JP2001257879A - Half-tone reproducing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To actualize tone reproduction with a natural feeling closer to an original image by removing the generation of unnatural color half-tones and widening their tone reproduction range. SOLUTION: The dot growth style of a dither matrix is made different corresponding to a reproduced tone area.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing technique used for printing or image display, and more particularly to a halftone reproducing method for quantizing multi-valued image data using a dither matrix.

[0002]

2. Description of the Related Art As a technique for reproducing a color halftone used for printing or image display, there are a dither method and an error diffusion method which is a kind of the dither method.

In the dither processing, the threshold is set to n for each pixel.
Using a dither matrix arranged in a matrix of × m, comparing multi-valued image data input in pixel units with each threshold value of the dither matrix and one pixel unit,
For example, if the input multi-valued image data ≧ the threshold value, the output image data is “1”, and if the input multi-valued image data <the threshold value, the output image data is “0”. Quantization conversion to an n value is performed to obtain a halftone image.

The dither method is classified into a random dither method and a regular dither method according to a dither matrix used. In the random dither method, a dither matrix is formed using random numbers having a uniform distribution of threshold values. However, in this method, it is difficult to obtain good image quality because granular noise is present on the entire image.

[0005] The regular dither method has many types, but is roughly classified into two types: a dot dispersion type and a dot set type. Here, FIG. 9 shows dots D formed by a concentrated 8 × 8 dither matrix.
The dot DT is a set of pixel-based output image data that has become “1” in the dither matrix comparison process in one dither matrix (cell).

[0006] The dot dispersion type dither matrix is characterized by adopting a threshold pattern in which threshold values are relatively evenly distributed in the matrix. That is, large thresholds tend to be located near small thresholds. The dot dispersion type has a surface suitable for halftone reproduction, but there is a problem in that a hard system in which individual pixels are so small that they cannot be faithfully reproduced cannot make an image look cloudy and dirty. Further, in this dot dispersion type, the reproduced image has a somewhat insufficient feeling in terms of sharpness.

[0007] As shown in FIG.
In order to express the gradation, the dots are grown so that the dots gradually increase with the center of the dither matrix as a nucleus. The dot set type is a method in which pixels are collected in a dither matrix to form dots, so that halftones are represented by the size of the dots. The dot set type has a coarser image than the dot dispersion type, but in a high-resolution system, the dither matrix itself is very small, so the image roughness does not matter.

The above-mentioned dot set type has various types depending on how the shape of the dot DT changes as the tone value increases. The dot growth shape includes a circle, an ellipse, a line (linear), a square, a cross, a rhombus, and the like.

[0009]

In the tone reproduction using the dot-assembled dither matrix, in the low tone region, even if the dot growth shape is different, there is no visually significant difference. Depending on the shape, the gradation value level-output density characteristic does not change linearly, but a periodic fluctuation pattern occurs. This is thought to be due to the fact that the dot shape collapses with the spread of the dot area, and is easily noticeable when the dot growth shape is a rhombus, cross, or square.

[0010] Further, also in the intermediate tone region and the high tone region, a difference occurs in tone reproduction characteristics due to a difference in dot growth shape.

For example, when the dot growth shape is a circle, the characteristics of the output density signal corresponding to the input multi-valued image data increase almost linearly and smoothly in the intermediate tone range, but large random values in the high tone range and higher. Fluctuations with fluctuations occur.

In the gradation value level-output density characteristic when the dot growth shape is a rhombus, a fluctuating portion having a sharp density fluctuation periodically occurs in the intermediate tone range and the high tone range. However, the gradation value level-output density characteristic when the dot growth shape is a rhombus has a larger gradient and a wider reproduction density width than the case where the dot growth shape is a circle, so that the contrast becomes clearer. A reproduced image can be obtained.

As described above, in tone reproduction using a dot concentration type dither matrix, differences in dot growth shape, such as shape collapse at the time of dot area change, regular fluctuation patterns, and instability of density change, occur due to differences in dot growth shape. And a discontinuous change in density occurs. That is, there are various dot growth shapes in the dot concentration type dither matrix, but each has its advantages and disadvantages.

