Image processing apparatus, threshold matrix generation apparatus, image processing method, method for generating threshold matrix, recording apparatus, and program

Abstract

To provide a technique to prevent the occurrence of moire, granularity, and the like due to overlapping of dots in a plurality of colors, in forming an image on a recording medium by using the dots in the plurality of colors.SOLUTION: An image processing apparatus receives input of a plurality of pieces of multi-valued data corresponding to a plurality of coloring materials, quantizes each of the plurality of pieces of multi-valued data, and generates a plurality of pieces of quantized data indicating recording or non-recording of dots in a recording medium with each of the plurality of coloring materials. Here, a plurality of dot patterns of the plurality of coloring materials recorded based on the plurality of generated quantized data have peaks in declination directions different from each other in the power spectrum of a frequency space, and have blue noise characteristics or green noise characteristics.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an image processing device, a threshold matrix generating device, a control method thereof, a recording device, and a program. [Background technology]

[0002] Image forming devices such as printers that print images represented by digital image data input via a computer or digital camera on a recording medium using a recording material are widely used. In digital image data, the pixel values ​​of each pixel constituting the image data are generally expressed in multiple gradations, such as 8 bits or 16 bits per color. On the other hand, in image forming devices, images are formed by turning on and off dots using a recording material, so the number of gradations that the image forming device can output is often lower than the number of gradations of the image data. Therefore, in general, halftone processing is performed on the digital image data in order to print a multi-gradation image using the image forming device. Halftone image data obtained by halftone processing represents a dot pattern that the image forming device outputs on a recording medium, and pseudo-expresses the gradations of the image.

[0003] One type of halftone processing is a method called dither processing that uses a threshold matrix. A threshold matrix is ​​a matrix in which thresholds are arranged two-dimensionally. Dither processing quantizes the pixel values ​​of each pixel that composes the image data by comparing the pixel value with the corresponding threshold value for each pixel that composes the image data, thereby reducing the number of gradations in the image data. The threshold matrix is ​​repeatedly arranged in a tile-like manner for the entire image data, and is set so that a threshold corresponds to every pixel. Threshold matrices used in dither processing are classified into two types. One is a dot-concentrated threshold matrix using the AM (Amplitude modulation) modulation method, in which the thresholds are arranged so that they increase in order from a certain position. The other is a dot-dispersed threshold matrix using the FM (Frequency modulation) modulation method, in which large and small thresholds are distributed.

[0004] While the AM modulation method can produce stable gradation output, in color printing, it is prone to moire phenomena due to the overlap of the C, M, Y, and K colors. In general, in printing technology, the screen angle is changed for each color to push the frequency of moire that occurs between colors to the high frequency side, making the moire less noticeable visually. For example, screen angles of 0° and 60° are given to the halftone dots of each color, with a center of 30°, to suppress the occurrence of moire due to overlapping colors. In this way, by making the screen angle different for each color, the occurrence of moire can be reduced. However, in the case of the AM modulation method, the reproducibility of high-frequency input images is low, and because the dot arrangement is regular, jaggies are visible at the edges of the image. In addition, even if moire is suppressed by introducing a screen angle, a regular pattern called a rosette pattern occurs due to the overlapping of colors.

[0005] As an approach to avoid the above problems, there are gradation expression methods using FM modulation such as the blue noise mask method and green noise mask method, which show frequency characteristics with good dispersion that reduce the low frequency range without having a peak at a specific angle. These methods have frequency characteristics that show high frequency patterns with good dot dispersion, so they are widely used in halftone screens in digital printing, which have good response to high frequency patterns, and in inkjet and thermal transfer printers. However, while FM modulation is less likely to cause moire due to color overlapping due to its non-periodic structure, low frequency components caused by interference of frequency components in the high frequency band can appear in the form of "granularity," which can cause deterioration of image quality.

[0006] Patent Document 1 describes a technique for suppressing low-frequency components that occur due to overlapping of colors by making the band width in frequency space different for each color. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2016-163197 A Summary of the Invention [Problem to be solved by the invention]

[0008] The frequency components of moire and graininess caused by overlapping colors appear in the frequency band of the convolution of the frequency components of each color. Therefore, as in the technology described in Patent Document 1, by making the bands of the colors different, it is possible to suppress interference in the ultra-low frequency band (near 0 frequency). However, when the bands of the colors are relatively close to each other, there is a problem that aliasing into the low frequency range cannot be sufficiently suppressed by the convolution of each band. On the other hand, when the bands of the colors are separated from each other in order to suppress interference in the low frequency range, problems such as a decrease in the sense of unity arise due to the different frequency bands for each color.

[0009] The present invention provides a technique for suppressing the occurrence of moire, graininess, and the like caused by overlapping dots of multiple colors when forming an image on a recording medium with dots of multiple colors. [Means for solving the problem]

[0010] An image processing device according to an aspect of the present invention includes the following arrangement: an input means for inputting a plurality of multi-value data corresponding to a plurality of color materials; a quantization unit that quantizes each of the plurality of multi-value data to generate a plurality of quantized data indicating recording or non-recording of dots on a recording medium with each of the plurality of color materials, The dot patterns of the plurality of color materials recorded based on the plurality of quantized data generated by the quantization means have peaks in different deflection angle directions in the power spectrum in frequency space, and have blue noise characteristics or green noise characteristics. Effect of the Invention

[0011] According to the present invention, when an image is formed on a recording medium using dots of multiple colors, the occurrence of moire, graininess, and the like caused by overlapping dots of multiple colors is suppressed. [Brief description of the drawings]

[0012] [Figure 1] 1 is a block diagram showing an example of the arrangement of an image forming system according to a first embodiment. [Diagram 2] FIG. 1 is a diagram explaining the characteristics of dot patterns using AM modulation. [Diagram 3] 1A to 1C are diagrams for explaining the characteristics of dot patterns in an FM modulation method. [Figure 4] 5A to 5C are diagrams illustrating features of a dot pattern according to the first embodiment. [Diagram 5] 5 is a flowchart illustrating an operation of an image forming process according to the first embodiment. [Figure 6] FIG. 2 is a block diagram showing an example of the arrangement of a quantization process according to the first embodiment. [Figure 7] 5 is a flowchart illustrating a dither matrix generation process according to the first embodiment. [Figure 8] FIG. 2 is a view for explaining a dot pattern according to the first embodiment. [Figure 9] 10A to 10C are diagrams for explaining dot patterns according to a second embodiment. [Figure 10] 5A and 5B are views for comparing the characteristics of dot patterns according to the first and second embodiments. [Figure 11] A diagram showing RAPS and FAPS obtained from the power spectrum of Figure 10. [Figure 12] A diagram showing RAPS and FAPS obtained from the power spectrum of Figure 9. [Figure 13] 10 is a flowchart illustrating a dither matrix generation process according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, the embodiments will be described in detail with reference to the attached drawings. Note that the following embodiments do not limit the invention according to the claims. Although the embodiments describe a number of features, not all of these features are essential to the invention, and the features may be combined in any manner. Furthermore, in the attached drawings, the same reference numbers are used for the same or similar configurations, and duplicated descriptions are omitted.

[0014] First Embodiment (Device configuration) 1 is a block diagram showing the configuration of an image forming system according to this embodiment. The image forming system of this embodiment has an image processing device 1 and a printer 2. The image processing device 1 can be implemented, for example, by a printer driver installed in a general personal computer. In this case, each unit of the image processing device 1 described below is realized by the computer executing a predetermined program. Alternatively, for example, the printer 2 may include the image processing device 1.

[0015] The image processing device 1 and the printer 2 are connected by a printer interface or a circuit. The image processing device 1 acquires image data to be printed from an image data input terminal 101. In this embodiment, the image data is an 8-bit RGB color image. The color matching processing unit 102 performs color matching processing on the input RGB image data and corrects the colors of the RGB image. By using the color matching processing, it is possible to obtain uniform color reproduction even when a printer or a recording medium having different color reproduction characteristics is used. The color matching processing unit 102 refers to a three-dimensional color matching LUT stored in the color matching LUT storage unit 103 during the color matching processing. In the color matching LUT, RGB values ​​are described on lattice points thinned to 17×17×17 points, and values ​​between the lattice points are calculated by linear interpolation.

