JP3444726B2 - Character processor - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device, and more particularly to a character processing device for extracting skeleton data which is a frame of a character from a character pattern represented by contour data.

[0002]

2. Description of the Related Art Here, terms used in this specification will be described. First of all, the outline has a certain thickness (thickness) in the strokes when looking closely at the typeface such as Mincho or Gothic in terms of the characters that are output.
The edge of the stroke that corresponds to. The skeleton represents the shape of a character and does not have any particular thickness, but the outline is represented by one line (= core line), and at least it can be distinguished what kind of character it is. Say. The constituent element of the character pattern refers to each stroke for forming a character, and is the shape of each stroke in the stroke order.

A contour point is a point on the edge line of a stroke and indicates a point on the contour. As a method of expressing the contour, the control point is a point (coordinates) provided on the contour curve or other than on the contour curve in order to approximate the contour by curve and express the contour.

An input / output vector is a direction vector seen from a control point immediately before (a relationship before and after the control point is arbitrary) at a certain control point. The direction vector that looks at the next control point is called the output vector. Therefore, the input / output vectors of a certain control point on the smooth curve match. The input / output angle is the difference between the angle of the output vector at a certain control point and the angle of the input vector at the next control point.

The striking portion is a portion where the entry of the brush at the first stroke portion is expressed when a character is written with the brush. It refers to the angle of the incoming Chikarabe vector control point located in Chikarabe vector and back out of the control point located in front of the control points in the point angle. The end point is a control point having a large point angle. The start point and the end point are determined by the relative positions of the two end point candidates.

[0006] A joint is a boundary point in the skeleton data extracted from the contour data of one component, which is located in the middle portion of the skeleton data and can be treated as skeleton data independent of each other with the middle portion sandwiched therebetween. For example, in a character pattern represented by a contour, it indicates a portion where the curvature (bent) of a curve is remarkable, or a portion where a stroke of a brush is formed in one character stroke and which has a characteristic shape in the contour data. A semi-joint is similar to a joint and has the same properties as a joint, although the skeletal data connecting the skeletal points does not have continuity.

Conventionally, in electronic equipment and computer terminals,
For display and printing, character patterns such as characters and symbols were previously stored in the memory as outline data called outline font and were read out from the memory. For example, about 7,000 Japanese kanji characters have all the outlines of one typeface. An extremely large amount of data is required just to store the data. Therefore, recently, a method has been adopted in which the constituent elements of the character pattern are stored in advance as skeleton data and the character pattern is constructed based on the skeleton data. According to this method, the amount of data to be stored can be extremely reduced. Therefore, a character processing device that extracts the skeleton data from the contour data of the character pattern is attracting attention.

Generally, in a character processing device for extracting skeleton data from contour data of a character pattern, a center line which is a skeleton of the character pattern is extracted as skeleton data. Specifically, a plurality of points at equal distances from both sides of the contour data of the character pattern are obtained, and these points are sequentially connected to form a center line. At this time, the minimum number of points is determined to form the center line, and this is the skeleton data. However, this method has a problem that the processing speed is slow because points at equal distances are obtained from both sides of the contour data. Also,
There is also a problem that skeleton data is erroneously extracted in a complicated character pattern.

As a device for solving such a problem, there is a stroke extracting device disclosed in Japanese Patent Laid-Open No. 2-244390. Here, after dividing the contour data into unit line segments, the shortest distance between the end point of the shortest unit line segment and the other unit line segment is assumed, and the sum of the two line segments is the smallest and opposite to each other. The unit line segment is detected and the center line (unit skeleton data) between these two unit line segments is extracted. Next, mutual correction between the unit skeleton data is performed. Finally, the unit skeleton data which can be connected is connected, and the final skeleton data is extracted.

[0010]

However, in the method disclosed in the above-mentioned publication, since the contour data is divided into unit line segments, the unit skeleton to be extracted is divided at the protruding portion and the intersecting portion. Therefore, the unit skeleton end points of the unit skeleton are moved and corrected, and the extracted unit skeleton data is converted or moved. Further, of the divided unit skeleton data, the portion that can be connected on a substantially straight line is corrected and sequentially connected to generate the skeleton data. Therefore, the skeleton end points of the skeleton data are moved, and the skeleton data is different from the originally extracted skeleton data.

Further, when various types of typefaces having a common skeleton are generated on the basis of the extracted skeleton data, various characteristic parts such as the writing portion and the bent portion of the brush of the generated character are changed. It is not possible to generate high-quality contour data with various changes. For example, it was not possible to generate a high-quality start or end decoration of a character pattern such as Mincho or Shojo typeface.

Therefore, an object of the present invention is to provide a character processing apparatus which can extract skeleton data which is a skeleton of a character from a character pattern represented by contour data without correcting it.

Another object of the present invention is to provide a character processing device capable of extracting high quality skeleton data which is a skeleton of a character from a character pattern represented by contour data.

