JP4921210B2 - Image processing apparatus and method - Google Patents

Description

  The present invention relates to image processing for binarizing image data.

  As information output devices for word processors, personal computers, facsimiles, and the like, there are various types of recording devices that record characters and images on a sheet-like recording medium such as recording paper or film. Among them, a method of forming text and images on a recording medium by attaching a recording agent to the recording medium has been put into practical use. A typical example of such a system is an ink jet recording apparatus.

  The ink jet recording apparatus uses a nozzle group in which a plurality of ink discharge ports (nozzles) capable of discharging the same color ink are integrated and arranged in order to improve the recording speed and improve the image quality. These nozzle groups are provided for inks having different colors. Further, there may be provided a nozzle group capable of ejecting ink having the same color but different density, and a nozzle group capable of ejecting by switching the ejection amount of the same color ink in several stages.

  The ink jet recording apparatus can express each pixel of the output image only with two gradations depending on the presence or absence of a recording agent. In order to form an image with a greater number of gradations than can be represented, the number of gradations of the image is converted to a gradation that can be represented by the recording device (referred to as quantization), and the gradation that the recording device cannot represent. Pseudo halftone processing is performed to simulate the key.

  An error diffusion method by R. Floyd et al. Is known as one method of pseudo halftone processing (see Non-Patent Document 1). The error diffusion method is a pseudo halftone process that diffuses an error (quantization error) caused by quantization of a certain pixel to the subsequent pixels and makes the average density of the output image equal to the density of the input image in a macro manner. It is.

  There are cases where an image is formed using different color inks, the same color and different density, or the same color and different ejection amounts. In this case, if input images corresponding to inks having different colors, densities, or ejection amounts (hereinafter, different inks) are independently pseudo-halftone, different inks may be recorded at the same position. That is, dots with different inks (hereinafter, different types of dots) overlap each other, the graininess is deteriorated, and good visual characteristics may not be obtained.

  In order to improve the above problem, Patent Document 1 discloses a pseudo halftone process that prevents overlapping of different types of dots by performing error diffusion processing by combining two or more colors corresponding to different inks.

  Further, the error diffusion method has a problem that when a uniform gradation image is quantized, a regular pattern called texture is generated in the quantized image. In particular, when an image having a uniform gradation of a specific gradation (around 64, 128, or 192 gradation in the case of an 8-bit image) is quantized, a texture is likely to be generated.

  As a method for suppressing the generation of textures and pseudo contours peculiar to the error diffusion method, there is a method of adding noise to input image data or a quantization threshold (see Patent Document 2).

  The invention of Patent Document 1 relates to a dot position control method based on an error diffusion method. Therefore, it is possible to prevent overlapping of different types of dots, but there is a problem that a texture peculiar to the error diffusion method occurs.

  In addition, as in the invention of Patent Document 2, when noise with a large amplitude is added to eliminate textures and pseudo contours, noise remains in the output image, and there is a problem that an image with poor image quality is formed with noticeable graininess. is there.

  Patent Documents 3 and 4 disclose pseudo-halftone processing that uses two or more colors in combination and uses an error diffusion method to prevent overlapping of dots of different colors. Since the invention of Patent Document 3 controls the dot position with reference to a lookup table (LUT), the amount of LUT data becomes enormous as the number of ink types increases. Further, in the invention of Patent Document 4, as the number of ink types increases, the number of times the LUT is referred to increases and the processing becomes complicated.

JP 2004-336570 A JP2003-069819 Japanese Patent Laid-Open No. 11-010918 JP 2004-229010 JP R. Floyd et al. “An Adaptive Algorithm for Spatial Gray Scale” in society for Information Display 1975 Symposium Digest of Technical Papers, 1975, pp. 36-37

An object of the present invention is to perform pseudo halftone processing that prevents deterioration of graininess and reduces the occurrence of texture .

  The present invention has the following configuration as one means for achieving the above object.

An image processing apparatus according to the present invention includes a holding unit that holds a binary pattern image having a predetermined average gradation value and a blue noise characteristic, and two binary values that respectively binarize image data of two channels. Each of the binarizing means includes an error memory for storing an error diffusion value, and an adding means for outputting an addition value obtained by adding the error diffusion value and the pixel value of the target pixel. A calculation means for calculating an error between the added value and the binarized value of the target pixel, and an error diffusion value stored in the error memory is updated so that the error is diffused to the pixel before binarization. For comparing the pixel value of the target pixel and the average gradation value of the pattern image, the pixel value of the pattern image corresponding to the position of the target pixel, and the comparison of the addition value and the quantization threshold Based on the above note And a quantization means for determining a binarization value of the pixel, the pattern image shows a determining condition of the binarized values indicating the dot arrangement in the two channels, the binarizing means, the pattern image Referring to the determination condition of said two channels is characterized that you control to have opposite to each other.

