JPH1127528A - Pseudo-medium gradation processing method, device and record medium therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent image degradation like connected dots, etc., from being generated, while maintaining preventing property on moires due to uneven texture and to hold high resolution due to adaptive modulation. SOLUTION: A cell is formed by successively gathering inputted pixels to which no binarization process is performed, adding values of the inputted pixels at the time and extending the inputted pixels until the total sum reaches a specified value by a cell setting part 2. A center of the cell is calculated, and a binarized and outputted pixel is set at the center of the cell by a pixel value setting part 3. An error in the cell caused by binarization is transferred to other cells in an update processing part 4. The cell is set for unprocessed pixels again, and a similar processing is repeated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a pseudo halftone processing method, apparatus and recording medium which do not cause image quality deterioration such as dot connection.

[0002]

2. Description of the Related Art Pseudo halftone processing of a digital image is as follows.
Continuous tone in which a pixel can have many values (for example, 256 values)
Data of a smaller number of gradations (eg, binary, quaternary, etc.)
This is the process of converting to.

A method of error diffusion processing, which is one of conventionally known pseudo halftone processing, will be described with reference to a block diagram shown in FIG. Here, it is assumed that 8-bit 256-value data is converted into binary data.

[0004] The error (err _{i, j} ) of the peripheral processed pixels recorded in the error buffer 21 is calculated by the product-sum unit 23 using a predetermined weighting coefficient w _{i, j} in the error mask 22. And the processed pixel error value (E _{i, j} ) is obtained.

[0005]

(Equation 1)

The processed pixel error (E _{i, j} ) is added to the input pixel value in _{i, j} by the adder 24, and the corrected input pixel value (su
m _{i, j} ) is calculated. The threshold processing unit 25 uses the first threshold (= 128) to calculate the corrected input pixel value (sum _{i, j} )
To obtain a binary output pixel value out _{i, j} .
In the subtracter 26, the corrected pixel value sum _{i, j} and the output pixel value o
The difference between ut _{i, j} is obtained, and the result err _{i, j} is recorded at a corresponding position in the error buffer 21.

[0007] Such an error diffusion method has the following two excellent features. Non-periodic; In pseudo gradation processing such as systematic dither processing, a periodic shading structure occurs in the output. If the image signal itself has a periodic structure, the image signal may interfere with the period of the pseudo halftone processing and cause moire. In this regard, the error diffusion processing has an advantage that moire hardly occurs because the processing itself does not have a periodic structure.

[0008] Adaptive modulation property: Each of the pseudo gradation processes applies some modulation to an input image signal. In the systematic dither processing, the cycle and phase (position) of the modulation are predetermined by a threshold matrix. On the other hand, in the error diffusion processing, the modulation period automatically changes when the modulation period is long in the low-density portion and the high-density portion and is short in the intermediate-density portion, and the phase (the position at which dots are formed) also changes in accordance with the input signal itself. Due to this property, the error diffusion processing has high resolution and gradation at the same time.

However, the error diffusion processing has the following disadvantages. Dot connection; for example, all input pixel values are 1/25
When a small number of pixels are turned on in order to reproduce a very thin input image such as 6, dots may be connected because the influence of pixels far away is difficult to reflect. This becomes low-frequency noise and is easily noticeable.

A mixture of periodic textures; for example, when irregular input is not always obtained for a flat input image having input pixel values of 128/256 or the like and periodic / aperiodic textures are mixed. There is. This is also noticeable as low frequency noise.

Calculation amount: Since the product sum of errors is calculated for each pixel, the calculation amount is large, and high-speed processing is difficult.

Decrease in stability; For example, in the case of an electrophotographic printer, an intermediate latent image potential is generated at an on / off switching portion, and an output result becomes unstable. Therefore, in such a printer, as the modulation period is shorter, the number of ON / OFF switching units is increased, and the output image becomes unstable as a whole.

[0013]

As a conventional method for solving the above-mentioned problem in the error diffusion method, there has been proposed a processing method in which dots are grouped into a halftone dot by periodically changing a threshold value (Japanese Patent Laid-Open No. HEI 9-163556). 62-239667). However, in this method, since the non-periodicity and the adaptive modulation property, which are the advantages of the error diffusion processing method, are weakened, the possibility of occurrence of moire is increased, and image quality deterioration such as reduction in resolution is caused.

An object of the present invention is to provide a pseudo-halftone processing method which does not cause deterioration of image quality such as dot connection while maintaining moiré prevention by irregular texture and high resolution by adaptive modulation, similarly to the error diffusion method. , An apparatus, and a recording medium.

