JPH06337928A - Picture processor - Google Patents

Abstract

PURPOSE:To suppress the unevenness of density and to obtain an easily readable character string picture at the time of arranging and displaying individually processed character patterns by finding out dots whose pixel value is maximum in each character and converting the pixel valve into a density value based on functional relation determined in accordance with the dot value. CONSTITUTION:A pixel value correcting parameter calculating means 105 is provided with a maximum pixel value register 116 and a pixel value comparing device 117 and determines the maximum value Vmax of pixel values successively inputted from a multileveling device 103. The initial value of the register 16 is set up to '0', the pixel value inputted from the device 103 is compared with the contents of the register 116 by a pixel value comparator 117, and when the inputted pixel value is larger, the inputted pixel value is stored in the register 16. At the time of ending scanning based on the device 103, the contents of the register 16 expresses the Vmax of each character pattern. Although the Vmax value is different in each character pattern, a picture pixel value converting means 106 makes the same density value to correspond to the Vmax values of all characters.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for converting a binary character pattern into a gradation (shade) pattern.

[0002]

2. Description of the Related Art As a prior art of the present invention, Japanese Patent Publication No. 63-1
There is an invention disclosed in Japanese Patent No. 4375 (hereinafter referred to as Document 1). A character pattern is stored in a character pattern ROM called a font memory as a bit pattern of monochrome data of 1 bit per address, and the character code of the character is read as an address. When the character pattern of one character is composed of, for example, 48 dots in the X direction and 48 dots in the Y direction, it is set to 1 in the X direction depending on the resolution of the output device (CRT display device or printer).
In some cases, it is necessary to perform conversion so as to form a dot pattern for one character with 2 dots and 12 dots in the Y direction.

In such a case, 12 dots in the X direction and Y
Converting to black and white data of 12 dots in the direction causes a problem that stepwise jaggedness is displayed on the contour line in an oblique direction with respect to the raster scanning. The invention disclosed in Document 1 is to solve this problem.

In the invention disclosed in Document 1, one character 48
Black and white data (binary data) of × 48 dots 12 for each character
12 characters per character, not converted to x12 black and white data
It was converted to × 12 grayscale data (multi-valued data) so that stepwise jaggedness was not displayed. When converting 48x48 dot data to 12x12 dot data, one dot after conversion contains 4x4 = 16 dots before conversion, so how many black dots are in this 16 dots? Depending on whether or not there is any, the density of the converted dot can be represented.
The number of black dots included in 16 dots is 0 to 16
There are 17 different stages.

[0005]

As described above, in the device disclosed in Document 1, the density level of each pixel of the display image and the logic "1" in the original pattern contained in the pixel are displayed.
Since the relation with the total number of bits of is fixed, there is a problem that the shading of the display image changes depending on the width of the line of the character pattern.

Consider, for example, the case where the original image portion of 4 × 4 dots is converted into pixels of 1 dot, and the line width is 4
In the case of dots, if the original image portion is a part of a straight line parallel to the X-axis, all 16 bits of the original image memory are logical “1”, but if the line width is 1 dot , The original image memory, which represents a portion of a straight line parallel to the X-axis, contains only four logic "1" bits. Therefore, when this image portion after conversion is represented by 1-dot pixels, the original image with a line width of 1 dot is represented with a density of 1/4 as compared with the original image with a line width of 4 dots. Will be done. As described above, it is not preferable that the density level of the converted image changes depending on the line width of the original image, and in particular, the density level of each character should be prevented from varying.

The present invention solves the above-mentioned problems in the device disclosed in Document 1, and in the character display, the lines oblique to the raster scanning direction are not jagged, and the lines of the original character pattern are An object of the present invention is to provide an image processing apparatus in which the density level after conversion does not change depending on the size.

[0008]

According to the present invention, the density of the dot having the maximum density of each character in all the character patterns, or the density of the dot having the highest appearance frequency in the character pattern is constant throughout all the characters. Controlled to be. That is, when the image memory of the original image represented by the binary signal is converted into the image memory in which one dot is represented by multi-valued density, the method of Document 1 is used, The number of black dots (bits of logic "1" in the memory) existing in each dot is the pixel value of the dot, and the dot having the maximum pixel value for one character (the pixel value is assumed to be V _max ) _Was calculated, and the pixel value was converted into the density value according to the functional relationship determined according to the value of V _max .