Still another object of the present invention is a character processing apparatus capable of extracting skeleton data which can be used for synthesizing a typeface having subtle changes or emphasized changes when synthesizing various typefaces from skeleton data. Is to provide.

[0015]

According to a first aspect of the present invention, there is provided a character processing apparatus, wherein a contour line of a character pattern is approximated by a curve or a straight line defined with a control point as a reference, based on contour data. And a storage unit for storing the contour data of the character pattern, the contour data of the character pattern is read from the storage unit, and end point information is added to the control point, which is the end point, for each constituent element of the character pattern. End point information adding means for doing, the control point corresponding to the bent portion of the contour shape of the character pattern of the contour data to which the end point information is added, the bending point information adding means for adding the bending point information, and the end point information, Skeleton extracting means for extracting the skeleton data of the character pattern from the contour data of each component of the character pattern to which the bending point information is added, end point information, and the bending point And distribution, and a skeleton data, and an output means for outputting as a scaffold information of the character pattern.

The character processing apparatus according to claim 1 is configured as described above, and the skeleton information of the character pattern is obtained for each constituent element of the character pattern by the end point information adding means, the bending point information adding means, and the skeleton extracting means. The obtained end point information, bending point information and skeleton data are obtained. Therefore,
It is possible to obtain accurate skeleton information that is more faithful to the original shape without correcting the skeleton information of the character pattern.

Further, the bending point information adding means has a characteristic or partial shape such as the shape of the brush of the original contour data and the curvature of the driving portion or the curve is large (the contour data is largely bent). It is possible to obtain high-quality skeletal information with bending point information added to.

According to a second aspect of the present invention, in the character processing apparatus according to the first aspect, the skeleton extracting unit of the apparatus according to the first aspect uses the joint data from the contour data for each component of the character pattern to which the end point information and the bending point information are added. It is configured to include means for extracting skeleton data that can have

The character processing device according to the second aspect is configured as described above, and the skeleton data capable of having a joint is obtained from the contour data for each constituent element of the character pattern to which the end point information and the bending point information are added. Is extracted. Therefore,
While dividing one skeleton data by joints to reduce the number of element shapes due to the combination of constituent elements, it is possible to synthesize various typefaces having subtle changes from one skeleton data.

In the character processing device according to a third aspect, the skeleton extracting means of the device according to the first aspect uses the half data from the contour data for each constituent element of the character pattern to which the end point information and the bending point information are added. It is configured to include means for extracting skeletal data that can have joints.

The character processing device according to claim 3 is configured as described above, and a skeleton capable of having a half joint based on the contour data for each constituent element of the character pattern to which the end point information and the bending point information are added. Data is extracted. Therefore, it is possible to reduce the amount of data while maintaining the high quality that one type of skeleton data can synthesize various typefaces having subtle changes as in the case of skeleton data that can have joints.

The character processing apparatus of claim 4, bending point information adding means character processing equipment according to any one of claims 1 3, target control points on the outline data of the character pattern bending point whether the bending point, and determining means for determining based on Saga示 to angle of the input vector of the target control point and the next control point, the determined control points to be bending points And means for adding information.

The character processing device according to claim 4 is configured as described above, and the bending point information adding means determines whether to add the bending point information according to the degree of curvature of the curve of the contour data of the character pattern. As a result, it becomes easy to judge whether or not to add the information about the joint.

[0024]

[0025]

[0026]

[0027]

A character processing device according to claim 5 is
The character processing device according to any one of 1 to 4 is further configured to include a bending point information removing unit for removing the above-described added bending point information.

The character processing device according to the fifth aspect is configured as described above and can remove the bending point information added by the bending point information removing means. Therefore, the bending point information of the contour data can be removed as necessary. By eliminating the bending point information and extracting the skeleton data, it is possible to obtain skeleton data closer to the original shape.

[0030]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a functional block diagram of a character processing device according to an embodiment of the present invention. In the figure, the character processing device includes a font memory 10, an end point addition unit 20, a bending point addition unit 30, a bending point removal unit 31, a program memory (not shown), a work memory 42, and a skeleton extraction unit 4.
1 including a CPU 40, an input unit 50 such as a keyboard, an input buffer 60 for temporarily storing input data from the input unit 50, an output buffer 70 for temporarily storing data generated as a result of processing by the CPU 40, and a bit of data in the output buffer 70. Bitmap memory 80 for developing into a map
And an output unit 90 for displaying or printing image data based on the data output from the bitmap memory 80.

FIG. 2 is a diagram showing a configuration example of contour data of a character pattern in the font memory 10 shown in FIG. FIG. 3 is a diagram showing a configuration example of the extracted skeleton data temporarily stored in the work memory 42 of FIG.

4, 5 and 6 are respectively the end point adding section 20, the bending point adding section 30 and the skeleton extracting section 4 of FIG.
It is a figure which shows the structural example of 1.