The present invention has been made in view of the above circumstances, and eliminates the occurrence of unnatural color halftones and expands the tone reproduction range by incorporating the characteristics and advantages of various dither patterns. Accordingly, it is an object of the present invention to provide a halftone reproduction method for realizing tone reproduction with a more natural feeling close to the original image.

[0015]

According to one aspect of the present invention, in a halftone reproduction method for quantizing multi-valued image data using a dither matrix, a dot growth pattern of the dither matrix is different depending on a reproduction tone range. It is characterized by being made to be.

According to the present invention, the dot growth form of the dither matrix (circle, ellipse, line (linear), square,
A dot growth shape such as a cross or a rhombus or a dispersed dither pattern) is made different to take advantage of various dither patterns to realize tone reproduction with a more natural feeling.

According to another aspect of the present invention, in a halftone reproduction method for quantizing multivalued image data using a dither matrix, a plurality of dither matrices having different dot growth modes are prepared and used according to a reproduction tone range. The selected dither matrix is selected from the plurality of dither matrices, and multi-valued image data is quantized using the selected dither matrix.

In the above invention, the reproduction tone area is divided into a low tone area, an intermediate tone area and a high tone area, and the dither matrix is formed by a first dot whose dot growth form is one of a set type circle, ellipse, and line. Group and a second group in which the dot growth form is one of a set square, cross, or diamond, and the first group is selected in a low tone area and a high tone area, and the first group is selected in a middle tone area. The second group may be selected.

In the present invention, quantization is performed using a dither matrix of a dot growth form of any of a circle, an ellipse, and a line in the low tone region and the high tone region, and a square, cross, and diamond in the middle tone region. Quantization using a dither matrix in any of the dot growth forms achieves smooth gradation reproduction with little fluctuation in the middle tone range, and a sharp contrast tone in the low and high tone ranges. Realize the reproduction.

According to another aspect of the present invention, there is provided a halftone reproduction method for quantizing multi-valued image data using an aggregated dither matrix. It is characterized in that the contact is grown in a pattern that occurs in a high tone region or higher.

According to the present invention, in a dot concentration type dither matrix, when a dot grows, the contact of the dot with an adjacent dot does not occur in the intermediate tone region but occurs in the high tone region or higher. The spread of toner or ink due to the contact of adjacent dots is particularly noticeable in the intermediate tone area,
Prevents sharp fluctuations in density due to overlap. In addition,
In the high tone region, the area of the toner or the ink becomes large as a whole, so that the contact with the adjacent dots does not occur as a rapid change in density.

[0022]

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

FIG. 1 shows an image processing apparatus 1 for a color laser printer to which the present invention is applied.

The image processing apparatus 1 receives R, G, and B (red, green, and blue) luminance data to be subjected to image processing. The RGB luminance data is, for example, 8 bits each.

The density converter 2 converts the input RGB luminance data into C, M, and Y (cyan, magenta, and yellow), which are the three primary colors for printing.

The black plate / UCR generator 3 converts the converted CM
UCR (under color removal) and black plate generation are performed using the Y data.
The generation of the UCR and the black plate is performed to save the color material (toner), increase the density of the high density region, secure the contrast, and the like. The black plate / UCR generation unit 3 outputs the generated K (black) data in addition to the CMY data.

The color correction unit 4 performs a masking process on the CMY data other than the K data, which is an achromatic component, to correct the influence of the unnecessary absorption band of each color material.

The YMC data color-corrected by the color correction unit 4 and the K data input from the black plate / UCR generation unit 3 are input to a gradation processing unit 5, and each of the YMCK colors has a gradation according to the present invention. After being processed, it is sent to the printer engine 6 of the color laser printer.

FIG. 2 conceptually shows the internal structure of the gradation processing section 5 for a certain color. In this case, the processing for magenta (M) will be described, but the other colors cyan (C) and black (K) have the same configuration and perform the same operation. However, with respect to yellow (Y), since there is no significant difference in density characteristics due to the difference in the dot growth shape of the dither matrix, the conventional gradation processing is performed in this case.

In FIG. 2, multi-valued image data Dm of magenta (M) in pixel units input from the color correction section 4 is input to the quantization processing section 10 and the reproduction tone area determination section 11.