An image processing method according to the present invention includes a holding unit that holds a binary pattern image having a predetermined average gradation value and a blue noise characteristic, an error memory that stores an error diffusion value, and image data of two channels. An image processing method of an image processing apparatus having binarization means for binarizing each of the binarization means, wherein the binarization means calculates an addition value obtained by adding the error diffusion value and a pixel value of a target pixel, and the addition An error between the value and the binarized value of the target pixel is calculated, and an error diffusion value stored in the error memory is updated so that the error is diffused to the pixel before binarization. Based on the comparison between the pixel value and the average gradation value of the pattern image, the pixel value of the pattern image corresponding to the position of the pixel of interest, and the comparison of the addition value and the quantization threshold, the binarization of the pixel of interest It performs a process to determine the value, Serial pattern image shows a determining condition of the binarized values indicating the dot arrangement in the two channels, the binarizing means, the determination condition of said two channels by reference are opposite to each other said pattern image control to characterized Rukoto as.

According to the present invention, with a simple configuration using a binary pattern image, it is possible to perform pseudo halftone processing that prevents deterioration of graininess and reduces texture generation .

  Hereinafter, image processing according to an embodiment of the present invention will be described in detail with reference to the drawings.

  Hereinafter, image processing for converting an input image having 8 bits and 256 gradations into an image having two gradations will be described. Of course, in the present invention, the number of input gradations is not limited to 256 gradations, and may be more or less than 256 gradations. In the following description, the horizontal size of the input image is W pixels and the vertical size is H pixels.

  The binary image has a binary value corresponding to 0 (off dot) or 255 (on dot). In a binary image, a case where pixels corresponding to a pixel value 255 are included in a ratio of g pixels in 256 pixels is expressed as g gradation.

  The image processing apparatus according to the embodiment uses a pattern image to create an input image of 2 or 3 channels corresponding to different inks so that dots of 2 or 3 channels do not overlap in an area where the dot recording rate is low. Convert to placed binary image.

  For example, the first channel represents an image recorded with, for example, cyan (C) ink, the second channel represents an image recorded with magenta (M) ink, and the third channel represents, for example, yellow (Y) ink Represents an image to be recorded. The combination of channels is an image recorded with dots having a large area, an image recorded with dots having a small area, an image recorded with dark ink of the same color, and an image recorded with light ink. Any combination is possible.

[Device configuration]
FIG. 1 is a block diagram illustrating a configuration example of the image processing apparatus according to the first embodiment.

  The input terminals 100, 109, and 116 are connected to a storage device such as a memory for storing input image data, or an input device such as a scanner or an imaging device. From these devices, the input terminal 100 receives the image data of the first channel in the raster scan order, and the input terminal 109 receives the image data of the second channel in the raster scan order. Note that the image data of both channels are input to the input terminals 100 and 109 in synchronization.

  The output terminals 108 and 115 are connected to a storage device or an output device such as a display device or a printing device. The output terminal 108 outputs binary image data of the first channel, and the output terminal 115 outputs binary image data of the second channel. Note that the image data of both channels are output from the output terminals 108 and 115 in synchronization.

  The accumulated error adding unit 101 (110) adds the error diffusion value stored in the accumulated error memory 107 (114) to the image data input from the input terminal 100 (109) for each pixel. The binary pattern memory 102 stores the binary pattern image and its gradation value, and stores the pixel value and gradation value of the binary pattern image corresponding to the position of the target pixel indicated by the signal input to the input terminal 116. Output.

  Although the details will be described later, the quantization unit 103 (111) calculates the quantization value (binarization value) of the pixel of interest in the first channel (second channel) based on the quantization threshold stored in the threshold memory 104 and the like. decide.

  The error calculator 105 (112) calculates a quantization error from the input / output image data of the quantizer 103 (111). The error diffusion unit 106 (113) updates the error diffusion value stored in the accumulated error memory 107 (114) so that the quantization error is diffused to a predetermined pixel before quantization.

  As described above, the image processing apparatus includes an input terminal, an accumulated error adding unit, a quantizing unit, an error calculating unit, an error diffusing unit, an accumulated error memory, and an output terminal independently for each channel. On the other hand, the binary pattern memory 102 and the threshold memory 104 are common to both channels.

[Binary pattern image settings]
The binary pattern memory 102 stores a binary pattern image having a predetermined gradation bg in advance. The image size Wb × Hb pixels of the binary pattern image is arbitrary. When the binary pattern image is smaller than the input image, the binary pattern image is repeatedly arranged in a tile shape so that the binary pattern corresponds to each pixel of the input image.

  The binary pattern image can take an arbitrary gradation bg and an arbitrary binary pattern. However, in order to reduce the occurrence of texture, it is preferable that the binary pattern image has high on-dot dispersibility and the spatial frequency has a blue noise characteristic. Any method may be used for creating a binary pattern image having blue noise characteristics. For example, the method disclosed in US Pat. No. 4,920,501 is known. The blue noise characteristic refers to a characteristic that has a low frequency component and a high frequency component.

  Furthermore, in order to reduce the occurrence of texture, the gradation value bg of the binary pattern image may be made equal to the gradation at which the texture is generated by error diffusion (such as 128 gradation when the input image is 8 bits). desirable.

