JPH0414193A - Italic character recognition method - Google Patents

Abstract

PURPOSE:To improve the rate of recognizing sfield charactors and recognition speed by counting the number of charactors at a specified angle to a main scanning direction in an input letter area and deciding the sfield charactors automatically. CONSTITUTION:The number of charactors in a main operating direction 202 is counted when a picture area is scanned in the main scanning direction and also the same counting is operated in a left down direction 302 to the direction 202 having tan theta=1/6, for example, so that the charactors are decided as the sfield charactors when the number of bit is more than a prescribed value. Whether the charactors are the sfield one or not is discriminated automatically after the same repetition in a subscanning direction 201. The sfield charactors are recognized in the high rate of recognition and speedily by referring to a corresponding dictionary based on this decision.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for recognizing character images read by an optical sensor or the like as characters.

[Prior art] In conventional means for recognizing italic characters and regular characters (Mincho typeface, Gothic typeface, etc.), when recognizing characters, character recognition is performed without determining whether or not they are italicized. Ta.

[Problem to be solved by the invention]

However, with this type of recognition method, the range of the recognition dictionary (pattern dictionary or peripheral feature dictionary) when recognizing characters extends to both italic and regular characters, resulting in a long recognition time. .

Furthermore, since many dictionary areas are searched, recognition fails and the recognition rate decreases.

Therefore, the present invention was made in view of such problems.
The purpose is to provide a recognition method that automatically determines whether a character (single writing) is italic, improves the character recognition rate, and shortens the recognition time.

[Means for Solving the Problems] The italic character recognition method of the present invention includes: a read image realized by an optical sensor or the like; a means for counting whether the input character area is a character in the main operation direction; means for counting whether or not the character area is a character at an angle downward to the left with respect to the main operation direction; means for repeating the counting means in the sub-operation direction; and characters divided into italic characters and other characters. a pattern dictionary or a peripheral feature dictionary for recognizing the character, a means for normalizing the input image for each character, a means for determining whether or not the normalized character area is an italic character; and means for comparing characters with the recognition dictionary.

〔Example〕

The details of the recognition means of the present invention will be explained below based on illustrated embodiments.

FIG. 1 is a flowchart of the character recognition means of the present invention. FIG. 2 is an example of a character image obtained by the image input 101.

For character recognition, first perform character line extraction 102 to extract a character line area in an image.
By counting the character bits in the sub-operation direction 201 of the image, an area where the frequency is high is determined to be a character line. Once a character line is determined, counting of character bits in the main operating direction 202 for that line is performed. This count is used for feature extraction for normal characters. First, set the sub-operation start line on the counter 103. 0th order counting start line 2
Set 03 on the counter 104. The counting start line is the top line of the character lines obtained in the character line area extraction 102.

The number of character bits is counted 105 in the main operation direction 202 from that line.The counting of 6 character bits is completed up to the end line 204 of the 0th line, which is the lowest line 204 obtained in the character line area extraction 102. Then, 106, the next main operation line is counted in the sub-operation direction. When counting is completed in the sub-operation direction as well (107), counting of character bits for feature extraction for italic characters is started. As shown in FIG. 3, the feature extraction character bits for italic characters are counted downwardly to the left by 301 degrees in the main operation direction 202. The counting direction component 302 for italic characters is usually approximately tanθ=1/6 303 based on experience. Counting is performed as follows. First, set the start line on the counter 108
The zeroth order counting start line 203 is set 109 on the counter. The counting start line is the top line of the character lines obtained in the extraction step 102 of one character line area. The number of character bits is counted 110 in the counting direction 302 for italic characters from that line.
When counting is completed up to the end line 204 of the 0th row, 1
11. Count the next main operation line in the sub-operation direction. When counting is completed in the sub-operation direction as well (112), counting of character bits is completed.

Next, the main routine for character recognition starts. Character recognition is performed on the character components of the character line area determined in character line extraction 102.

Move the character line area from the sub-operation start line to the sub-operation direction 20
1, cut out each character, scale the size of each character, and normalize it. The method of cutting and compressing characters here is realized by a method of extracting a single character outline, etc. When a character is cut out, the area where the character exists (sub-operation direction 201)
is determined.

In that character area, the character bit values counted at 105 and 110 are compared. It is determined 114 whether the extracted character is an italic character or a normal character based on the maximum value of the count value and the count distribution characteristic. Here, the details of the count distribution characteristics are shown in Figure 4 (
Shown in a) and (b).

FIGS. 4(a) and 4(b) are diagrams showing count distribution characteristics according to the letter "D". Ordinary characters shown in Figure 4(a)
D"401 has a distribution characteristic when counting character bits in the main operation direction of each bit as shown in 402. Also, the count distribution characteristic when counting with an angle of °゛θ"301 in the main operation direction. will look like 403. The italic character ° "D" 404 shown in FIG. 4(b) has a distribution characteristic as shown in 405 when character bits in the main operation direction of each bit are counted. Further, the count distribution characteristic when counting is performed with an angle of "°θ" 301 in the main operation direction is as shown in 406. As described above, since the count distribution characteristic clearly changes the reproduction rate of a character depending on the count of the "θ" angle, it is determined 114 whether the character is an italic character or not based on the shape distribution of the characteristic value and the maximum value of the count value. Next, based on the character type determined in step 114, the corresponding recognition dictionary is compared with the normalized character to recognize the character (115). Here, character types are narrowed down based on the shape using the shape distribution value in step 114 (narrowing down within the same character type) to shorten the recognition time.Repeat 1 or more until the final line in the sub-operation direction to finish character recognition. .

[Effects of the Invention] As described above, according to the present invention, italicized characters and other normal characters can be recognized in a short recognition time.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of a flowchart configuration according to the present invention, FIG. 2 is a diagram showing an example of an input character image,
FIG. 3 is a diagram showing an example of the counting angle in the main operation direction according to the present invention, and FIGS. 4(a) and (b) are diagrams showing an example of character bit count distribution characteristics according to the present invention. Normal letter font Normal count distribution characteristics Count distribution characteristics for italics Italic letter font Normal count distribution characteristics Count distribution characteristics for italics More than i1 person foot 2o3 Consideration No. 1 mouth

Claims (1)

[Scope of Claims] (a) A read image realized by an optical sensor or the like; (b) Means for counting whether the input character area is a character in the main operation direction; (c) The input (d) means for repeating the counting means in the sub-operation direction; (e) italicized characters; a pattern dictionary or peripheral feature dictionary for recognizing characters divided into other characters; (f) means for normalizing the input image character by character; (g) means for comparing the normalized characters with the recognition dictionary.