JPH045781A - Character recognizing system - Google Patents

Abstract

PURPOSE:To improve the accuracy of recognition without enlarging a dictionary capacity hardware by extracting the point of an outline form a character pattern data binarized after reading from a document and based on the result, calculating the amount of feature as a line segment by required directions. CONSTITUTION:An outline pattern is extracted from the character pattern data binarized after reading form a slip stored in a pattern register 1 by an outline extraction part 2 to decide the direction of the line segment dividing 360 deg. into 8 equal divisions from the (x, y) coordinate of the outline pattern in a direction decision part 3. Based on this result of the decision, a feature amount calculating part 4 calculates the feature amount composed of the relative position of the line segment form the (x, y) coordinate by directions to recognize the character in comparison with the amount of feature stored in an identification part 5. By such a recognition method for recognizing the positions of distributing the line segments of the character pattern by the directions of the line segments, the character is recognized with high accuracy without enlarging the storage capacity and the size of the hardware.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a character recognition method, and particularly to a character recognition method suitable for improving character recognition accuracy in a character recognition device.

(Prior Art) Conventionally, various methods have been proposed as character recognition methods for character recognition devices. Stroke analysis methods and pattern matching methods are often used as character recognition methods for alphanumeric characters and katakana, and methods that analyze the line structure of patterns and background structure are methods that better absorb character deformation. There are methods to analyze the

(Invention or problem solving problem) However, in the conventional method described above, for example, in the case of the stroke analysis method, the characteristics of each character, including the deformations, are described in the dictionary, resulting in a huge dictionary and a long processing time. There is a problem with it being long.

On the other hand, if an attempt is made to shorten the processing time, the hardware scale will become large, making it unsuitable for simple applications.

In addition, in the conventional background structure analysis method described above, characters are identified by focusing on the background part of the character and extracting features such as loops of character lines or concave or convex shapes, or black and white points are determined. There is a problem in that the processing method is complicated.

An object of the present invention is to solve such conventional problems and provide a character recognition method that improves character recognition accuracy in a character recognition device without increasing the dictionary capacity or hardware scale. There is a particular thing.

(Means for Solving the Problems) In order to achieve the above object, the character recognition method of the present invention includes means for reading characters on a form and storing binary character pattern data, and means for extracting the outline of a pattern; means for determining the direction of a line segment from the outline data of the character pattern extracted by the extraction means; and a predetermined reference point for the outline segment for each of the determined directions. means for calculating a feature quantity consisting of a relative position from , and a means for comparing a feature quantity stored in advance with a feature quantity calculated by the calculation means, It is characterized by character recognition by identifying the position according to the direction of the line segment.

(Operation) In the present invention, means for determining the direction of a line segment from the outline data of the character pattern extracted by the extraction means;
means for calculating a feature amount consisting of the relative position of the contour line segment from a predetermined reference point for each of the determined directions; and means for comparing the feature amount stored in advance with the feature amount calculated by the calculation means. The system recognizes characters by identifying the distribution positions of line segments in an input character pattern according to the direction of the line segments. This can improve your character recognition skills.

(Example) Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram for explaining a character recognition method showing an embodiment of the present invention. This is a block diagram of a character recognition device to which the character recognition method of this embodiment is applied.

In FIG. 1, 1 is a pattern register that stores a binarized character pattern, 2 is an outline extractor that extracts the outline of the character pattern based on the contents of pattern register 1, and 3 is an x extracted by the outline extractor 2. , the line segment is 8 from the y contour coordinates
Direction (→, /, ↑.

\, ←, /, ↓, -); 4 is a feature calculation unit that calculates a feature amount consisting of the relative position of the line segment of the direction-based character pattern determined by the direction determination portion 3; This is an identification section that identifies character patterns.

Pattern register 1 reads the characters on the form and registers 2.
The digitized character pattern is input by a character pattern input section (not shown). Here, it is assumed that the character pattern "A T1" shown in FIG. 11, 24) as a starting point to extract contour features.There are various contour extraction methods, but in this example, this known contour extraction method is used. , the details are omitted. Note that the contour tracking starting point may be any point.

Next, the direction determining section 3 determines the direction of the line segment from the x and y contour coordinates extracted by the contour extracting section 2. Here, eight directions are determined as shown in FIG. 4. That is, for the 0 degree direction (→), X l <X; + 1 and y = 3'; + t
--- (1) For 45 degree direction (1), xl<Xl+1 or y, < yt, +
--- (2) For the 90 degree direction (↑), X ; = X i + 1 and yt<yz+1 ---
(3) For the 135 degree direction (\), x 1 > X i + and yz < yi,!
---(4) Regarding the 180 degree direction (←), X i > X i * (and 3'i = !/i+
+ --- (5) Regarding the 225 degree direction (〆), X□>Xl+1 and y, > Yi++
---(6) For the 270 degree direction (↓), X i ''X i * 1 and yI >y; ++
---(7) Regarding the 315 degree direction (\), each condition of Xi<Xiφ1 and y□> y, -1 ---(8) is determined. By determining each of the above conditions, the direction of each line segment is determined.