[0015]

In order to achieve the above object, according to the present invention, N-valued digital image data having N (> 2) or more levels is input to each pixel, and A pseudo-grayscale processing method for converting the digital data into binary digital image data having two levels and outputting the binary digital image data, wherein input pixels are searched for in a predetermined order from a predetermined start point of the input data,
A pixel set (hereinafter, cell) composed of adjacent unprocessed pixels is generated based on the sum of the unprocessed input pixel values, and the output pixel value in the cell is determined based on the position of the input pixel and the input value. It is characterized in that it is set based on the sum of pixel values.

The invention according to claim 2 is characterized in that an error generated as a result of binarizing the cell is propagated to other cells to be processed thereafter.

The invention according to claim 3 is characterized in that an input pixel is searched in the order recorded in the table with reference to a table recording the relative position from the start point.

The invention according to claim 4 is characterized in that a plurality of tables having different search orders are prepared, and a table to be used is sequentially or randomly switched for each cell.

According to a fifth aspect of the present invention, the cell is generated based on the sum of the input pixel values and the number of pixels.

According to a sixth aspect of the present invention, the output pixel value in the cell is set based on the position of the center of gravity of the input pixel value and the sum of the input pixel values.

According to the seventh aspect of the present invention, N (>)
2) A pseudo halftone processing device comprising means for inputting N-valued digital image data having more than one kind of level, and means for converting and outputting binary digital image data having two kinds of levels for each pixel. Means for searching for input pixels in a predetermined order from a predetermined start point of the input data, means for adding values of unprocessed input pixels among the searched input pixels, and comparing the addition result with a predetermined threshold value Means for generating a pixel set (hereinafter referred to as a cell) consisting of unprocessed pixels adjacent to each other, and means for setting an output pixel value in the cell based on the sum of the input pixel position and the input pixel value And means for propagating an error generated as a result of binarizing the cell to other cells to be processed thereafter.

According to the eighth aspect of the present invention, N (>)
2) The input data for outputting pseudo halftone images by inputting N-valued digital image data having more than two kinds of levels and converting them into binary digital image data having two kinds of levels for each pixel A function of searching for input pixels in a predetermined order from a predetermined start point, a function of adding the values of unprocessed input pixels among the searched input pixels, and a function of comparing the addition result with a predetermined threshold value to be adjacent to each other. A function of generating a pixel set (hereinafter referred to as a cell) composed of the unprocessed pixels, a function of setting an output pixel value in the cell based on the sum of the input pixel position and the input pixel value, and A computer-readable recording medium on which a program for causing a computer to realize a function of propagating an error resulting from the binarization processing to another cell to be processed thereafter is recorded. That.

[0023]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. <Embodiment 1> In this embodiment, 8 bits (2
The input image data (56 gradations) is converted into 1-bit binary data. Here, the input pixel value 0 is white and the input pixel value 25
5 is black. The output pixel value 0 converted to binary is white, and the output pixel value 1 is black.

In the gradation processing according to the present invention, the following two-stage processing is repeated. That is, (1) setting of cells The pixels that have not been subjected to the binarization processing are put together to form a pixel set (hereinafter referred to as a cell). According to the content of the input image data, the input pixel value in the cell is 0 or 255
When the pixel value approaches the intermediate value 128, a large cell is created.

(2) Pixel Price in Cell According to the sum of the input image data in the cell, the output pixel value (0 or 1) in the cell is set so that the average value of the input / output in the cell becomes substantially equal. Is determined, and the values are put together at almost the center of the cell.

FIG. 1 shows the configuration of the first embodiment of the present invention.
In the figure, 1 reads image data of 256 gradations,
An input unit for preparing an array of various data, a cell setting unit for setting a cell which is a set of unprocessed pixels, a pixel value setting unit for setting an output pixel value in the cell, an error array, An update processing unit 5 for updating the array is an output unit for outputting binary image data.

FIG. 2 shows a processing flowchart of the first embodiment of the present invention. Hereinafter, the detailed operation of the present invention will be described.

Input unit 1: Read and prepare (Steps 101 and 102) Read 8-bit input image data into array In [i] [j]. An array Out [i] [j] for output image data having the same size as the input data, an array proc [i] [j] for recording a processed pixel position, and a binarization error recording array err [i] [j Is prepared. proc [i]
[J] and err [i] [j] are all initialized to zero.

Cell setting section 2: Search for unprocessed pixels (step 103) The array proc is scanned to find unprocessed pixels having an initial value (0). This is the starting point of the cell.

