JP3111811B2 - Apparatus and method for adjusting character spacing - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for adjusting the distance between characters in a character string to be printed in a computer typesetting system, a phototypesetting machine, a word processor or the like so that the character string is easy to read and beautiful.

[0002] In this specification, "letter" means will,
Includes all types of letters, symbols, signs, numbers, and figures used to express thoughts, emotions, etc. Representative characters may include kanji (including kanji), Japanese hiragana and katakana, Hangul, alphabet, Greek, Arabic, Roman, hyphen, comma, colon, etc.

[0003]

2. Description of the Related Art Adjustment of character spacing in a character string arranged horizontally or vertically is also called kerning, and is one of important tasks in typesetting for printing. Kerning must be performed appropriately in sentences containing kanji, hiragana, and katakana, especially in headlines and advertisements in leaflets, since the impression given to those who read them is very important.

Heretofore, character spacing adjustment work has been performed by skilled workers. Since this work cannot be performed unless the person has sufficient experience, there is a problem that work efficiency is poor.

[0005] As a device for automatically adjusting the character spacing, there is a device proposed in Japanese Patent Application Publication No. 3-244542. A virtual area (zone) having a constant width is set outside the character along the outline. The space between two adjacent characters is determined so that a part of the outer edge of the virtual area is in contact with each other. Alternatively, the character spacing is adjusted so that the overlapping area of the virtual regions of two adjacent characters maintains a constant value.

However, in the automatic character spacing adjustment, the character spacing is adjusted based only on the outline of the character, and other elements of the character are not considered at all. For example, the size and complexity of a character's face, its relevance to adjacent characters, etc. are out of consideration, despite being important factors affecting character spacing.

[0007]

DISCLOSURE OF THE INVENTION The present invention provides an apparatus and a method for automatically adjusting a character spacing in consideration of the impression given by a character (for example, the area of a character surface) and the relationship between adjacent characters. With the goal.

Another object of the present invention is to make it possible to perform automatic character spacing adjustment in consideration of the position of a particularly small character in consideration of the size of the character.

[0009] The present invention further enables the overall balance to be achieved in consideration of the length of the arrangement of a plurality of characters.

[0010] The character spacing adjusting apparatus according to the present invention generates a first characteristic amount for each character in character dot data of a plurality of characters developed on a memory. A second feature value for generating, for each character, a second feature value representing a relationship between the character and a character adjacent thereto in the character dot data of a plurality of characters expanded on the memory; Generating means, and character spacing fuzzy inference means for inferring an appropriate character spacing for each character by fuzzy inference by applying the first feature amount and the second feature amount to knowledge on character spacing adjustment. Is what it is.

Representative examples of the most basic first feature amount include:
There is character density data representing the area of the character face.

Representative examples of the most basic second feature amount include:
In the state where the target character and the adjacent character are in contact with each other, the overlapping character area data indicating the area of the character surface existing in the overlapping area of the two characters, or in the state where the target character and the adjacent character are in contact with each other, There is blank area data representing the area of a blank generated between characters.

According to the present invention, an appropriate character spacing is inferred based on the characteristic amount of the character itself and the characteristic amount generated between two adjacent characters. In sentences with mixed kanji, hiragana, and katakana, two adjacent
The number of single letter combinations is extremely large. It is almost impossible to determine the optimum intervals in advance for all such combinations. According to the present invention, each time a character string is formed and a combination is determined, an appropriate character interval is determined in consideration of not only the feature amount of each character but also the mutual relevance of two adjacent characters.

In the fuzzy inference of the character spacing, it is preferable that data representing the degree of the character spacing that the operator desires and inputs is also considered.

More preferably, first ratio data relating to the ratio between the minimum character interval and the character width representing the character interval when the target character and the adjacent character are in contact with each other, and the target character and the adjacent character are in contact with each other. Ratio data on the ratio between the minimum character spacing and the character height representing the character spacing when the character is focused, and the complexity of overlapping both characters in the state where the character of interest and the character adjacent thereto are in contact. (The number is expressed by the number of character portions that bite into the alternated area and the degree of bite (the length of the bite portion))
Is considered in the fuzzy inference of character spacing.
By performing the fuzzy inference processing in consideration of these, the characters and the characteristics between the characters are more finely reflected in the character spacing.

When the adjacent character is a small character,
Preferably, the above-described character spacing inference is performed for each of the position expanded into the dot data of the small character and the position shifted by a predetermined number of dots up and down.
The smallest one of a plurality of character intervals obtained by this character interval inference is determined to be an appropriate character interval. This causes a small character to accompany the preceding (or following) character at the narrowest interval.
This is an interval adjustment in consideration of the mode of use of small characters.

Once the appropriate spacing for all characters in a given character string or at least one line of character strings has been determined by the above inference, spacing correction that takes into account the overall balance in the character string is: It is preferably performed as follows.

The length data relating to the length of the arrangement of a plurality of characters including the character of interest, arranged in a line with the character spacing obtained by the above-described character spacing inference, is generated.
By applying the generated length data to knowledge about character spacing correction given in advance, the amount of spacing correction is fuzzy inferred for each character of interest. The final character spacing is generated by adding the spacing correction amount obtained by the fuzzy inference to the character spacing obtained by the character spacing fuzzy inference.

The present invention also provides a character spacing adjusting method.

