JPH0993446A - Image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress production of granular noise or regular stripe pattern when an image is processed by using the error spread method. SOLUTION: An error component A error for each color set by the error spread method is added to each of input signals Ay, Am, Ac and a random number component (asynchronous component) th1 and a periodic component th2 are added by a prescribed rate and binarization is applied as to the sum. The periodic component th2 is set, based on comparison between a threshold level of a matrix representing a checkered pattern in the unit of 2×2 picture elements and the input signal.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method, and more specifically, it expresses a color gradation image by a combination of density gradation (intensity modulation) and area gradation (area modulation) having a smaller number of levels than the original. And an image processing method for doing so.

[0002]

2. Description of the Related Art Conventionally, in a digital printer, a digital facsimile, etc., a method for reproducing a halftone by a combination of density gradation (intensity modulation) and area gradation (area modulation) having a smaller number of levels than the original. As the above, a dither method and an error diffusion method are known (Japanese Patent Laid-Open No. 61-35).
676, JP-A-63-217768, etc.).

In the dither method (binary dither method), the value of each matrix of the dither matrix is used as a threshold value, and compared with the density of the pixel at the corresponding coordinate point, 1 (printing or light emission), 0 (no printing). (Or no light emission) is determined and binarized, and binarized data for area gradation can be obtained only by comparing and computing the original image data and a threshold, and high-speed computation is possible. Further, in the error diffusion method, the error when the gradation image data is converted into the data of a smaller number of levels (for example, binary) is dispersed to the neighboring pixels, and the gradation representation error is performed together with the neighboring pixels. (Reference: “RW”).
Floyd and L.D. Steinberg "An Adaptive Algorithm for
Spatial Gray Scale ", SID 75 Digest (1976)")
.

[0004]

However, the dither method has a problem that the gradation and the resolution cannot be compatible because the gradation and the resolution directly depend on the size of the dither matrix. Further, the error diffusion method has a problem in image quality that graininess noise in a highlight portion or a shadow portion of an image is conspicuous and a regular stripe pattern called a texture is generated mainly in a halftone.

The present invention has been made in view of the above problems, and improves the problem of image quality in the error diffusion method,
It is an object of the present invention to provide an image processing method capable of reproducing a high quality image.

[0006]

Therefore, the invention according to claim 1 is an image processing method for expressing a color gradation image by a combination of density gradation and area gradation having a smaller number of levels than the original. An error component, a periodic component, and an aperiodic component are added to the input signal to be processed.

According to this structure, by adding the periodic component in addition to the error component due to the error diffusion, it is possible to cancel the regular striped pattern generated in the halftone regardless of the original image. On the other hand, in the highlights and shadows, on the contrary, this periodicity is perceived as noise, so adding non-periodic components makes the noise appear lower,
By adding the non-periodic component, that is, the random number component, it is possible to suppress the granular noise in the highlight and shadow portions. Further, the regular striped pattern that occurs in the halftone irrespective of the original image due to error diffusion is canceled by the addition of the periodic component, thereby suppressing the occurrence of the striped pattern.

According to the second aspect of the invention, the distribution ratio of the periodic component and the aperiodic component is changed depending on the output medium of the image signal. With such a configuration, it is possible to appropriately change the distribution ratio of the periodic component and the non-periodic component according to the characteristics of the output medium such as a printer, and effectively suppress the occurrence of grainy noise and stripe patterns.

According to the third aspect of the invention, the distribution ratio of the periodic component and the non-periodic component is changed according to the input signal. According to this structure, for example, the ratio of the non-periodic component is increased in the highlight and shadow portions, while the ratio of the periodic component is increased in the halftone, so that it is possible to effectively suppress the generation of the granular noise and the striped pattern. .

In the invention of claim 4, the periodic component is
The configuration is different for each input signal. According to such a configuration,
The periodic component can be set in a form in which the change of the input image signal is added to the basic periodic change, and the occurrence of the striped pattern can be suppressed without deteriorating the image quality. According to the fifth aspect of the invention, the periodic component is different for each color.