  The image processing apparatus according to the first embodiment outputs a pattern of a binary pattern image when a uniform gradation image having a gradation equal to the gradation value bg of the binary pattern image is input. That is, when an image having a gradation value bg is input, a binary pattern image is output, so that the generation of texture can be prevented.

  In addition, it is possible to arrange dots by preventing dot superimposition up to the input gradation value that is smaller of the gradation value bg of the binary pattern image and the value obtained by reversing it from black and white (255-bg). It is. Therefore, in order to prevent the overlapping of the two-channel quantization result dots up to a higher gradation, the gradation value bg of the binary pattern image is preferably a value close to 128.

[Quantization]
First Channel Quantization FIG. 2 is a flowchart for explaining the quantization process.

  First, image data is input to the input terminal 100 in the scanning order (S201).

  FIG. 3 is a diagram illustrating scanning of an input image. Image scanning starts from the pixel 400 at the upper left corner of the image, and proceeds rightward by one pixel (main scanning). When the right end of the image is reached, the pixel moves to the left end pixel one line down (sub-scanning). The main scanning and the sub-scanning are repeated, and when the pixel 401 at the lower right end is reached, the image scanning is finished.

  The accumulated error adding unit 101 adds the error diffusion value corresponding to the pixel position in the accumulated error memory 107 to the input pixel data (the value of the target pixel before quantization) (S202).

FIG. 4 is a diagram for explaining error diffusion values stored in the cumulative error memory 107. FIG. The cumulative error memory 107 has a storage area E ₁₀ corresponding to one pixel and a storage area E (x) ₁ (x = ₁ to W) of W pixels (the same number as the horizontal size of the image) in the main scanning direction. The error diffusion value is stored by a method described later. All the storage areas E ₁ are initialized to the initial value 0 before the processing is started.

The accumulated error adding unit 101 adds the error diffusion value stored in the storage area E ₁ (x) corresponding to the horizontal pixel position x of the target pixel to the target pixel, as shown in Expression (3).
I ₁ '= I ₁ + E ₁ (x)… (1)
Where I ₁ is the value of the pixel of interest
I ₁ 'is the value of the pixel of interest to which the error diffusion value is added (hereinafter referred to as the added value)

  Next, the binary pattern memory 902 outputs the pixel value B of the binary pattern image corresponding to the position of the target pixel and the gradation value bg of the binary pattern image (S203).

If the binary pattern image size Wb × Hb is smaller than the input image size W × H, the binary pattern image coordinates (xb, yb) are changed to the input image coordinates (x, y) as follows: Make it correspond.
xb = x% Wb
yb = y% Hb (2)
Where% is the modulo operation

Next, the quantization unit 103 includes the pixel value I ₁ , the addition value I ₁ ′, the pixel value B of the binary pattern image, the tone value bg of the binary pattern image, and the quantum stored in the threshold memory 104. The output threshold value th is compared to determine the output pixel value O ₁ (S204). Details of the quantization will be described later.

Next, the error calculator 105 calculates a quantization error E ₁ that is a difference between the added value I ₁ ′ and the output pixel value O ₁ (S205).
E ₁ = I ₁ '-O ₁ … (3)

Next, the error diffusion unit 106 performs error diffusion processing and updates the error diffusion value stored in the cumulative error memory 107 (S206). The error diffusion unit 106 diffuses the error according to the following equation according to the horizontal position x of the target pixel.
If x <W, E ₁ (x + 1) = E ₁ (x + 1) + E ₁ × 7/16
When x> 1, E ₁ (x-1) = E ₁ (x-1) + E ₁ × 3/16
If 1 <x <W, E ₁ (x) = E ₁ 0 + E ₁ × 5/16
When x = 1, E ₁ (x) = E ₁ 0 + E ₁ × 8/16… (4)
When x = W, E ₁ (x) = E ₁ 0 + E ₁ × 13/16
If x <W, E ₁ 0 = E ₁ × 1/16
If x = W, E ₁ 0 = 0

That is, if the position x of the pixel of interest is other than both ends of the line, and distributed to the pixel immediately after the (x + 1) of the target pixel 7/16 of the quantization error E _1, the pixel of interest belongs to 9/16 remaining Distributes to the next line of lines. In the next line, 5/16 of the quantization error E ₁ is distributed to the pixel at the same position x as the target pixel (the pixel immediately below the target pixel). Then, 3/16 and 1/16 of the quantization error E ₁ are distributed to the immediately preceding pixel (the pixel at the lower left of the pixel of interest) and the immediately following pixel (the pixel at the lower right of the pixel of interest).

In addition, when the position x of the target pixel is the head of the line, 7/16 of the quantization error E _{1 is} the pixel immediately after the target pixel (x + 1), 8/16 is the pixel immediately below the target pixel, 1/16 is distributed to the lower right pixel of the target pixel.

The position x of the pixel of interest is the case at the end of the line, to distribute all of the quantization error E ₁ for the next line. That is, the lower left pixel of the target pixel 3/16 of the quantization error E _1, distributes the 13/16 to the pixel directly below the pixel of interest.