[0016] The color separation processing unit 104 generates, from the image data corrected by the color matching processing unit 102, a four-plane 8-bit ink value image corresponding to the four colors of ink provided in the printer 2. In this embodiment, the printer 2 is equipped with four colors of ink, cyan (C), magenta (M), yellow (Y), and black (K), in the print head 201. The color separation processing unit 104 refers to a three-dimensional color separation LUT stored in the color separation LUT storage unit 105 during color separation processing. In the color separation LUT, color material amount values ​​(ink values) of four colors of ink are described on lattice points thinned out to 17×17×17 points, and values ​​between the lattice points are calculated by linear interpolation.

[0017] The quantization processing unit 106 converts the ink value image of each color obtained by the color separation processing unit 104 into a quantized image having a binary or more than binary number of gradations less than the number of input gradations. In this embodiment, the quantization processing unit 106 quantizes the ink value image by a comparison process with a dither matrix for each ink color stored in a dither matrix storage unit 107. The quantized image data generated by the quantization processing unit 106 is output to the printer 2 from an output terminal 108. The threshold matrix generating device 301 generates a dither matrix for each ink color and stores it in the dither matrix storage unit 107. The threshold matrix generating device 301 may be a separate configuration from the image processing device 1, or may be incorporated as a part of the functions executed by the image processing device 1. A method of generating a dither matrix will be described later.

[0018] The printer 2 forms the quantized data generated by the image processing device 1 on a recording medium. In this embodiment, an inkjet type or the like is used as the recording head 201. The recording head 201 is, for example, a long line head in which a plurality of nozzles for ejecting color materials (ink) are arranged to cover the entire range of a drawing area in the width direction of the recording medium. A printed image is formed by ejecting ink while moving the recording medium relative to the recording head 201. In this embodiment, the recording head 201 is equipped with four colors of ink: cyan (C), magenta (M), yellow (Y), and black (K).

[0019] An input terminal 204 receives the quantized image data formed by the image processing device 1 and passes it to an ink color selection unit 203. The ink color selection unit 203 selects an ink color corresponding to the quantized image data from among the ink colors mounted on the print head 201. The head control unit 202 generates a drive signal for controlling the print head 201 based on the quantized image data. The print head 201 drives the print head of the selected ink color based on the drive signal, and actually prints each ink dot on a print medium.

[0020] (Relationship between dot superposition and power spectrum) In a configuration like printer 2 that records by overlapping dots of multiple colors in the same drawing area on a recording medium, the overlapping of colors produces a power spectrum that is not present in the original dot pattern of each color. Below, we will explain the relationship between the overlapping of dots of two types of ink and the power spectrum that is produced by this.

[0021] FIG. 2 is a diagram showing the characteristics of a dot pattern by the AM modulation method. FIG. 2(a) shows an example of a dot pattern (shown as dots with vertical stripes) for recording a first color ink (e.g., cyan). FIG. 2(b) shows an example of a dot pattern (shown as dots with horizontal stripes) for recording a second color ink (e.g., magenta). FIG. 2(c) shows a dot pattern in which the ink dots of the first color in FIG. 2(a) and the ink dots of the second color in FIG. 2(b) are superimposed. In the dot pattern in FIG. 2(c), the dots shown in the vertical and horizontal stripes are dots in which two types of ink are superimposed. Such superimposed dots generally do not have a density obtained by simply adding the densities of two types of dots due to the phenomenon of two ink droplets joining together on a printing medium, the nonlinearity of color development due to the mixing of the inks, and the like. As a result, the superimposed dot pattern has frequency components of the dot pattern shown as the logical product or logical sum of the dot patterns of the two types of ink.

[0022] The frequency and direction of the frequency components generated by overlapping two types of ink can be obtained as the result of calculating the convolution integral of the frequency components in frequency space, based on the relationship between convolution and product in the nature of the Fourier transform. Figures 2(d),(e),(f) show the power spectrum obtained by performing a two-dimensional Fourier transform on the dot pattern shown in Figures 2(a),(b),(c), and expressing the power (square of the absolute value) as shades of gray. In Figures 2(d),(e),(f), the center is the DC component, the u direction corresponds to the frequency components in the y direction in Figures 2(a),(b),(c), and the v direction corresponds to the frequency components in the x direction in Figures 2(a),(b),(c), and the whiter the area, the greater the power. In the case of a periodic line screen pattern as shown in Figures 2(a)(b), in frequency space, a peak appears in the direction of the deflection angle perpendicular to the screen angle, and at the coordinate of a frequency inversely proportional to the screen period, as shown in Figures 2(d)(e).

[0023] In addition, the frequency components that arise from the logical product or logical sum (logical product of inverted patterns) of two dot patterns appear at the position where the convolution integral of the frequency components in frequency space is calculated. Therefore, a power spectrum such as that shown in Figure 2(f) can be calculated as the frequency components that arise from the overlapping of colors. In frequency space, the same characteristics appear from 0 to 180 degrees and 180 to 360 degrees, so the characteristics from 0 to 180 degrees will be explained here. As shown in Figure 2(f), frequency components that do not exist in the original pattern are generated by the overlapping of two colors. This phenomenon is called the moiré phenomenon. In general, the closer the difference in screen angle between colors is to 90 degrees, the more the frequency components that arise are pushed to high frequencies, and the occurrence of moiré of low frequency components can be suppressed. In actual AM screen printing, the four colors of CMYK are often printed with angles such as C at 15 degrees, M at 75 degrees, Y at 0 degrees, and K at 45 degrees.

[0024] Next, the relationship between the overlap of dots when a blue noise pattern is printed as a dot pattern of two types of ink and the resulting power spectrum will be described.

[0025] FIG. 3 is a diagram for explaining the characteristics of the dot pattern of the FM modulation method. FIG. 3(a) shows an example of a blue noise dot pattern for recording the first color ink, and FIG. 3(b) shows an example of a blue noise dot pattern for recording the second color ink. Furthermore, FIG. 3(c) shows a dot pattern in which the first color ink dots of FIG. 3(a) and the second color ink dots of FIG. 3(b) are superimposed. In the dot pattern of FIG. 3(c), the dots represented by vertical and horizontal stripes are dots in which two types of ink are superimposed. FIG. 3(d), (e), and (f) show the state of the power spectrum in which the dot patterns such as those of FIG. 3(a), (b), and (c) are two-dimensionally Fourier transformed, and the power (square of the absolute value) is expressed by the shade. The center of FIG. 3(d), (e), and (f) corresponds to the DC component, the u direction corresponds to the frequency component in the y direction of FIG. 3(a), (b), and (c), and the v direction corresponds to the frequency component in the x direction of FIG. 3(a), (b), and (c), and the whiter the part, the greater the power. In the case of a blue noise pattern with dots distributed non-periodically as shown in Figures 3(a) and (b), the frequency space does not have a peak at a specific deflection angle, as shown in Figures 3(d) and (e), and the frequency characteristics are such that the low-frequency components are suppressed and the high-frequency components have power.Furthermore, it can be predicted that the frequency components generated by the logical product or logical sum (logical product of inverted patterns) of two types of dot patterns will be aliased high-frequency components of each color as low-frequency components, as shown in Figure 3(f).

[0026] As explained with reference to FIG. 2, when an AM modulation dot pattern having power at a specific frequency and deflection angle is used, the power of aliasing caused by interference between multiple ink dots is pushed to a high frequency, and the occurrence of moire of low-frequency components can be suppressed. On the other hand, the periodic structure of the dot pattern makes it easy for interference with the input image to occur, and a rosette pattern can also occur due to overlapping of colors. In addition, when an FM modulation blue noise pattern is used, it has the characteristic that interference with the input image is unlikely to occur and a rosette pattern is unlikely to occur due to a non-periodic structure. On the other hand, as explained with reference to FIG. 3, when a blue noise pattern is used, low-frequency components are generated by convolution of frequency components in the high frequency range (FIG. 3(f)), which appears in the form of graininess and can cause image quality degradation in mixed-color images.

