JPH04159862A - Picture processing system - Google Patents

Abstract

PURPOSE:To improve the gradation and contrast of an output picture without lowering the resolution so that a visually smooth picture can be outputted by making the distributing method of a picture element matrix which is converted in density as the gradation of input picture signals increases different between low-and high-density input picture signals. CONSTITUTION:Input picture signals A are converted into picture data D which represent an intermediate tone picture by using a picture element matrix composed of multi-gradation picture elements after picture-processing is performed by a picture processing system. The distributing method of picture elements which are converted in density as the gradation of the signals A increases is made different depending upon the used picture element matrix between low- and high-density input picture signals. When, for example, a picture element concentrating type distributing method is used, the picture elements which are converted in density in the matrix are outputted in a concentrated state and, when a picture element scattering type distributing method is used, the picture elements which are converted in density in the matrix are outputted in a discrete state. Therefore, the output picture becomes a visually smooth picture, since the gradation and contrast become excellent without deteriorating the resolution.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is used in an image forming apparatus such as a copying machine, and performs image processing on a digital image signal to output image data representing a halftone image. Method [Prior Art] Conventionally, in an image processing method for an image forming apparatus such as a digital copying machine, an input image density signal is quantized to a 2'' level as is well known, and a digital multivalued image is created in which the density is expressed using n bits. Various types of image processing are performed after converting the image into .

Among them, the dither method and density pattern method use the integral effect of the human eye to effectively convert digital multi-value image data into binary image data representing halftones to express light and shade. It is frequently used in the above devices because it has good restoration accuracy and is easy to handle.

[Problem to be solved by the invention]

However, either the dither method or the density pattern method has the disadvantage that when expressing a halftone image by changing the number of recorded dots, the ability to express gradation and the resolution are not compatible. In other words, if the size of the matrix pattern is increased in order to increase the number of gradations of image density, the resolution will decrease, and if the size of the matrix pattern is decreased to improve the resolution, it will not be possible to increase the number of gradations of image density. It is a point.

Also, if we pay attention to the type of matrix pattern, the number of black pixels (less than 100x) increases as the density increases.If the dot concentration type grows around a certain core pixel, the number of black pixels increases. Although the change in the density gradation of the output image corresponding to the number is close to linear, the resolution is slightly lowered. On the other hand, in the case of the dot dispersion type, which uniformly increases the number of black pixels without creating a core pixel, the resolution does not decrease much7, but on the other hand, the number of black pixels in the matrix and the density gradation of the output image The linearity of the change tended to deteriorate.

As one method to solve these problems, a multi-level dither method has been developed that enables multi-level density output by setting multiple darkness values for one pixel, but the root of the above problems is No solution has been reached.

In other words, if a dot-concentrated matrix pattern based on method 1 is used, the resolution problem cannot be solved, and if a dot-distributed matrix pattern is used, the recorded image of fine pixels becomes unstable. The problem was that it was easy to become

The present invention has been made in view of the above circumstances, and uses an image processing method that expresses halftone images using a matrix pattern of multi-tone pixels without degrading the resolution and improving the gradation and contrast of the output image. The purpose of the present invention is to provide an image processing method that can output an excellent and visually smooth image.

[Means to solve the problem]

In order to solve the above-mentioned problems, the present invention has a first means that performs image processing on an input image signal to convert it into image data representing a halftone image using a pixel matrix consisting of multi-tone pixels. In this method, the distribution method of the pixels of the pixel matrix whose density is converted as the gradation level of the input image signal increases is different between a low-density input image signal and a high-temperature input image signal.

Further, the second means is the method for changing the density in the first means.
There are two methods for distributing pixels to be converted: a pixel concentrated distribution method in which pixels whose density is converted as the gradation level of the input image signal increases are distributed sequentially to a predetermined pixel and pixels adjacent to the pixel; This method includes a pixel dispersion type distribution method in which pixels are sequentially distributed to pixels located spatially apart from each other.

Further, in the third means, in the second means, the pixel distributed distribution method configures the pixel matrix with a plurality of sub-matrices, and the pixels whose density is converted as the gradation level of the input image signal increases are transferred to the sub-matrix. The distribution is made sequentially for each time.

Also, as for the fourth means, what is the pixel concentrated distribution method in the second means? The distribution is made for each pixel in order from the pixel with the lowest agricultural gradation.