  This completes the binarization processing for one pixel of the first channel. Then, when the binarization process is performed on all the pixels of the image of the first channel, the process ends. When there is an unprocessed pixel, the process returns to step S201, and the above operation is repeated (S207).

Quantization unit 103
FIG. 5 is a flowchart for explaining the processing of the quantization unit 103.

The quantization unit 103 compares the value I _{1 of the} pixel of interest with the gradation value bg of the binary pattern image (S301). The added value I ₁ ′ may be compared with the gradation value bg of the binary pattern image. Subsequently, it is determined whether or not the pixel value B of the binary pattern image indicates an on dot (S302, S307).

Then, when I ₁ > bg and B = on dot, the quantization unit 103 sets the output pixel value O ₁ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S306). .
When I ₁ > bg and B = 255, O ₁ = 255 (5)

In addition, when I ₁ > bg and B = off dot, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (for example, 128) (S303), as shown in Expression (6) The output pixel value O ₁ is determined (S304, S305).
I ₁ > bg and B = 0
And if I ₁ '<th, O ₁ = 0… (6)
And if I ₁ '≥ th, O ₁ = 255

Further, when I ₁ ≦ bg and B = off dot, the quantization unit 103 sets the output pixel value O ₁ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S311) .
When I ₁ ≤bg and B = 0, O ₁ = 0 (7)

In addition, when I ₁ ≦ bg and B = on dot, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (S308), and outputs an output pixel value O as shown in Equation (8). ₁ is determined (S309, S310).
I ₁ ≤bg and B = 255
And if I ₁ '<th, O ₁ = 0… (8)
And if I ₁ '≧ th, O ₁ = 255

● Quantization of the second channel Figure 2 also shows the flow of quantization processing for the second channel.

  First, image data is input to the input terminal 109 in the scanning order (S201).

  The accumulated error adding unit 110 adds the error diffusion value corresponding to the pixel position in the accumulated error memory 114 to the input pixel data (the value of the target pixel before quantization) (S202).

FIG. 6 is a diagram for explaining error diffusion values stored in the cumulative error memory 114. FIG. The cumulative error memory 114 has a storage area E ₂₀ corresponding to one pixel and a storage area E (x) ₂ (x = 1 to W) of W pixels (the same number as the horizontal size of the image) in the main scanning direction. The error diffusion value is stored by a method described later. All storage area E ₂ is initialized to an initial value 0 before treatment start.

The accumulated error adding unit 110 adds the error diffusion value stored in the storage area E ₂ (x) corresponding to the horizontal pixel position x of the target pixel to the target pixel, as shown in Expression (9).
I ₂ '= I ₂ + E ₂ (x)… (9)
Where I ₂ is the value of the pixel of interest
I ₂ 'is the value of the pixel of interest to which the error diffusion value is added (hereinafter referred to as the added value)

  Next, the binary pattern memory 902 outputs the pixel value B of the binary pattern image corresponding to the position of the target pixel and the gradation value bg of the binary pattern image (S203).

Next, the quantization unit 111 includes the value I _{2 of} the target pixel, the addition value I ₂ ′, the pixel value B of the binary pattern image, the gradation value bg of the binary pattern image, and the quantum stored in the threshold memory 104. The output threshold value th is compared to determine the output pixel value O ₂ (S204). Details of the quantization will be described later.

Next, the error calculator 112 calculates a quantization error E ₂ that is a difference between the added value I ₂ ′ and the output pixel value O ₂ (S205).
E ₂ = I ₂ '-O ₂ … (10)

Next, the error diffusion unit 113 performs error diffusion processing and updates the error diffusion value stored in the cumulative error memory 114 (S206). The error diffusion unit 113 diffuses an error in the same manner as the error diffusion unit 106 according to the horizontal position x of the target pixel.
If x <W, E ₂ (x + 1) = E ₂ (x + 1) + E ₂ × 7/16
When x> 1, E ₂ (x-1) = E ₂ (x-1) + E ₂ × 3/16
If 1 <x <W, E ₂ (x) = E ₂ 0 + E ₂ × 5/16
When x = 1, E ₂ (x) = E ₂ 0 + E ₂ × 8/16… (11)
When x = W, E ₂ (x) = E ₂ 0 + E ₂ × 13/16
If x <W, E ₂ 0 = E ₂ × 1/16
E ₂ 0 = 0 when x = W

  This completes the binarization process for one pixel of the second channel. Then, when the binarization process is performed on all the pixels of the image of the second channel, the process ends. When there is an unprocessed pixel, the process returns to step S201, and the above operation is repeated (S207).

Quantization unit 111
FIG. 7 is a flowchart for explaining the processing of the quantization unit 111.

The quantization unit 111 compares the value I ₂ of the target pixel with the gradation value bg of the binary pattern image (S601). The added value I ₂ ′ may be compared with the gradation value bg of the binary pattern image. Subsequently, it is determined whether or not the pixel value B of the binary pattern image indicates an on dot (S602, S607).