The end point adding section 20 of FIG. 4 adds the end point information of the contour data representing the start point and the end point for each constituent element of the character pattern to the control points on the contour data of the character pattern, in order to add the control point extracting section 201. , Point angle recognition unit 2
02, an end point candidate detection unit 203, and a control point relative position recognition unit 204.

The inflection point adding section 30 in FIG. 5 is a control point extracting section for adding inflection point information to control points on the contour data of a character pattern where the contour shape of the character pattern corresponds to the inflection point. 301, input / output angle recognition unit 302, inner bending start point detection unit 303, inner bending end point detection unit 30
4, an outer bending end point detection unit 305, an outer bending start point detection unit 306, and a first stroke bending point detection unit 307.

The bending point removing unit 31 removes the bending point information added to the control points by the bending point adding unit 30 as desired.

The CPU 40 is for performing all the operations for processing the contour data of the character pattern, and has the skeleton extracting section 41 of FIG. The skeleton extracting unit 41 extracts the skeleton information from the contour data of the character pattern, the component extracting unit 411, the contour dividing unit 412, the skeleton point detecting unit 413, the joint detecting unit 414, the skeleton point connecting unit 415, and the half joint detecting unit. 416 is included.

The input section 50 is for inputting a character code or a typeface code of a character pattern to be processed, and for inputting an instruction for removing the extracted bending point information.

Here, a method of approximating a curve of the contour shape of the character pattern will be described. In this method, (1) a method of expressing with a circular arc that uses a part of the circumference, (2) a method of expressing with a spline curve connected by a smooth curve that passes a specified point, (3) of each curve There is a method in which one unit (= 1 segment) is composed of a start point, an end point and two control points, and the control point expresses the direction of the slope of the curve at the start point and the end point and the bulge of the curve by a Bezier curve. . In the present embodiment, the contour shape of the character pattern is curve-approximated by the method using the cubic Bezier curve in (3) above. However, it is not limited to this.

Next, an example of the configuration of the font memory 10 will be described with reference to FIG. In the font memory 10, a typeface code 101 for distinguishing the typeface of the outline data of the character pattern to be extracted, a character code 102, and a constituent element code 103 for distinguishing the constituent elements of the corresponding character pattern for each character code, Component code 103 (element)
Control point code 10 for indicating control point information belonging to each
4, the control point coordinate data 105, the input / output angle 106 of the control point, and the division attribute 107 indicating whether or not the control point is divided as an end point. The arrangement of the control points forming the contour data is clockwise (clockwise). The division attribute 107 includes end point information 2 indicating the start point and end point of each contour data.
00, bending point information 201, and starting-stroke bending point information 202 are stored.

Next, referring to FIG. 3, the work memory 42
An example of the storage content of the extracted skeleton data temporarily stored in the following will be described. The extracted skeleton data is a typeface code 1 that distinguishes the typeface of the outline data of the extracted character pattern.
11 and a plurality of character codes 1 corresponding to the typeface code 111
12 and. Further, for each character code 112, the number of constituent elements 113 and the number of skeleton points constituting the character pattern are 11
Including 7. A component code 114 corresponding to each of the character codes 112 that distinguishes the components by the corresponding number 113 of components, the number 115 of skeleton points included in one component, and a pointer 116 to a skeleton point code described later. It contains a set of information consisting of and. Further, the skeleton point code 11 of the component code for each component is equal to the number of component elements 113.
8, a skeleton point coordinate data 119, a skeleton point attribute 120, and a skeleton point input / output angle 121.

The skeleton point attribute 120 stores point attributes such as curves, corners, and tangents for expressing skeleton data, and joint information that is a joint or a half joint.

Next, the operation of the character processing device of FIG. 1 will be described with reference to the processing flowchart of the character processing device of FIG.

[0044] First, while adding recognizes and bending point information the angle of the incoming out Chikarabe vector at the control points on the outline data of the character pattern, the procedure for extracting without correcting the skeleton data described. FIG. 8 is a diagram illustrating an example of contour data of a character pattern in the font memory 10 of FIG. First, the character code of the character pattern “A” in FIG. 8 from which the skeleton data is extracted from the keyboard of the input unit 50 (for example, the JIS code “25
22 ") and a font code (for example, font name) are input (S1). The input character code and typeface code are separately stored in the input buffer 60 (S2). According to the character code and the typeface code stored in the input buffer 60, the contour data of the corresponding character pattern "a" in the font memory 10 is read (S3) and transferred to the work memory 42 of the CPU 40 (S4).

From the contour data transferred to the work memory 42, the constituent elements forming the character pattern are taken out one by one (S5). If the extraction order is, for example, the arrangement order of the components of the stored outline data, the outline data (the shape is “F”) of the first stroke of the character pattern “A” is extracted first here. The retrieved data is
In FIG. 2, the character code 102 part is the JIS code “2522” of the character pattern “a”, and the component code 103 is “1” (= shape is “fu”) and “2” (= shape is “no”). ) And the component code 103 is “1” and the control point code 104 of “2”,
It has contour data of control points “A” to “J”. Here, the process for the component code 103 of “1” (= shape is “F”) will be described.