Further, in the fifth means, in the second means, the pixel concentrated distribution method for a low-density input image signal continuously converts pixels whose density is converted as the gradation level of the input image signal increases into the same pixel. This is a distribution method to distribute the data.

[Effect]

The input image signal is subjected to image processing by the image processing method described in item 1, and converted into image data representing a halftone image using a pixel matrix consisting of multi-tone pixels. The distribution method of pixels whose density is converted as the gradation level of the input image signal increases by the pixel 7 matrix used at that time differs between a low-density input image signal and a high-density input image signal.

When the gradation level of the input image signal increases, for example, in the pixel concentrated distribution method, the pixels whose density is converted in the matrix are intensively output, and in the pixel distributed distribution method, the pixels whose density is converted in the matrix are outputted intensively. Output discretely.

Further, in a pixel distributed distribution method in which a pixel matrix is composed of a plurality of submatrixes, pixels whose density is converted as the gradation level of an input image signal increases are sequentially output for each submatrix.

In addition, in the pixel concentrated distribution method, pixels are distributed sequentially from pixel to pixel with the lowest density gradation, or pixels whose density is changed as the gradation of the input image signal increases for a low density input image signal. Continuously distribute to the same pixel.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 5 is a schematic block diagram of an image processing circuit of a digital copying machine according to an embodiment of the present invention. In the figure, 1 is a line image sensor (LIS) that reads a document image line by line and converts it into an analog image signal A, and 2 is an A/D converter that converts the converted analog image signal into a digital image signal. , 3 is a softening correction circuit that corrects image distortion and image density variation caused by uneven illuminance on the document surface and sensitivity variation of the line image sensor 1.

4 is a multi-value processing unit that performs multi-value processing to bit-convert the digital image signal D' into a multi-value number using a multi-value dither, a multi-value density pattern method, etc., and converts the digital image signal D'' into image data D'' corresponding to the image density. 5 is an output control circuit, which outputs a reproduced image signal C according to the input image data D#. 6 is a multilevel printer, which outputs a reproduced image signal C according to the input image data D#. 7 is a control unit which outputs control signals such as synchronization signals to each of the circuit units 1 to 6 to control image processing operations.

FIG. 6 is a diagram showing control signals for timing control in control unit No. 7 in correspondence with the reading operation of a document image, where M is a document, FGATE is a signal representing the effective document width in the sub-scanning direction, and LGATE is a signal representing the effective document width in the sub-scanning direction. is a signal representing the effective document width in the main scanning direction, and LSYNC is a synchronization signal for reading in the main scanning direction.

The original M image is read line by line in the main scanning direction in synchronization with L S Y N C, and the FGATE and LGAT
It becomes a valid image signal only when both E are HIGH. The read image signal is output from the line image sensor 1 pixel by pixel in synchronization with the reference signal CLK of the control unit 7.

Fig. 1 is a block diagram showing the internal circuit of a multi-value processing unit 4 using the multi-value dither method, and Fig. 2 shows the value of the output image data D'' of the multi-value processing unit 4 and the value of the input image data D'. In this embodiment, the output image data D'' has five values from 0 to 4.

A main scanning counter 41 counts pixels in the main scanning direction according to CLK and outputs an address signal CNTl. A sub-scanning counter 42 similarly counts lines in the sub-scanning direction according to LSYNC and outputs an address signal CNT2. 4o is a teaser ROM (DROM) that determines and outputs the value of the output image data D'' based on the magnitude relationship between the redither threshold TH and the value of the input image data D'. 42 has a ring counter configuration, and CNT1 and CNT2 correspond to each image address in the DROM 40. Therefore, the DROM 40 stores the threshold data THz (i-1 to 4) at each address and the value of the input image data D'. By comparing the magnitude relationships between the two, multivalued image data D'' corresponding to the recording density is generated according to the relationship diagram shown in FIG.

Darkness value data TI-(.

It is possible to form various dither matrices by changing .

Figure 3 shows 4X4=16 pixels P□, +(i, j=1 to 4)
Then, one pixel is further divided into four fine pixels DPijk (k-1~
4) shows the dither matrix constructed in step 4). In this example, one pixel p is divided into four fine pixels Dp,k (k=1 to 4) by pulse width modulation, and one pixel outputs gradation of five values (D''-〇 to 4). It makes it possible.