According to this structure, it is possible to preferentially generate any one of the complementary color combinations in the gray halftone portion and suppress the occurrence of color unevenness. In the invention according to claim 6, the periodic component has a checkered pattern. With this configuration, the checkerboard pattern cancels the regular striped pattern. Here, for example, checkerboard pattern is 2 × 2
The best image quality can be obtained by using pixel units.

According to a seventh aspect of the present invention, the non-periodic component has a table. According to such a configuration,
By having the non-periodic component in the table, the processing can be simplified by repeatedly using the table data. Here, it is preferable that the size of the table is as large as possible in order to avoid that a certain pattern is generated by the table.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below. In the present embodiment, in a printer (or display) configured to input color gradation signals Ay, Am, Ac (8-bit data) of three primary colors (yellow Y, magenta M, cyan C) corresponding to color materials. , A gradation image signal is processed in order to express a color gradation image with a binary value (with or without dots) having a smaller number of levels than the original, and gradation is expressed by an area gradation (area modulation) method. To do. However, the color gradation signal may be given by the three primary colors of red R, green G, and blue B, and this may be converted into the Y, M, C system for use.

As a first embodiment, a color gradation signal A
Error diffusion is performed independently for each of y, Am, and Ac, and Y,
An image processing method for individually determining M and C dot striking is described below.
A description will be given based on the flowchart of FIG. In the flowchart of FIG. 1, first, as preprocessing, color gradation signals Ay, Am, A as input signals (original signals)
In order to add the random number component th1 (non-periodic component) to c, a random number table is created by the random number generator (S31). The size of the random number table is from 64 × 64 to 25
About 6 × 256 is suitable, but if the cost increase is ignored, a larger size is desirable to avoid the occurrence of a fixed pattern.

Incidentally, the random number component th1 is multiplied by a coefficient coeff as will be described later, and is added to the input signal (original color signal) as th1 × coeff, and the table created in S31 is used. After that, it is commonly used for processing the color gradation signals Ay, Am, Ac in the image. Next, a reference matrix of the periodic component th2 to be added together with the random number component th1 to the input signal is created (S32).

The matrix is the periodic component th.
2 is required to exhibit a periodic change in order to have a periodicity. For example, a Bayer type matrix (see FIG. 2) showing a checkered pattern of 1 × 1 pixel unit or 2 × 2 based on this Bayer type matrix. A checkered matrix in pixel units (see FIG. 3) or 4 × 4 pixel units (see FIG. 4) can be used, but the checkered matrix in 2 × 2 pixel units is most preferable. Further, the size of the matrix is preferably 16 × 16 when the input signal is 8 bits, but may be smaller than this.

The calculation of the periodic component th2 based on the checkered reference matrix is performed based on the input signal and the reference matrix as follows, and the periodic component th2 differs depending on the input signal. . For example, when an input signal (original color signal) with coordinates (x, y) is B0, B0 ≦ matrix (x mod 16, y
If mod 16), th2 =-(32 + B0) and B0
> Matrix (x mod 16, y mod 16), th2 =
(32+ (255-B0)).

Further, B0≤matrix (x mod 16, y mod 1
If 6), th2 = −32 and B0> matrix (x mod
If 16, y mod 16), th2 = 32 may be set. With the configuration for calculating the periodic component th2 as described above, the periodic component th2 reflecting the gradation characteristics of the original image is used while the checkerboard pattern as the periodical change is basically used.
Can be set. Therefore, if the configuration is such that the periodic component th2 is added to the input signal and then binarization processing is performed using the error diffusion method, a striped pattern is generated due to the error diffusion without impairing the gradation of the original image. Can be canceled.