Cell Enlargement (Step 104) Unprocessed pixels around the above-mentioned starting point are sequentially put into the cell. (A) The cell search order is determined in advance, and the relative position from the start point is stored in a table. The table has the contents as shown in FIG. This indicates that the cells are enlarged in the order of the arrows in FIG. The unprocessed (0) pixels are sequentially recorded by referring to the processed array according to this table.

(B) Another table having a different search order as shown in FIG. 6 is prepared, and the table to be used is randomly switched for each cell. This prevents cells having the same shape from continuing, and makes the output image irregular. Further, a plurality of tables may be sequentially switched. Even in such a case, since the cells are not regularly arranged in the image, a similar effect can be obtained by substantially randomly switching the cells. This prevents image quality degradation due to regular mixture of textures.

End of enlargement (step 105) Input pixel value (In [i] [j]) at the pixel position entering the cell
And the error value (err [i] [j]) are sequentially added. If the pixel value at the start point is less than 128, the sum is 255 or more. , The cell is enlarged until the sum becomes (number of pixels−1) × 255 or less. When the size of the cell is determined in this way,
If only one pixel in a cell is set to 1 (when the pixel value at the start point is less than 128) or 0 (when it is also 128 or more), the average value of the input / output pixel values in the cell becomes substantially equal. .

For example, input image data (all 50) as shown in FIG. 7 (a), the already processed state of (b), (c)
(20, -10 are errors from separate cells), the search is performed in the order of FIG. According to the table of FIG. 5, (1), (2),
(3). . . The search proceeds in the order of. The addition result at each position is as shown in FIG.

Since the pixel value at the start point is 50, the sum of the input pixel value and the error becomes 260 when up to 5 pixels are incorporated into the cell at the position (6), and the enlargement ends when the sum exceeds 255. A cell composed of pixels is set. Thus, the size of the cell can be determined according to the average value of the input image.

Pixel value setting unit 3; Calculation of cell center position (step 106) When the pixel value at the starting point is less than 128, only one pixel is set to 1 (other pixels are set to 0), but this pixel position Is determined by the average value of the input pixel positions in the cell as in the following equation.

[0036]

(Equation 2)

Here, (x _i , y _i ) is the x and y coordinates of the i-th pixel in the cell, and n is the number of pixels in the cell. Also, i
nt (x) represents the largest integer not exceeding x.

When the pixel value at the start point is 128 or more, the center position of the 0 output pixel (1
One pixel is set to 0 and the other pixels are set to 1).

Output pixel value determination (step 107) If the center position is within the cell, the pixel is set to 1 (if the input pixel value of the starting point is less than 128, 0 if it is 128 or more, 0;
However, if the center position is outside the cell, a pixel in the cell is searched around that position. In the search, the order is determined in advance as shown in FIG. As a result, one 1 is located at a substantially central position in the cell.
(Or 0). Output pixel values at other positions in the cell are set to 0 (or 1).

For example, in the case of FIG. 7D, as a result of the calculation of the above equation, the center position is: x = int ((0 + 1 + 0-1-1) /5+0.5) = 0 y = int ((0 + 1 + 1 + 1 + 0) / 5 + 0 .5) = 1, and the output pixel value (1) is set as shown in FIG.

For example, when a pixel having an input pixel value of 16 (low-density pixel) continues, one cell is generated by 16 pixels, and 15 pixels become 0 as an output pixel value in the cell.
One pixel at the center of the cell is set to 1. As a result,
The average value (16/256) of the input pixel values in the cell is equal to the average value (1/16) of the output pixel values.

Similarly, when a pixel having an input pixel value of 150 (> 128) continues, one cell is generated by three pixels (sum of input pixel values = 450 <(3-1) × 255), and
As output pixel values in the cell, two pixels are 1 and one pixel is 0
And the average value of the input pixel values in the cell (150/2
56) and the average value (2/3) of the output pixel values are substantially equal.

Update processing unit 4 Update error array (step 108) The difference between the sum of the input pixel values in the cell and 255 times the sum of the output pixel value (1) occurs in the binarization processing in this cell. Error. If this error is ignored, the average pixel value that can be reproduced in the output image is limited to 1 / N. Therefore,
In order to compensate for this error in other peripheral cells, the error is recorded in an error array of adjacent unprocessed pixel positions. Here, this error is recorded at an unprocessed pixel position immediately below the center position. In the example of FIG. 10, the input pixel value in the cell is 260, and the total 255 times of the output pixel value (1) is 255, so the error is 5, and this error is recorded as shown in FIG. . Then, this value is used in the enlargement processing of the cell including the pixel position. As a result, it is possible to reproduce a smooth gradation without gradation jump.