4(a) to (11) show density patterns of pixels Pij of a recorded image formed in accordance with multi-value image data D'' output from the multi-value processing unit 4.

FIG. 7 shows the dither matrix ■ when outputting the halftone output image data D'' of the DROM 40 according to the first embodiment. In the figure, the large frame indicates the pixel PiJ, and the narrow frame indicates the fine pixel DP. , , k are respectively represented, and the numerical values in the frames indicate the threshold data THiO value corresponding to each fine pixel.
It takes a value of ~64, with 0 and 64 corresponding to white and black densities, respectively.

Therefore, in the dither matrix I, if we pay attention to the pH pixel, for example, when the input image data D' is 19, the fine pixel DP
When iz is 20 to 32, fine pixel DPz+ and DPI
When Ig is 33 to 48, fine pixel DP+z to DP113
is 49 or more, the fine pixels DPz+ to DpH4 are output as black density output image data D#.

The feature of the dither matrix I of this embodiment is that when the value of the input image data D' is relatively small in the range of 0 to 32, the density of one pixel p becomes half black (average density 50%). This is a so-called dot concentration type dither matrix array in which the density is concentrated for each pixel, and the density of adjacent pixels is increased sequentially around a core pixel.When the number is in the rather large range of 33 to 64, one pixel p is used. without successively increasing the density gradation of pixel Pij, and sequentially increasing the density gradation of pixel Pij so that the next pixel whose density gradation is to be increased is not spatially close to each other;
The point is that they are arranged in a so-called dot-distributed dither matrix arrangement.

Figures 1θ (a) to (h) show the output density patterns of the matrix (directly corresponding to the input image data D' with uniform density) by changing the values of the input image data D',
(a), (b), (c), (d), (e),
(f), (g) and (h) show output density patterns when the values of input image data D' are 4.8.16.32.36.40.48 and 64, respectively.

First, when the value of input image data D' is relatively small, (
a) As shown in (D'=4), two fine pixels DPtti, DPz□, DP, 3°, DP
z3z is output as the black density. When D′=8 (
b) As in the above 2 Nishimoto P. , two pixels P directly above Po
l! + Fine pixel DP in P13, z+, DPrt
z+DP+2+rDPi3z is output as black density.

Furthermore, when D'=16, as shown in (C), the 2 pixels P R7, P 13 near the center 6 pixels P I
t+ P I 3+P! Ii P t4r P s
Fine pixel DPjjk (k=1
．． 2> is output as the black density. At D'=32, (d
As shown in FIG. 1, all the pixels P, DPii*□c=1,2) in the left half of the pixels P44 are output as black density.

In this way, when D'≦32, the value of input image data D' is 0.
As the density is gradually increased from 1 to 3, the fine pixels in the left half of the adjacent pixel are changed to black density one by one, with pixel PtZ as a core, while the right half of pixel Pij is left blank. Next, when the value of the input image data D' increases and exceeds 32, which is half of the total black, each time the value of D' increases by one, the number of pixels in the submatrix that is divided into four adjacent pixels is increased. The white density ? located on the left is sequentially changed pixel by pixel by changing the submatrix and covers all pixels PiJ. The density of IF, pixel DP, . . . is sequentially changed to black. (e) to (hl show representative examples thereof.

FIG. 8 shows the dither matrix ■ when outputting the halftone output image data D'' of the DROM 40 according to the second embodiment. The illustrated frame and numerical values are as explained in the first embodiment. It is the same as the thing.

The feature of the dither matrix (■) of this embodiment is that when the value of the input image data D' is relatively small in the range o to 32, the first Contrary to the embodiment described above, instead of increasing the density gradation of one pixel continuously, the submatrix is subdivided one pixel at a time so that the pixels P and J whose density gradation is to be increased next are not spatially close to each other. Change the matrix and sequentially all pixels p, \si
A so-called dot-distributed dither matrix arrangement is used in which the density of the fine pixels DP, . Furthermore, the value of D′ is slightly larger 33
~64, the neighboring pixels p are concentrated around the core pixel.
, the density of the surrounding pixels is increased in sequence, and the density gradation is increased by one for each pixel P4, which is a dither matrix arrangement similar to the so-called concentrated dot type.