The periodic component th2 is multiplied by a coefficient coeff2 as described later and added to the input signal as th2 × coeff2. In S33, the target value Ctarget for each color
-y, Ctarget-m, Ctarget-c are calculated. The target value Ctarget-y, Ctarget-m, Ctarget-c for each color is an error for each color assigned to each of the original color gradation signals Ay, Am, Ac from surrounding pixels for each color by the error diffusion method. It is calculated by adding the components Aerror-y, Aerror-m, Aerror-c, the random number component (non-periodic component) th1 × coeff and the periodic component th2 × coeff2 of the random number table.

Ctarget-y = Ay + Aerror-y + th1 ×
coeff + th2 × coeff2 Ctarget-m = Am + Aerror-m + th1 × coeff + th
2 x coeff 2 C target-c = Ac + A error-c + th 1 x coeff + th
2 × coeff2 The periodic component th2 is set based on the comparison between the matrix threshold value and the input signal as described above.

Here, the target value Ctarget-y,
In the calculation of Ctarget-m and Ctarget-c, the random number component th
Since 1 and the periodic component th2 are added, the random number component t
By h1 it is possible to avoid the bias of the dither image and suppress the generation of graininess noise particularly in the highlight and shadow parts, and the basic component is a checkerboard pattern to reflect the characteristics of the original image. It is possible to suppress the occurrence of a striped pattern.

The random number component th1 and the periodic component th
The coefficients coeff and coeff2 multiplied by 2, in other words, the distribution ratio of the random number component th1 and the periodic component th2 may be fixed values, but for each medium (printer) from which the processed image signal is output. By selecting the optimum value, the random number component th1 and the periodic component th2 can be added at an appropriate ratio according to the characteristics of the output medium, and thus the image quality can be effectively improved for each output medium. For example, if the grainy noise is more noticeable than the striped pattern on the hard copy due to the characteristics of the printer,
The proportion of the random number component th1 may be increased, or conversely, when a regular striped pattern appears prominently on the hard copy, the proportion of the periodic component th2 may be increased.

The distribution ratio may be changed according to the input signal. For example, the target value (Ctarget-y, Ctarget-m, Ctarget-
c) can be set. Target value = input signal + error component + (| 128-B0 | / 12
8) x th1 x coeff + {1- (| 128-B0 | / 128
)} × th2 × coeff2 If the configuration is such that the random number component th1 and the periodic component th2 are added to the input signal based on the above equation, the ratio of the random number component is increased in the highlight and shadow portions, while in the halftone. The ratio of the periodic component can be increased, so that it is possible to effectively suppress the grainy noise in the highlight and shadow portions and the stripe pattern in the halftone.

In S34, the target value Ctarget- for each color is
y, Ctarget-m, and Ctarget-c are compared with a fixed threshold value, and binarized for each color. In S35, the error due to the binarization is calculated for each color. In S36, the calculated error for each color is distributed to the surrounding pixels for each color, and the error components Aerror-y, Aerror-m,
Determine Aerror-c.

An example of distribution pixels and weighting in the above error distribution is shown in FIG. 5 or FIG. FIG. 5 shows 3 × 3
An error diffusion matrix of pixels is shown. In this example,
The error in the central pixel of the matrix is 7/16, 5/1
It is configured to be distributed to surrounding pixels with constant weighting of 6, 3/16, and 1/16.

Further, the example shown in FIG. 6 shows a randomized error diffusion matrix, and the weighting of the error distribution to each surrounding pixel is a / 16, b / 16, c / 16, d / 16, The numbers a, b, c and d are determined by random numbers. Here, if a function for generating an integer random number is rand () and an operator for calculating the remainder of division is%, then a,
b, c, d are determined as a = rand ()% 17 b = rand ()% (17-a) c = rand ()% (17-ab) d = 16-abc be able to.