In this character spacing adjusting method, character dot data representing a character constituting a specified character string is expanded in a memory, and the character dot data of a plurality of characters expanded in the memory is used as the character dot data. For each character, a first feature quantity relating to the character is generated, and in the character dot data of the plurality of characters developed on the memory, the relationship between the character and a character adjacent thereto is represented for each character. By generating a second feature quantity and applying the first and second feature quantities to knowledge on character spacing adjustment, an appropriate character spacing is fuzzy inferred for each character.

[0021]

The character spacing adjusting apparatus and method according to the present invention can be applied to a computer typesetting system, a phototypesetting machine, a word processor, and the like. In these systems or machines, character dot data representing characters constituting a specified character string is stored in a memory at a specified character size and with a character spacing corresponding to the obtained character spacing. Expanded, and the expanded character dot
Characters will be printed using the data.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an outline of the overall configuration of a computer typesetting system.

The overall operation of the computer typesetting system is performed by the CPU 11
Is governed by Like a normal computer system, this computer typesetting system uses a ROM as a storage device.
12, RAM 13 and hard disk drive 14,
A keyboard 15 is provided as an input device, and a CRT 16 and a printer 17 are provided as output devices.

The ROM 12 generally stores fixed data. The RAM 13 stores variable data and is used as a work area. It is used to store large-capacity data on the hard disk of the hard disk drive device 14. Instead of the hard disk drive 14,
Alternatively or additionally, a floppy disk drive device can be provided. For example, a program for controlling the CPU 11, a fuzzy inference rule and a membership function for adjusting the character spacing, which will be described later, are stored in the ROM 1.
2, stored in RAM 13, hard disk or floppy disk. Data representing the document to be printed is stored on a hard disk or floppy disk.

The keyboard 15 is used to input data representing characters constituting a document and various commands. A mouse, a light pen, and the like are further provided as necessary as input devices. The CRT 16 displays an input document, a character string with adjusted character spacing, a graphic, and the like.

Printer 17 is most commonly an electrophotographic printer (such as a laser printer). Inkjet /
Printers and thermal transfer printers are also used as needed.
The typesetting result is printed by the printer 17. The data constituting the character is stored in advance in a storage device built in the printer 17 in the form of an outline vector font or a bit map font. font·
The data may be stored on a hard disk.

The computer typesetting system has many functions, such as a function of inputting document data, a function of editing an input document, a function of adjusting character spacing in a document, and a function of printing an edited document with adjusted character spacing. Has a function.
Among these functions, the part that realizes the character spacing adjustment function is a character spacing adjustment device. Computer typesetting systems can be thought of as being assembled by devices that perform many of these types of functions.

FIG. 2 shows the operation flow of the character spacing adjusting device in the computer typesetting system. A series of processing shown in FIG. 2 is mainly executed under the control of the CPU 11.
Hereinafter, description will be given according to the flow of this operation.

It is assumed that data (character code string) representing a document to be printed has already been input and stored on the hard disk. A character code string (for example, one line or a plurality of lines) is read from the hard disk, and dot data (bit map character data) representing a character specified by the character code is read according to the font data. It is developed in the work area of the RAM 13 (step 21). FIG. 3 shows an example of the expanded bit map character data.

Generally, kanji, Japanese hiragana and katakana,
Korean Hangul characters are arranged not only horizontally but also vertically to compose sentences. Therefore, a frame (or a body, or a virtual frame, or a virtual body) F which is a basic design of these characters is a square. The size of the frame F of the bit map character data developed for adjusting the character spacing is fixedly determined in advance (for example, 256 × 256 dots (pixels)). In FIG. 3, the frame F is 24 × 24 dots for convenience of drawing. The width of the frame F is indicated by FX and the height is indicated by FY (in this embodiment, FX is
= FY).

Within these frames F, characters are represented by dots. For convenience of explanation, the dots forming the character surface are referred to as black dots, and the dots representing portions other than the character surface are referred to as white dots. As in the most common screen scanning, the horizontal direction (horizontal direction in FIG. 3) and the vertical direction (vertical direction in FIG. 3) are determined for the bit map character data.

In FIG. 3, five types of characters (Japanese katakana) are arranged in the horizontal direction. Although the present invention is of course applicable to the adjustment of the character spacing arranged in the vertical direction, the adjustment of the spacing of the characters arranged in the horizontal direction will be described consistently. In the development of the bit map character data in step 21, a plurality of frames are arranged in contact with each other. I, (i + 1), (i + 2), (i + 3) and (i + 3)
A sequential number of (i + 4) is assigned.

Each character and the inter-character features in these character data strings subjected to bit map development are extracted, calculated or generated (step 22).

The feature amount of each character includes a character (character surface) width C
There are W (i), the height of the character (face) CH (i), the left margin CSL (i), and the right margin CSR (i) of the character. Here, (i) indicates that it is a feature amount for the character i. The character i (the i-th character) is referred to as a target character as needed.

FIG. 4 shows two characters extracted from the bit map character data string shown in FIG. 3 and enlarged. Referring to this figure, the width CW of the character (character surface) is the length between the leftmost dot and the rightmost dot of the black dots constituting the character (character surface) (the number of dots: the leftmost dot and the rightmost dot). (Including dots). For example, in the i-th character, the character width CW (i) is represented by the number of dots from the leftmost black dot e3 to the rightmost black dot e4 (CW
(i) = 24). Similarly, at the (i + 1) th character,
The character width CW (i + 1) is the number of dots between the leftmost black dot e5 and the rightmost black dot e6 (including the dots e5 and e6) (CW (i + 1) = 18).