FIGS. 11(a) to (h) show output density patterns of a matrix corresponding to input image data D' of uniform density by changing the values of the input image data D', and (al
, (b), (cl, (di, (e), (rl
, fg) and (h) are input image data D'
The output density patterns when the value of is 4.8.16°, 32.3.6, 40.48, and 64 are shown.

First, the value of input image data D' is relatively small D'=4
In this case, as shown in (a), each pixel PfJ in the submatrix, which is obtained by dividing the 4×4 matrix into four adjacent pixels, is divided into the white density minute pixels located on the left. The density of pixels DPiJk is sequentially changed to black. That is, starting from the upper leftmost pixel P, the pixels Pza,
The leftmost fine pixels DPz+, D in the order of pHPz+
Pzn+, DPI31, and DPffil+ are output as black density. When D'-8 is reached, the black density conversion goes through one cycle for each pixel p in the next submatrix, and a mesh density pattern is output as shown in (b). D'=1
6, all pixels Pi = (i, j = 1 ~
The leftmost fine pixel DPijk ck=1) in 4) has a black density, and a density pattern of a thin vertical stripe pattern is output. D'=
In 32, as shown in (d), all pixels P are positive, (i, j=1
The left half fine pixel DP,k (k=1.2) in ~4) has a black density and outputs a thick vertical striped density pattern.

In this way, when D'≦32, the value of input image data D' is O
When gradually increasing from 1 pixel p in the submatrix, the fine pixel DPij of white density located at
This is a so-called dot-dispersed dither matrix arrangement that changes the density of * to black.

Next, when the value of the input image data D' increases and exceeds 32, which is half of the total black, each time the value of D' increases by one, the pixel PR The fine pixels DPi,k within the pixel are changed to black density one by one. fel is four adjacent pixels P I! , P L3+
P2ffi, P! : fine pixel DP+zs in l.

DP113, DPzz3. DPt3s indicates the density pattern newly output as the black density.

(f) is a further adjacent four pixels P! I+P
z4+P ff2+P3! This shows that the density of the micropixels within has newly changed to black. Typical example (the illusion is ultimately the same as the first example. In this way, when D'≧32, as the value of input image data D' gradually increases from 0, the nuclear pixel This is a dither matrix arrangement similar to a so-called concentrated dot type, in which the DP'th fine pixel in the adjacent pixel Pij is changed to black density one by one with P0 as the center.

FIG. 9 shows the dot-concentrated dither matrix ■ when outputting the halftone output image data D'' of the DROM 40 according to the reference example, and FIG. 12 (al to (h) shows the uniform density input image data The output density pattern of the matrix by the dither matrix ■ corresponding to D' is shown by changing the value of the input image data D', (al, (b)
, fc), (dl, (e), (f), (g)
and (h) show output density patterns when the values of input image data D' are uniformly 4.8, 16, 32, 36, 40.48, and 64, respectively.

In this dither matrix ■, each time the value of D' increases by one, the black density is changed one by one for each fine pixel DPzI (k=1 to 4) within the pixel, with the pixel P2t on the upper left side of the center as the core. When pixel P2□ becomes completely black, adjacent pixel p i
This is a typical dot concentration type arrangement in which the density gradation of pixel p is successively raised in the same way as pixel p.

In such a dither matrix ■, the density gradation output method is similar to the binary dither method, so the number of black density fine pixels and the number of gradations form a linear pattern, improving the contrast of the output image, but the intermediate Since the tonal image is expressed by a collection of dots, the grain of the output image becomes rough, the resolution deteriorates, and the line image is more likely to have jitter, etc., and the advantages of the multilevel dither method are not fully utilized.

゛On the other hand, in the density gradation output using the dither matrix I of the first embodiment, a different sub Since it moves to the pixels of the matrix, it is possible to compensate for the unstable spread of minute recording pixels of one output minute pixel, and the non-linearity of the number of pixels converted to black density and recording density, improving gradation reproducibility and It becomes possible to output recorded images with good contrast. In addition, in the range of high-density input image data, by using a dot-distributed method for outputting density gradations,
It is possible to prevent line images from blurring and output visually smooth recorded images with excellent resolution.

On the other hand, the density l by the dither matrix ■ of the second embodiment
By making the density gradation output method dot-distributed in the range of low-density input image data, it is possible to prevent the background part of the document from being output as a recorded image with a regular array of dots. , can output visually smooth recorded images with excellent resolution.