[0027] Therefore, in this embodiment, a dot pattern that combines the characteristics of the AM modulation method and the FM modulation method is used. By using such a dot pattern, it is possible to suppress the occurrence of low-frequency components due to convolution of frequency components in the high-frequency range while suppressing interference with the input image and the occurrence of rosette patterns due to color overlap.

[0028] Using FIG. 4, the relationship between the overlap of dots when a dot pattern having the above-mentioned characteristics is printed as a dot pattern of two types of ink and the resulting power spectrum will be described.

[0029] FIG. 4(a) shows an example of a dot pattern for recording a first color ink, and FIG. 4(b) shows an example of a dot pattern for recording a second color ink. Furthermore, FIG. 4(c) shows a dot pattern in which the first color ink dots of FIG. 4(a) and the second color ink dots of FIG. 4(b) are superimposed. In the dot pattern of FIG. 4(c), the dots represented by vertical and horizontal lines are dots in which two types of ink are superimposed. As shown in FIG. 4(a), the dot pattern for recording a first color ink is a pattern that has a tendency to continue in a predetermined angular direction and also has irregular characteristics. As shown in FIG. 4(b), the dot pattern for recording a second color ink is a pattern that has a tendency to continue in an angle away from the angular direction of the dot pattern of the first color ink shown in FIG. 4(a) and also has irregular characteristics.

[0030] Figures 4(d), (e), and (f) show the power spectrum obtained by performing a two-dimensional Fourier transform on the dot patterns shown in Figures 4(a), (b), and (c) and expressing the power (square of the absolute value) as a shade of gray. The center of Figures 4(d), (e), and (f) corresponds to the DC component, the u direction corresponds to the frequency component in the y direction in Figures 4(a), (b), and (c), and the v direction corresponds to the frequency component in the x direction in Figures 4(a), (b), and (c), and the whiter the part, the greater the power. As shown in Figure 4(d), the dot pattern for recording the first color ink is a pattern with blue noise characteristics having a peak in a predetermined deflection angle direction in the power spectrum. As shown in Figure 4(e), the dot pattern for recording the second color ink is a pattern with blue noise characteristics having a peak in a deflection angle direction that is a certain value or more away from the above-mentioned predetermined deflection angle direction. That is, in the power spectrum of the frequency space, the deflection angle direction of the peak of the dot pattern of the first color and the deflection angle direction of the peak of the dot pattern of the second color have an angle difference of a certain value or more. This suppresses the low-frequency components in the power spectrum that arise from the convolution of the dot pattern of the first ink color with the dot pattern of the second ink color.The power spectrum shown in Figure 4(f) can be predicted as the frequency components that arise from the logical product or logical sum (logical product of inverted patterns) of the two dot patterns.

[0031] As shown in Figure 4 (a), (b), (d), and (e), a dot pattern for recording the first and second color inks uses a pattern with blue noise characteristics that has a peak at a deviation angle of a certain value or more in frequency space. This allows a pattern in which the dot pattern of the first color ink and the dot pattern of the second color ink are superimposed to shift part of the aliasing power to high frequencies like the AM modulation method, and suppresses the generation of low frequency components. Furthermore, because it also has irregular blue noise characteristics like the FM modulation method, it is possible to suppress interference with the input image and the generation of rosette patterns due to color superposition.

[0032] In the case of a line screen as shown in FIG. 2, since peaks generally occur in the screen angle and the angle in the vertical direction in the light and dark gradations, angles are often allocated within the range of 0 to 90 degrees. However, in this embodiment, since a pattern is generated to have blue noise characteristics, which are irregular dot arrangements, even in the light and dark gradations, no peak occurs in the vertical angle, and the deflection angle can be distributed and allocated in the range of 0 to 180 degrees. For example, when there are n colors of color materials, the peak direction (deflection angle) can be distributed and allocated to 180 degrees / n. Specifically, the deflection angles can be allocated to the four colors of CMYK as follows: C is 45 degrees, M is 135 degrees, Y is 0 degrees (180 degrees), and K is 90 degrees. Of course, the present invention is not limited to this, and the peak deflection angle directions may be distributed and allocated, such as C at 150 degrees, M at 30 degrees, Y at 75 degrees, and K at 105 degrees (in this case, an angle difference of 30 degrees or more is obtained for the deflection angle directions of each color). Even in the case of a line screen as shown in FIG. 2, a pattern can be generated so as to have blue noise characteristics in the light and dark gradations, but irregular and regular patterns are mixed depending on the gradation, which reduces the sense of unity. On the other hand, according to this embodiment, the loss of the sense of unity can be suppressed by using a pattern with blue noise characteristics having a peak in a specific deflection angle direction in the intermediate gradations.

[0033] (Flow of image formation process) Next, the image forming process in the image processing device 1 of this embodiment that realizes the above-mentioned image generation will be described with reference to the flowchart of FIG.

[0034] The image processing device 1 acquires an input image from the image data input terminal 101 (S501). Next, the color matching processing unit 102 performs color matching processing on the acquired input image (S502). In the color matching processing, the color matching processing unit 102 refers to a three-dimensional color matching LUT stored in the color matching LUT storage unit 103. Next, the color separation processing unit 104 performs color separation processing on the image data processed by the color matching processing unit 102 to generate an ink value image (S503). In the color separation processing, a three-dimensional color separation LUT stored in the color separation LUT storage unit 105 is referred to. Next, the quantization processing unit 106 performs quantization processing to convert the data after the color separation processing into binary data, and obtains binary image data (S504). The image processing device 1 outputs the binary image data after the quantization processing from the output terminal 108 in any size, such as the entire image or the bandwidth of each unit recording area.

[0035] The printer 2 receives the quantized image data (binary image data) from the input terminal 209 and performs image formation (S505). In image formation, the ink color selection unit 203 selects an ink color based on the binary image data after quantization processing, generates scanning data, and sends it to the head control unit 202. The head control unit 202 drives the recording head 201 in accordance with the received operation data to record a dot pattern on the recording medium. While moving the recording medium relative to the recording head 201, the printer 2 drives each nozzle of the recording head 201 at regular drive intervals, and records an image on the recording medium. This completes a series of image formation processes.

[0036] (Explanation of quantization processing unit) Fig. 6 is a block diagram for explaining the details of the quantization process performed by the quantization processing unit 106 of this embodiment. The quantization process of this embodiment is performed by a dither method. In the quantization process of this embodiment, a process related to a threshold is performed, and a process of comparing an input value with a threshold is performed. This series of processes is processed in parallel for each color (each channel). Each process will be described in detail below with reference to Fig. 6.

[0037] The quantization processing unit 106 is capable of receiving a maximum 16-bit signal. FIG. 6 shows a state in which 16-bit data (multi-value data of the first to fourth colors) of inks of the first to fourth colors are input in parallel. The threshold acquisition unit 601 acquires a threshold corresponding to the pixel position of the data to be processed from the dither matrix storage unit 107 configured by a memory such as a ROM. In this embodiment, the dither matrix 610 is a threshold matrix in which thresholds of 0 to 65535 are arranged so as to have a predetermined characteristic, and can have various sizes and shapes such as 512×512 pixels, 256×256 pixels, and 256×512 pixels. The dither matrix storage unit 107 stores a plurality of threshold matrices corresponding to each ink color in advance, and the threshold acquisition unit 601 selects a threshold matrix corresponding to the ink color from among these. Then, from the plurality of thresholds arranged in the selected threshold matrix, a threshold corresponding to the pixel position (x, y) of the data to be processed is provided to the comparison processing unit 602. The characteristics of the threshold matrix having the predetermined characteristic will be described in detail later. A comparison processing unit 602 compares the multi-valued data to be processed with the threshold value acquired by the threshold value acquisition unit 601, and determines whether to record (1) or not record (0) the pixel position.