The height CH of the character (character surface) is the length (the number of dots; including the upper dot and the lower dot) between the upper dot and the lower dot of the black dots constituting the character (character surface). ). For example, in the i-th character, the character height CH (i) is the number of dots from the upper black dot e1 to the lower black dot e2 (CH (i) = 20).

The left margin CSL of a character is represented by the number of dots between the vertical line on the left side of the character frame and the leftmost black dot of the character face in the character frame (excluding the leftmost black dot). The right margin CSR of a character is represented by the number of dots between the vertical line on the right side of the character frame and the rightmost black dot of the character face in the character frame (excluding the rightmost black dot). For example, in the i-th character, CSL (i) = 0 because the leftmost black dot e3 is in contact with the left vertical frame line. Similarly, CSR (i) = 0. For the (i + 1) -th character, CSL (i + 1) = 3 and CSR (i + 1) = 3.

For each of the five characters shown in FIG.
FIG. 6 shows the feature amounts CW, CH, CSL, and CSR.

Basic character features and inter-character features used for fuzzy inference of appropriate character spacing include character density DS, overlapping character area SB and blank area SW.

The character density DS is a characteristic amount of a character, and is represented by the number of black dots constituting a character surface in one frame. For example, the density DS (i) of the i-th character is 90,
The density DS (i + 1) of the (i + 1) -th character is 45.

The overlapping character area SB and the blank area SW are characteristic amounts between the target character and the character adjacent to the right side. The overlapping face area SB (i) and blank area SW (i) for the i-th character are generated in association with the i-th character and the (i + 1) -th character.

As shown in FIG. 5, the (i + 1) -th character is moved to the left, and is brought closer until a part of the (i + 1) -th face touches a part of the i-th face (i The (i + 1) th character
May be closer to the second letter). In the example shown in FIG.
The black dot e5 at the left end of the (+1) th character is in contact with the black dot e2 at the lower end of the i-th character (generally, the left end of the (i + 1) -th character is a part of the i-th character. (It does not necessarily touch, and part of the (i + 1) -th character does not necessarily touch the lower end of the i-th character.)

As described above, when the i-th character and the (i + 1) -th character are moved toward each other, the width CW (i) of the i-th character and the width of the (i + 1) -th character And CW (i + 1). The sum of the number of black dots of the i-th character and the number of black dots of the (i + 1) -th character existing in the area where the character widths overlap is the overlapping character surface area SB (i ).

In FIG. 5, the character face (black dot) of the i-th character and the character face (black dot) of the (i + 1) -th character are indicated by double (crossed) hatching. Of these, the overlapping character area SB (i) is blacked out. In this example, SB (i) = 37. The blank area SW (i) described below is indicated by single hatching. In this example, SW (i) = 129.

The blank area SW (i) is defined as follows: when a part of the i-th character is in contact with a part of the (i + 1) -th character,
This is the number of white dots sandwiched in the horizontal direction between the character face (black dot) of the i-th character and the character face (black dot) of the (i + 1) -th character.

After obtaining such a characteristic amount, an appropriate character spacing is inferred using the obtained characteristic amount (step 2).
3). In the basic fuzzy inference and the auxiliary fuzzy inference to be described later, it is sufficient if the feature amount of the target character and the character adjacent thereto is obtained. However, in the re-adjustment (correction) process described later, in order to adjust the overall balance between a plurality of characters, a feature amount and a character interval of a plurality of characters near the target character are required. Therefore, the calculation of the characteristic amount is preferably performed on at least one line of characters developed in the bit map character data.

The character interval CS (i) for the i-th character obtained by the inference operation is, as shown in FIGS. 7 to 9, the black dot e4 at the right end of the i-th character and the (i + 1) -th character. Is the distance (the number of dots) between the character and the leftmost black dot e5.

FIG. 7 shows a state where the data is developed into bit map data in step 21. The character spacing at this time becomes the maximum. This maximum character spacing is set as CSmax (i). Character spacing adjustment is performed in a direction to reduce the spacing between two characters, and the adjusted character spacing does not become larger than the maximum character spacing. The maximum character spacing CSmax (i) is i
It is represented by the sum of the right margin CSR (i) of the th character and the left margin CSL (i + 1) of the (i + 1) th character. That is, CSmax (i) = CSR (i) + CSL (i + 1) (1)

FIG. 8 shows a state in which the (i + 1) -th character is moved to the left and a part of the character width of both characters overlaps.
Thus, when the character widths of adjacent characters overlap, the character space CS (i) takes a negative value.

FIG. 9 shows a state where the (i + 1) -th character is moved until a part of the (i + 1) -th character touches a part of the i-th character. The character spacing CS (i) then becomes minimum. Write the minimum character spacing as CSmin (i). The minimum character spacing is generally zero or a negative value.

For fuzzy inference, the i-th character and (i
The character spacing degree CSD (i) representing the degree of the character spacing between the (+1) th character and the character is defined by the following equation.

CSD (i) = [CS (i) −CSmin (i)] / [CSmax (i) −CSmin (i)] (2)

CSmin (i) ≦ CS (i) ≦ CSmax (i) (3)

The degree of character spacing CSD (i) is 0-1.
Take a continuous value between.

In the fuzzy inference shown below, the degree of character spacing CSD (i) is obtained, and finally the character spacing CS (i) is obtained according to equation (2).

Fuzzy inference is basically based on character density DS
(i), the overlapping character area SB (i) and the blank area SW (i) are used as antecedent variables. The consequent part is the character spacing open condition CSD (i).

An example of the basic rule group is as follows.