Updating of the processed array (step 109) Finally, 1 indicating that processing has been completed is recorded in the processed array corresponding to each pixel position of the cell processed this time. The process is repeated from the cell setting until all the pixels of the binary image data have been processed. When all pixels have been processed, 2
The value image data is output (step 110).

In this embodiment, since the pixels which have been incorporated into the cells and which have been subjected to the binarization processing are recorded in the processed array, they are not used in the subsequent processing (other than referring to the processed array). Therefore, the operation required for the processing is as follows: for each pixel, two additions in the cell enlargement step, two additions in the center position calculation step, and for each cell having a smaller number of pixels than in the center position calculation step Only two additions and two divisions are required.

In the conventional error diffusion processing, the product sum of a large number (normally 5 to 10) of errors is calculated for each pixel (5 to 1).
0 multiplications and additions). Therefore, the processing method according to the present invention requires less computation than the error diffusion processing.

The sequence Out [i] [j], proc
[I] [j] and err [i] [j] do not always need to have the same size as the input data, and only the data covering the search range can be prepared and used while being shifted.

<Embodiment 2> This embodiment is an embodiment in which the present invention is implemented as a program executed in a general computer system as shown in FIG. The program is executed as follows.

1. A program created in advance is loaded into the computer system via an input device such as a floppy disk / communication device. 2. The multivalued digital image data to be processed is taken into a computer system via an input device such as a floppy disk / communication device. 3. By executing the program using a CPU, a memory, a hard disk, or the like, the binarization processing of the present invention described in the first embodiment is executed, and the result is recorded in the memory. 4. The processing result on the memory is used as follows.

Output from an image output device such as a display / printer. . Output to the outside via a data output device such as a floppy disk / communication device.・ Record in an external storage device such as a hard disk. Further, a processor, a memory in which a program is recorded, and a data recording memory are built in the printer, and the binarization process of the present invention can be realized by the printer alone.

Third Embodiment In the first embodiment, the end of cell expansion is determined by the sum of an input pixel value and a binarization error. For this reason, for example, when the input pixel value is 128, most of the cells have a size of two pixels, one of which is 0 and the other is 1. However, such an output binary image is ON / O.
Since the number of FF switching is increased, the output result tends to be unstable when output is performed by an electrophotographic printer or the like.

The present embodiment is intended to improve the instability due to such a too short modulation period. Therefore, such a problem can be improved by limiting the minimum cell size.

For example, as an end condition for cell expansion, FIG.
As shown in (1) that the number of pixels in the cell is 10 or more (step 20)
1) It is assumed that the sum of the input pixel value and the error value exceeds 255 (step 202). Steps other than steps 201 and 202 in FIG. 11 are the same as those in FIG. 2 and thus are omitted in FIG.

Thus, the sum of the input pixel values in the cell becomes 25
There is a possibility that it greatly exceeds 5. In that case, by searching in the order of FIG. 9 and setting a plurality of pixels in the cell to 1 one by one for 255 of the sum (for example, when 10 pixels having the input pixel value 128 in the cell continue, A process is performed in which one pixel is set to 1 for two pixels having a sum of 256, and 10
Among the pixels, 5 pixels are 1), and the average density can be maintained. When the search is performed in the order shown in FIG.
Pixels that become (or 0) are concentrated near the center of the cell, so that ON / OFF switching can be reduced.

Embodiment 4 In Embodiment 1, the center position of the output pixel in the cell is determined from the center position of the pixel in the cell. However, when the input pixel values in the cell are large or small, for example, when the input pixel values in the cell are distributed as shown in FIG.
It is considered that the reproducibility of the input image is better and the resolution of the output image is higher when 1 is set at the right end than when 1 is set at the center of the cell.

Therefore, in order to determine the position of the output pixel according to the magnitude of the input pixel value, the center of gravity weighted by the input pixel value can be obtained as in the following equation.

[0057]

(Equation 3)

[0058] Here (x _i, y _i), the cell in the i-th pixel of the x, y-coordinate, v _i is the pixel value of the i th pixel, n represents the number of pixels in the cell. Int (x) represents the largest integer not exceeding x.

When the pixel value at the start point is 128 or more, the position of the center of gravity of the 0 output pixel can be determined by the following equation in the same way, considering black and white in reverse.

[0060]

(Equation 4)

The periphery is searched around the weighted center of gravity in the same manner as in the first embodiment, and the output pixel value in the cell is determined. Even in this calculation, when the input pixel values are distributed flat as shown in FIG. 7A, it is the same as the simple center position, and 1 is placed near the center of the cell, so that dot connection can be prevented. As a result, an output with higher resolution can be obtained.