In S37, it is determined whether or not the processing has been completed for all pixels, and the processing of S31 to S36 is repeated until the processing is completed. Next, a second embodiment of the configuration in which the corresponding color space is specified from the color gradation signals Ay, Am, Ac and dot striking for each color is determined will be described. Here, as shown in FIG. 7, the color space is a cube having eight primary colors (basic colors) white W, yellow Y, magenta M, cyan C, black K, blue B, green G, and red R as vertices. Further, the color space is defined and further divided into three small color spaces as shown in FIG.

The three small color spaces are the first small color space represented by W, Y, M, and C primary colors, Y, M, C, R, and
A second small color space represented by the primary colors of G and B, R, G,
It is composed of a third small color space expressed by B and K primary colors, and the combination of primary colors is different in each small color space. When each of the small color spaces is defined on the gradation signal, A
Assuming that y, Am, and Ac represent Y, M, and C density data (8-bit data), Ay + Am + Ac ≤25
The condition of 5 corresponds to the area of the first small color space, and 255 <Ay
The condition of + Am + Ac <510 corresponds to the region of the second small color space, and the condition of Ay + Am + Ac ≥510 corresponds to the region of the third small color space.

Here, the applied color gradation signals Ay, A
m, Ac (color image signal) is represented by the primary colors W, Y, M, and C (where the vertices are W, Y, M, and C) under the condition that Ay + Am + Ac ≤255. When the color belongs to the small color space, the dot of any one of the four primary colors W, Y, M, and C expressing the first small color space is used as the output.

Further, given color gradation signals Ay, Am,
Ac (color image signal) is 255 <Ay + Am + Ac <510
Which is represented by the primary colors of Y, M, C, R, G, B (the vertices are Y, M, C, R, G, B).
Of the six primary colors Y, M, C, which represent the second small color space as an output,
Any one of R, G, and B dots is used.

Further, given color gradation signals Ay, Am,
Ac (color image signal) is a third small color space represented by primary colors of R, G, B, and K (where vertices are R, G, B, and K) under the condition that Ay + Am + Ac ≥ 510. If it is a color belonging to, the dot of any one of the four primary colors R, G, B, K expressing the third small color space is used as the output.

That is, the kind of assigned color is limited depending on which of the small color spaces that divide the color space belongs. However, the white W dots indicate that dots are not actually printed, the yellow Y, magenta M, and cyan C dots indicate dots of each color ink, and the red R and green G dots are also printed. , Blue B dots are Y + M, Y + C, M
+ C overlay, black dots are Y, M,
It indicates the superposition of C or black dot striking with pure black ink.

It should be noted that the setting of the color space and the division of the color space are not limited to those shown in FIGS. 7 and 8, and it is obvious that various modifications are possible. Here, the color gradation signals Ay, Am, Ac that are performed as described above are used.
The distribution of the primary color signals for (color image signal)
The details will be described with reference to the flowcharts of FIGS. 9 and 10.

In the flow chart of FIG. 9, S1, S
2, in the same way as S31 and S32 in the flowchart of FIG. 1, the table creation of the random number component th1 (non-periodic component) and the reference matrix of the periodic component th2 (for example, 2 ×
2 pixel unit checkered pattern) setting. Next, a parameter for determining which small color space shown in FIG. 8 the original color belongs to by adding together the density data for each color of the given three primary color gradation signals Ay, Am, Ac Cto
tal (← Ay + Am + Ac) is calculated (S3).

Next, the target value Ctarget is determined (S
4). The target value Ctarget is the target value Ctarget for each color.
-y, Ctarget-m, Ctarget-c are added together and determined (Ctarget ← Ctarget-y + Ctarget-m + Ctarget-c),
Target values Ctarget-y, Ctarget-m, Ctarget for each color
-c is an error component Aerror-y, Aerror-m, Aerror-c for each color allocated from surrounding pixels by the error diffusion method to each of the original color gradation signals Ay, Am, Ac, and a random number in the random number table. Component (non-periodic component) th1 x coef
It is calculated by adding f 2 and the periodic component th2 × coeff2.

