JPS6075865A - Character data compression system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a compression method for character data, and more particularly to a data compression method that compresses data by storing its contour specifying information.

It is well known that binary data obtained by dividing characters into dots has extremely high redundancy.

Conventionally, various data compression methods have been proposed to reduce this redundancy.

One such data compression method is the so-called contour method, which compresses the amount of data by determining the shape of a character by its outline and storing information specifying the outline.

As data compression methods based on the contour method, the vector approximation method as shown in FIG. 1 and the method of approximating the outline of a character using a polynomial as shown in FIG. 2 are known.

The vector approximation method shown in FIG.
This method is disclosed in Japanese Patent Application Laid-Open No. 55-79154, etc. The outline of this vector approximation method is to approximate the contour 1 of an arbitrary character indicated by a dotted line with a set of two-dimensional vectors 2, ) as storage data, data compression is possible.

Furthermore, the method of approximating the outline of a character using a polynomial as shown in FIG. 2 is not disclosed in Japanese Patent Laid-Open Nos. 58-134745 and 1982-134746, etc. by the present applicant.

The outline of this method is as shown by the solid line in Figure 2. First, a block B is created in which the character outline X is divided into functions, and within that block B, multiple sample points P
Select L and divide it into multiple segments SL. After that, ①0 wheel contour body is continuous.

■The first derivative of the 0 contour changes continuously.

The system calculates the coefficients of a polynomial that approximates a segment that satisfies the following conditions, and stores the coefficients.

These contour method-based data compression methods have the characteristic of being able to handle character image reproduction at various magnifications by performing interpolation processing or vector magnification conversion processing when the compressed data is recalled and characters are reproduced. are doing.

However, on the other hand, the vector approximation method does not guarantee optimal results in terms of contour smoothness and continuity of contour slope, unless the approximation is performed based on character pixel data that has been decomposed into dots with a sufficiently large matrix size. First, it has the disadvantage that the roughness of the outline is noticeable especially when character images are reproduced at high magnification.

The present invention is based on the above points, and by using the same mask font, small-sized characters can be reproduced by vector approximation, and large-sized characters can be reproduced by polynomial approximation, thereby improving the smoothness of character outlines. The object of the present invention is to provide a data compression method that can faithfully reproduce the text and obtain a sufficient data compression rate without reducing the reproduction speed of characters. Its feature is that in a character data compression method that compresses the amount of data by storing character outline specific information, the sample points used in the first approximation method to approximate the character outline are expanded into a matrix. A first step of extracting from the outline of the character; and at least reusing the sample points extracted in the first step to form sample points to be used in a second approximation method different from the first approximation method into a matrix. a second step of extracting sample points from the contours of the developed characters;
When the sample points are the first
, the sample point is stored in the character data along with information identifying which of the second approximation methods the sample point is used for.

The present invention will be explained in detail below using the drawings.

The method according to the present invention is applied to binary character data in the form of a matrix and dot matrix supplied from, for example, a scanner device. In yet another example, the present invention is applied to encoded character data using any method that can specify a matrix-like dot pattern.

That is, for example, characters are decomposed into dots in the form of a matrix by raster scanning using a scanner device, and the bit pattern data obtained thereby is supplied as original character data to be processed by the method of the present invention.

The mask font data created according to the present invention includes a plurality of blocks 81 divided by points P so that they can be expressed by a function using the character outline X as a variable, as shown in FIG. 3, for example.
~ B4, and one character's worth of data is constructed as a set of information on each divided block. One block is further divided into multiple segments at the sample point P' (
Point P is also a sample point). For each sample point, an identification code that identifies whether the point is used for both vector approximation and polynomial approximation, or for only one, sample point coordinate information, and for each segment during polynomial approximation. It is equipped with coefficient information, etc.

An embodiment of the mask font data creation method according to the present invention will be described below. Figure 4 (1) shows the outline of the original character.
FIG. 3 is a diagram schematically showing a block.

First, the outline is developed to correspond to the most frequently used small character size, for example, NXN dot size 71 to 90x (FIG. 4 (2)). The curved line indicated by the dotted line is the developed contour.