As described above, the end point adding section 20 adds end point information of the start point and the end point to the end points which are the start point and the end point among the contour points of the constituent elements extracted from the character pattern (S6). Here, the procedure of adding the endpoint information by the endpoint adding unit 20 will be described with reference to the processing flow chart of adding the endpoint information of the endpoint adding unit 20 in FIG.

In FIG. 9, the control point extraction unit 201 of the end point addition unit 20 extracts control points from the components extracted in advance from the character pattern (S6a), and the points of the respective control points to be extracted are extracted. The angle is determined by the point angle recognition unit 20.
2 is obtained (S6b). Here, the order of extracting the control points and obtaining the point angles is, for example, the arrangement order in the contour data.

After the point angles have been obtained for all the control points as described above, the two control points having the first and second largest point angles are detected by the end point candidate detection unit 203 as end point candidates (S6c). . Here, in FIG. 8, point D
The point angle at θ is θ1, and the point angle at point F is θ2. Therefore, in FIG. 8, points A and F are end point candidates.

Next, the control point relative position recognition unit 204 determines the start point and the end point based on the relative positions of the two end point candidates (S6d). From the general writing order of the character patterns, it can be determined by utilizing that the starting point (starting stroke part) is located to the left of the end point (ending stroke part) and above the bottom. Therefore, the end point information indicating that the point A is the start point and the point F is the end point is added. The end point information 200 of the start point and the end point added as described above is stored in the division attribute 107 corresponding to the contour point of the contour data of the work memory 42 (S).
7).

At the control points on the contour data of the constituent elements of the work memory 42 in the CPU 40 to which the end point information is added,
The bending point addition unit 30 adds the bending point information 201 (S8). The portion to which the bending point information is added is the shape of the brush of the contour data of the constituent elements of the character pattern, and is the control point on the contour data where the curvature of the curve is tight (large). Here, the bending point information 20 is generated by the bending point adding unit 30.
Regarding the procedure for adding 1, the bending point adding section 30 of FIG.
This will be described with reference to the processing flow chart for adding the bending point information.

However, it is determined whether or not the skeleton data between the start point and the end point of the contour data should be divided by joints.
It will be decided in-house. Therefore, the skeleton data may be contour data of a shape that is not divided by joints, but here, the contour data to which the bending point information 201 is added will be described.

As described above, the control point extraction unit 301 extracts the control points on the contour data of the constituent elements whose end point information 200 is stored in the division attribute 107 one by one (S8).
a), the difference between the input Chikarabe vector control points and the next control point in the process, determined by the output angle recognition unit 302 (S8
b). The control points are taken out in the order in which the contour data are arranged, for example, the point B next to the point A which is the start point. Then, the inner bending start point detection unit 303 detects the contour starting point bent to the left at an angle equal to or larger than a threshold value having a vector difference as an inner bending start point (S8c). this is,
This is because when the contour data has a large curvature in the shape of a curve or the like, this vector difference becomes large and can be detected as an inner bent portion.

FIG. 11 is a diagram for explaining a point angle according to the embodiment of the present invention. In FIG. 11, point G
The input and Chikarabe difference vector is .theta.3 becomes the following: H, a point G and the inner bending start point as a threshold angle or more which is .theta.3. Clockwise from this point G will take out one by one contour point in turn (clockwise) (S8a), determines the difference between incoming Chikarabe vector control points and the next control point during processing (S8b), The inner bending end point detection unit 304 detects a contour start point that bends to the left at an angle equal to or less than a certain threshold as the inner bending end point (S8d).

[0054] Input of 11, the point H and the next point I
Difference .theta.4 next vector, the point H and inner bending end point as an angle of less than a certain threshold .theta.4. Next, the outer bending end point detection unit 305 obtains the control point that passes through the point G and is closest to the intersection of the normal line of the input vector of the point G and the contour data on the opposite side, that is, the outer bending end point (S8).
e). In FIG. 11, it becomes a point E. Then, the control point is the closest to the intersection of the normal line of the input vector of this point and the contour data on the opposite side, that is, the outer bending start point is obtained by the outer bending start point detection unit 306 (S8).
f). In FIG. 11, the point is D.

Therefore, the inward bending start point information and the point H are set to the point G.
The inside bending end point information is added to the point, the outside bending end point information is added to the point E, and the bending point information 201 which is the outside bending start point information is added to the point D. In the bending point adding section 30, the condition for adding the bending point can be changed by setting the angle threshold value in advance (S8). For example, it is possible to determine whether or not to add the bending point information 201 according to the degree of curvature of the curve, and as a result, it is possible to determine whether or not to add the joint information. That is, in the contour, a portion having a particularly large curvature (bent) of the curve can be recognized as the bending point portion, and the bending point information 2 can be recognized as the contour point or the control point.
01 is added. When the bending point information 201 is added,
When recognizing the skeletal information, it is possible to set a condition for adding the joint information and determine whether or not to add the joint information.