Ctarget-y = Ay + Aerror-y + th1 ×
coeff + th2 × coeff2 Ctarget-m = Am + Aerror-m + th1 × coeff + th
2 x coeff 2 C target-c = Ac + A error-c + th 1 x coeff + th
2 × coeff2 Ctarget = Ctarget-y + Ctarget-m + Ctarget-c Subsequently, it is determined which small color space it belongs to based on the parameter Ctotal, and one of the primary colors expressing the small color space to which it belongs is set to the target value. A process of allocating based on Ctarget is executed (S5).

Here, in the calculation of the target value Ctarget-y, Ctarget-m, Ctarget-c for each color which is the basis of the target value Ctarget, since the random number component th1 and the periodic component th2 are added, the random number The component th1 can prevent the bias of the dither image to suppress the generation of the grainy noise particularly in the highlight and shadow portions, and the periodic component th in which the characteristics of the original image are reflected with the basic checkerboard pattern.
By 2, it is possible to suppress the occurrence of a striped pattern that is not present in the original image.

The random number component th1 and the periodic component th
The coefficients coeff, coeff2 multiplied by 2, in other words, the distribution ratio of the random number component th1 and the periodic component th2 may be a fixed value as described above, but the processed image signal is output. It is preferable to select an optimum value for each medium (printer) to be used, and it may be changed according to an input signal.

Here, the contents of the distribution process in S5 will be described in detail with reference to the flowchart of FIG. In the flowchart of FIG. 10, first, based on the parameter Ctotal, which of the three small color spaces shown in FIGS. 7 and 8 belongs is determined.

Specifically, the parameter Ctotal is
Ctotal ≤255, 255 <Ctotal <510, Ctotal ≥51
It is determined whether it is 0 (S11). Ctotal ≤
255, the first expressed in W, Y, M, C primary colors
If it belongs to the small color space of
It is determined whether it is 180 or less (12). The target value C
If the target is 180 or less, it is determined that the pixel belongs to the vicinity of the highlight in the first small color space, and white W, which is one of the primary colors of the small color space belonging to the pixel, is selected. Sort (S13).

The above-mentioned 180, which is a reference for the distribution of the white W, is an example, and is not limited to such a value, but the first represented by the primary colors W, Y, M, and C.
It is preferable to set a value such that the volume of the small color space is divided substantially evenly in correspondence with the distribution of the primary colors. On the other hand, when the target value Ctarget exceeds 180, the target value Ctarget-y for yellow Y is greater than or equal to the target value Ctarget-m for magenta M, and the target value Ctarget-y for yellow Y is cyan C
It is determined whether or not the target value Ctarget-c is greater than or equal to (S1
Four).

The target value C target-m for magenta M and the target value C target-c for cyan C are the target values C for yellow Y.
If target-y is not exceeded, yellow Y, which is one of the primary colors of the small color space belonging to the pixel, is distributed (S15). If there is a target value that exceeds the target value Ctarget-y for yellow Y, the target value Cta for magenta M
It is determined whether rget-m is greater than or equal to the target value Ctarget-c for cyan C (S16).

When the target value Ctarget-m of magenta M is greater than or equal to the target value Ctarget-c of cyan C, magenta M, which is one of the primary colors of the small color space belonging to the pixel, is allocated. (S17). When the target value Ctarget-m of magenta M is less than the target value Ctarget-c of cyan C, 1 of the primary color of the small color space belonging to the pixel is
Cyan C, which is one of the two, is distributed (S18).

That is, when the belonging to the first small color space represented by the primary colors W, Y, M, and C is determined by the parameter Ctotal, the four primary colors (colors corresponding to the vertices of the space) are determined. The primary color having the highest degree of proximity among them is distributed, and the dots of the distributed primary color are applied to the pixel. As described above, when the belonging to the first small color space is determined, the colors are distributed only among the four primary colors W, Y, M, and C that represent the first small color space. , The dots belonging to the first small color space near the highlight, but having a dark color (dots of K, R, G, B) are suddenly struck due to the area gradation processing. You can avoid that.