In the reproduction tone area determination section 11, a plurality of pixels (n corresponding to the dither matrix cell) are inputted.
× m) of the multi-valued image data Dm, the tone level (tone value) of the dot including the pixel is obtained, and the obtained tone level is calculated for a plurality of pre-divided tone areas. It is determined which of these areas is to be entered.

FIG. 3 is a graph showing the relationship between the tone level (1 to 100) of one dot formed by dither processing and the output density (0 to 2.5). The gradation level area is divided into four gradation levels d1 to d1.
By d4, for example, the image is divided into five regions, a very low gradation region, a low gradation region, an intermediate gradation region, a high gradation region, and a super high gradation region. For example, the very low gradation range (0 to d1) corresponds to the very low density range (0.2 or less), and the low gradation range (d1 to d2) corresponds to the low density range (0.2 to 0.4). Corresponding, intermediate gradation range (d2-d
3) corresponds to the intermediate density range (0.4 to 1.0), the high gradation range (d3 to d4) corresponds to the high density range (1.0 to 1.5), and the ultra-high gradation range (d4 ~) Is the ultra-high concentration range (1.5 or more)
Shall correspond to This region division is performed based on the result of examining human visual perception characteristics for identifying tone reproduction of an image.

As described above, the reproduction tone area determination unit 11 determines in advance, based on the multi-valued image data Dm of the pixel of interest and the multi-valued image data Dm of peripheral pixels included in n × m surrounding cells. By performing the set predetermined operation, the final gradation level of the dot including the target pixel is obtained. For example, the tone level of the upper left dot in FIG. 9 shown above is 25 (the highest tone level value is 100), and the tone level of the upper right dot is 37.5. Then, the reproduction tone area determination unit 11 compares the calculated gradation level with the four boundary gradation levels d1 to d4 to determine the gradation level of the dot including the target pixel. It is determined to which of the plurality of gradation areas classified by the boundary values d1 to d4 belongs, and the selection signal SL indicating the determination result
Is output to the selector 12.

The selector 12 includes a reproduction tone range determination unit 11
One of a plurality of dither matrices DZ1 to DZ5 is selected based on the selection signal SL input from the CPU, and the quantization processing by the quantization processing unit 10 using the selected one dither matrix is performed. I do.

That is, the dither matrix DZ1 stores a dither pattern of a dot growth shape exhibiting optimum characteristics in an ultra-low gradation range (d1 or less). Similarly, the dither matrices DZ2, DZ3, DZ4, DZ
5 has a low gradation range (d1 to d2) and an intermediate gradation range (d2 to d2).
d3), a dither pattern of a dot growth shape showing optimum characteristics in a high gradation range (d3 to d4) and an ultra-high gradation range (d4 or more) are stored.

The quantization processor 10 quantizes the input multi-valued image data Dm of the pixel of interest into binary or multi-valued data using the dither matrix selected by the selector 12. In this quantization process, an appropriate screen angle is set for each dither matrix in order to avoid moire between colors. The quantization processing unit 10 also performs output γ correction for correcting the output density.

FIG. 4 shows the gradation level-output density characteristics of black (K) dots for six different dot growth shapes. As the dot growth shape,
Six types of circle, diamond, ellipse, line, square, and cross (cross) are shown. In this case, the screen frequency, that is, the number of lines is 120 lines / inch.

Similarly, FIG. 5 shows gradation level-output density characteristics for six types of dot growth shapes (circle, rhombus, ellipse, line, square, and cross) for cyan (C) dots.

FIG. 6 shows the gradation level-output density characteristics of magenta (M) dots for two types of dot growth shapes (circle and rhombus).

As can be seen from the characteristics shown in FIGS. 4 to 6, according to the study by the present inventors, the dither pattern has a frequency of occurrence of an irregular pattern at tone reproduction, a gradient of density change, and a reproduction density width. (Difference between maximum output density and minimum output density)
From these results, it was found that (a) a group of circles, ellipses, and lines, (b) a group of squares, crosses, and diamonds, and (c) a group of dispersed dots. The groups (a) and (c) are suitable for smooth tone reproduction (the gradient of the density change is small), and the group (b) performs enhanced tone reproduction, that is, a reproduced image with clear contrast. Was found to be suitable.