Returning to FIG. 7, the added bending point information 20.
1 is stored in the division attribute 107 at the control point of the contour data of the work memory 42 in the CPU 40 (S9).

End point information 200 in the work memory 42
The skeleton information capable of having a joint is extracted by the skeleton extraction unit 41 from the contour data to which the and the bending point information 201 are added (S10). Here, a procedure for extracting the skeleton data capable of having a joint by the skeleton extracting unit 41 will be described with reference to the drawings.

FIG. 12 is a diagram for explaining a first example of the skeleton extracting method by the skeleton extracting unit 41 of FIG. FIG. 13 is a flowchart showing a first example of the skeleton extraction procedure of the skeleton extraction unit 41 of FIG.

First, the component extraction unit 41 of the skeleton extraction unit 41
The contour data of the component is extracted by 1 (S10a).

Next, referring to FIG. 12, the bending point information 201
The contour dividing unit 412 equally divides the contour data between the start point A and the end point F other than between the points D and E and between the points G and H (S10b). The number of divisions is the input unit 5
It is input as 0 and stored in the CPU 40 in advance.

Next, the skeleton point detector 413 obtains skeleton points (S10c). Specifically, point A and point D in FIG.
, Point E and point F, point F and point G, and point H and point A are divided into three parts, respectively, and point P1, point P2, point P3, point P
4, point P5, point P6, point P7, and point P8 are generated. The starting point A is a skeleton point K0, and points P1, P8, and P2
And point P7, point D and point H, point E and point G, point P3 and point P6,
The midpoints of the line segments connecting the points P4 and P5 are the skeleton points K1 and K.
2, K3, K4, K5 and K6, and the end point F is the skeleton point K7.

Next, the joint detecting section 414 obtains a skeleton point to be a joint (S10d). That is, the skeleton point K3 generated from the outer bending start point D and the inner bending end point H
And the closest skeleton point K in the direction of the starting point A from the skeleton point K3
2 and a skeleton point K4 generated from an outer bending end point E and an inner bending start point G, and a straight line connecting a skeleton point K5 closest to the end point F direction from the skeleton point K4. The skeletal points that become joints can be obtained. Therefore, it is possible to extract skeleton information that can have a joint.

Returning to FIG. 7, the skeleton information obtained as described above is stored in the work memory 42 as skeleton data (S11), and the processing for the component extracted this time is completed.

Next, a series of processing (S5 to S1) is performed for all the constituent elements of the contour data for one character read this time.
It is determined whether or not step 1) has been performed (S12), and if the process is completed, the process proceeds to the next process (S13). If the process is not completed, the contour data of the component to be processed next is taken out (S12).
5) The same process is repeated.

The skeleton data of the character pattern for which the series of processing has been completed is output to the output buffer 70 (S1).
3). The skeleton data in the output buffer 70 is transferred to the bitmap memory 80 (S14), the bitmap data in the bitmap memory 80 is transferred to the output unit 90, and the process for the character ends.

FIG. 14 is a diagram for explaining a second example of the skeleton extracting method by the skeleton extracting unit 41 of FIG. FIG. 15 is a flowchart showing a second example of the skeleton extraction procedure of the skeleton extraction unit 41 of FIG. The skeleton extraction procedure described with reference to FIGS. 14 and 15 is an improvement of the skeleton extraction procedure of FIG. 13 described above. The following is a procedure for adding bending point information 201 from the contour data (the shape is “F”) of the first stroke of the character pattern “A” shown in FIG. 8 and extracting two skeleton data with joints. Will be described.

Here, it is possible to extract a plurality of skeleton data from the contour data of one constituent element and to handle them collectively. These skeleton data do not necessarily have continuity as described in FIG. This is because in the process of FIG. 13, the shape of the outline data of the original character pattern and the shape of the extracted skeleton data may be significantly different. For example, the extracted skeleton data may deviate from the contour data in a portion of the contour data of the character pattern where the curvature of the curve is extremely large. This can be dealt with by increasing the number of skeleton points, but the amount of data increases accordingly. Therefore, in the procedure shown in the flowchart of FIG. 15, the data amount is reduced to extract higher quality skeleton data.

A procedure for extracting skeleton data capable of having a half-joint from the contour data (the shape is “F”) of the first stroke of the character pattern “A” shown in FIG. 14 will be described with reference to the flowchart of FIG. .

First, the skeleton extracting unit 4 from the contour data to which various information is added in the work memory 42 of the CPU 40.
The skeleton data which can have a half joint is extracted by 1 (S10 '). Here, a procedure for extracting the skeleton data capable of having a half joint by the skeleton extracting unit 41 will be described below.