[0038] (How to create a dither matrix) A method for generating a dither matrix according to this embodiment will be described. In the following description, the dither matrix in the generation process or after generation is designated as M. The dither matrix M is a two-dimensional array with a size of Sx rows in the x direction (the conveying direction of the recording medium) and a size of Sy columns in the y direction (the width direction of the recording medium) (Sx and Sy are natural numbers). The size (Sx, Sy) of the dither matrix M is arbitrary, but typically each side has a length that is a power of 2, and preferably each side has a length of 256 pixels or more (for example, 256×256 pixels, 512×512 pixels, 256×512 pixels). In this embodiment, Sx is 256 pixels, and Sy is 256 pixels.

[0039] The Void & Cluster method is known as a method for generating a dot-dispersed dither matrix. The Void & Cluster method applies a low-pass filter to obtain a smoothed density image, determines the layout of dots to be added so as to suppress local density fluctuations, and generates a dither matrix that expresses each tone by repeating this process. In this embodiment, a similar method is used to generate a dither matrix having dot-dispersed blue noise characteristics or green noise characteristics.

[0040] The dot pattern generated in the process of generating the dither matrix is ​​d(x, y). d(x, y) is a two-dimensional array, and its size is the same as that of the dither matrix M(x, y). The value of each pixel in d(x, y) is 1 if a dot exists, and 0 if no dot exists. The dot pattern d(x, y) changes in the iterative process for generating the dither matrix M. According to the iterative process, Sx×Sy+1 dot patterns are generated, ranging from a dot pattern with 0 dots to a dot pattern with Sx×Sy dots. Therefore, when the number of dots in the dot pattern d(x, y) is g, the dot pattern at a certain point in time in the iterative process can be specified by using this g. In the following description, the number of dots g is called the gradation value g, and the dot pattern d(x, y) when the gradation value is g is expressed as d(g, x, y), or d(g) with x and y omitted.

[0041] Also, the density fluctuation map used to evaluate the density of dots generated in this iterative process is n(g). The smaller the value in the density fluctuation map n(g), the lower the smoothing density and the sparser the dots are evaluated to be, and conversely, the larger the value in the density fluctuation map n(g), the higher the smoothing density and the denser the dots are evaluated to be. In S103 in FIG. 7 described later, dots are added to positions in the density fluctuation map n(g) where the value is small, that is, positions where the dots are sparse. This suppresses density fluctuations and obtains a dot pattern that achieves low graininess. The density fluctuation map n(g) is a two-dimensional array of the same size as the dot pattern d(g), and the values ​​of the array change depending on the gradation value g, just like the dot pattern d(g). In the following description, n(g) is also written as n(g, x, y). In S102 in FIG. 7 described later, the density fluctuation map n(g) is obtained by applying a filter to the dot pattern d(g). Since it is assumed that the dither matrix is ​​applied periodically to the input image, the density variation map n(g) is generated by adding the results of a cyclic convolution operation between the dot pattern d(g) and the filter coefficients. The cyclic convolution operation is an operation in which a normal convolution operation is performed between the dot pattern d(g) with a periodic boundary condition set and the filter coefficients. The details of the filter coefficients used in S102 will be described later.

[0042] Hereinafter, the generation process of the dither matrix according to the present embodiment will be described in detail with reference to FIG. 7. FIG. 7 is a flowchart showing the overall flow of the dither matrix generation method. Hereinafter, an example of generating a dither matrix for four ink colors, cyan (C), magenta (M), yellow (Y), and black (K), will be described. The threshold matrix generating device 301 first generates a density variation map when the gradation value g is 0 as an initial density variation map n(0) (S100). The initial density variation map n(0) is a density variation map configured with initial values. Here, in the dither matrix corresponding to the four colors of CMYK, density variations are given to the initial density variation map n(0) in advance so that the peak of the power spectrum has an angle. In this example, the angle of the peak of the power spectrum is set to 45 degrees for cyan, 135 degrees for magenta, 0 degrees for yellow, and 90 degrees for black. Furthermore, random numbers with an amplitude equal to or less than the above-mentioned density variation are given so that the pattern does not become regular. For example, an initial density fluctuation map n(0) for each of the colors CMYK is generated by the following formulas (1) to (4).

[0043] For Cyan:

number

number

number

number

[0044] In the above formulas (1) to (4), rand is a random number whose value differs for each pixel position (h, w), and its amplitude is smaller than the density difference given to the initial density variation map. In this example, the amplitude of the random number rand is 0 to 1 / 64. Furthermore, % represents a remainder, for example, "(x)%4" represents the remainder when x is divided by 4. Furthermore, formulas (1) to (4) provide density variations that are periodically repeated in a predetermined direction. For example, in formula (1), while (x+y)%4 periodically changes to 0, 1, 2, and 3, the value of the density difference repeats between 0 and 1 / 64, and the direction of the repetition is a 45-degree direction.

[0045] In this manner, the initial density fluctuation is given to the initial density fluctuation map n(0) at different angles for each color. In S104 described later, dots are added at positions where the density fluctuation map n(g) has a small value, so that such initial density fluctuation can realize a dot arrangement in which the dots are continuous in a predetermined angle direction for each color. In the above-mentioned formulas (1) to (4), the density difference of the initial density fluctuation is set to 1 / 32 and 1 / 64, but is not limited to this. The density difference of the initial density fluctuation may be adjusted by the user as a parameter. The larger the density difference given in advance, the larger the peak value in the spatial frequency becomes, resulting in a regular arrangement in which the characteristics of the AM modulation method become stronger. On the other hand, the smaller the density difference given, the smaller the peak value in the spatial frequency becomes, resulting in an irregular arrangement in which the characteristics of the FM modulation method become stronger. In addition, it is preferable to set the density difference to be relatively smaller as the frequency of the given density fluctuation becomes higher.

[0046] S101 to S106 are processes for repeating dot addition. S101 and S106 are loop ends, and the processes from S102 to S105 are repeated until the tone value g changes from 0 to g MAX In the process from S102 to S105, one dot is added to the dot pattern d(g) with a tone value of g, and a dot pattern d(g+1) with an adjacent tone, that is, a tone value of g+1, is generated. Then, the tone value g is set in the dither matrix M at a position corresponding to the position where the dot was added. In this embodiment, MAX=Sx×Sy-1=65535. By repeating the process of adding dots from 1 to 65536 in this manner, dot patterns for all gradations and dither matrices that correspond to all gradations are generated.

[0047] The threshold matrix generating device 301 performs a convolution operation on the dot pattern d(g) corresponding to the gradation value g using a low-pass filter f described later, and calculates the density fluctuation map n(g) by adding an initial density fluctuation map n(0) (S102). Specifically, the density fluctuation map n(g,x,y) at all positions (x,y) is calculated by the following equation. Note that in the convolution operation, for the portion where the filter coefficient f protrudes from the dot pattern d(g), the calculation is performed by cyclically referencing the dot pattern so that the top, bottom, left, and right are connected. Note that when g=0, n(0) is used as is, and 0 is set at all positions in the dot pattern d(0).

number

[0048] Next, the threshold matrix generating device 301 selects the pixel position (x) at which the density fluctuation map n(g) has the smallest value among the pixel positions that are OFF in the dot pattern d(g). MIN , y MIN ) is specified as a dot addition position. Then, the threshold matrix generating device 301 calculates the dot pattern d(g) by dividing the specified dot addition position (x MIN , y MIN ), a dot pattern d(g+1) for a gradation value g+1 is generated (S103). MIN , y MIN The value of the dither matrix M corresponding to (M(x MIN , y MIN) is set as the gradation value g (S104). In order to process the next gradation value, the threshold matrix generating device 301 increments the value of the gradation value g to g+1 (S105). When the above process is repeated for g=0 to 65535, the values ​​of 0 to 65535 are set in the dither matrix M.