(I) Rules for character density DS IF DS (i) is large, AND DS (i + 1) is large, TH
EN CSD (i) is large IF DS (i) is medium, AND DS (i + 1) is medium, THEN CSD (i) is medium IF DS (i) is small, AND DS (i + 1) Is small, TH
EN CSD (i) is small

A character having a high density has a large number of black dots. Therefore, when such two characters are arranged adjacent to each other, if the character spacing is reduced, it becomes difficult to read, so that the character spacing is increased to make it easier to read. Conversely, when characters with low density are arranged side by side, it is easier to read if these characters are closer to each other.

(II) Rule group for overlapped character area SB IF SB (i) is large, THEN CSD (i) is large IF SB (i) is medium, THEN CSD (i) is medium IF SB (i) Is small, THEN CSD (i) is small

When the overlapping character area is large, the character spacing is widened to make it easier to read, and when the overlapping character area is small, the character spacing is made small to make it easier to read.

(III) Rule group concerning blank area SW IF SW (i) is large, THEN CSD (i) is small, IF SW (i) is medium, THEN CSD (i) is medium IF SW (i). Small, THEN CSD (i) is large

If the blank area is large, the character spacing is perceived to be wide. Therefore, it is easier to read when the character spacing is reduced. Conversely, when the blank area is small, as in the case where the overlapping character area is large, it is easier to read when the character spacing is widened.

An example of the antecedent membership function used in such a basic rule group is shown in FIGS. 10 shows a membership function for the character density DS, FIG. 11 shows a membership function for the overlapping character area SB, and FIG. 12 shows a membership function for the blank area.

FIG. 13 shows the membership function related to the character spacing degree CSD in the consequent part.

Fuzzy inference using the above rule group and membership function may be performed according to an already established inference operation method. For example, the consequent part membership function is truncated according to the obtained antecedent part conformity (grade) for each rule. MAX of truncated consequent part membership function of all rules
An operation is performed. The center of gravity of the MAX calculation result is obtained. The calculated center of gravity is the final degree of character spacing. The character spacing is obtained by substituting the degree of opening of the character spacing into equation (2).

In the above rule group and FIGS. 10 to 13, three types of membership functions,
Although "small", "medium" and "large" are shown, membership functions representing linguistic information such as "slightly small" and "slightly large" may be added as necessary. Needless to say.

In the above-described rule group, the type of the antecedent part variable is one, but a rule having two or more types of variables as the antecedent part can also be created. For example, IF DS (i) is large, AND SB (i) is large, AND
It is also possible to set rules such that SW (i) is small and THEN CSD (i) is extremely large. In the antecedent part, character density DS, overlapping character area SB and blank area S
Any combination of W can be described.

The above-described inference operation for determining the character spacing is sequentially executed for all characters included in a given character string. Eventually, proper character spacing for all characters will be obtained.

The above-described basic rule groups (I), (II) and (I)
II) relates to the minimum antecedent variables DS, SB and SW necessary to infer an appropriate character spacing.
Additional antecedent variables can be used as needed.

These auxiliary antecedent variables include a user input interval UCS, a variable KS relating to a ratio between a minimum character interval and a character width, a variable KH relating to a ratio between a minimum character interval and a character height, and character overlap complexity. CPX and the like. Hereinafter, these auxiliary variables will be described.

The user input interval UCS is a character interval specified by the user for a specific character. The user can specify a specific character using the keyboard 15 and input a desired character interval. The character spacing may be entered in other ways, for example, by the number of dots, by the ratio of the particular character displayed on the CRT 16 to the character spacing. An example of a rule regarding the user input interval UCS is as follows.

(IV) Rule group for user input interval UCS IF UCS (i) is large, THEN CSD (i) is large IF UCS (i) is medium, THEN CSD (i) is medium IF UCS (i) Is small, THEN CSD (i) is small

If the user desires to increase the character spacing, the character spacing is increased accordingly, and if the user desires a small character spacing, the character spacing is reduced.

An example of the membership function for the user input interval UCS is shown in FIG. Needless to say, more types of membership functions (for example, “very large”) can be set as needed.

In the inference (step 23) in consideration of the user input interval UCS, in addition to the above-described basic rule groups (I), (II) and (III), the above-mentioned UCS-related rule group (IV) includes:
Set as rules for fuzzy inference. An inference operation is performed according to all these rules, and
Eventually, an optimum character interval in which the user input interval is considered is obtained.

Needless to say, it is possible to create a rule group having any combination of DS, SB, SW, and UCS as antecedent parts.

The variable KS (i) is defined by the following equation.

KS (i) = | CSmin (i) | / MIN [CW (i), CW (i + 1)] {Equation (4)

Here, MIN [CW (i), CW (i + 1)] means that the smaller one of CW (i) and CW (i + 1) is selected and adopted.

An example of a rule regarding the variable KS is as follows.

(V) Rule (group) concerning variable KS IF KS (i) is large, THEN CSD (i) is large

An example of a membership function for the variable KS is shown in FIG. Here, the language information "big"
Only the membership function is defined for. Therefore, the above rule has no effect in a region in which the membership function “large” is not defined (range where the grade is zero). Of course, it is needless to say that two or more types of membership functions can be created as needed, and more rules besides the above rules can be set.