First, the component extraction unit 4 of the skeleton extraction unit 41
The control points of the constituent elements are extracted by 11 (S10).
a '). Next, the contour dividing unit 412 equally divides the contour data between the start point A and the end point F between the points D and E to which the bending point information 201 is added and between the points G and H (S10b). ′). Here, the number of divisions is input by the input unit 50 and stored in the CPU 40 in advance.

Next, as shown in FIG. 14, assuming that points A and D, points E and F, points F and G, points H and A are divided into three parts, respectively, and point P1 , Point P2, point P3, point P
4, point P5, point P6, point P7 and point P8 are generated. Then, the skeleton point detector 413 sets the starting point A to the skeleton point K.
0, points P1 and P8, points P2 and P7, points D and H, points E and G, points P3 and P6, and points P4 and P.
The midpoint of the line segment connecting with 5 is the skeleton point K1, K2, K3, K
4, K5, K6, and the skeleton point with the end point F as the skeleton point K7 (S10c ').

Next, the skeleton point K4 generated from the outer bending end point E and the inner bending start point G, and the skeleton point K4 to the end point F.
The closest skeleton point K5 in the direction to the skeleton point connection portion 415
By a straight line (S10d '), and the semi-joint detecting unit 416
Is extended in the direction opposite to the end point F to obtain the intersection with the contour data. This intersection is defined as a half joint (S10e '). Therefore, it is possible to extract skeleton information capable of having a half joint according to the flowchart of FIG.

Here, a straight line connecting the skeleton point K3 generated from the outer bending start point D and the inner bending end point H and the closest skeleton point K2 in the starting point A direction from the skeleton point K3 is set to the start point A. It is also possible to extend in the opposite direction to the half joint at the intersection with the contour data.

Next, a third example of the skeleton extracting method by the skeleton extracting unit 41 will be described with reference to the flowchart shown in FIG. FIG. 16 is a diagram illustrating a third example of the skeleton extracting method by the skeleton extracting unit 41 of FIG. 1.

In the character pattern according to the typeface such as Mincho typeface and regular typeface, the second part of the character pattern "A" shown in FIG.
There is a case where the driving part is seen in the shape of a brush like the contour data (the shape is “No”) of the stroke. The third of this skeleton extraction method
In the example, a procedure for extracting skeleton data capable of having a joint at a characteristic portion having a striking portion in the shape of a brush with contour data of vertical bars of a character pattern is shown. If such a characteristic part is added to the skeleton data as information, when synthesizing a typeface using this skeleton data,
It is possible to create a typeface with a different impression by simply changing the shape of the driving part. Moreover, since the skeleton data can be shared among a plurality of typefaces, an increase in the amount of data can be suppressed.

The case where the vertical bar contour data has a brush-shaped driving portion and the like is described here.
In the case of a horizontal bar, the description is omitted because it is almost the same except that the coordinate axes are changed.

Based on the flowchart shown in FIG. 7, FIG.
A procedure for extracting skeleton data capable of having a joint from the contour data (the shape is “No”) of the second stroke of the character pattern “A” shown in 6 will be described. In this procedure as well, the processing relating to S1 to S7 and S12 to S15 in FIG. 7 is the same as that described above, so the processing from S21 to S24 will be described.

Here, as preprocessing, the end point information 200 is added as in the above (S6), and in FIG. 16, the end point information 200 of the start point is added to the point A and the end point information 200 is added to the point F. To do.

First, since the constituent elements are extracted in advance (S5), the inflection point information 202 is added to the control point by the inflection point adding section 30 (S21). Here, a procedure for adding the starting-stroke bending point information 202 by the bending point adding section 30 will be described.

FIG. 17 is a processing flowchart for adding the first stroke bending point information by the bending point adding section 30 in FIG.

In FIG. 17, first, the control point extraction section 301
The control points of the constituent elements are taken out one by one (S21).
a). The order of the control points to be taken out is, for example, the order of data arrangement from the start point A.

Next, Shihitsu bending point difference is largest contour point of the incoming Chikarabe vector control point of Chikarabe vector and back out of the front of the control points in the following conditions are satisfied control point by the detector 307 Is calculated as the bending point of the first stroke.

The condition of the starting stroke point is to search for a control point having a Y coordinate larger than the middle of the Y-axis height of the component being processed (S
21b). The control points among them are taken out one by one (S21
a). Then, a control point having a Y coordinate smaller than the Y coordinate of the starting point is searched (S21c). In addition, the control points among them are taken out one by one (S21a). Then, a control point having an X coordinate larger than the X coordinate of the starting point is searched for (S21d). In addition, the control points among them are taken out one by one (S21a). A control point whose Y coordinate is smaller than the previous control point and larger than the rear control point is searched for (S21e). In addition, the control point in that is 1
Take out one by one (S21a). A control point in which the control point being processed is curved to the right with respect to the direction of the contour data is searched (S21f). Take out the control points one by one (S2
1a). Yet, the difference between the input Chikarabe vector of control points of Chikarabe vector and back out of the front of the control point is look for the largest contour points (S21c). This control point becomes the starting stroke point (S21).
g). In FIG. 16, the control point B becomes the starting stroke point, and the starting stroke point information 202 is added to the control point B.