Then, when I ₂ > bg and B = off dot, the quantization unit 111 sets the output pixel value O ₂ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S606). .
When I ₂ > bg and B = 0, O ₂ = 255 (12)

Further, when I ₂ > bg and B = on dot, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (for example, 128) (S603), as shown in Expression (13) The output pixel value O ₂ is determined (S604, S605).
I ₂ > bg and B = 255
And if I ₂ '<th, O ₂ = 0… (13)
And if I ₂ '≥ th, O ₂ = 255

Further, when I ₂ ≦ bg and B = on dot, the quantization unit 111 sets the output pixel value O ₂ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S611) .
When I ₂ ≤ bg and B = 255, O ₂ = 0… (14)

In addition, when I ₂ ≦ bg and B = off dot, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (S608), and outputs an output pixel value O as shown in Expression (15). ₂ is determined (S609, S610).
I ₂ ≤ bg and B = 0
And if I ₂ '<th, O ₂ = 0… (15)
And if I ₂ '≥ th, O ₂ = 255

Thus, in the quantization of the first channel, when I ₁ of the target pixel is equal to or lower than the gradation value bg of the binary pattern image, the corresponding binary pattern image pixel value B is 255 (on dot). A pixel that forms an on-dot from the pixel is selected. The pixel value I ₁ is the case where more than the tone value bg, the pixel value B of the corresponding binary pattern image 0 selects a pixel which does not form a dot-on from the pixels of the (off-dots).

On the other hand, in the quantization of the second channel, in order to arrange the on-dots of the second channel so as not to overlap the on-dots of the first channel, control opposite to that of the first channel is performed. That is, if the pixel value I ₂ of the pixel of interest is less than the gradation value bg, the pixel value B of the corresponding binary pattern image selects a 0 pixel forming the dot-on from the pixels of the (off-dots). Also, if the pixel value I ₂ of the pixel of interest is greater than the tone value bg, the pixel value B of the corresponding binary pattern image to select the pixels that do not form a dot-on from the pixels 255 (ON dots). In other words, in the binarization of the first channel and the binarization of the second channel, the determination conditions for the binarization value indicating the on dot are opposite to each other.

In other words, in the quantization of the first channel, when the value I ₁ of the target pixel is equal to or lower than the gradation value bg (the dot recording rate is low), the pixel value B of the binary pattern image is 255 (on dot). On dots can be formed only in certain pixels. On the other hand, in the quantization of the second channel, when the value I ₂ of the target pixel is equal to or lower than the gradation value bg (the dot recording rate is low), the pixel value B of the binary pattern image is 0 (off dot). On dots can be formed only in certain pixels. Therefore, when the dot recording rates of both the first channel and the second channel are low, the pixels that can form on-dots are different, so that the overlap of the dots of both channels can be prevented.

  Furthermore, in both channels, pixels that form on-dots or do not form on-dots are not freely selectable from the entire image, but the pixel values B of the corresponding binary pattern image are from dots-on pixels or , Dot-off pixels are selected. For this reason, by using a pattern with high dispersibility of dot arrangement as the binary pattern image, it is possible to prevent the regular formation of dots and suppress the occurrence of texture.

  The image processing according to the second embodiment of the present invention will be described below. Note that the same reference numerals in the second embodiment denote the same parts as in the first embodiment, and a detailed description thereof will be omitted.

  In the first embodiment, the binary pattern image for controlling the dot formation position is commonly used for the quantization of the first channel and the quantization of the second channel. In the second embodiment, an example in which different binary pattern images are used for the quantization of the first channel and the quantization of the second channel will be described.

[Device configuration]
FIG. 8 is a block diagram illustrating a configuration example of the image processing apparatus according to the second embodiment. 1 is different from the configuration shown in FIG. 1 in that binary pattern memories 802 and 811 and threshold memories 804 and 813 are provided for each channel instead of the binary pattern memory 102 and threshold memory 104 common to both channels.

[Binary pattern image settings]
The binary pattern memories 802 and 811 previously store binary pattern images having predetermined gradations bg1 and bg2, respectively. The gradation values bg1 and bg2 have the following relationship, and the gradation values bg1 and bg2 can take arbitrary values as long as the following relationship is satisfied. The tone values bg1 and bg2 are proportional to the appearance rate of on dots (B = 255) in the binary pattern image.
0 ≤ (bg1 + bg2) ≤ 255… (16)

  Also, the binary pattern images stored in the binary pattern memories 802 and 811 need to have an exclusive arrangement of pixels with a pixel value B = 255 (on dots). In other words, the pixel of the binary pattern image stored in the binary pattern memory 811 at the same position as the pixel of the pixel value B1 = 255 (dot on) of the binary pattern image stored in the binary pattern memory 802 is the pixel value. It is necessary that B2 = 0 (off dot off). Similarly, the pixel of the binary pattern image stored in the binary pattern memory 802 at the same position as the pixel of the pixel value B2 = 255 (on dot) of the binary pattern image stored in the binary pattern memory 811 is the pixel value. B1 = 0 (off dot) needs to be set.

  If the dot arrangement is exclusive, the binary pattern images stored in the binary pattern memories 802 and 811 may be arbitrary patterns. However, in order to reduce the occurrence of texture, it is preferable that the binary pattern image has high on-dot dispersibility and the spatial frequency has a blue noise characteristic.