The minimum character spacing CSmin is equal to the i-th character and (i
Indicates the maximum possible overlap with the +1) th character. According to the basic rules (I), (II) and (III) described above, i
If a character spacing is obtained such that the i-th character and the (i + 1) -th character overlap considerably, it may be difficult to see. In particular, when the character width CH of one of the two characters is narrow, it may give a feeling that one character is dependent on the other character. Minimum character spacing CSmin
If the absolute value of is significantly large, or the smaller of the two character widths is much smaller,
According to the above, the variable KS becomes considerably large. In such a case, the character spacing is increased by the rule (V). Rule (V) is for correcting the excessively narrowing of the character spacing by the basic rule group.

Fuzzy inference of character spacing (step 23)
Is performed in accordance with a rule group formed by adding the rule (group) (V) to the basic rule groups (I), (II) and (III). As a result, an appropriate character spacing in consideration of the variable KS is obtained. The feature amount MIN [CH (i), CH (i
+1)] is calculated prior to inference (step 22). D
It goes without saying that a rule having any combination of S, SB, SW and KS in the antecedent part can be set.

All the above rule groups (I), (II), (I)
Fuzzy inference may be performed according to a rule group composed of (II), (IV) and (V). In this case, the user input interval UCS and the variable KS are considered in inferring the character interval. DS, SB, SW and UCS
Of course, a rule having an arbitrary combination of KS and KS in the antecedent part may be set.

The variable KH (i) is defined by the following equation.

KH (i) = | CSmin (i) | / MIN [CH (i), CH (i + 1)] {Equation (5)

Examples of rules regarding the variable KH are as follows.

(VI) Rule (Group) Concerning Variable KH IF KH (i) is large, THEN CSD (i) is large

FIG. 16 shows an example of a membership function relating to the variable KH. Also in this membership function, the grade is zero in the range where the value of KH is small,
In this range, the above rules have no effect on the whole inference. Of course, it goes without saying that two or more types of membership functions can be created for KH and rules other than the above can be set.

Even if the character height CH of one of the i-th character and the (i + 1) -th character is low, if the two characters overlap considerably, one of the lower characters Gives the feeling of being subordinate to the other character.
There is a rule (VI) to prevent such a result from being generated by the basic rule groups (I) to (III). Rule (VI)
Also works to correct the inference result by the basic rules.

It is for this reason that the membership functions for variables KS and KH are defined only in regions where the values of these variables are large (see FIGS. 15 and 16).

Fuzzy inference of character spacing (step 23)
Is performed according to a rule group formed by adding the above rule (VI) to the above basic rule groups (I), (II) and (III). As a result, an appropriate character spacing in consideration of the variable KH can be obtained. Prior to inference, MIN [CH (i), CH
(i + 1)] is calculated (step 2
2). It goes without saying that a rule having any combination of DS, SB, SW and KH in the antecedent part can be set.

The above rule groups (I), (II), (I)
Fuzzy inference may be performed according to a rule group configured according to any combination of (II), (IV), (V) and (VI). However, the basic rules (I), (II)
And (III) are always used. One or more of the rule groups (IV), (V) and (VI) are added to the basic rule group. In this case, DS, SB and S
A rule having an arbitrary combination of W, UCS, KS, and KH in the antecedent part may be set.

The character overlap complexity CPX (i) is defined as an overlap area (i-th character) formed by bringing the i-th character and the (i + 1) -th character close to each other until a part of them touches each other. Area where the width of the (i + 1) th character overlaps the width of
Is the average value of the number of times that the vertical scanning line for each dot (pixel) transitions from the face of one character to the face of the other character, or from the face of the other character to the face of one character.

This will be specifically described with reference to FIG. There are eleven vertical scanning lines V1 to V11 in the overlapping area of the i-th character and the (i + 1) -th character. Each of these vertical scanning lines is from the face of the i-th character to the face of the (i + 1) -th character.
Transitions times. Therefore, the average value is also 1, and the character overlap complexity CPX is 1.

In the example shown in FIG. 18, the i-th character
There are six vertical scan lines in the overlap area with the (i + 1) th character. Of these, the vertical scanning lines V21 to V23 shift from the face of the i-th character to the face of the (i + 1) -th character, and further from the face of the (i + 1) -th character to the face of the i-th character. Move on to Therefore, there are two transition circuits. On the other hand, the number of transition circuits for the vertical scanning lines V24 to V26 is four. Therefore, the character overlap complexity is calculated according to the following equation.

CPX (i) = (2 + 2 + 2 + 4 + 4 + 4) / 6 = 3 Equation (6)

The character overlap complexity CPX (i) is calculated in the feature amount generation processing (step 22).

The rule group relating to the character overlap complexity CPX is, for example, as follows.

(VII) Rule group for character duplication complexity CPX IF CPX (i) is large, THEN CSD (i) is large IF CPX (i) is medium, THEN CSD (i) is medium IF CPX (i) ) Is small, THEN CSD (i) is small

FIG. 19 shows an example of a membership function relating to the character overlap complexity CPX.

The high degree of character duplication complexity means that if two characters are close to each other, some of them are considerably duplicated and difficult to read. Therefore,
In this case, the character spacing is increased to make it easier to read. If the character overlap complexity is small, there is no problem even if the character spacing is reduced.

The fuzzy inference of the character spacing in consideration of the character overlap complexity CPX (step 23)
This is performed according to a rule group constituted by adding the above-mentioned rule group (VII) to (I), (II) and (III). DS and S
Needless to say, a rule having an antecedent part of any combination of B, SW, and CPX can be created.