The starting-stroke inflection point information 202 is stored in the division attribute 107 of the control point which has become the initial-stroke inflection point as described above (S22).

Next, the work memory 42 in the CPU 40
The skeleton extracting unit 41 extracts the skeleton data from the contour data to which various kinds of information are added. This extraction procedure will be described with reference to a flowchart showing a fourth example of the skeleton extraction procedure by the skeleton extraction unit 41 in FIG. 1 shown in FIG.

In the fourth example of the skeleton extraction procedure, skeleton information capable of having a first stroke joint can be extracted.

First, the control point B in FIG. 16 which is determined to be the first stroke bending point as described above is taken out as a point P1 (S23).
a) The control point P0 is generated at the intersection where the normal line of the input vector of the point P1 is set to the contour data on the opposite side (S23).
b). Then, the skeleton point K1 that serves as a joint is set at the midpoint between the two control points (S23c).

Here, the control point B and the end point F in FIG.
Meanwhile, the contour data between the point F and the control point P0 is equally divided (S23d). The number of divisions is input by the input unit 50 and the CPU 40
It should be stored in advance. In FIG.
It is assumed that the points B and F and the points F and P0 are divided into three parts, and points P2, P3, P4 and P5 are generated. The starting point A is a skeleton point K0, and points P1, P0, and P2
Is the skeleton point K1, K2 and K3, respectively, and the end point F is the skeleton point K.
The skeleton point is determined as 4 (S23e). In addition, this skeleton point K1 becomes a starting stroke joint.

The skeleton information obtained as described above is C
The data is stored in the skeleton data of the work memory 42 in the PU 40 (S24). The subsequent processing is S12 to S of FIG.
It is similar to the process of 15.

Next, the procedure for extracting the skeleton data by eliminating the bending point information of the contour data will be described. In the skeleton extraction procedure of the first to fourth examples described above, the skeleton data is extracted by adding the bending point information, but if the skeleton data is extracted without the bending point information of the contour data, the original skeleton data is extracted. Skeleton data close to the shape can be obtained. Here, a procedure for extracting the skeleton data by once extracting the skeleton data by the procedure of S21 to S24 in FIG. 7 and then arbitrarily removing the bending point information in the extracted skeleton data will be described. Referring to the flowchart of FIG. 7, the bending points are removed from the skeleton data obtained in S23 in the contour data (the shape is “No”) of the second stroke of the character pattern “A” shown in FIG. A procedure for newly extracting skeleton data will be described. Since the processing relating to S1 to S7, S21 to S24, and S12 to S15 in FIG. 7 is the same as that described above, the description thereof will be omitted, and the skeleton data extraction procedure of S31 to S36 will be described.

Here, the contour data in which the end point information 200 is stored in the division attribute 107 is extracted from the work memory 42 (S22). Further, skeleton data having skeleton information is also stored.

First, a control point code (for example, B) having a division attribute 107 (corresponding to point B in this case) in which bending point information to be removed on the contour data extracted in advance from the input unit 50 is stored is input. Yes (S31). The input control point code is stored in the input buffer 60 (S3
2) is transferred to the work memory 42 in the CPU 40 (S33).

From the contour data extracted in advance in S22, S
The division attribute 107 of the control point code 104 input at 31
The bending point information stored in is removed by the bending point removing unit 31 (S34). That is, the division attribute 10 of the control point B
The starting-stroke bending point information 202 is stored in 7, but the starting-point bending point information 202 is cleared by the bending-point removing unit 31, and the content of the division attribute 107 becomes blank. Then, the control point B becomes the same as the control point on the normal contour data, and is extracted as the conventional skeleton data having no joint.

The skeleton extracting unit 41 extracts skeleton data having no joint from the contour data in the work memory 42 of the CPU 40 from which the bending point information is removed (S35).

First, the contour data between the starting point A and the ending point F and between the ending point F and the starting point A in FIG. 16 are equally divided. Here, it is assumed that the contour data is divided into four. P1 ', P2', P3 'are generated between the start point A and the end point F, and P4', P5 ', P6' (not shown) are generated between the end point F and the start point A. The starting point A is a skeleton point K0, and points P1 ', P6', and P2 '
And the point P5 ', and the midpoints of the line segments connecting the points P3' and P4 'are skeleton points K1', K2 ', K3' (not shown), the end point F is the skeleton point K4, and each skeleton point is Ask.

The skeleton information composed of the skeleton points obtained as described above is stored in the skeleton data of the work memory 42 (S36). The subsequent process continues to S12 of FIG.