On the other hand, the parameter Ctotal is Ctotal
Third ≥510 and expressed in K, R, G, B primary colors
If it belongs to the small color space of
It is determined whether it is 585 or more (S19). When the target value Ctarget is 585 or more, it is determined that the target value Ctarget belongs to the vicinity of the shadow in the first small color space, and black that is one of the primary colors of the small color space that belongs to the pixel. Distribute K (S20).

When the target value Ctarget is less than 585, the target value Ctarget-y for yellow Y, the target value Ctarget-m for magenta M, and the target value Ctarget-c for cyan C are compared (S21), and yellow Y When the target value Ctarget-y of C is the smallest, blue B (a superposition of magenta M and cyan C) which is a complementary color of yellow Y is distributed (S22), and when the target value Ctarget-c of cyan C is the smallest (S2
3), red R (compound of yellow Y and magenta M), which is a complementary color of cyan C, is distributed (S24), and when the target value Ctarget-m of magenta M is the smallest, magenta M
The complementary color of green G (yellow Y and cyan C are superimposed) is distributed (S25).

As described above, when the affiliation to the third small color space is determined, the color distribution is performed only among the four primary colors K, R, G, B expressing the third small color space. Since it is performed, although it is a color belonging to the third small color space near the shadow, dark dots (W, Y, M, C dots) that are light and bright due to the area gradation processing are sudden. You can avoid being hit by.

Further, the Ctotal is 255 <Ctotal <51
If it is 0 and belongs to the second small color space represented by the primary colors of Y, M, C, R, G, and B, the target value Ctarg
By determining whether et is 383 or less (S26),
In the second small color space, it is determined whether the area is close to the first small color space or the area close to the third small color space.

When it is determined that the target value Ctarget is 383 or less, the target value Ctarg of yellow Y is
By comparing the target value Ctarget-m of et-y and magenta M and the target value Ctarget-c of cyan C, the primary color having the highest proximity among yellow Y, magenta M, and cyan C is distributed (S14 to S18). ). If the target value Ctarget exceeds 383, the target value Ctarget-y for yellow Y, the target value Ctarget-m for magenta M, and the target value Ctarget-c for cyan C are compared to obtain yellow Y, magenta. The primary color having the lowest degree of proximity among M and cyan C is discriminated, and one of red R, green G, and blue B having a complementary color relationship with the discriminated primary color is distributed thereafter (S21 to S25).

When belonging to the second small color space, that is, neither near the shadow nor near the highlight,
As described above, Y, which represents the small color space to which it belongs,
Since the distribution is performed only among the primary colors of M, C, R, G, and B, white W dots and black K dots are not printed. As described above, when any one of the primary colors of white W, yellow Y, magenta M, cyan C, black K, blue B, green G, and red R is distributed for each pixel, the distribution (binarization) is performed. ) Error is calculated for each color (Y, M, C) (S6).

Then, the calculated error is distributed to surrounding pixels based on the error diffusion method, and the error components Aerror-y, Aerror-m and Aerror-c in the periodic pixels are set (S7),
Until the processing is completed for all pixels (S8), the above S1
To S7 are repeated. By the way, the periodic component th2
The reference matrix used for the setting may be common to each color, but a different matrix may be used for each color.

For example, while a 16 × 16 basic matrix matrix having a checkered pattern of 2 × 2 units is created (see FIG. 11), a matrix matrix1 (see FIG. 12) obtained by shifting the basic matrix matrix by two to the right is used. Basic matrix ma
A matrix matrix2 (see FIG. 13) is created by shifting trix downward by two. And the three matrices ma
trix, matrix1 and matrix2 are made to correspond to each color. For example, when the original color signals are R, G and B,
When R, the basic matrix matrix, when G, the matrix matrix1 and when B, the matrix matrix 1.
Try to use trix2.