[0049] After completing the loop of S101 to S106, the threshold matrix generating device 301 adjusts the value range of the dither matrix M according to the range of pixel values ​​of the input image (S107). Before performing S107, the dither matrix M(x, y) includes a range from 0 to 65535 (m MIN From m MAX ) is stored. When the input image is 8-bit, the range of the input image is 0 to 255 (th MIN From th MAX ), even if a dither matrix in which values ​​from 0 to 65535 are stored is used, an appropriate halftone processing result cannot be obtained. Therefore, in S107, the range of values ​​of the dither matrix M(x, y) is adjusted so as to match the input image. For example, the range of values ​​of the dither matrix is ​​adjusted to th MIN From th MAX If you want to change the dither matrix to the value a×M(x,y)+b, the adjusted dither matrix value will be the integer part of a×M(x,y)+b. MAX -th MIN )÷(m MAX -m MIN ), b=th MIN -a×m MIN The adjusted dither matrix M is stored in the dither matrix storage unit 107. The image processing device 1 may include the function of the threshold matrix generating device 301. In that case, for example, a user who operates the image processing device 1 may be allowed to generate (update) a dither matrix at any timing. In that case, the user may be allowed to set the density difference (equations (1) to (4)) to be given to the initial density variation map as a parameter.

[0050] (Details of filter coefficients) The low-pass filter coefficient f used in S102 of this embodiment will be described. This filter coefficient is used to calculate the density fluctuation map n(g). f is a two-dimensional array, and is also written as f(fx, fy). In this embodiment, the array size of f(fx, fy) is the same as the dither matrix M. That is, when the filter size in the fx direction is Sfx and the filter size in the fy direction is Sfy, these values ​​are 256. In S104, dots are added so as to relax the density fluctuation map n(g), thereby reducing the density of dots and achieving low graininess. In order to achieve this favorably, it is necessary to extract the density between dots. The density between dots can be extracted, for example, by applying a low-pass filter to the dot pattern. In this embodiment, the filter f is further made a function related to the inverse of the distance r from the center of the filter as shown in Equation (6), thereby further improving the dispersibility. In this embodiment, 1 is added to the denominator to avoid division by zero at the distance r=0.

number

[0051] In formula (6), x0 and y0 are the center positions of the filter, and x0=Sfx÷2 and y0=Sfy÷2. The filter function indicates a two-dimensional convolution filter, and performs two-dimensional convolution filter processing based on the kernel coefficients with fx and fy as the centers. Here, the filter processing based on the kernel is called the first filter processing. The kernel is, for example, a Gaussian filter with a size of 3×3 or 5×5. With respect to the frequency band of the blue noise characteristics of the generated dot arrangement, the smaller the kernel size is, the higher the peak start can be shifted to the high frequency side, and the larger the kernel size is, the lower the frequency side can be shifted. The filter f in formula (6) performs convolution processing (hereinafter, second filter processing) using a coefficient generated based on a value related to the inverse of the distance from the center of the filter. The convolution operation (second filter processing) of formula (5) is performed on the processing result of the first filter processing. By the second filter process, a larger value is added to the density fluctuation map n(g) the closer the dot is to the dot placed in the dot pattern d(g). Therefore, it is possible to obtain a density fluctuation map that prevents a dot from being placed near a dot that has already been placed, and improves the dispersion of the dots. Note that the filter f used in S102 is not limited to this, and a low-pass filter that can extract frequency components that are perceived as graininess may be used. Also, the second filter process may be performed first, followed by the first filter process.

[0052] According to the process flow of FIG. 7 described above, it is possible to generate a dither matrix that realizes a dot arrangement that is continuous in a predetermined angular direction for each CMYK color, while also having irregular characteristics with little repetition of the same pattern within the range of the dither matrix.

[0053] (Characteristics of generated dot patterns) FIG. 8 shows a dot pattern generated using a dither matrix created by the process shown in FIG. 7, and its power spectrum. In the power spectrum of FIG. 8, it can be confirmed that the dot pattern has blue noise characteristics with suppressed low-frequency components, and that C has peaks at 45 degrees, M has peaks at 135 degrees, Y has peaks at 0 degrees (180 degrees), and K has peaks at 90 degrees. Here, it is desirable that the positions of the frequency and deflection angle of the peaks of each color are located at positions that do not fold back into the low-frequency band when the convolution integral of the frequency components in the frequency space is calculated. As shown in the power spectrum in the lower part of FIG. 8, in the frequency space, each color has peaks in the deflection angle direction that is separated by a certain value or more. This allows a part of the folding power to be pushed to high frequencies like the AM modulation method, and the occurrence of low-frequency components can be suppressed. In addition, since the dot pattern has blue noise characteristics with suppressed low-frequency components, it is possible to suppress interference with the input image and the occurrence of a rosette pattern due to color overlap.

[0054] <Second embodiment> In the first embodiment, an example was shown in which power is concentrated at a specific frequency / angle in the power spectrum. In the second embodiment, an example is described in which power is concentrated by distributing it at a specific frequency / angle and around it. Since the configuration is similar to that of the first embodiment, the dither matrix generation method of the second embodiment will be described in detail below with reference to FIG. 7.

[0055] FIG. 7 is a flowchart showing the overall flow of the dither matrix generating method. The threshold matrix generating device 301 generates an initial density fluctuation map n(0) when the gradation value g is 0 (S100). In this embodiment, in generating a dither matrix corresponding to the four colors of CMYK, density fluctuations of a predetermined angle as in the first embodiment are not given, and only random numbers (rand) of small amplitude are given so as not to form a regular pattern. The initial density fluctuation map for each color of CMYK is set by the following formula (7). The subsequent processes in S101 to S107 are the same as those in the first embodiment.

number

[0056] (Details of filter coefficients) The low-pass filter coefficient f used in S102 of the second embodiment will be described. In this embodiment, the filter coefficients are given different anisotropies for each color by the following formulas (8) to (12), so that the peaks of the power spectrum are given angles such as C at 45 degrees, M at 135 degrees, Y at 0 degrees, and K at 90 degrees. In formulas (9) to (12), so that the peaks in the power spectrum appear in a predetermined direction, each area divided based on the predetermined direction is given different characteristics (for example, the kernel size is different for each area). Note that, as in the first embodiment, the dispersion is further improved by making it a function related to the inverse of the distance r (formula (8)) from the center of the filter.

number

number

number

number

number

[0057] In Equation (8), x0 and y0 are the center positions of the filter, where x0 = Sfx÷2 and y0 = Sfy÷2. Also, the filter functions in Equations (9) to (12) represent a two-dimensional convolution filter. In the first filter process indicated by Kernel, two-dimensional convolution filter processing based on the coefficients of the kernel is performed at each position (fx, fy) of the filter. The kernel is, for example, a Gaussian filter with a size of 5×5 or 9×9. By setting the kernel sizes as kernel1 < kernel2 < kernel3, anisotropy can be imparted to the frequency characteristics of the arranged dots. In this embodiment, kernel1 has a size of 1×1 (no filter), kernel2 has a size of 5×5, and kernel3 has a size of 9×9. Thereby, utilizing the characteristic that the frequency band of the blue noise characteristic shifts to the high-frequency side as the kernel size becomes smaller and to the low-frequency side as the kernel size becomes larger, power can be concentrated in the low-frequency side direction where the dot arrangement has less constraints. For example, in the case of cyan, according to Equation (9), by using kernel1 in the first and third quadrants of the filter, a peak appears in the 45-degree direction in the power spectrum. Also, rather than setting conditions for a specific angle as in the conditional expressions within Equations (9) to (12), by setting conditions with a certain angular width in terms of quadrants, the power is not concentrated too much at a specific frequency and declination of frequency cavitation, but can be distributed around it and concentrated. The second filter process is a filter process using a coefficient related to the reciprocal of the distance r from the central part of the filter, and the details and effects of this process are as described in the first embodiment.

