character size of the recognized character is corrected at the time of character recognition for character recognition based on the character pattern. It is about.

[0002]

2. Description of the Related Art In recent years, character recognition devices have been actively introduced as one of input terminals of various devices, and are being put to practical use. In the conventional character correction method of the character recognition device, the character pattern is classified based on the vertical or horizontal length of the character pattern. Was.

[0003]

However, the above-mentioned conventional method is effective to some extent in limiting the target typeface or character style, but is difficult to correct when the target type is expanded. Had. This will be described using a specific example.

FIG. 6 shows six types of kana characters (ya, ya, u, yu, u) with different typefaces and the same number of points.
And four types of kanji (smoke, swallow, monkey, rim). The column direction is the same font. In the figure, for example, c1 is a lowercase “u” and c2 is a uppercase “yu”. In this case, the vertical length is c1 which is a lowercase letter and c is an uppercase letter.
It can be seen that it is difficult to correct even if there is an error in the character code of lowercase or uppercase during recognition by the conventional method using only the size.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to provide a method for correcting a recognized character in which even if there is an error in a character code relating to a character size in character recognition, the error is surely corrected.

[0006]

This onset bright in order to achieve the above object, according to the solution to ## in recognizing a character from the document patterns to be recognized, the maximum value the longest side length of the character in the recognized character string calculated as, for having both upper and lower case in a recognized character by calculating the area ratio of the length to the said maximum value of the vertical or horizontal the longer side, of case To determine the font size.
This is to make corrections .

[0007]

According to the recognition character correcting method of the present invention according to the above-described procedure, the first and second maximum values, which are the maximum values of the vertical and horizontal sizes of the characters in the character string, are obtained. From the comparison with the longer side of the kana character having both, it is possible to accurately distinguish between uppercase and lowercase characters by making use of the characteristics of the character typeface or character type.

The principle of the present invention will be described below with reference to the drawings. In FIG. 6, the typeface and font are different, but the number of points is the same. Six types of kana characters (ya, ya, u, yu, yo, yo) having both uppercase and lowercase and four types of kanji (smoke, swallow) , Monkey, rim). The column direction is the same typeface or font, that is, the same font. From this, the following can be said.

(1) In the same font, kana uppercase letters are always smaller than kana lowercase letters. (2) In the same font, the ratio between the vertical and horizontal lengths of uppercase and lowercase characters of kana characters (hereinafter, aspect ratio) is almost equal.

(3) In the same font, Kana capital letters and long sides of Chinese characters are almost equal. In the present invention, for characters with different fonts and fonts,
Effective classification is performed by using the properties of (1), (2) and (3). That is, if it is known that the input character is a character that may be both uppercase and lowercase (for example, “tsu” and “tsu”), this character and the character assumed Calculates the ratio of the longer one of the horizontal lengths (hereinafter referred to as the long side), and determines uppercase and lowercase characters based on the input character area based on this ratio and the uppercase area estimated from the aspect ratio of the input character. I do. That is,
When the same typeface or font is the same, all characters in the character string to which the input character belongs, including kanji, are based on the fact that the aspect ratio is almost the same between uppercase letters and the length ratio between the long sides is specified. The maximum value among the long sides of the pattern is obtained and is used as a reference.

Now, the long side of the kanji having the maximum value and the vertical and horizontal lengths of the input kana character are respectively represented by M, a, and
In the case of b, if a <b, the area of a kana uppercase letter is estimated to be a value obtained by multiplying the area ab of the input kana character by (M / b) · (M / b). Therefore, the ratio of the area of the input kana character to the kana capital letter is (b / M) · (b / M). Similarly, when a ≧ b, (a / M) · (a / M). When a = b, the same result is obtained with either equation.

As described above, the ratio of the lengths of the long sides is specified as the area ratio, and it is determined from the value that the case is uppercase or lowercase.

[0013]

FIG. 1 is a functional block diagram of a character recognition apparatus according to an embodiment using a method for correcting a recognized character according to the present invention. In FIG. 1, reference numeral 1 denotes a character input unit for photoelectrically converting a document and storing it as binary data in a document pattern memory; 2, a character extraction unit for extracting characters from the binary data; A character recognizing unit that outputs a character code corresponding to a character pattern, 4 is a character in the same character string given from the character cutout unit 2 when the character recognized by the character recognizing unit 3 is a character having both uppercase and lowercase characters. A character size determination unit that determines the maximum value of the vertical and horizontal lengths of the pattern and determines whether the character pattern is uppercase or lowercase based on this value and the vertical and horizontal lengths of the input pattern. When the character size of the recognized character from the character recognizing unit 3 is incorrect according to the result of the size determining unit 4, the correcting unit corrects the character size. 6 is a display unit that outputs the result of the correcting unit 5. Ri out of all characters in the resulting string of 2 vertical and horizontal which is the maximum long side calculation unit for obtaining the maximum of the long side from the length.

FIG. 2 is a block diagram showing the configuration of the character recognition device of the present embodiment. Here, a scanner 21 reads a document to be recognized, and outputs the read document as bit data. Reference numeral 22 denotes a readable / writable memory (hereinafter abbreviated as RAM), which includes a document pattern area 23 for storing bit data from the scanner 21 and a character string cut out from a document pattern in the document pattern area 23. A character size area 27 for storing the size of the character pattern to be stored, a maximum long side area 25 for storing the maximum value of the vertical and horizontal lengths in the character size area 27, and a character pattern area 24 for storing the character pattern. A character code area 26 for storing a character code obtained by recognizing the character pattern, and a register area 28 used for processing. 29 is a read-only memory (hereinafter abbreviated as ROM), a character size dictionary area 30 storing upper and lower case character codes and a threshold value unique to the character codes;
And a program storage area 31 storing a control program according to the flowchart shown in FIG. Reference numeral 32 denotes a processing circuit that performs processing according to the control program stored in the program storage area 31. Reference numeral 33 denotes a keyboard for inputting data. Reference numeral 34 denotes a display unit for displaying a character pattern in the character pattern area 24 and a character code in the character code area 26, or a character font corresponding thereto.