[Quantization]
The quantization of the first channel is performed in the process of the quantization unit 103 described in the first embodiment by converting the gradation value bg of the binary pattern image to bg1 and the pixel value B of the binary pattern image to B1. The threshold th may be replaced with th1. Similarly, the quantization of the second channel is performed by changing the gradation value bg of the binary pattern image to bg2 and the pixel value B of the binary pattern image to B2 in the processing of the quantization unit 111 described in the first embodiment. The quantization threshold th may be replaced with th2. However, both the quantizing units 103 and 113 are different from the first embodiment in that the output pixel values O ₁ and O ₂ are determined by the processing shown in FIG. 5 (or FIG. 8).

  In the second embodiment, the arrangement of on dots is exclusive for the binary pattern image for the first channel and the binary pattern image for the second channel. Therefore, when the dot recording rate of both channels is low, it is possible to prevent the overlapping of dots of both channels as in the first embodiment.

  Hereinafter, image processing according to the third embodiment of the present invention will be described. Note that the same reference numerals in the third embodiment denote the same parts as in the first and second embodiments, and a detailed description thereof will be omitted.

  In the first embodiment, the case where the input image has two channels has been described. If a multi-value pattern is used as a pattern for controlling the dot formation position, it is possible to prevent dots from being superimposed on an input image exceeding two channels. For example, a ternary pattern image is used to prevent overlapping of three-channel dots.

[Device configuration]
FIG. 9 is a block diagram illustrating a configuration example of the image processing apparatus according to the third embodiment.

  In addition to the configuration shown in FIG. 1, the image processing apparatus according to the third embodiment quantizes the terminals 914 and 920 for inputting and outputting the image data of the third channel and the image data of the third channel (corresponding channel). The structure to be provided is provided. The configuration of the quantization process is an adder 915, a quantization unit 916, an error calculation unit 917, an error diffusion unit 918, and an accumulated error memory 919. Further, a ternary pattern image 92 and a ternary pattern memory 921 for storing the gradation values are provided.

[Ternary pattern image settings]
The ternary pattern memory 921 stores in advance a ternary pattern image having a predetermined gradation bg. The image size Wb × Hb pixels of the ternary pattern image is arbitrary. When the ternary pattern image is smaller than the input image, the binary pattern image is repeatedly arranged in a tile shape so that the binary pattern corresponds to each pixel of the input image.

Each pixel of the ternary pattern image takes one of three values of 0, 128, and 255. The ratios (appearance rates) of pixels whose values in the ternary pattern image are 0, 128, and 255 are bg1, bg2, and bg3, respectively. The ratios bg1, bg2, and bg3 are arbitrary values (0 ≦ bg1, bg2, bg3 ≦ 1, real number), but the relationship of the following equation exists.
bg1 + bg2 + bg3 = 1… (17)

Then, bg represented by Expression (18) is used as the gradation value of the ternary pattern image, and is stored in the ternary pattern memory 921 together with the ternary pattern image. That is, the gradation value bg of the ternary pattern image is equal to the sum of the products of the pixel values of the ternary pattern image and the appearance rate of the pixel values.
bg = 0 x bg1 + 128 x bg2 + 255 x bg3
Where 0 ≤ bg ≤ 255, integer ... (18)

  A ternary pattern image is an image in which pixels having values 0, 128, and 255 are arranged in an arbitrary pattern. However, in order to reduce the occurrence of texture, it is preferable that the ternary pattern image has high on-dot arrangement dispersibility and the spatial frequency has a blue noise characteristic. Any method may be used for generating a ternary pattern image having a blue noise characteristic. For example, there is a method in which a 256-level blue noise mask is created using the Bayer-type perturbation systematic dither method disclosed in Japanese Patent Laid-Open No. 2000-059626, and then ternarized using a two-stage threshold.

  In addition to the ternary pattern image and its gradation value bg, the ternary pattern memory 921 stores ternary ratios bg1, bg2, and bg3. Then, the pixel value B and the ratios bg1 to bg3 of the ternary pattern image corresponding to the position of the target pixel indicated by the signal input to the input terminal 116 are output.

[Quantization]
Quantization of the first channel The quantization of the first channel may be performed by replacing the tone value bg of the pattern image with bg1 × 255 in the process shown in FIG. 2 described in the first embodiment.

  FIG. 10 is a flowchart for explaining the processing of the quantization unit 103.

The quantization unit 103 compares the value I ₁ of the target pixel with the normalized value bg1 ′ = bg1 × 255 of the appearance rate bg1 of the pixel having a value of 0 (S1301). The added value I ₁ ′ and the value bg1 ′ may be compared. Subsequently, it is determined whether or not the pixel value of the ternary pattern image is B = 0 (S1302, S1307).