[0062]

As described above, claims 1 and 3,
According to the inventions described in the seventh and eighth aspects, by setting the pixel near the center of each cell to 1 (or 0), dots are almost evenly arranged in an input image close to white or black to prevent connection. Therefore, a high-quality binary image with little low-frequency noise can be obtained. Further, since the size of the cell is determined by the input pixel value, an output image with a high resolution can be obtained by adaptive modulation as in the error diffusion method.
Furthermore, it can be realized by a simple calculation as compared with the conventional method.

According to the second aspect of the invention, since the error generated in one cell is propagated to another cell, a smooth gradation can be reproduced.

According to the fourth aspect of the invention, by changing the search direction for cell expansion sequentially or randomly,
The output pixel distribution can be made irregular for any input, and a stable and irregular texture output can be obtained.

According to the fifth aspect of the present invention, the lower limit is set for the cell size and the output value in the cell is fixed at 1 (or 0), so that the ON / OFF switching of the pixel is reduced and the stable output is achieved. Can be obtained.

According to the sixth aspect of the present invention, the output pixel position is determined according to the distribution of the input pixel values, so that a high-resolution output can be obtained.

[Brief description of the drawings]

FIG. 1 shows a configuration of a first exemplary embodiment of the present invention.

FIG. 2 shows a processing flowchart of Embodiment 1 of the present invention.

FIG. 3 shows a configuration of a second exemplary embodiment of the present invention.

FIG. 4 shows a search order of cell pixels.

FIG. 5 shows an example of a cell pixel search position table.

FIG. 6 shows another example of a search order of cell pixels.

FIGS. 7A to 7D show an example of a cell setting process.

FIG. 8 shows a process of expanding the cell of FIG. 7;

FIG. 9 shows a search order of output value setting pixels.

FIG. 10 shows an example of setting a pixel value in a cell.

FIG. 11 shows a part of a processing flowchart of the second embodiment.

FIG. 12 shows an example in which input pixel values in a cell are offset.

FIG. 13 shows a configuration of a conventional error diffusion process.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input part 2 Cell setting part 3 Pixel value setting part 4 Update processing part 5 Output part

Claims (8)

[Claims] 1. A pseudo intermediate for inputting N-value digital image data having N (> 2) or more levels for each pixel, converting the data into binary digital image data having two levels for each pixel, and outputting the converted data. A tone processing method, wherein an input pixel is searched for in a predetermined order from a predetermined start point of the input data, and a pixel set of unprocessed pixels adjacent to each other based on the sum of unprocessed input pixel values ( (Hereinafter referred to as a cell), and an output pixel value in the cell is set based on the sum of the input pixel position and the input pixel value. 2. The pseudo halftone processing method according to claim 1, wherein an error generated as a result of binarizing the cell is propagated to another cell to be processed thereafter. 3. The pseudo halftone processing method according to claim 1, wherein a table in which a relative position from the start point is recorded is referred to, and input pixels are searched in the order recorded in the table. 4. The pseudo halftone processing method according to claim 3, wherein a plurality of tables having different search orders are prepared, and a table to be used is switched sequentially or randomly for each cell. 5. The pseudo halftone processing method according to claim 1, wherein the cell is generated based on the sum of the input pixel values and the number of pixels. 6. The pseudo halftone processing method according to claim 1, wherein the output pixel value in the cell is set based on the position of the center of gravity of the input pixel value and the sum of the input pixel values. 7. A means for inputting N-valued digital image data having N (> 2) or more levels for each pixel, and converting the output into binary digital image data having two levels for each pixel and outputting the same. Means for searching for input pixels in a predetermined order from a predetermined start point of the input data, and adding a value of an unprocessed input pixel among the searched input pixels. Means for generating a pixel set (hereinafter referred to as a cell) composed of unprocessed pixels adjacent to each other by comparing the addition result with a predetermined threshold value, and outputting an output pixel value in the cell as a position of the input pixel. A pseudo-halftone processing device, comprising: means for setting based on the sum of input pixel values; and means for propagating an error generated as a result of binarizing the cell to other cells to be processed thereafter. . 8. Pseudo halftone by inputting N-value digital image data having N (> 2) or more levels for each pixel and converting it into binary digital image data having two levels for each pixel A function of searching for input pixels in a predetermined order from a predetermined start point of the input data, a function of adding values of unprocessed input pixels among the searched input pixels, and And a function of generating a pixel set (hereinafter referred to as a cell) consisting of unprocessed pixels adjacent to each other by comparing the output pixel value in the cell with the sum of the position of the input pixel and the input pixel value And a computer-readable recording recording a program for causing a computer to implement a function of setting a function based on the function and a function of transmitting an error generated as a result of binarizing the cell to other cells to be processed thereafter. Medium.