Furthermore, in the range of high-density input image data, by changing the density gradation output method to a pseudo-dot concentration type, it is possible to output a recorded image with good contrast without significantly deteriorating the resolution.

FIGS. 15(i), (ii), and (ii) show the output density patterns of the matrix by the dither matrices ■, ■, and ■ of the almost uniform low-density input image data D' shown in FIG. 13(a). , and Fig. 16(i),(ii)
.

(iii) shows the output density pattern of the matrix based on the dither matrices ■, ■, and ■ of the almost uniform high-density input image data D' shown in FIG. 13 (bl).

As is clear from these figures, for the low-density input image data D' shown in FIG. The image becomes rough. On the other hand, the output density pattern (i) of the matrix produced by dither matrix I becomes a density pattern that is somewhat similar to a dot-like pattern (array) and causes image blur, but since there is a white background between images, the roughness of the wood grain is considerably improved. Ru. In addition, since the output density pattern of the matrix by the dither matrix H is uniformly dispersed, a smooth output image with fine grains is produced without image distortion, but the stability of the recorded pixels is somewhat lacking.

The output density pattern (iii) of the matrix based on the dither matrix ■ of the high-density input image data D' shown in FIG. This results in an image with rough grain, similar to that for the density input image data.

On the other hand, the output density patterns (i) and (ii) of the dither matrices I and H produce smooth output images because the white areas between the images are almost evenly distributed. In particular, the output density pattern (ii) by the dither matrix H is a pseudo-dot concentrated type, so it looks a little rough visually, but it has excellent density gradation and contrast.

Figure 17 (i), (ii), and (iii) are the first
Input image data D representing low-density diagonal lines shown in Figure 4 (a)
' dither matrix 1. The output density pattern of the matrix according to II and ■ is also shown in FIG. 18(i), (
ii).

(iii) are the dither matrices I and II of the input image data D' representing the high-density diagonal lines shown in FIG. 14 (1)).
This figure shows the output density pattern of the matrix according to and ■.

In these figures, similarly to the output density pattern of the uniform input image data D', the low density input image data D
′, the output density pattern (iii) by the dither matrix ■ becomes a dot-like (array) density pattern and causes obvious image (line) toggle, whereas the output density pattern (iii) by the dither matrix x0)tg degree pattern (i) In this case, the output density pattern becomes somewhat dot-like (array) and the image (line)
Although some jerking occurs, the resulting output image has excellent recording pixel stability. The output density pattern (ii) using dither matrix ■ hardly causes image (line) toggle, but
The stability of recording pixels becomes slightly unstable.

The matrix output density pattern shown in FIG. 18(iii) based on the dither matrix 3 of the input image data D' representing a high-density diagonal line shown in FIG. On the other hand, in the output density patterns of 7 trisoxes based on dither matrices I and 2 shown in FIGS. 18(i) and (ii), not only no image (line) oscillation occurs, but also a smooth line image is output.

In the explanation of the above embodiment, the pixel Pi of the matrix 7 is converted into black density as the value of the input image data D' increases.
The method of arranging J has been explained according to a dither matrix that is divided into two stages, a low density region and a high density region, but of course the division of the density region is not limited to two stages. An arrangement may be adopted in which the area is divided into three or more stages.

In addition, the method of arranging the pixels pij of the matrix whose density is converted in the dither matrix has been explained using either the dot concentrated type or the dot dispersed type depending on the density range, but the method is not limited to this. However, other arrangement methods may be used, or a matrix arrangement of the same type but with different patterns may be used. The dither matrix size is not limited to 4×4 pixels, but can be implemented in the same manner even if it is of another size.

Furthermore, although the image processing method for converting the image density signal into digital image data representing a halftone image has been explained using the dither method as an example, it is also possible to use other image processing methods such as the density pattern method. I do not care.

In addition, we have explained how to make the pixel density multivalued by pulse width modulation that divides the lighting time for one pixel to form fine pixels. Multi-value means such as intensity modulation for controlling the amount of light may also be used.

〔Effect of the invention〕

As explained above, according to the present invention, a pixel matrix is used in which the distribution method of pixels whose density is converted as the gradation level of the input image signal increases is different for low-density input image signals and high-density input image signals. As a result, the input image signal is subjected to image processing and converted into multi-tone image data that represents a half-tone image, so the output image does not deteriorate in resolution, has excellent gradation and contrast, and is visually smooth. The result is a beautiful image.