All the above rule groups (I), (II), (I
(II), (IV), (V), (VI) and (VII) in any combination (however, rule groups (I), (II) and (III) are always included) Therefore, fuzzy inference may be performed. That is, the basic rule groups (I), (II), and (III) have the auxiliary rule group (I
One or more of (V), (V), (VI) and (VII) will be added. In this case, DS, S
A rule having an arbitrary combination of B, SW, UCS, KS, KH and CPX in the antecedent part may be created.

The handling of small characters will be described. The small character is, for example, (i +
3) A character, such as the third character, whose height CH and width CW are both smaller than other characters. When the CPU 11 determines whether a character is a small character, a reference value for this determination (for example, the number of bits of the upper limit of the height and the number of bits of the upper limit of the width, or the ratio of these to the frame) is used. It is stored in the RAM 13 in advance, or input by the user. In the character code string stored in the hard disk 14, a bit (or code) indicating a small character may be added to a small character code. You may make a user specify on a display screen that it is a small character. In any case, if there is a small character, the following processing is performed in relation to the character before it (the (i + 2) th character in FIG. 3 or FIG. 6). .

FIGS. 20 to 24 show the (i + 2) th character and the (i + 3)
And the second character. FIG. 22 shows the bit map data expanded in step 21. Fig. 21
Indicates a state in which the entire character surface of the (i + 3) th small character is displaced upward by two dots as compared to FIG. Figure 20
This figure shows the entire character surface of the (i + 3) -th character shifted upward by four dots from the state shown in FIG. Further, FIGS. 23 and 24 show the (i + 3) -th small characters displaced downward by 2 dots and 4 dots from the state of FIG. 22, respectively.

As described above, when there is a small character, the inference of the character spacing CS (i + 2) for the character (i + 2) immediately before the small character (i + 3) causes the small character (i +3) is maintained at the bitmap-expanded height, and the appropriate character spacing is inferred, and the small character (i + 3) is displaced upward and downward by an appropriate amount. Infer the appropriate character spacing with the small character at the displaced position. In these inferences, the above-described basic rule group (and the necessary auxiliary rule group) are used. The smallest character spacing is selected from the appropriate character spacings at various heights of the small character (i + 3) thus obtained (step 2).
Four). In addition, the height position of the small character that causes the smallest character spacing is determined to be the height of the character. Figure
In the example shown in FIGS. 20 to 24, the character spacing CS (i
+2) is the smallest. In FIG. 23, a part of the black dot of the (i + 3) -th character protrudes downward from the frame, but there is no problem. Since the purpose of character spacing adjustment is to make the spacing between two characters as small as possible, the smallest character spacing is selected. In FIGS. 20 to 24, small characters (i + 3) are displaced up and down by two dots at a time, but the displacement amount can be arbitrarily determined.

In general, small characters in Japanese are associated with characters preceding them. In the example described above, by utilizing such a characteristic of Japanese, it is determined that the interval between a small character and a character preceding it is as small as possible. In a language in which a small character has a property that accompanies the character that follows, the character spacing between the small character and the character arranged thereafter is adjusted to be as small as possible.

As described above, the appropriate character spacing for all the characters constituting the text represented by the character code string or at least one line of characters is obtained by fuzzy inference. Thereafter, a process of correcting (re-adjusting) the character spacing of the target character in consideration of the overall balance with the other characters arranged before and after the target character (step 25). This correction process is performed for all characters using all the characters as the target character.

Fuzzy inference is also used in the correction processing. As variables in the antecedent part of the inference rule, CI (i) representing the width of two characters including the target character and CL (i) representing the width of four characters including the target character are used. These variables are defined by:

CI (i) = CW (i) + CS (i) + CW (i + 1) (7)

CL (i) = CW (i-1) + CS (i-1) + CW (i) + CS (i) + CW (i + 1) + CS (i + 1) + CW (i + 2) {Formula (8) )

In these equations, CW (i-1) and CW (i)
, CW (i + 1) and CW (i + 2) are the widths of the (i-1) -th, i-th, (i + 1) -th and (i + 2) -th characters. CS (i-1), CS (i) and CS (i + 1) are the (i-1) -th, i-th, and
The appropriate character spacing for the (i + 1) th character.

The consequent part of the fuzzy inference rule is the character spacing C
This is the correction amount ΔCS (i) of S. When this correction amount is obtained in fuzzy inference, the final optimal character spacing is SC
(i) + ΔCS (i).

An example of a rule group for correcting the character spacing is as follows.

(VIII) Rule group IF CI (i) for correcting character spacing is small, THEN ΔCS (i) is slightly large IF CL (i) is small, THEN ΔCS (i) is large

An example of the membership function for the antecedent variables CI and CL and the antecedent variable ΔCS is shown in FIG.
25, 26 and 27, respectively.

These rule groups basically improve the overall balance by slightly increasing the character spacing CS (i) when the width of two consecutive characters or the width of four consecutive characters is too small. Things. Therefore, the antecedent membership functions for variables CI and CL are defined only in regions where the values of these variables are small. The consequent membership function for the correction amount ΔCS is defined only for the region where ΔCS is large.

More or different membership functions for CI, CL and ΔCS may be set as needed. The correction rules are not limited to those described above. The width of three consecutive characters including the character of interest or the width of five or more characters may be used as the antecedent variable. C in the antecedent
A proposition regarding the combination of I and CL can also be described.

When the character spacing is determined as described above, the characters arranged at the determined character spacing are printed. The character spacing is inferred based on a frame (character face) having a predetermined size as described above, and a character spacing suitable for this size is obtained.