[0058] (Characteristics of the generated dot pattern) Figure 9 shows the dot pattern generated using the dither matrix created by the flow shown in the second embodiment and its power spectrum. In the power spectrum at the lower part of Figure 9, it can be confirmed that it has a blue noise characteristic with suppressed low-frequency components and has peaks at C = 45 degrees, M = 135 degrees, Y = 0 degrees (180 degrees), and K = 90 degrees. Also, compared with Figure 8, by having a distribution in the peaks, more power can be concentrated at a predetermined frequency and declination and around it.

[0059] (Characteristics of power spectrum) Here, the characteristics of the power spectrum of the dot patterns generated in the first and second embodiments will be explained in comparison with examples of the conventional AM modulation method and FM modulation method.

[0060] Fig. 10 shows dot patterns and power spectra obtained by subjecting the patterns to a two-dimensional Fourier transform for an example of a general AM modulation method, an example of a general FM modulation method, the magenta example of the first embodiment, and the magenta example of the second embodiment. As described above, it can be seen from Fig. 10 that the dot patterns of the first and second embodiments have both the characteristics of the AM modulation method and the FM modulation method.

[0061] Regarding the lower part of Fig. 10, if the coordinate positions centered on the origin of the frequency space are u and v, the frequency components of the power spectrum can be expressed as F(u,v). In this case, if the frequency r and argument θ are defined as in the following equation (13), the frequency components of the power spectrum can also be expressed as polar coordinates (circular coordinates) F(r,θ). Note that π is the ratio of the circumference of a circle to its diameter, and arctan is a function that obtains the arc tangent (arc tangent). The frequency r and argument θ can be calculated precisely based on the position (u,v) of the power spectrum, but for example, the decimal points can be rounded off and a certain range can be treated as the same frequency and argument.

number

[0062] In this case, the power of the same frequency r in the polar coordinate representation of the power spectrum F(r,θ) is averaged to obtain the radian averaged power spectrum (RAPS) for each frequency. RAPS is shown in equation (14). Here, n is the number of pixels in F(u,v) that have the same frequency r.

number

[0063] The power of the polar coordinate representation F(r,θ) at the same argument angle θ is averaged to obtain the frequency averaged power spectrum (FAPS) for each angle. FAPS is shown in equation (15). Here, m is the number of pixels in F(u,v) that have the same argument angle θ.

number

[0064] Note that the above conversion to polar form F(r,θ) is expressed for the purpose of explanation, and in reality, it is possible to obtain the frequency r and argument θ of each coordinate (u,v) from the Cartesian coordinate form F(u,v) and directly calculate RAPS and FAPS. Also, since the power spectrum is point-symmetric with respect to the origin, the following description will explain the characteristics in the range of θ=0 degrees to 180 degrees.

[0065] FIG. 11 shows RAPS and FAPS calculated from the power spectrum in the lower part of FIG. 10. As shown in FIG. 11, in the example of the AM modulation method, the power is modulated at a predetermined frequency (17 cycles / mm) and a predetermined deflection angle (135 degrees). In addition, in the example of the FM modulation method, it is understood that there is no peak in the angle direction, and the power in the low frequency band is suppressed, and the power is modulated to the high frequency band, resulting in a blue noise (green noise) characteristic. On the other hand, in the example of the first embodiment, while having a peak at a predetermined frequency (17 cycles / mm) and angle (135 degrees), the power in the low frequency band is suppressed, and the power is modulated to the high frequency band, resulting in a blue noise (green noise) characteristic. Compared to the conventional FM modulation method, it is understood that the power in the high frequency band is suppressed, and the power is concentrated at a specific frequency and deflection angle. Similarly, in the example of the second embodiment, it is understood that while having a peak at a predetermined frequency (17 cycles / mm) and deflection angle (135 degrees), the power in the low frequency band is suppressed, and the power is modulated to the high frequency band, resulting in a blue noise (green noise) characteristic. Furthermore, it can be seen that the peaks in the example of the second embodiment are broadened compared to the example of the first embodiment. In this way, it can be seen that by broadening the peaks, the power of more high frequency bands is suppressed and concentrated in a specific frequency band and deflection angle band.

[0066] Next, the relationship between colors in the second embodiment will be described with reference to FIG. 12. FIG. 12 shows RAPS and FAPS calculated from the power spectrum in the lower part of FIG. 9. In the frequency average power spectrum (FAPS) for each angle, it is preferable that the power other than the peak is suppressed to a degree smaller than half the power (maximum value) at the peak. Regarding the spread of the peak, if the width at half the power (maximum value) at the peak is the full width at half maximum, it is preferable that the full width at half maximum is 45 degrees or less, which is 180 degrees divided into four. Furthermore, when allocating to three colors, it is preferable that the full width at half maximum is 60 degrees or less, which is 180 degrees divided into three. Furthermore, when allocating to two colors, it is sufficient that the full width at half maximum is 90 degrees or less, which is 180 degrees divided into two. It is preferable to allocate the deflection angle of the peak within 0 degrees to 180 degrees so that the deflection angle regions of the full width at half maximum of the peak do not overlap with each other for each color. The full width at half maximum may be calculated directly from the shape of the FAPS, or may be calculated after approximating the FAPS to a continuous function such as a Gaussian function.

[0067] Also, only two colors (e.g., cyan and magenta) that are significantly affected by overlapping colors may be treated as blue noise patterns having peaks at angles that are apart by a certain value or more, while yellow and black may be treated as normal blue noise patterns, etc. Also, although the overlapping of dot patterns of different colors has been described, the same effect can be obtained even in the case of dot patterns for each scan when an image is formed by multiple scans using ink of the same color.

[0068] <Third embodiment> In the first and second embodiments, an example has been described in which a dither matrix is ​​generated in which values ​​from 1 to 65536 are stored by repeating the addition of dots from 1 to 65536. In the third embodiment, an example will be described in which a dither matrix is ​​generated by generating an initial dot pattern of intermediate gradations and adding and deleting dots from the initial dot pattern.

[0069] The dither matrix generating method of this embodiment will be described in detail below with reference to Fig. 13. The processes from S101 to S107 are the same as those in the flowchart of Fig. 7, except that the initial gradation is set to g0.

[0070] In S200, an initial dot pattern d(g0) is generated when the gradation value g is g0. Here, an initial dot pattern is generated that has blue noise characteristics with suppressed low-frequency components for the four colors of CMYK, and the deviation angles of the peaks of each color are dispersed. In this example, the deviation angles of the peaks of CMYK are 135 degrees, 45 degrees, 0 degrees (180 degrees), and 90 degrees, respectively. For example, the dot pattern is subjected to a discrete Fourier transform to calculate a power spectrum, and an evaluation filter that provides an arbitrary frequency characteristic is multiplied, and the dot arrangement is repeatedly rearranged so that the total value of the power is minimized. Alternatively, a dot pattern of intermediate gradations may be generated based on the methods of the first and second embodiments, and the resulting pattern may be used as the initial dot pattern.

[0071] In steps S201 to S206, the process of deleting dots from the dot pattern is repeated. S201 and S206 are the ends of a loop, and indicate that the process of steps S202 to S205 is repeated until the gradation value g reaches 0 from g0. The process of steps S202 to S205 is a process of deleting one dot from the dot pattern with a gradation value of g+1, and generating the adjacent gradation, i.e., the dot pattern with a gradation value of g, and the corresponding dither matrix.