According to this structure, any one of the complementary color combinations is preferentially generated in the gray halftone portion, and the uneven color is less likely to occur. Also, for example, when the original color signals are four colors of Y, M, C, and K, matrix is set for K and matrix1 is set for Y.
, And matrix 2 for M, and matrix 2 for C, respectively.

[0054]

As described above, according to the image processing method of the first aspect of the present invention, it is possible to suppress the grainy noise in the highlight and shadow portions, and the original image is displayed in halftone by error diffusion. It is possible to suppress the occurrence of a regular striped pattern irrespective of the above, and thus it is possible to improve the image quality of an image processed by the error diffusion method.

According to the second aspect of the present invention, the distribution ratio of the periodic component and the non-periodic component is appropriately changed according to the characteristics of the output medium such as the printer, and the granular noise and the regular striped pattern are generated. Is effectively suppressed for each output medium. According to the third aspect of the present invention, for example, the proportion of the non-periodic component is increased in the highlight and shadow portions, while the proportion of the periodic component is increased in the halftone to avoid the occurrence of the granular noise. There is an effect that a non-periodic component and a periodic component for avoiding the occurrence of a regular striped pattern can be added to each as needed.

According to the fourth aspect of the present invention, since the periodic component can be set in a form in which the change of the input image signal is added to the basic periodic change, a regular striped pattern can be formed without degrading the image quality. The effect is that the generation can be suppressed. According to the invention described in claim 5, it is possible to preferentially generate any one of the combinations of complementary colors in the gray halftone portion, so that it is possible to suppress the occurrence of color unevenness.

According to the invention of claim 6, there is an effect that a regular striped pattern can be effectively erased by the checkered pattern. According to the invention described in claim 7, by having the non-periodic component as a table, there is an effect that the processing can be simplified.

[Brief description of drawings]

FIG. 1 is a flowchart showing a first form of image processing.

FIG. 2 is a diagram showing an example of a 16 × 16 Bayer matrix.

FIG. 3 is a diagram showing a checkerboard matrix of 2 × 2 pixel units.

FIG. 4 is a diagram showing a checkerboard matrix of 4 × 4 pixel units.

FIG. 5 is a diagram for explaining how error diffusion is weighted.

FIG. 6 is a diagram for explaining how error diffusion is weighted.

FIG. 7 is a diagram showing a color space division configuration according to a second embodiment.

FIG. 8 is a diagram showing a color space division configuration according to a second embodiment.

FIG. 9 is a flowchart showing a second mode of image processing.

FIG. 10 is a flowchart showing allocation of primary colors in the second mode.

FIG. 11 is a diagram showing a checkerboard-like reference matrix in units of 2 × 2 pixels.

FIG. 12 is a diagram showing a matrix obtained by shifting the array of reference matrices by two pixels to the right.

FIG. 13 is a diagram showing a matrix in which an array of reference matrices is shifted downward by two pixels.

Claims (7)

[Claims] 1. An image processing method for expressing a color gradation image by a combination of density gradations and area gradations having a smaller number of levels than the original, wherein an error component is generated for an input signal to be processed. An image processing method characterized by adding a periodic component and an aperiodic component. 2. The image processing method according to claim 1, wherein the distribution ratio of the periodic component and the aperiodic component is changed depending on the output medium of the image signal. 3. The image processing method according to claim 1, wherein the distribution ratio of the periodic component and the non-periodic component is changed according to the input signal. 4. The image processing method according to claim 1, wherein the periodic component differs for each input signal. 5. The image processing method according to claim 1, wherein the periodic component is different for each color. 6. The image processing method according to claim 1, wherein the periodic component has a checkered pattern. 7. The image processing method according to claim 1, wherein the non-periodic component is included in a table.