In printing, the size (point) of a character is specified. The resolution of the printer 17 (DPI:
The dot per inch is predetermined (even if the resolution is variable, one of the resolutions is specified for printing). The number of dots per character (the number of vertical and horizontal dots in the frame) is determined according to the designated character size and the resolution of the printer 17. In the buffer memory of the printer 17, the characters to be printed are bit-mapped to the determined frame size. At this time, the character spacing determined by the inference is converted according to the size of the frame for printing, and black dots constituting each character are arranged at the converted character spacing. Printing of characters is performed based on the dot data developed on the buffer memory.

In the above embodiment, the character space inference processing is executed by software, but may be executed by a hardware circuit. In particular, many types of hardware circuits dedicated to fuzzy inference are already known. Needless to say, the present invention can also be applied to the adjustment of the character spacing arranged in the vertical direction.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an overall electrical configuration of a computer typesetting system.

FIG. 2 is a flowchart illustrating a flow of a character spacing adjustment process.

FIG. 3 shows a character string expanded into bit map data.

FIG. 4 is a diagram for explaining a feature amount of each character.

FIG. 5 is a diagram for explaining a feature amount between characters.

FIG. 6 shows a feature amount of the character string shown in FIG.

FIG. 7 shows the maximum character spacing.

FIG. 8 is a diagram for explaining character spacing.

FIG. 9 shows a minimum character spacing.

FIG. 10 is a graph showing an example of a membership function relating to a character density DS.

FIG. 11 is a graph showing an example of a membership function relating to the overlapping character area SB.

FIG. 12 is a graph showing an example of a membership function relating to a blank area SW.

FIG. 13 is a graph showing an example of a membership function relating to a character spacing degree CSD.

FIG. 14 is a graph showing an example of a membership function for a user input interval UCS.

FIG. 15 is a graph showing an example of a membership function for a variable KS.

FIG. 16 is a graph showing an example of a membership function related to a variable KH.

FIG. 17 is for explaining the character duplication complexity.

FIG. 18 is for explaining the character duplication complexity.

FIG. 19 is a graph showing an example of a membership function relating to character overlap complexity.

FIG. 20 shows a state in which the (i + 3) th character is shifted upward by four dots.

FIG. 21 shows a state in which the (i + 3) -th character is shifted upward by two dots.

FIG. 22 shows characters expanded to bit map data.

FIG. 23 shows a state in which the (i + 3) -th character is shifted downward by two dots.

FIG. 24 shows a state in which the (i + 3) -th character is shifted downward by four dots.

FIG. 25 is a graph showing an example of a membership function for a variable CI representing a width of two characters.

FIG. 26 is a graph showing an example of a membership function for a variable CL representing a width of four characters.

FIG. 27 is a graph showing an example of a membership function relating to a character space correction amount ΔCS.

[Explanation of symbols]

 11 CPU 12 ROM 13 RAM 14 Hard disk drive 15 Keyboard 16 CRT 17 Printer

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-3-244542 (JP, A) European Patent Application Publication 465704 (EP, A1) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/21 G09G 5/32 B41B 27/00 B41J 19/00 G06F 9/44

Claims (22)