Sample points used for vector approximation are extracted using an arbitrary method on this matrix size. In Figure 4 (2)
The points indicated by marks are sample points used for vector approximation.

After extracting sample points by vector approximation from a small matrix in the first step as described above, in the next second step, it is expanded into a large-sized matrix. FIG. 4(3) is an enlarged view of FIG. 4(2) (for example, 5N x 5N dots), and the .marks correspond to the sample points in FIG. 4 (■). Further, the outline of the original character in FIG. 4(1) is directly enlarged into a matrix of the same size as that in FIG. 4(3), and superimposed on the matrix of sample points enlarged to the human size. The part indicated by the dotted line is the enlarged outline.

Next, sample points used for polynomial approximation are additionally extracted by reusing at least the sample points (marked with *) used for vector approximation. The sample points are extracted by the method disclosed in, for example, Japanese Patent Application Laid-Open No. 134747/1983 by the present applicant. The points indicated by square marks in FIG. 4(3) are additionally extracted sample points. When reproducing in polynomial approximation, the points indicated by . and the points indicated by mu are used as sample points.

Finally, each segment t' 8 divided by sample points
Calculate the coefficients of each term of the approximate polynomial for each of 1 to S9. For example, the calculation of this coefficient is described in Japanese Patent Application Laid-Open No. 1983-1999 by the present applicant. -1347
It can be determined by the method described in Japanese Patent No. 45 or the like. In this way, in the second step, the stored data is obtained by polynomial approximation from a large matrix.

FIG. 5 is a diagram showing the basic structure of master font data created according to the present invention. As shown in the figure, one character data is
Consists of multiple pieces of block information.

The format of one block information is shown in B and has a structure.

First, the block header stores information about the entire block. First, the X coordinate information of the end point of the block, and second, when the block is turned from OFF to ON when playing characters.
Information for identifying whether the outline of a character changing from ON to OFF is shown or the outline changing from ON to OFF is stored, and thirdly information about the number of segments of the block is stored. The block header is followed by an information section regarding each segment indicated by A+ + A2 r . . . . The segment header of the information section regarding the segment stores the coordinates of the sample point, the degree of the polynomial, and a code for identifying whether the sample point is used for vector approximation. If we now set the identification code as 0: sample point not used for vector approximation, 1: sample point used for vector approximation,
Point P1 in Figure 4 (3) is used for vector approximation, so it is 01
'' (note that the last sample point is used for polynomial approximation). Point P2 is not used for vector approximation, so 110
It becomes 1+.

The segment information section stores coefficients of each term in polynomial approximation.

In this way, one block information is constructed from segment information as many as the number of segments stored in the block header.

A plurality of pieces of one block information constitute master font data for one character.

Next, a case will be described in which a character image is reproduced based on the storage gater using the character data compression method described above.

[Small-sized characters] Since the outline of small-sized characters is reproduced by vector approximation, only the segment information with the identification code "1" used for vector approximation is extracted from the 1 block information in Figure 5. Interpolated values are calculated for each segment, and vector approximation for one block is performed. Note that no coefficient information is necessary at this time. The above operations are performed for one character block.

[Large-sized characters] Since the contours of large-sized characters are reproduced by polynomial approximation, the identification codes '1' and '1' used for polynomial approximation, that is, all segment information, are extracted from the 1 block information in Figure 5. Polynomial approximation is performed by calculating interpolated values using The above operations are performed for one character block.

Next, another embodiment of the present invention will be described.

FIG. 6 is a diagram illustrating a method for creating master font data according to another embodiment of the present invention. The procedure will be explained below.

First, as shown in Figure 6 (1), the outline of the character is developed into a medium-sized matrix size. The curve in FIG. 6(1) shows one block of the contour. Next, sample points used for polynomial approximation are extracted.・The points indicated by marks are sample points. Then, calculate the coefficients of each term in the polynomial approximation for each segment. The above is the first step. As a second step, sample points to be used for vector approximation are additionally extracted using the same matrix size. At this time, the sample points used for the multi-line approximation that have already been extracted are always used and extracted. ,
The points shown in the middle in FIG. 6 are additionally extracted sample points used for vector approximation.

