JPH0126589B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image display method for displaying a halftone image on a device that can express only two levels (white, black) at each point.

In this case, one method is to make one pixel of the halftone image correspond to a block consisting of multiple pixels, and express different densities depending on the number of black points in the block.If the size of the block is M points, then This means that (M+1) different gradations can be expressed.

However, when the number of black dots in a block is K, there are _MCK ways of arranging the black dots, and even if you exclude _identical patterns, it is necessary to select one of the many patterns. The conventional method for this kind of work is to frequency-decompose each pattern, select the one with the lowest power at the lowest frequency, and if multiple values of the same value can be considered, select the one with the next lowest frequency. The idea was to compare and choose the one with lower power consumption.

FIG. 1 is an example of a conventional pattern when the block is a 4×4 square.

When performing frequency decomposition, the coordinates (x,
It is known that when the value of y) is f _x,y (1 for black and 0 for white), f _x,y is Fourier expanded as shown in the following equation.

Here, the Fourier expansion coefficient F _u,v is However, j=√−1.

That is, A _u,v =√ _u,v・_-u,-v (3) Then, each dot pattern is decomposed into a sine wave of amplitude A _u,v and wavelength λ _u,v . Figure 2 shows the amplitude of each wavelength for each pattern (displayed at 16x magnification).

Note that when determining the pattern, the points that are already black in the patterns corresponding to lower gradations are left as they are, and one point that is white changes to black. This condition is
This is necessary in order to avoid unnecessarily emphasizing the boundaries between gradations.

Now, even if images are represented using such a pattern, sufficient quality cannot necessarily be obtained, so improvements have been made since then through trial and error, but none of them are sufficient, and more universal image display is possible. A method was needed.

The present invention was made under these circumstances, and is based on the results of experiments that confirmed that it is not the amplitude of the lowest frequency component of each pattern that determines image quality, but the similarity in frequency components between patterns. It is based on a specific frequency that is determined in advance.
It is an object of the present invention to provide an image display method capable of displaying a halftone image with high quality by distributing black points and determining a pattern so that the frequency component thereof is maximized for each tone.

Embodiments of the present invention will be described below with reference to the drawings.

The present invention is an image display method in which one pixel of a halftone signal corresponds to a fixed block consisting of a plurality of points, each point within the block is expressed in two levels, white and black, and the block as a whole represents a halftone pixel. , the arrangement pattern of the black dots in the above block is
When the white and black values of each point in the above block are subjected to Fourier expansion, the sum of the amplitudes of the Fourier expansion coefficients at a specific frequency becomes the maximum, and the sum of the amplitudes of the Fourier expansion coefficients at frequencies below that becomes small. It was designed to be established as follows.

Figures 3 and 5 show each point (x,
Based on the wavelength λ _{u,v = 4} when the white and black values f
This is the arrangement pattern when the black dots are arranged in the block so that the sum of the amplitudes A u _,v of F _u,v is maximum. Figures 4 and 6 show the arrangement of each frequency component for each pattern. This is the amplitude value (displayed at 16x). In addition, in Figure 3, the number of sunspots is increased in order from 1 to 8.
9 and above are the patterns obtained when the amplitude values are chosen to be symmetrical, and Figure 5 shows the pattern obtained when the optimal pattern is selected with a sunspot of 8, and then the sunspots are gradually reduced. 9 and above are selected so that the amplitude values are symmetrical.

Next, the derivation of the pattern shown in FIG. 3 will be explained in more detail.

In this case, since the reference wavelength λ _u,v = 4 is selected, it is clear that the black nuclei are clustered together. Therefore, for blacks 1, 2, and 3, the pattern shown in FIG. 3 is clearly obtained (because patterns that become the same after processing such as parallel movement, rotation, and reversal are considered to be the same).
For black 4, there are three possible patterns shown in FIG. 14a, and since the amplitudes of each pattern are as shown in the table, pattern (4-) is selected. For black 5, one method is determined as shown in FIG. 3 for the same reason as for blacks 1 to 3. Seven patterns shown in Figure 14b are considered for Black 6, and the pattern (6-) is selected based on the amplitude, and pattern (7-) is selected from the four patterns shown in Figure 14C for Black 7. . In all of these patterns, the positions where black is added are limited due to the condition that the black nuclei do not disperse in a lump, and the same amplitude when translated, rotated, or turned over has the same amplitude, so the numbers are All you have to do is look at the pattern of the species.