(57) [Claims] 1. A character dot data creating means for expanding character dot data representing a character constituting a specified character string on a memory, wherein the character dot data expanded on the memory includes: A first feature amount generating means for generating character density data representing an area of a character face of the character; in the character dot data expanded on the memory, for each character, the character and a character adjacent thereto are in contact with each other; In the state, the second character plane data generating unit generates overlapping character plane area data representing the area of the character plane existing in the overlapping area of the two characters.
In the character dot data expanded on the memory, for each character, blank area data representing a blank area generated between two characters in a state where the character is in contact with the character adjacent to the character is generated. The third feature value generating means to be created, and the character density data, the overlapped character area data and the blank area data respectively generated by the first, second and third feature value generating means are converted into a predetermined character spacing. A character spacing adjusting device comprising: character spacing inferring means for inferring an appropriate character spacing by fuzzy inference for each character by applying the knowledge to adjustment. 2. An interval input means for inputting user input interval data representing a desired degree of a character interval, and said knowledge including knowledge for inferring a character interval using said user input interval data, said user input interval being included in said knowledge. The character spacing adjusting device according to claim 1, further comprising: the character spacing inferring means for fuzzy inferring an appropriate character spacing by further applying data. 3. A character dot developed on the memory.
In the data, a fourth ratio data generating unit generates first ratio data relating to a ratio between a minimum character interval and a character width representing a character interval when the character and an adjacent character contact each other.
The first ratio data generated by the fourth characteristic amount generating means is further added to the knowledge including the characteristic amount generating means and the knowledge for inferring the character spacing using the first ratio data. The character spacing adjusting device according to claim 1 or 2, further comprising: the character spacing inferring means for performing fuzzy inference on an appropriate character spacing by applying the character spacing. 4. A character dot developed on the memory.
A fifth feature amount generating means for generating, for each character, second ratio data relating to a ratio between a minimum character interval and a character height representing a character interval when the character and an adjacent character are in contact with each other; And further applying the second ratio data generated by the fifth feature value generation means to the knowledge including the knowledge for inferring the character spacing using the second ratio data, to obtain an appropriate character. The character spacing adjusting device according to any one of claims 1 to 3, further comprising: the character spacing inferring means for performing fuzzy inference of spacing. 5. A character dot developed on the memory.
In the data, for each character, when the character and the adjacent character are in contact with each other, the overlap complexity represented by the number of character portions that bite into the overlapping area of both characters and the degree of bite is shown. A sixth feature value generating means for generating character overlap complexity data, and the above-mentioned knowledge including knowledge for inferring a character interval using the character overlap complexity data, generated by the sixth feature value generating means. The character spacing adjusting device according to any one of claims 1 to 4, further comprising: the character spacing inferring means for further applying the character overlap complexity data obtained and fuzzy inferring an appropriate character spacing. 6. When the adjacent character is a small character, the character spacing inferring means determines the character spacing for each of the position developed in the dot data of the small character and the position shifted up and down by a predetermined number of dots. 6. A device according to claim 1, further comprising: means for controlling to perform fuzzy inference, and means for determining the smallest one of a plurality of character spacings obtained from said character spacing inference means as an appropriate character spacing. The character spacing adjusting device according to claim 1. 7. A seventh feature quantity generating means for generating length data relating to the length of a sequence of a plurality of characters including a character of interest, arranged in a line with the character spacing obtained by the character spacing inference means. By applying the length data generated by the seventh feature value generating means to knowledge about character spacing correction given in advance, the interval correction amount is inferred by fuzzy inference for each target character. Inference means, and correction means for generating a final character spacing by adding the space correction amount obtained by the space correction amount inference means to the character spacing obtained by the character space inference means. Item 7. The character spacing adjusting device according to any one of Items 1 to 6. 8. A character dot data representing a character constituting a specified character string is expanded in a memory with a specified character size and with a character interval corresponding to the obtained character interval, The character spacing adjusting device according to claim 1, further comprising a printer that prints a character using the developed character dot data. 9. A computer typesetting system comprising the character spacing adjusting device according to claim 1. Description: 10. A word processor comprising the character spacing adjusting device according to claim 1. Description: 11. A first feature amount generating means for generating, for each character, a first feature amount of a character in character dot data of a plurality of characters expanded on a memory, wherein the first feature amount is expanded on the memory. Character dots of multiple characters
In the data, for each character, a second feature value for generating a second feature amount indicating a relationship between the character and a character adjacent thereto is provided.
Characteristic amount generating means, and the first characteristic amount and the second characteristic amount
A character spacing adjustment device comprising: character spacing fuzzy inference means for inferring an appropriate character spacing by fuzzy inference for each character by applying the feature amount of (1) to knowledge on character spacing adjustment. 12. The character space adjusting device according to claim 11, wherein said character space fuzzy inference means infers a character space in consideration of input data representing a desired degree of character space. 13. When the adjacent character is a small character, the character spacing fuzzy inference means sets the character spacing fuzzy inference for each of the position developed into the dot data of the small character and the position shifted vertically by a predetermined number of dots. Means for controlling the character spacing to be fuzzy inferred, and means for determining the smallest one of a plurality of character spacings obtained from the character spacing fuzzy inference means as an appropriate character spacing. The character spacing adjusting device as described. 14. A third feature value generating means for generating length data relating to the length of a plurality of characters including a character of interest arranged in a line with the character spacing obtained by the character spacing fuzzy inference means. By applying the length data generated by the third feature value generating means to knowledge about character spacing correction given in advance, the spacing correction amount is inferred for each character of interest by fuzzy inference. Fuzzy inference means, and correction means for generating a final character spacing by adding the space correction amount obtained by the fuzzy inference means to the character spacing obtained by the character space fuzzy inference means. 14. The character spacing adjusting device according to claim 11, further comprising: 15. Character dot data representing a character constituting a specified character string is converted into a character string having a specified character size.
12. The printer according to claim 11, further comprising a printer that expands the data on a memory while maintaining a character interval corresponding to the obtained character interval, and prints a character using the expanded character dot data.
15. The character spacing adjusting device according to any one of items 14 to 14. 16. A computer typesetting system comprising the character spacing adjusting device according to claim 11. Description: 17. A word processor comprising the character spacing adjusting device according to claim 11. Description: 18. Character dot data representing a character constituting a designated character string is expanded in a memory, and in character dot data of a plurality of characters expanded in the memory, for each character, the character And generating a first feature value for each of the plurality of characters,
In the data, for each character, generate a second feature quantity representing a relationship between the character and a character adjacent thereto, and apply the first feature quantity and the second feature quantity to knowledge on character spacing adjustment. A character spacing adjustment method in which an appropriate character spacing is fuzzy inferred for each character. 19. The character space adjusting method according to claim 18, wherein in the fuzzy inference of the character space, data representing a degree of a desired character space inputted is considered. 20. When an adjacent character is a small character, the character spacing is fuzzy inferred for each of the position developed in the dot data of the small character and the position shifted by a predetermined number of dots up and down. 20. The character spacing adjusting method according to claim 18, wherein a smallest one of a plurality of character spacings obtained by the spacing fuzzy inference is determined as an appropriate character spacing. 21. Generates length data relating to the length of a sequence of a plurality of characters including a character of interest arranged in a line at a character interval obtained by fuzzy inference, and gives the generated length data in advance. The fuzzy inference of the amount of space correction for each character of interest is performed by applying it to the knowledge about character spacing correction that has been performed, and the character spacing obtained by the above character space fuzzy inference The space correction amount obtained by the above fuzzy inference 21. The character spacing adjusting method according to claim 18, wherein a final character spacing is generated by adding an amount. 22. Character dot data representing a character constituting a specified character string is converted by a specified character size into
22. The method according to claim 18, further comprising developing the data on a memory while maintaining the character spacing corresponding to the obtained character spacing, and printing the character using the developed character dot data. Character spacing adjustment method.