[0072] The threshold matrix generating device 301 applies a low-pass filter to a dot pattern d(g) corresponding to the gradation value g, and adds the result to an initial density fluctuation map n(0), thereby calculating a density fluctuation map n(g) (S202). In this embodiment, the low-pass filter used in S202 can be, for example, one shown in equation (6) or equations (9) to (12). The threshold matrix generating device 301 calculates the position (x MAX , y MAX ) corresponding to the dot d(g, xMAX , y MAX ) is deleted (S203). The dot pattern obtained by deleting the ON-state dots from the dot pattern d(g) is called d(g-1). The threshold matrix generating device 301 deletes the positions (x MAX , y MAX ) to the gradation value g-1 (S204). The threshold matrix generating device 301 decrements the value of the gradation value g to g-1 (S205). In S207, the threshold matrix generating device 301 combines the dither matrix generated in S101 to S106 and the dither matrix generated in S201 to S206 to obtain one dither matrix corresponding to all gradations.

[0073] In this way, we generate an initial dot pattern d(g0) that has an arbitrary power spectrum, and then subtract dots from g0 to g MAX A dither matrix is ​​generated by repeating the addition of up to g0 and the subtraction from g0 to 0. In other words, a dither matrix is ​​generated by generating a dot pattern by changing the gradation value in ascending and descending order from a predetermined gradation value (g0). This makes it possible to generate a dither matrix with the desired frequency characteristics for at least the initial dot pattern d(g0). Also, in highlight and shadow gradations, restrictions on dot arrangement are reduced, making it possible to achieve a dot arrangement with higher dispersion blue noise characteristics.

[0074] <Modification> In the first and second embodiments, an example has been described in which the quantization processing unit 106 generates a dot pattern represented by two values, that is, dot on / off, but the present invention is not limited to this. The effects of the technology disclosed herein can also be obtained when generating a dot pattern with three or more values. For example, an input image can be divided into multiple data (e.g., data of dark dots and data of light dots), and the multiple data are quantized using the same threshold matrix, and the quantization results are added to obtain a quantization result with more than two values. Even in this case, a dot pattern can be generated that has blue noise characteristics or green noise characteristics in which low-frequency components are suppressed in the power spectrum, and has a peak at a predetermined deflection angle.

[0075] (Other embodiments) The present invention can also be realized by a process in which a program for implementing one or more of the functions of the above-described embodiments is supplied to a system or device via a network or a storage medium, and one or more processors in a computer of the system or device read and execute the program. The present invention can also be realized by a circuit (e.g., ASIC) that implements one or more of the functions.

[0076] The disclosure of this specification includes the following image processing device, recording device, threshold matrix generating device, image processing method, threshold matrix generating method, and program. [Item 1] An input means for inputting a plurality of multi-value data corresponding to a plurality of color materials; a quantization unit that quantizes each of the plurality of multi-value data to generate a plurality of quantized data indicating recording or non-recording of dots on a recording medium with each of the plurality of color materials, 13. An image processing apparatus comprising: a plurality of dot patterns of the plurality of color materials recorded based on the plurality of quantized data generated by the quantization means, the plurality of dot patterns having peaks in different deflection angle directions in a power spectrum of a frequency space, and having blue noise characteristics or green noise characteristics. [Item 2] The image processing device according to item 1, wherein the deflection angle directions of the multiple dot patterns are in the range of 0 degrees to 180 degrees. [Item 3] The image processing device according to item 2, wherein when the plurality of color materials are n colors, the deflection angles of the plurality of dot patterns differ by 180 degrees / n. [Item 4] In the polar coordinate representation F(r, θ) of the power spectrum in the frequency space of each of the plurality of dot patterns, when the average power at frequency r is expressed as the radian average power spectrum RAPS(r) for each frequency, and the average power at an argument angle θ ranging from 0 degrees to 180 degrees is expressed as the frequency average power spectrum FAPS(θ) for each angle, 4. The image processing device according to any one of items 1 to 3, wherein RAPS(r) of each of the plurality of dot patterns has blue noise characteristics or green noise characteristics and has a peak in a band corresponding to the peak in the deflection angle direction, and FAPS(θ) of the plurality of dot patterns has blue noise characteristics or green noise characteristics and has a peak in the deflection angle direction. [Item 5] The image processing device according to Item 4, wherein the deflection angle regions which are full widths at half maximum of the peaks in the FAPS(θ) of the respective dot patterns do not overlap with each other. [Item 6] The image processing device described in any one of Items 1 to 5, characterized in that the quantization means has a plurality of threshold matrices corresponding to the plurality of color materials, and generates quantized data by comparing each of the plurality of multi-value data with the threshold of the threshold matrix for the corresponding color material. [Item 7] An image processing device according to any one of items 1 to 6, a recording unit for recording each of the plurality of color materials on a recording medium in accordance with the plurality of quantized data. [Item 8] A threshold matrix generating device for generating a threshold matrix, comprising: a first generating means for generating a density fluctuation map representing the density of dots in the dot pattern of the first gradation value based on a result of filtering the dot pattern of the first gradation value and an initial density fluctuation map; a second generating means for generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; a setting means for setting the second gradation value at a position of the threshold matrix corresponding to a dot position where a change occurs between the dot pattern of the first gradation value and the dot pattern of the second gradation value, The threshold matrix generating device according to the present invention is characterized in that the initial density fluctuation map is a map obtained by applying density fluctuations that are periodically repeated in a predetermined direction to a two-dimensional array made up of random numbers. [Item 9] The threshold matrix generating device according to Item 8, wherein the amplitude of the random number is smaller than the difference in density fluctuation given to the initial density fluctuation map. [Item 10] A threshold matrix generating device for generating a threshold matrix, comprising: a first generating means for generating a density variation map representing the density of dots in the dot pattern of a first gradation value based on a result of filtering the dot pattern of the first gradation value; a second generating means for generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; a setting means for setting the second gradation value at a position of the threshold matrix corresponding to a dot position where a change occurs between the dot pattern of the first gradation value and the dot pattern of the second gradation value, a filter having different characteristics for each area divided based on a predetermined direction, so that a peak appears in a predetermined direction in a power spectrum, said filter processing being performed using said filter having different characteristics for each area divided based on the predetermined direction. [Item 11] The threshold matrix generating device according to Item 10, wherein filters with different kernel sizes are used as the filters with different characteristics. [Item 12] The threshold matrix generating device according to any one of Items 8 to 11, wherein the filtering process is a process using a low-pass filter. [Item 13] The first gradation value is smaller than the second gradation value, The second generating means generates the dot pattern of the second gradation value by adding a dot at a position where no dot exists in the dot pattern of the first gradation value and where the value of the generated density variation map is minimum. The threshold matrix generating device described in any one of items 8 to 12, [Item 14] The first gradation value is a gradation value greater than the second gradation value, The second generating means generates the dot pattern of the second gradation value by deleting dots from positions where dots exist in the dot pattern of the first gradation value and where the value of the generated density variation map is maximum. [Item 15] Set a dot pattern at a given gradation value, 13. The threshold matrix generating device according to any one of items 8 to 12, characterized in that the second generating means generates a dot pattern of the second gradation value by changing the gradation value in ascending and descending order from the predetermined gradation value. [Item 16] The threshold matrix generating device according to any one of items 8 to 15, characterized in that the filtering process includes a first filtering process in which the dot pattern of the first gradation value is processed by a low-pass filter, and a second filtering process in which the result of the first filtering process is processed by a filter having a coefficient whose value becomes smaller the farther from the center of the filter. [Item 17] An input step of inputting a plurality of multi-value data corresponding to a plurality of color materials; a quantization step of quantizing each of the plurality of multi-value data to generate a plurality of quantized data indicating recording or non-recording of dots on a recording medium by each of the plurality of color materials, a plurality of dot patterns of the plurality of color materials recorded based on the plurality of quantized data generated in the quantization step, each having a peak in a different deflection angle direction in a power spectrum in frequency space, and having blue noise characteristics or green noise characteristics. [Item 18] A method for generating a threshold matrix, comprising: a first generation step of generating a density fluctuation map representing the density of dots in the dot pattern of the first gradation value based on a result of filtering the dot pattern of the first gradation value and an initial density fluctuation map; a second generation step of generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; a setting step of setting the second gradation value at a position of the threshold matrix corresponding to a dot position where a change occurs between the dot pattern of the first gradation value and the dot pattern of the second gradation value, The threshold matrix generating method according to the present invention, wherein the initial density fluctuation map is a map obtained by applying density fluctuations that are periodically repeated in a predetermined direction to a two-dimensional array of random numbers. [Item 19] A method for generating a threshold matrix, comprising: a first generation step of generating a density variation map representing dot sparseness or density in the dot pattern of a first gradation value based on a result of filtering the dot pattern of the first gradation value; a second generation step of generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; a setting step of setting the second gradation value at a position of the threshold matrix corresponding to a dot position where a change occurs between the dot pattern of the first gradation value and the dot pattern of the second gradation value, a filter having different characteristics for each area divided based on a predetermined direction is used in the filtering process so that a peak appears in the predetermined direction in a power spectrum. [Item 20] A program for causing a computer to execute the method described in any one of Items 17 to 19.