The mask font data created in the first and second steps as described above is stored in the format shown in FIG. 5, as in the previous embodiment. However, in this embodiment, the identification code is a code for identifying whether a sample point is a sample point used for polynomial approximation. Here, the identification code is set as O: not a sample point used for polynomial approximation, and 1: sample point used for polynomial approximation (all sample points are used for vector approximation). Therefore, Fig. 6 (
Point R1 in 2) is used for polynomial approximation, so the identification code is 1.
It becomes 111+.

Since point R2 is not used for polynomial approximation, it becomes "0".

The other difference from the previous example is that all the sample points fitted in the previous example were used for polynomial approximation, so the fifth
The sample point coordinate information stored in the segment header in the figure is always followed by a segment information section storing coefficient information.

However, in this embodiment, not all the sample points fitted are used for polynomial approximation, so some segment information parts do not necessarily have coefficient information following the segment header.

For character reproduction, [Small-sized characters] For small-sized characters, vector approximation is performed using all sample points.

[Large size characters] Polynomial approximation is performed using only segment information with an identification code of '1'.

As explained above, according to the present invention, using the same mask font data, an approximation method suitable for reproducing each character size is used, that is, vector approximation is used for small-sized characters, and polynomial approximation is used for large-sized characters. Since the mask font data is configured to be able to reproduce the outline of The effect is remarkable.

Specifically, it is sufficient to display small-sized characters that are frequently used in the main text, etc., in a correspondingly small dot matrix. Since it is necessary to output a large number of characters in the main text, etc., an approximation method with a high output speed is desirable. In this respect, in the present invention, since the algorithm for reproduction is simple, it is possible to increase the output speed using vector approximation, which has a high output speed.

On the other hand, dog-sized characters do not appear frequently but are large, so even a slight drop in character quality will be noticeable, as in reproduction using vector approximation. Therefore, polynomial approximation can be used, which provides good quality of reproduced characters, although the output speed is slightly lower than vector approximation. Normally, large-sized characters appear much less frequently than small-sized characters, so they have little effect on the overall character output speed.

Furthermore, it is already well known from the above-mentioned publications that both vector approximation and polynomial approximation have high data compression rates, and even if mask font data is configured to use these two approximation methods separately, it is easy to distinguish between the two. The data compression ratio is still high even in the method according to the present invention, since the number of codes to be processed is only increased.

The present invention has remarkable effects as described above.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining vector approximation, Figure 2 is a diagram explaining polynomial approximation, Figure 3 is a diagram explaining mask font data, and Figure 4 is a master font creation method according to an embodiment of the present invention. FIG. 5 is a diagram showing the basic structure of mask font data, and FIG. 6 is a diagram showing a method for creating a master font according to another embodiment of the present invention. Patent applicant Shaken Co., Ltd. Figure 1, Figure 2, Figure 3, Figure 4, Figure 5

Claims (3)

[Claims] (1) In a character data compression method that compresses the amount of data by storing the outline specific information of one character, the sample points used in the first approximation method to approximate the character outline are expanded into a matrix to create the character outline. a first step of extracting from the first step; and a character developed into a matrix by at least reusing the sample points extracted in the first step and using the sample points for a second approximation method different from the first approximation method. A second step of extracting sample points from the contour of . A character data compression method, characterized in that the sample points are stored in the character data along with information identifying whether the sample points are used in a first or second approximation method. (2), the vector approximation method is the first approximation method;
A polynomial approximation method is used as the second approximation method, and in the second step, sample points are extracted using a matrix larger in size than the matrix from which the first sample points were extracted. A character data compression method according to claim (1). (3) A polynomial approximation method is used as the first approximation method, a vector approximation method is used as the second approximation method, and the first approximation method is used as the first approximation method. Extraction of sample points used in the second approximation method is as follows:
The character data compression method according to claim 1, wherein the character data compression method is performed using matrices of the same size.