Also, using the wavelength λ _{u,v =} 2√2 as a reference, the sum of the amplitudes A u, _{v of the Fourier expansion coefficients F u,} v for all u and _v where the wavelength λ u,v = 2√2 is For the component of wavelength λ _u,v = 4 whose frequency is lower than the wavelength λ _u, v = 2√2, the sum of the amplitudes of this component is as small as possible. Once the arrangement is selected, the pattern shown in FIG. 7 and the amplitude values shown in FIG. 8 are obtained. When images are expressed using such patterns, evaluation results are obtained that exceed those of conventional patterns.

According to the method of the present invention, an optimal pattern can be selected regardless of the block size, and is extremely universal and is also compatible with visual experimental results.

Further, regarding the selection of the reference wavelength, since the tone of the image differs depending on the reference wavelength, the selection may be made according to the desired tone.

Next, the configuration of the present invention will be explained.

FIG. 12 is a block diagram showing a configuration for implementing the present invention, and in the figure, 1 is a quantizer;
2 is a threshold value, 3 is a pattern memory, 4 is a buffer memory, and 5 is an address controller.

The input signal is led to a quantizer 1, where a comparison with a threshold value is performed and a quantized signal is obtained. This quantized signal represents one of the 17 patterns shown in FIG. 3, FIG. 5, or FIG. 7, respectively. This pattern is stored in the pattern memory 3, and the contents of the pattern memory 3 are read out using the quantization signal, and the read data is written into the buffer memory 4. At this time, the address controller 5 controls the horizontal coordinates. If the buffer memory 4 has four lines, pattern signals are written into the buffer memory in units of four lines in the vertical direction. When one line of analog input is scanned, the binary output becomes a signal for four lines and is output from the buffer memory 4.

Next, the second invention of the present application converts halftone image signals into two
The first invention of the present application is applied when displaying on a value display device. As a method for displaying a halftone image signal on a binary recording device in this way, there is a dither method in which one halftone pixel corresponds to one recording point. This method divides one screen into blocks of a certain size, assigns a different threshold value depending on the position within each block, and converts the halftone signal to black and white binary values for each point and displays it. .

To explain this in more detail, for example, if the input halftone image signal is 100 x 100 pixels and the display device is
If the device can display 400 x 400 pixels, by displaying one pixel of the input image using 4 x 4 pixels on the display device, high resolution display can be achieved using the method of the first invention, but the input intermediate The image signal is 100
x 100 pixels, and if the display device is capable of displaying 100 x 100 pixels, it will not be possible to display the high resolution as described above. each 4×4
It attempts to display a half-tone image by dividing the pixel into a total of 25 x 25 blocks and giving different threshold values depending on the position within each block.

Even in this method, if the halftone pixel signal within a block is constant, the signal converted according to its density will take a fixed pattern, and this will allow a halftone image to be displayed. However, the same problem as that of the method of arranging black dots according to the first invention of the present application described above occurs. Therefore, it is only necessary to determine the order of the threshold values so that the optimal pattern in the above sense is obtained for each halftone image signal, and for the pattern in FIG. 3, the order in FIGS. Figure 10 for the pattern in Figure 7, Figure 11 for the pattern in Figure 7.
It is sufficient to set the threshold values in the order shown in the figure.

Next, a configuration for implementing this method will be described.

FIG. 13 is a diagram showing a configuration for implementing the second invention of the present application, in which 6 is a comparator, 7 is a matrix threshold memory, 8 is an X address counter, and 9 is a Y address counter. .