Further, the pixel value V _max of the dot having the maximum pixel value
Instead of using the functional relation associated with, the functional relation associated with the most frequently occurring pixel value in the character pattern may be used. When the normalization is performed by the above method, the average density changes depending on the line width of the character pattern, and the problem that the density changes for each character does not occur.

[0010]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the present invention. In the figure, a binary pattern input means 101 includes an address calculation means 111, a character pattern ROM 112, and a character pattern transfer device 113. The address in the character pattern ROM 112 of the character pattern corresponding to the character code input from the outside is calculated by the address arithmetic unit 111, and the ROM 1
12 is accessed, and the read data (bit pattern representing a character pattern) is transferred to the character pattern transfer device 113.
Transfer to the value pattern storage device 102.

The binary pattern storage device 102 includes a pattern size storage unit 114 and a pattern storage unit 115.
FIG. 3 is a diagram showing an example of the stored contents of the binary pattern storage device 102. The character pattern of the Katakana character A is decomposed into 48 × 48 pixels, and 1-bit data (black and white binary data ) Is stored. Numerical values in the storage device are represented using binary numbers, but in FIG.
It is a hexadecimal number of two digits and is represented by a two-digit hexadecimal number. That is, the numerical value 30 of the pattern size storage unit 114 becomes 48 when converted into a decimal number, which indicates a size of 48 × 48 pixels.

The pattern storage unit 115 has 48 sections in the vertical direction and represents 48 bits corresponding to 48 pixels, but in the horizontal direction, 8-bit bit patterns are collectively represented by 1 section, and 6 sections of 48 bits. It represents a bit pattern. For example, in the 11th line of the pattern storage unit 115, 3,
83, ff, ... means "0000001
1,1000,0011,1111,1111, ...
"" Represents a bit pattern. The values of the numbers 30 (generally Y _IN ) and 30 (generally X _IN ) in the pattern size storage unit 114 are equal to the size of the character pattern stored in the character pattern ROM 112.

Next, in the multi-value quantization apparatus 103, the binary pattern of the pattern storage unit 115 is scanned with an opening of 4 × 4 (generally M × N) pixels, and the number of black pixels in the opening is counted. This is output to the pixel value correction parameter calculation means 105 and the multi-valued pattern storage device 104 as a pixel value. Here, the pattern storage is performed by moving the aperture scanning in the multi-value quantization apparatus 103 at M pixel intervals in the main scanning direction (X direction) and at N pixel intervals in the sub scanning direction (Y direction). All bits on section 115 will be scanned. Therefore, the pattern sizes stored in the pattern size storage unit of the multi-valued pattern storage unit 104 are X _OU = X _IN / M and Y _OU = Y _IN / N.

FIG. 4 is a multi-value conversion device 1 from the binary pattern of FIG.
FIG. 3 is a diagram showing an example of a multi-valued pattern generated by 03, in which the pattern storage unit 119 stores the pixel value of each pixel in a section of 1 section and 1 pixel. For example, the pattern storage unit 119 of FIG.
The numerical value 6 in the third row, second column of is the pattern storage unit 11 in FIG.
5, bit pattern "000" in the second half of the 9th row to the 1st column
It is the total number of "1" included in "0", "0011", "0011", and "0011".

The pixel value correction parameter calculation means 105 is
The maximum pixel value register 116 and the pixel value comparison device 117 are provided, and the maximum value V _max of pixel values is determined from the pixel values sequentially input from the multi-value quantization device 103. That is, the initial value of the maximum pixel value register 116 is set to 0, and the multi-value quantization apparatus 1
The pixel value input from 03 and the content of the register 116 are compared by the pixel value comparison device 117, and when the input pixel value is larger than the content of the register 116, the input pixel value is stored in the register 116. . By doing so, the content of the register 116 at the time when the scanning by the multi-value quantization apparatus 103 is completed is the V for each character pattern.
will represent _max . After processing all pixel values on the multi-valued pattern storage means 104, V _max is output to the pixel value conversion means 106.

Although the value of V _max differs depending on the character pattern, the pixel value converting means 106 associates the same density value D with V _{max of} all characters. In this case, D is the maximum density value D _max . That is, the density level is 0, 1,
If four gradations of 2 and 3 are output, the density for V _max is 3 (= D _max ). The concentration D _x of the pixel a pixel value is _{x, D x = x (D} max / V max) ··· (1)
It is calculated by. The pixel value conversion unit 106 calculates D _x corresponding to all pixel values (x) by the calculation using the equation (1) and creates the pixel value conversion table 120.