Although the control point from which the bending point information should be removed is selected in S31 of FIG. 7, the user can arbitrarily select the control point to which the bending point information is added as the selection criterion. You may make it select using the input part 50, such as a keyboard.

Next, the procedure for extracting the skeleton data may be different for each constituent element of one character pattern. That is, the process (the process of S1 to S11 of FIG. 7) of extracting the skeleton data having the joint in the contour data (the shape is “F”) of the first stroke of the character pattern “A” shown in FIG. 8 is performed. Next, a process (S21 to S24) of extracting higher-quality skeleton data having the first stroke joint from the contour data (the shape is “No”) of the second stroke of the character pattern “A” is performed. In other words, since the outline data of the second stroke of the character pattern "A" includes the entry of the brush of the first stroke portion, it is preferable to extract higher-quality skeleton data having information about the first stroke joint.

When extracting the skeleton information of each constituent element of the character pattern, which of the processing steps S8 to S11 and S21 to S24 in FIG. The user may arbitrarily decide whether or not to perform it, and process based on the decision.

[Brief description of drawings]

FIG. 1 is a functional block diagram of a character processing device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration example of contour data of a character pattern in the font memory 10 of FIG.

3 is a diagram showing a configuration example of extracted skeleton data temporarily stored in a work memory 42 in FIG.

FIG. 4 is a diagram showing a configuration example of an end point adding section 20 in FIG.

5 is a diagram showing a configuration example of a bending point adding section 30 in FIG.

6 is a diagram showing a configuration example of a skeleton extraction unit 41 in FIG.

7 is a process flowchart of the character processing device of FIG.

8 is a diagram illustrating an example of contour data of a character pattern in the font memory 10 of FIG.

9 is a processing flowchart for adding endpoint information by an endpoint adding unit 20 in FIG. 1. FIG.

FIG. 10 is a processing flowchart for adding bending point information in the bending point adding section 30 of FIG.

FIG. 11 is a diagram for explaining a point angle according to the embodiment of the present invention.

FIG. 12 is a diagram illustrating a first example of a skeleton extracting method by a skeleton extracting unit 41 of FIG. 1.

13 is a first skeleton extraction procedure of a skeleton extraction unit 41 of FIG.
It is a flowchart which shows an example.

FIG. 14 is a diagram illustrating a second example of a skeleton extracting method by the skeleton extracting unit 41 of FIG. 1.

15 is a second skeleton extraction procedure of the skeleton extraction unit 41 of FIG.
It is a flowchart which shows an example.

16 is a diagram illustrating a third example of a skeleton extracting method by the skeleton extracting unit 41 of FIG. 1.

FIG. 17 is a processing flowchart of adding the inflection point information of the first stroke by the inflection point adding section 30 of FIG. 1;

FIG. 18 is a flowchart showing a fourth example of a skeleton extraction procedure by the skeleton extraction unit 41 of FIG. 1.

[Explanation of symbols]

Memory for 10 fonts 20 End point addition part 30 Bending point addition part 31 Bending point removal section 40 CPU 41 Skeleton extractor 42 Work memory 50 Input section 107 split attribute 200 End point information 201 Bending point information 202 Inflection point information In the drawings, the same reference numerals indicate the same or corresponding parts.

─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A 1-166967 (JP, A) JP-A 10-39851 (JP, A) JP-A 6-131499 (JP, A) JP-A 7- 140958 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G09G 5/24

Claims (5)

(57) [Claims] 1. A character processing device for processing a character on the basis of contour data obtained by approximating a contour line of a character pattern by a curve or a straight line defined with reference to a control point, the contour data of the character pattern. Storage means for storing, and reading the contour data of the character pattern from the storage means,
For each constituent element of the character pattern, end point information adding means for adding end point information to the control point that is an end point, and the control point corresponding to the bent portion of the contour shape of the character pattern of the contour data with the end point information added, Bending point information adding means for adding the bending point information, and skeleton extracting means for extracting the skeleton data of the character pattern from the contour data of each constituent element of the character pattern to which the end point information and the bending point information are added, A character processing device, comprising: output means for outputting the end point information, the bending point information, and the skeleton data as skeleton information of a character pattern. 2. The skeleton extracting means includes means for extracting skeleton data capable of having a joint from contour data for each constituent element of a character pattern to which end point information and bending point information are added. The character processing device according to claim 1. 3. The skeleton extracting means includes means for extracting skeleton data capable of having a half joint from the contour data of each constituent element of the character pattern to which the end point information and the bending point information are added. The character processing device according to claim 1. 4. The bending point information adding means determines whether or not the target control point on the contour data of the character pattern is a bending point, by determining the difference between the input vector of the target control point and the input vector of the next control point.
Determination means for determining based on the shown angle formed, said and means for adding the inflection point information to the control point which is determined to be bending points by determining means, any one of claims 1 to 3 The character processing device according to item 1. 5. further comprising a bending point information removing means for removing the added bending point information, claims 1 to 4 Neu
The character processing device according to item 1 .