[0077] The invention is not limited to the above-described embodiments, and various modifications and variations are possible without departing from the spirit and scope of the invention. Accordingly, the following claims are appended to apprise the public of the scope of the invention. [Explanation of symbols]

[0078] 1: image processing device, 2: printer, 102: color matching processing unit, 103: color separation processing unit, 106: quantization processing unit, 107 dither matrix storage unit, 301: threshold matrix generating device

Claims

1. Input means for inputting a plurality of multi-value data corresponding to a plurality of color materials; Quantization means for quantizing each of the plurality of multi-value data to generate a plurality of quantization data indicating dot recording or non-recording on a recording medium by each of the plurality of color materials, and The plurality of dot patterns of the plurality of color materials recorded based on the plurality of quantization data generated by the quantization means have peaks in different angular directions in the power spectrum in the frequency space, and have blue noise characteristics or green noise characteristics. An image processing apparatus characterized by this.

2. The image processing apparatus according to claim 1, wherein the angular directions of the plurality of dot patterns are in the range of 0 degrees to 180 degrees.

3. The image processing apparatus according to claim 2, wherein when the plurality of color materials are n colors, the angular directions of the plurality of dot patterns are different by 180 degrees / n.

4. In the polar coordinate representation F(r, θ) of the power spectrum in the frequency space of each of the plurality of dot patterns, the average of the power at frequency r is the average power spectrum per frequency RAPS(r), and the average of the power at the angular deviation θ in the range of 0 degrees to 180 degrees is the average power spectrum per angle FAPS(θ). When expressed as, The RAPS(r) of each of the plurality of dot patterns has blue noise characteristics or green noise characteristics and has a peak in the band corresponding to the peak in the angular direction, and the FAPS(θ) of the plurality of dot patterns has blue noise characteristics or green noise characteristics and has a peak in the angular direction. The image processing apparatus according to claim 1, characterized by this.

5. The image processing apparatus according to claim 4, wherein the angular regions that are the full width at half maximum of the peak in the FAPS(θ) of each of the plurality of dot patterns do not overlap each other.

6. The quantization means has a plurality of threshold matrices corresponding to the plurality of color materials, and generates quantization data by comparing each of the plurality of multi-value data with the thresholds of the threshold matrices of the corresponding color materials. The image processing apparatus according to claim 1, characterized by this.

7. The image processing apparatus according to any one of claims 1 to 6, and A recording apparatus comprising: recording means for recording each of the plurality of color materials on a recording medium according to the plurality of quantization data.

8. A threshold matrix generation device for generating a threshold matrix, a first generation means for generating a density variation map representing the density of dots in the dot pattern of the first gradation value based on the result of filtering the dot pattern of the first gradation value and the initial density variation map; a second generation means for generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; setting means for setting the second gradation value at a position corresponding to a position of a dot where a change has occurred between the dot pattern of the first gradation value and the dot pattern of the second gradation value in the threshold matrix; and having, The threshold matrix generation device, wherein the initial density variation map is a map obtained by giving density variations that are periodically repeated in a predetermined direction to a two-dimensional array composed of random numbers.

9. The threshold matrix generation device according to claim 8, wherein an amplitude of the random number is smaller than a difference in density variation given to the initial density variation map.

10. A threshold matrix generation device for generating a threshold matrix, a first generation means for generating a density variation map representing the density of dots in the dot pattern of the first gradation value based on the result of filtering the dot pattern of the first gradation value; a second generation means for generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; setting means for setting the second gradation value at a position corresponding to a position of a dot where a change has occurred between the dot pattern of the first gradation value and the dot pattern of the second gradation value in the threshold matrix; and having, In the filtering process, a filter having different characteristics is used for each region divided based on the predetermined direction so that a peak appears in a predetermined direction in the power spectrum. Threshold matrix generation device.

11. The threshold matrix generation device according to claim 10, wherein filters having different kernel sizes are used as the filters having different characteristics.

12. The threshold matrix generation device according to claim 8 or 10, wherein the filtering process is a process by a low-pass filter.

13. The first gradation value is a gradation value smaller than the second gradation value, The second generation means generates a dot pattern of the second gradation value by adding dots at positions where there are no dots in the dot pattern of the first gradation value and where the value of the generated density variation map is minimized. The threshold matrix generation device according to claim 8 or 10.

14. The first gradation value is a gradation value larger than the second gradation value, The second generation means generates a dot pattern of the second gradation value by deleting dots at positions where dots exist in the dot pattern of the first gradation value and where the value of the generated density variation map is maximized. The threshold matrix generation device according to claim 8 or 10.

15. Set a dot pattern at a predetermined gradation value, The second generation means generates a dot pattern of the second gradation value by changing the gradation value in ascending and descending order from the predetermined gradation value. The threshold matrix generation device according to claim 8 or 10.

16. The filtering process includes a first filtering process of processing the dot pattern of the first gradation value by a low-pass filter, and a second filtering process of processing the result of the first filtering process by a filter having a coefficient such that the value becomes smaller as it is farther from the center of the filter. The threshold matrix generation device according to claim 8 or 10.

17. An input step of inputting a plurality of multi-value data corresponding to a plurality of color materials, A quantization step of quantizing each of the plurality of multi-value data to generate a plurality of quantization data indicating recording or non-recording of dots on a recording medium by each of the plurality of color materials, The plurality of dot patterns of the plurality of color materials recorded based on the plurality of quantization data generated in the quantization step have peaks in different angular directions in the power spectrum in the frequency space and have blue noise characteristics or green noise characteristics. An image processing method characterized by this.

18. A method for generating a threshold matrix, A first generation step of generating a density variation map representing the density and sparseness of dots in the dot pattern of the first gradation value based on the result of filtering the dot pattern of the first gradation value and the initial density variation map; A second generation step of generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; A setting step of setting the second gradation value at a position corresponding to the position of a dot where a change has occurred between the dot pattern of the first gradation value and the dot pattern of the second gradation value in the threshold matrix, and having: The initial density variation map is a map obtained by giving density variations that are periodically repeated in a predetermined direction to a two-dimensional array composed of random numbers, and is a method for generating a threshold matrix.

19. A method for generating a threshold matrix, comprising: A first generation step of generating a density variation map representing the density and sparseness of dots in the dot pattern of the first gradation value based on the result of filtering the dot pattern of the first gradation value; A second generation step of generating a dot pattern of a second gradation value adjacent to the first gradation value based on the generated density variation map and the dot pattern of the first gradation value; A setting step of setting the second gradation value at a position corresponding to the position of a dot where a change has occurred between the dot pattern of the first gradation value and the dot pattern of the second gradation value in the threshold matrix, and having: In the filtering process, a filter having different characteristics for each region divided based on the predetermined direction is used so that a peak appears in a predetermined direction in the power spectrum, and is a method for generating a threshold matrix.

20. A program for causing a computer to execute the image processing method according to claim 17.

21. A program for causing a computer to execute the method for generating a threshold matrix according to claim 18 or 19.

Images