The dot-dispersion type (c) is effective for use in a low-to-medium density region where periodic pattern fluctuations are conspicuous, but it is necessary to emphasize tone reproduction in a high-density region. Not enough.

For the purpose of reducing overexposure at ultra-low densities (0.2 or less) and underexposure at ultra-high densities (1.5 or more), selection should be made from group (b). In the intermediate density range, it is considered that the group (a) is effective for obtaining smooth tone reproduction.

Based on the analysis results as described above, a plurality of dither matrices DZ1 to DZ5 of FIG. 1 store a dither pattern having an optimum dot growth shape for each gradation area.

FIG. 7 specifically explains the concept of the present invention.

FIG. 7A shows the visual characteristics of the output density when the dot growth shape is a circle of the group (a). As can be seen from FIG. 7A, the dot growth shape is a circle. In the case, the density change is smooth and nearly linear in the intermediate density range, but the fluctuation varies in the high density range.

FIG. 7B shows the visual characteristics of the output density when the dot growth shape is the rhombus of the group (b). As can be seen from this figure, the dot growth shape is a rhombus. In such a case, in the intermediate density range and the high density range (with a gradation level of 60, 85, etc.), a fluctuating portion as indicated by an arrow occurs intermittently.

Therefore, in this case, as shown in FIG. 7 (c), a circular dither pattern is selected in the intermediate density region, and a rhombic dither pattern is selected in the high density region, thereby forming both dot shapes. It incorporates the advantages of each.

In the characteristic shown in FIG. 7C, a fluctuating portion still remains in a portion surrounded by a broken line in the high density region. It is considered that the cause of the fluctuation portion is dot overlap between adjacent cells. That is, in the case of a rhombic dot, as shown in FIG. 8A, as the dot grows, it comes into contact with the adjacent dot at the corner of the rhombus, and as a result, the toner joins and overlaps between the adjacent dots. In addition, the dot area becomes larger than a predetermined size defined by the gradation level signal.

In order to prevent this, at the gradation level at which the density changes, the dot growth shape is adjusted so that the end of the dot does not come into contact with the adjacent dot, for example, as shown in FIG. By deforming the rhombus into a shape in which the corners are shifted from each other, a periodically fluctuating portion is eliminated.

By performing such dot shape correction, as shown in FIG. 7D, it is possible to obtain a visual characteristic in which a fluctuation portion in a high density region is eliminated.

By the way, in a dot growth shape such as a cross or a rhombus, as described above, as the dot grows, the dot comes into contact with an adjacent dot, and variation in density occurs due to overlap of toner or ink. This variation variation is visually noticeable in the intermediate density range, but is less noticeable in the high density range because the toner or ink area is generally large.

Therefore, as shown in FIG. 8B, a set type dither matrix having a dot growth shape such as a cross or a rhombus is formed so that the contact between adjacent dots does not occur in the intermediate density region. If the growth is performed in a pattern that occurs in the high-density region or higher, the periodic density fluctuation in the intermediate-density region can be suppressed.

In the above embodiment, the dither matrix to be used in accordance with the reproduction density area is selected from a plurality of dither matrices having different dot growth forms. The dot growth shape changes to the dot growth shape. In order to prevent the density characteristics from irregularly changing due to the change in the dot growth shape, a dither pattern that grows with an intermediate shape between the two types of dot growth shapes is used at the boundary of the density region. The gradation processing may be performed using a pattern.

In the above-described embodiment, the dither processing is performed in which one pixel of input data is compared with one threshold value of an n × m dither matrix cell. The present invention may be applied to the gradation processing of the density pattern method for comparing with the (n × m) threshold values. In the case of the density pattern method, since the input data directly corresponds to the gradation level (tone value), a corresponding gradation area is identified by the input data, and a dither pattern corresponding to the gradation area is identified based on the identification result. May be selected.

Further, the human visual perception characteristics are examined in detail for each color, and the type of dither pattern to be used and the method of dividing the density region (so as to obtain the optimum output density characteristics for each color based on the result). The division boundary value, the number of divisions, and the like may be different for each color.