Then, when I ₁ > bg1 ′ and B = 0, the quantization unit 103 sets the output pixel value O ₁ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S1306). .
When I ₁ > bg1 'and B = 0, O ₁ = 255 (19)

In addition, when I ₁ > bg1 ′ and B ≠ 0, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (for example, 128) (S1303), as shown in Expression (20) The output pixel value O ₁ is determined (S1304, S1305).
I ₁ > bg1 'and B ≠ 0
And if I ₁ '<th, O ₁ = 0… (20)
And if I ₁ '≥ th, O ₁ = 255

Further, when I ₁ ≦ bg1 ′ and B ≠ 0, the quantization unit 103 sets the output pixel value O ₁ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S1311). .
When I ₁ ≤ bg1 'and B ≠ 0, O ₁ = 0… (21)

Further, when I ₁ ≦ bg1 ′ and B = 0, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (S1308), and outputs an output pixel value O as shown in Expression (22). ₁ is determined (S1309, S1310).
I ₁ ≤ bg1 'and B = 0
And if I ₁ '<th, O ₁ = 0… (22)
And if I ₁ '≥ th, O ₁ = 255

Second Channel Quantization The second channel quantization may be performed by replacing the tone value bg of the pattern image with bg2 × 255 in the process shown in FIG. 2 described in the first embodiment.

  FIG. 11 is a flowchart for explaining the processing of the quantization unit 111.

Quantization unit 111 compares the value I ₂ of the pixel of interest, the normalized value bg2 '= bg2 × 255 of incidence bg2 value is 128 pixels (S1601). Note that the added value I ₂ ′ may be compared with the value bg2 ′. Subsequently, it is determined whether or not the pixel value of the ternary pattern image is B = 128 (S1602, S1607).

Then, when I ₂ > bg2 ′ and B = 128, the quantization unit 111 sets the output pixel value O ₂ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S1606). .
When I ₂ > bg2 'and B = 128, O ₂ = 255 (23)

In addition, when I ₂ > bg2 ′ and B ≠ 128, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (for example, 128) (S1603), as shown in Expression (24). The output pixel value O ₂ is determined (S1604, S1605).
I ₂ > bg2 'and B ≠ 128
And if I ₂ '<th, O ₂ = 0… (24)
And if I ₂ '≥ th, O ₂ = 255

Further, when I ₂ ≦ bg2 ′ and B ≠ 128, the quantization unit 111 sets the output pixel value O ₂ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S1611) .
When I ₂ ≤ bg2 'and B ≠ 128, O ₂ = 0… (25)

Further, when I ₂ ≦ bg2 ′ and B = 128, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (S1608), and outputs an output pixel value O as shown in Equation (26). ₂ is determined (S1609, S1610).
I ₂ ≤ bg2 'and B = 128
And if I ₂ '<th, O ₂ = 0… (26)
And if I ₂ '≥ th, O ₂ = 255

Quantization of the third channel The quantization of the third channel may be performed by replacing the tone value bg of the pattern image with bg3 × 255 in the process shown in FIG. 2 described in the first embodiment. Note that the third channel accumulated error memory 919 has the same structure as the first and second channel accumulated error memories 107 and 114, as shown in FIG. Therefore, error diffusion may be performed by replacing the subscript “1” or “2” with “3” in the methods shown in the equations (4) and (11).

  FIG. 13 is a flowchart illustrating the processing of the quantization unit 916.

The quantization unit 916 compares the value I ₃ of the pixel of interest with the normalized value bg3 ′ = bg3 × 255 of the appearance rate bg3 of the pixel whose value is 255 (S1701). Note that the added value I ₃ ′ may be compared with the value bg3 ′. Subsequently, it is determined whether or not the pixel value of the ternary pattern image is B = 255 (S1702, S1707).

Then, when I ₃ > bg3 ′ and B = 255, the quantization unit 916 sets the output pixel value O ₃ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₃ ′ (S1706). .
When I ₃ > bg3 'and B = 255, O ₃ = 255 (27)

In addition, when I ₃ > bg3 ′ and B ≠ 255, the quantization unit 916 compares the added value I ₃ ′ with the quantization threshold th (for example, 128) (S1703), as shown in Expression (28). The output pixel value O ₃ is determined (S1704, S1705).
I ₃ > bg3 'and B ≠ 255
And if I ₃ '<th, O ₃ = 0… (28)
And if I ₃ '≧ th, O ₃ = 255

Further, when I ₃ ≦ bg3 ′ and B ≠ 255, the quantization unit 916 sets the output pixel value O ₃ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₃ ′ (S1711). .
When I ₃ ≤ bg3 'and B ≠ 255, O ₃ = 0… (29)

Further, when I ₃ ≦ bg3 ′ and B = 255, the quantization unit 916 compares the added value I ₃ ′ with the quantization threshold th (S1708), and outputs an output pixel value O as shown in Expression (30). ₃ is determined (S1709, S1710).
I ₃ ≤ bg3 'and B = 255
And if I ₃ '<th, O ₃ = 0… (30)
And if I ₃ '≥ th, O ₃ = 255

  Hereinafter, image processing according to the fourth embodiment of the present invention will be described. Note that the same reference numerals in the fourth embodiment denote the same parts as in the first to third embodiments, and a detailed description thereof will be omitted.