The X address counter 8 changes depending on the X coordinate,
The Y address counter 9 changes depending on the Y coordinate.
In the example shown in FIG.
Address counter 8 and Y address counter 9 are 4
Constructed as a digit counter. The contents of the matrix threshold memory 7 are determined according to the threshold order according to the present invention, and one of the 16 thresholds is selected based on the outputs of the X address counter 8 and the Y address counter 9, and the comparator 6 A comparison is made and a binary output is obtained.

In addition, in the pattern given in the drawing explained above,
It goes without saying that those obtained by performing processes such as parallel translation, rotation, mirroring, and inversion have equivalent effects and are included in the present invention.

As described above, according to the present invention, one pixel of a halftone signal corresponds to a certain block consisting of a plurality of points, each point within the block is expressed in two levels, white and black, and the block as a whole is In an image display method that expresses tone pixels, the sum of the amplitudes of the Fourier expansion coefficients of a specific frequency is the maximum when the black dot arrangement pattern in the block is subjected to Fourier expansion of the white and black values of each point in the block. Then,
Since the sum of the amplitudes of the Fourier expansion coefficients of frequencies below this frequency is determined to be small, the placement of the black dots within the block takes into account visual characteristics, and it is said that an image of extremely high quality can be obtained. have an effect.

[Brief explanation of drawings]

Figure 1 is a diagram showing the arrangement pattern of black dots according to the conventional image display method, Figure 2 is a diagram showing the amplitude of each wavelength (displayed at 16 times magnification) for each pattern in Figure 1,
5 and 7 are diagrams showing examples of black dot arrangement patterns according to the image display method of the present invention.
Figure 8 shows the amplitude of each wavelength (16
Figures 9, 10, and 11 are diagrams showing examples of threshold order in the second invention of the present application, and Figure 12 is a diagram showing a configuration for carrying out the first invention of the present application. FIG. 13 is a block diagram showing a configuration for implementing the second invention of the present application, and FIG. 14 is a diagram for explaining selection of the pattern shown in FIG. 3.

Claims (1)

[Claims] 1 One pixel of a halftone signal corresponds to a certain block consisting of a plurality of points, each point within the block is expressed in two levels, white and black, and the halftone pixel is expressed as a whole block. In the image display method,
When the black dot arrangement pattern in the above block is Fourier expanded on the white and black values of each point in the above block, the sum of the amplitudes of the Fourier expansion coefficients at a specific frequency becomes the maximum, and the Fourier expansion coefficients at frequencies below An image display method characterized in that the sum of amplitudes of is determined to be small. 2 The size of the above block is 4 points both vertically and horizontally, and there are 17 patterns corresponding to the number of black points in the block from 0 to 16, with the black point being “1” and the white point being “0”. An image display method according to claim 1, characterized in that the image display method is defined as follows. 3 The size of the above block is 4 points both vertically and horizontally, and there are 17 patterns corresponding to the number of black points in the block from 0 to 16, with the black point being “1” and the white point being “0”. An image display method according to claim 1, characterized in that the image display method is defined as follows. 4 The size of the above block is 4 points both vertically and horizontally, and there are 17 patterns corresponding to the number of black points in the block from 0 to 16, with the black point being “1” and the white point being “0”. An image display method according to claim 1, characterized in that the image display method is defined as follows. 5 Divide one screen of the image display device into blocks of a predetermined size, give each point in each block a different threshold depending on its position, and divide the halftone signal into binary values of white and black for each point. An image display method in which the binary pattern that occurs when the halftone signal in the block is at a certain level is identified when the white and black values of each point in the block are Fourier expanded. The threshold values in the block are set so that the sum of the amplitudes of the Fourier expansion coefficients of the frequencies becomes the maximum, and the sum of the amplitudes of the Fourier expansion coefficients of the frequencies below it is determined to match the arrangement pattern of the black dots in the block. An image display method characterized by determining an order. 6. The image display method according to claim 5, wherein the size of the block is four points both vertically and horizontally, and the threshold order of each point is set as follows. [Table] 7. The image display method according to claim 5, wherein the size of the block is 4 points both vertically and horizontally, and the threshold order of each point is set as follows. [Table] 8. The image display method according to claim 5, wherein the size of the block is 4 points both vertically and horizontally, and the threshold order of each point is set as follows. 【table】