Furthermore, for example, a pixel distributed distribution method configured with multiple sub-matrices, a pixel-by-pixel distribution method starting from the lowest density gradation, or a density conversion method for low-density input image signals as the gradation level increases. By using a pixel matrix based on a pixel concentration distribution method in which pixels are continuously distributed to the same pixel, a recorded image that is adapted to the characteristics of the input image signal can be obtained.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the internal circuit of a multi-value processing unit of a digital copying machine according to an embodiment of the present invention, and FIG. 2 shows how the value of output image data of the multi-value processing unit is adjusted according to the value of input image data. Figure 3 is an explanatory diagram showing a dither matrix using the multi-value dither method composed of 4 x 4 pixels, and Figures 4 (a) to (e) are diagrams showing the relationship between the multi-value image data. The density pattern of the pixel formed by 1. FIG. 5 is a schematic block diagram of an image processing circuit of a digital copying machine according to an embodiment of the present invention, and FIG. 6 is an explanatory diagram showing control signals for timing control in correspondence with document image reading operations. , FIG. 7 and FIG. 8 are explanatory diagrams showing dither matrices according to the first and second embodiments, respectively, FIG. 9 is an explanatory diagram showing a dither matrix according to a reference example, and FIG. 10 ( a) to (h) and FIG. 11 (al to (h) are explanatory diagrams showing output density patterns corresponding to values of uniform density input image data according to the first and second embodiments, respectively; Figures 12(a) to th+ are explanatory diagrams showing output density patterns corresponding to input image data with uniform density according to a reference example, and Figures 13(a) and (b) are almost uniform low density and high density patterns. Figure 14 (al) is an explanatory diagram showing a dither matrix of input image data, (b) is an explanatory diagram showing a dither matrix of input image data representing diagonal lines of low density and high density, i), (ii
), (iii) and Figure 16 (i), (ii
) + (iii) Figure 13 (a) Oyohi (b)
An explanatory diagram showing output density patterns of embodiments and reference examples of almost uniform low-density and high-density input image data shown in D;
Figure 17 (i), (ii), (iii) and Figure 18 (i), (ii), (iii) are shown in Figure 14 (a).
) and (t) are explanatory diagrams showing output density patterns of an example and a reference example of input image data representing diagonal lines of low density and high density shown in l. 4... Multivalue processing unit, 40... Dither ROM,
41... Main scanning counter, 42... Sub-scanning counter. Figure 1, 4th? Figure (C no.
(b)-(c)
(d) Figure to (e) (
f) D'=7+8D=64 (O)
(D) (c)
(d) D'=1
6 D', 32 Figure 11 (e)
(f)(a),
(b) [)'
4 D'=8(C)

(d) D'=16D door 32 (g)
(h) D 48 1)
Hiroshi Fig. 13 (a) (b) Fig. 14 (O) (b) (fil No. 16 (j)
(it) (77j) (I)
(if) (i)
(H)(
iH)

Claims (5)

[Claims] (1) Using a pixel matrix consisting of multi-tone pixels,
In an image processing method that performs image processing on an input image signal and converts it into image data representing a halftone image, the difference between a low-density input image signal and a high-density input image signal increases as the gradation of the input image signal increases. An image processing method characterized in that the pixels of the pixel matrix subjected to density conversion are distributed in different ways. (2) In the description of claim 1, the distribution method of pixels subjected to density conversion is such that pixels subjected to density conversion as the gradation level of an input image signal increases are divided into predetermined pixels and pixels adjacent to the pixels. An image processing method comprising: a pixel concentrated distribution method that sequentially distributes the pixels; and a pixel distributed distribution method that sequentially distributes the pixels to a certain pixel and pixels located spatially apart from the pixel. (3) In the description of claim 2, the pixel distributed distribution method is such that a pixel matrix is composed of a plurality of sub-matrices, and pixels whose density is converted as the gradation level of an input image signal increases are arranged in the sub-matrix. An image processing method characterized by sequentially distributing each image. (4) The image processing method as set forth in claim 2, characterized in that the pixel concentrated distribution method sequentially distributes each pixel starting from a pixel with a low density gradation. (5) In the description of claim 2, the pixel concentration distribution method for low-density input image signals continuously converts pixels whose density is changed as the gradation level of the input image signal increases into the same pixel. An image processing method characterized by being a distribution method.