FIG. 7 shows an example of the pixel value conversion table.
For the pixel value, all 16 bits of the original binary pattern are "0" (x
= 0) 16 bits are all "1" (x = 16) 17
There can be types, but in the example shown in FIG. 4, V _max = e
(= 14). Therefore, if the pixel value is 14 or more (x = 15, 16 does not exist in the example of FIG. 4), the density value is 3
Becomes The density value for the pixel value x is calculated from the equation (1), but since the density value has only four gradations, the content of the pixel value conversion table 120 changes slightly depending on how to process the decimal point of the calculation result. However, this change has no substantial effect. That is, the relationship between x and D _x does not have to be a proportional relationship, and may be a functional relationship that simply increases.

Next, the pixel value writer 121 converts each pixel value in the pattern storage unit 119 of the multi-valued pattern storage means 104 according to the pixel value conversion table 120. Figure 4
FIG. 5 shows a multi-valued pattern after density conversion, which has been converted by the pattern storage unit 119 of FIG. The gradation pattern output means 107 reads the multi-valued pattern shown in FIG. 5 and converts the density value into an analog signal by the D / A converter 122 and displays it on the display device 108.

FIG. 2 is a block diagram showing another embodiment of the present invention. In the figure, the same reference numerals as those in FIG. 1 indicate the same parts, and the operation thereof is also the same, so that the description thereof will be omitted. In the embodiment of FIG. 1, the pixel value correction parameter calculation means makes the maximum pixel value for each character pattern correspond to a constant maximum density value through all character patterns, but in the embodiment shown in FIG. 2, the appearance frequency for each character pattern. The maximum pixel value was made to correspond to a constant density value through all character patterns.

The pixel value correction parameter calculation means 201 comprises a histogram generation device 211 and a median value register 21.
2 and the histogram totaling device 213. The histogram generator has 1 corresponding to 17 kinds of pixel values of 0 to 16.
It has seven counters, all counters are cleared at the time of initialization, and each time the pixel value is determined by the multilevel halftoning device 103, a numerical value 1 is added to the counter corresponding to that pixel value.
Is added. The count value of each counter of the histogram generation device 211 after all the pixel values of the pattern storage unit 119 of FIG. 4 are determined is as shown in FIG. 6, for example.

The pixel with the highest appearance frequency except the pixel value 0 is the pixel value 8, and the pixel value 8 may be simply stored in the median value register 212. In addition, the histogram aggregation device 213 calculates that the frequency of appearance of pixel values of pixel value 7 or less (excluding pixel value 0) and the frequency of appearance of pixel values of pixel value 7 or more are the same (substantially the same). However, the pixel value 7 may be stored in the median value register 212.

The pixel value conversion means 202 inputs the median value from the median value register 212, determines the density value corresponding to this median value, and creates the pixel conversion table 213. In the example of the pixel value histogram shown in FIG. 6, the density value corresponding to the pixel value 7 or 8 is set to 2. Excluding density level 0 among density levels 0, 1, 2, and 3,
It is advisable to set the density level 2 to the density having the highest appearance frequency. If the pixel value of the median value register is V _mid and the density corresponding thereto is D _mid , the density value D _x corresponding to the pixel value _x
Becomes D _x = x (D _mid / V _mid ) ... (2). However, in the above example, the maximum value of D _x is 3.

In the above description, numerical examples are used for convenience of explanation, but it is needless to say that the present invention is not limited to these numerical examples. Further, the case has been described in which the 4 × 4 bit data of the binary pattern storage device 102 is used as the 1-pixel data of the multi-value pattern storage device 104, but this assumes that the M × N bit data is the 1-pixel data. In the case of M = N = 4 in the case, more generally speaking, a plurality of sampling points are selected on the binary pattern storage device, the binary pattern is resampled, and the neighborhood of the sampling point is selected. Is a numerical example of a method of determining the pixel value of the sample point by the number of bits of logic "1". Further, the pixel value correction parameter calculation unit 105 sets the density corresponding to the maximum pixel value or the pixel value with the maximum appearance frequency in each character pattern to be the same for all character patterns. In addition, the density corresponding to the pixel value in a predetermined order in each character pattern can be the same for all character patterns.