FIG. 3 shows a method for correcting a recognized character according to the present invention using the character recognition device of the present embodiment configured as described above.
This will be described according to the flowchart of FIG. First, step S
1, a character is cut out from a document pattern read from a document pattern to be recognized by the character input unit 1 by the processing of a character cutout unit 2, and the vertical and horizontal lengths of all character patterns existing in the same character string are extracted. To character size area 2
Write to 7. In the next step S2, the maximum long side calculation unit 7
The maximum value M is searched for in the vertical and horizontal lengths in the character size area 27, and the result is stored in the maximum long side area 25.
Next, in step S3, a character recognition process is performed by the character recognition unit 3, and the recognized character and the character code are stored in the character code area 26. In step S4, it is determined whether or not the recognized character in the character code area 26 is registered in the character size dictionary area 30. If the character is registered, the character has a lowercase letter, so that a character size determination process is performed in step S4. If not, the process ends. Step S
In step 5, first, the magnitude relation between the lengths a and b of the input character pattern is checked. If a ≧ b, step S6, a <b
In the case of, the process proceeds to step S7. In step S6, (a /
M) is calculated, and in step S7, the square of (b / M) is calculated. Here, M is the value of the maximum long side area. In step S8, if the result of step S6 or S7 is smaller than a predetermined value registered in the character size dictionary area 30, it is determined that the character is a small letter, and if it is larger, it is determined to be a capital letter. Finally, in step S9, the correction unit 6 corrects the contents of the character code area 26 when the result of the determination is different from the character code of the recognized character in the character recognition unit 3 and the character size of the recognized character is wrong.

Next, the operation will be described below by taking as an example a case where the recognition target example character string pattern "National Electric" is read. First, in the character input unit 1,
A recognition character string pattern (hereinafter, referred to as an input character) is input as a binary image and stored in the document pattern area 23. Next, after the input character is cut out by the character cutout unit 2, the character cutout unit 2 cuts out one character at a time. Write 27. FIG. 4A shows the character size area 27 at this point. According to the maximum long side calculation unit 7, since the horizontal length of “R” is the largest, the maximum long side M
= 61 is obtained. As shown in FIG. 4A, it is assumed that the character recognizing unit 3 recognizes the input character "National Electric" as "National Electric" (this is called a recognized character). The character size determination unit 4 first determines whether the recognized character is a character having both uppercase and lowercase characters. Here, three of "yo", "e", and "tu" correspond.
"Y" is longer than horizontal, so the square of (a / M)
Since "e" and "tsu" have a large horizontal length, the square of (b / M) is obtained as shown in FIG. On the other hand, R
The character size dictionary area 30 of the OM 29 stores uppercase and lowercase letters and predetermined values for kana characters having uppercase and lowercase letters, as shown in FIG. Is a boundary value that specifies Therefore, if the value of the input character is larger than the value of the corresponding character, it is determined to be uppercase, and if it is smaller, it is determined to be lowercase. The correcting unit 6 corrects the character size of the recognized characters for "Y", "E", and "T" because they are wrong, and corrects the recognized characters as shown in FIG. 4B.

As described above, according to this embodiment, by providing the maximum long side calculation unit, the character size determination unit, and the correction unit, the character size error among the recognition characters output from the character recognition unit is corrected. In this case, it can be corrected with high accuracy.

Although the kanji and kana characters have been described in the present embodiment, it is needless to say that the same effect can be obtained for alphabets and the like whose fonts can be determined.

[0019]

As is clear from the above description, the area ratio is calculated based on the maximum length of the recognized character and the character sizes are compared to correct errors in the character size of the recognized character. It can be corrected frequently.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a method for correcting a recognized character according to the present invention for each function.

FIG. 2 is a block diagram of a character recognition device showing the embodiment as a hardware configuration;

FIG. 3 is a flowchart showing a procedure of the embodiment.

FIGS. 4A and 4B are explanatory diagrams showing specific character string procedures in the embodiment.

FIG. 5 is a layout diagram of storage contents stored in a character size dictionary area according to the embodiment;

FIG. 6 is a character pattern diagram for explaining a conventional recognition character correction method.

[Explanation of symbols]

 1 character input section 2 character cutout section 3 character recognition section 4 character size determination section 5 correction section 6 display section 7 maximum long side calculation section

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06K 9/46 G06K 9/62

Claims (2)

(57) [Claims] 1. A reads a document pattern to be recognized, the document pattern cut out characters from, calculate the maximum value of the vertical size or horizontal size of the character pattern among string cut, character codes from the character pattern were read out, the character code determines the possibility of having both upper and lower case, to compare the longitudinal length and lateral length of the character pattern in the case of having the potential, the vertical length determined case by the value of the area ratio was calculated from the vertical length and the maximum value is longer, if the horizontal length is long by the lateral length and the value of the area ratio was calculated from the maximum value A method of correcting a recognized character , wherein upper and lower case characters are determined, and the character code is corrected if the result of the determination is different from the read character code. The maximum value of the vertical size or horizontal size of 2. A character pattern is the maximum value of the long side of the character pattern, this
2. The method according to claim 1, further comprising the steps of :