Further, in the above embodiment, the present invention is applied to a color laser printer, but the present invention may be applied to other ink jet printers, thermal printers and the like. Further, the present invention may be applied to image display such as a display. In the case of display control, the density is replaced by the brightness.

[0057]

As described above, in the present invention, since the dot growth form of the dither matrix is made different according to the tone range to be reproduced, the advantages of various dither patterns can be taken in, and Thus, tone reproduction with a natural feeling can be realized.

According to the present invention, quantization is performed using a dither matrix of a dot growth form of a circle, an ellipse, or a line in the low tone region and the high tone region, and a square, cross, or diamond in the middle tone region. Quantization is performed using the dither matrix of any of the dot growth forms described above, so that gradation reproduction with smooth fluctuation is realized in the middle tone range, and the contrast is sharp in the low and high tone ranges. It is possible to realize a reproduced tone.

Further, according to the present invention, in a dot concentration type dither matrix, when a dot grows, the contact of the dot with an adjacent dot does not occur in the intermediate tone region but occurs in the high tone region or higher. The spread of toner or ink due to the contact of adjacent dots is particularly noticeable in the intermediate tone range,
It is possible to prevent a sharp change in the density due to the overlap.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an image processing system of a color laser printer to which the present invention is applied.

FIG. 2 is a block diagram illustrating a conceptual internal configuration of a gradation processing unit.

FIG. 3 is a graph showing a relationship between a gradation level and an output density which are divided into groups.

FIG. 4 is a graph showing gradation-output density characteristics for various dot growth shapes of black dots.

FIG. 5 is a graph showing gradation-output density characteristics for various dot growth shapes of cyan dots.

FIG. 6 is a graph showing gradation-output density characteristics for various dot growth shapes of magenta dots.

FIG. 7 is a graph for explaining an embodiment of the present invention.

FIG. 8 is a diagram showing a growth mode of a dither pattern according to the embodiment of the present invention.

FIG. 9 is a diagram illustrating an example of a concentrated dither pattern.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image processing apparatus 2 Density conversion part 3 Black board / UCR generation part 4 Color correction part 5 Tone processing part 10 Quantization processing part 11 Reproduction tone range judgment part 12 Selector DZ1-DZ5 Dither matrix DT dot

Continuation of the front page (72) Inventor Keisuke Hirai 1583 Iiyama, Atsugi-shi, Kanagawa F-term in the Faculty of Engineering, Tokyo Polytechnic University 5B057 CA01 CA08 CA12 CA16 CB01 CB07 CB12 CB16 CC02 CE13 CE18 DA08 DB02 DB06 DB09 DC22 5C055 AA14 BA08 EA05 EA06 EA16 HA36 HA37 5C066 AA05 AA11 BA13 CA05 CA17 EC06 GA01 GA24 HA03 KG01 5C077 MP01 MP08 NN09 NN19 PQ08 RR02 TT03

Claims (4)

[Claims] 1. A halftone reproducing method for quantizing multi-valued image data using a dither matrix, wherein a dot growth form of the dither matrix is made different according to a reproduced tone range. How to reproduce. 2. A halftone reproduction method for quantizing multi-valued image data using a dither matrix, wherein a plurality of dither matrices having different dot growth forms are prepared, and the plurality of dither matrices to be used in accordance with a reproduction tone range are prepared. A multi-valued image data is quantized using the selected dither matrix and the selected dither matrix. 3. A reproduction tone area is divided into a low tone area, a middle tone area, and a high tone area, and a dither matrix is formed by a first group whose dot growth form is one of a set circle, ellipse, and line. , The dot growth mode is divided into a second group of any of a set square, cross, and diamond, and the first group is divided into a low tone region and a high tone region.
3. The halftone reproducing method according to claim 2, wherein a group of the halftone image is selected, and the second group is selected in an intermediate tone range. 4. A halftone reproduction method for quantizing multi-valued image data using an aggregated dither matrix, wherein the aggregated dither matrix has a dot abutment with an adjacent dot occurring in a high tone region or higher. A halftone reproduction method characterized in that the pattern is grown in such a pattern as to produce a halftone.