[0024]

As described above, the image processing apparatus according to the present invention generates a multi-valued pattern by re-sampling a binary pattern and calculating an output pixel value from the values of pixels in the vicinity of the sample point. By suppressing the unevenness of the density of each input pattern by converting the pixel value of the multi-valued pattern into the density value based on the distribution of the appearance frequency for each value of the pixel value of the multi-valued pattern, especially in a character image, Even when the character patterns that have been individually processed are displayed side by side, it is possible to suppress unevenness in density and generate an easily readable character string image.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram showing another embodiment of the present invention.

FIG. 3 is a diagram showing an example of stored contents of a binary pattern storage device of the present invention.

FIG. 4 is a diagram showing an example of pixel values stored in a multi-valued pattern storage means of the present invention.

FIG. 5 is a diagram showing an example of density values stored in a multi-valued pattern storage means of the present invention.

FIG. 6 is a diagram showing an example of a pixel value histogram of the present invention.

FIG. 7 is a diagram showing an example of a pixel value conversion table of the present invention.

[Explanation of symbols]

 101 binary pattern input means 102 binary pattern storage device 103 multi-value conversion device 104 multi-value pattern storage means 105 pixel value correction parameter calculation means 106 pixel value conversion means 107 gradation pattern output means 108 display device 111 address calculation device 112 characters Pattern ROM 113 Character pattern transfer device 114 Pattern size storage unit 115 Pattern storage unit 116 Maximum pixel register 117 Pixel comparison device 119 Pattern storage unit 120 Pixel value conversion table 121 Pixel value writer 122 D / A converter 201 Pixel value correction parameter calculation Means 202 Pixel value conversion means 211 Histogram generation means 212 Median value register 213 Pixel value conversion table

Claims (6)

[Claims] 1. An image pattern representing data of each pixel is generated and output by a binary signal of logic "0" or "1".
Value pattern input means, inputting the image pattern from the binary pattern input means,
A binary pattern storage device that stores a binary signal that constitutes the image pattern at an address position corresponding to the coordinate position of a pixel in the original image, and each point determined by design on this binary pattern storage device is a sample point, For each sample point, a multi-valued device that resamples the total number of bits of logic "1" in a binary signal within a predetermined range from the sample point as a pixel value of a pixel corresponding to the sample point, Input the pixel value resampled by the quantizer,
A multi-valued pattern storage means for storing each pixel value at an address position corresponding to the coordinate position of the sample point where the pixel value is calculated, and a pixel in one image pattern regarding the pixel value re-sampled by the multi-valued device. Pixel value correction parameter calculation means for determining a pixel value / density value conversion parameter by associating values in a predetermined order with a predetermined density value, and pixels stored at each sample point of the multi-valued pattern storage means. An image processing apparatus comprising: a pixel value conversion means for converting a value into a corresponding density value using the pixel value / density conversion parameter. 2. The image processing apparatus according to claim 1, wherein the pixel value correction parameter calculation means associates the maximum pixel value in the image pattern with the maximum density value in the entire image pattern. 3. The pixel value correction parameter calculation means according to claim 2, comprising a maximum pixel value register and a pixel value comparison device,
At the time of initialization, the content of the maximum pixel value register is set to 0,
The pixel value output by being resampled by the multi-valued device and the content of the maximum pixel value register are compared by the pixel value comparison device, and when the pixel value output by the multi-valued device is larger, An image processing apparatus, wherein a pixel value is input to the maximum pixel value register. 4. The image processing apparatus according to claim 1, wherein the pixel value correction parameter calculation means associates a pixel value having the maximum appearance frequency in the image pattern with a constant density value throughout the entire image pattern. 5. The pixel value correction parameter calculation means according to claim 4 comprises a histogram generation means, a median value register, and a histogram aggregation device, and a histogram of pixel values output by being resampled by the multivalued device is output. Generate, calculate the median value from this histogram,
An image processing apparatus, wherein the calculated median value is set as the pixel value having the maximum appearance frequency. 6. The image processing apparatus according to claim 1, wherein the pixel value conversion means comprises a pixel value conversion table and a pixel value writer, and obtains a density value corresponding to each pixel value according to the pixel value / density value conversion parameter. It is determined and stored in the pixel value conversion table, and the pixel value writer refers to the pixel value conversion table and rewrites the value of the pixel value stored in the multi-valued pattern storage means to a corresponding density value. Image processing device.