FIG. 5 shows the x and y coordinates and directions of the character pattern. Fifth
In the direction (tilt) of the diagram, "-" means 0 degree direction, "/pa" means 45
degree direction, ゛↑pa is 90 degree direction, “-” is 135 degree direction,
← Pa indicates a 180 degree direction, "〆" indicates a 225 degree direction, "↓゛' indicates a 270 degree direction, and "-°'' indicates a 315 degree direction.

The feature value calculation unit 4 calculates x for each direction determined by the direction determination unit 3.
, y-contour coordinates to calculate the feature amount consisting of the relative position of the line segment. This feature quantity calculation will be explained in detail below.

Here, the position where each line segment exists is used as the feature amount. The position in which a certain line segment exists in the x and y directions is expressed numerically on the x and y component side.

First, consider a line segment in the 0 degree direction (→).

The relative position on the X-axis is: The relative position on the y-axis is: Here, calculations in parentheses are performed only for the component in the 0 degree direction, and for components in other directions, the value is calculated as ○. Also, H
x and Hy indicate the character width of the character pattern.

In the same way, formulas for each direction component are obtained as shown below.

Regarding the 45 degree direction (1), Regarding the 90 degree direction (↑), Regarding the 270 degree direction (↓), Regarding the 135 degree direction (\), For the 315 degree direction (-), Regarding the 180 degree direction (←), Regarding the 225 degree direction (end), It can be found from the formula.

The above Qlx"Q8x and QIy-Q8V are the lower left of the rectangle circumscribing the character pattern (coordinates (1, 1 in Figure 3)
) is the origin and performs calculations using a coordinate system. This results in
As for the position where the x+V component of each line segment exists, the lower it is on the y-axis, and the further left it is on the X-axis, the smaller the feature amount becomes. In addition, in the method of this embodiment, all weights related to the positions of all line segments are added, so
This appears as a small difference in the line segment change in the area close to the origin, and is influenced by the change in the line segment in the upper right part of the rectangle circumscribing the character pattern. Therefore, in the upper right corner of the rectangle circumscribing the character pattern (in the coordinate system whose origin is
'□Q8x -Qly '' - Perform the calculation regarding Qay and send the results to Q9 ~Q16X, Q9Y ~Q1
Let it be 6y.

Calculations are performed using the x, y coordinates and direction data shown in FIG. 5 using the equations Qlx-Q16 and Q1y to Q+ay explained above. That is, calculations are performed on the outer edge contour shown in FIG. 5(a) and the inner edge contour shown in FIG.
tex, QIY-QIBY is obtained. In this manner, the above calculation is performed for all line segments of a character that has multiple inner edges.

The identification unit 5 stores Q 1 x ” Q I for each character to be recognized.
6x, Q131"QI6y are stored. The identification unit 5 stores the feature quantities QIX-Q168, Qty"0+sy of the input character pattern calculated by the feature calculation unit 4, and the case Qh, ~ Match Q, 6, and so on.

Let the difference in these matchings be the matching distance,
This calculation is performed for each character to be recognized stored in the identification unit 5, and sorting is performed based on the matching distance.

According to the rearrangement result, the character with the minimum matching distance is selected as the recognition result.

(Effects of the Invention) As explained above, according to the present invention, characters are treated as figures, and the special amount drawing is performed in each direction by considering the direction of the line segment, so that the hardware configuration is increased. It is possible to improve character recognition accuracy in a character recognition device without any problems.

[Brief explanation of drawings]

Fig. 1 is a diagram for explaining a character recognition method showing an embodiment of the present invention, Fig. 2 is a diagram showing an example of a character pattern, Fig. 3 is an outline extraction diagram of a character pattern, and Fig. 4 is a line segment. Figure 5 is a contour feature diagram showing coordinates and directions. 1: Pattern register, 2: Contour extraction unit, 3 Two-way discrimination unit, 4: Feature extraction unit, 5: Identification unit.

Claims (1)

[Scope of Claims] A character recognition device comprising means for reading characters on a form and storing binary character pattern data, and means for extracting the outline of the character pattern from the data, the extraction means means for determining the direction of the line segment from the contour data of the character pattern extracted by the method; and means for calculating a feature amount consisting of the relative position of the contour line segment from a predetermined reference point for each of the determined directions; and a means for comparing the feature quantity calculated by the calculation means with the feature quantity calculated by the calculation means, and the character is recognized by identifying the distribution position of the line segments of the input character pattern according to the direction of the line segments. A character recognition method that uses