  In Example 3, the multi-value pattern image for controlling the dot formation position is commonly used for the quantization of the first to third channels. In the fourth embodiment, an example in which different multi-value pattern images are used for quantization of the first to third channels will be described.

[Device configuration]
FIG. 14 is a block diagram illustrating a configuration example of the image processing apparatus according to the fourth embodiment.

  In addition to the configuration shown in FIG. 8, the image processing apparatus according to the fourth embodiment includes terminals 914 and 920 that input and output third-channel image data, and a configuration that performs quantization processing on the third-channel image data. . The configuration of the quantization process includes an adder 915, a quantization unit 916, an error calculation unit 917, an error diffusion unit 918, an accumulated error memory 919, a binary pattern image, and a binary pattern memory 922 that stores the gradation value thereof, A threshold memory 923.

[Binary pattern image settings]
The binary pattern memories 802, 811 and 922 previously store binary pattern images of predetermined gradations bg1, bg2 and bg3, respectively. The gradation values bg1, bg2, and bg3 have the following relationship, and the gradation values bg1, bg2, and bg3 can take arbitrary values as long as the following relationship is satisfied.
0 ≤ (bg1 + bg2 + bg3) ≤ 255… (31)

  In addition, the binary pattern images stored in the binary pattern memories 802, 811 and 922 need to have an exclusive arrangement of pixels having a pixel value B = 255 (on dots). In other words, the binary pattern image pixels stored in the binary pattern memory 811 and 922 at the same position as the pixel of the binary pattern image pixel value B1 = 255 stored in the binary pattern memory 802 have the pixel value B2 = B3 = 0 needs to be set. Similarly, the pixels of the binary pattern image stored in the binary pattern memories 802 and 922 at the same position as the pixel of the pixel value B2 = 255 (on dot) of the binary pattern image stored in the binary pattern memory 811 are It is necessary that the pixel value B1 = B3 = 0 (off dot).

  If the dot arrangement is exclusive, the binary pattern images stored in the binary pattern memories 802, 811 and 922 may be arbitrary patterns. However, in order to reduce the occurrence of texture, it is preferable that the binary pattern image has high on-dot dispersibility and the spatial frequency has a blue noise characteristic.

[Quantization]
The quantization of the first channel is performed in the process of the quantization unit 103 described in the first embodiment by converting the gradation value bg of the binary pattern image to bg1 and the pixel value B of the binary pattern image to B1. The threshold th may be replaced with th1. Similarly, the quantization of the second channel is performed by changing the gradation value bg of the binary pattern image to bg2 and the pixel value B of the binary pattern image to B2 in the processing of the quantization unit 111 described in the first embodiment. The quantization threshold th may be replaced with th2. Similarly, for the quantization of the third channel, the gradation value bg of the binary pattern image is set to bg3 and the pixel value B of the binary pattern image is set to bg3 in the processing of the quantization unit 103 or 111 described in the first embodiment. In B3, the quantization threshold th may be replaced with th3. Then, the quantization units 103, 113, and 916 all determine the output pixel values O ₁ , O ₂ , and O ₃ by the process shown in FIG. 5 (or FIG. 8).

  Hereinafter, image processing according to the fifth embodiment of the present invention will be described. Note that the same reference numerals in the fifth embodiment denote the same parts as in the first to fourth embodiments, and a detailed description thereof will be omitted.

  In Examples 1 and 3, the pattern image having the number of gradations equal to the number of input channels was used to prevent the overlapping of dots in each channel. In the fifth embodiment, a method of generating a pixel value every time each pixel of an input image is quantized, that is, in units of pixels will be described. Although not a mathematically correct expression, the generated pixel value will be called a random number below.

  In the following, for the sake of simplicity, a case where the input is two channels will be described as in the first embodiment, but the expansion to three or more channels can be performed in the same manner as in the third embodiment. That is, for N-channel input, an N-value random number is generated and used instead of the pixel value of the N-value pattern image.

[Device configuration]
FIG. 15 is a block diagram illustrating a configuration example of the image processing apparatus according to the fifth embodiment. A difference from the configuration shown in FIG. 1 is that a binary random number generator 1202 for generating a binary random number is provided instead of the binary pattern memory 102.

[Generation of binary random numbers]
The random number generated by the binary random number generator 1202 in units of pixels is one of binary values of 0 and 255. The ratio (appearance rate) of the value 255 generated by the binary random number generator 1202 is rg255, and the gradation value of the pixel pattern (binary random number pattern) is rg = 255 × rg255. The gradation value rg can take any value.

  The pixel pattern generated by the random number generator 1202 may be arbitrary, but in order to reduce the generation of texture, it is preferable that the pixel pattern has high dispersibility and the two-dimensional spatial frequency has blue noise characteristics.

  Furthermore, in order to reduce the occurrence of texture, the gradation value rg of the pixel pattern is set to the gradation at which texture is generated by error diffusion (when the input image is 8 bits, 64 gradations, 128 gradations, 192 gradations, etc. ).

[Quantization]
First Channel Quantization FIG. 16 is a flowchart for explaining the operation of the image processing apparatus. A difference from the operation shown in FIG. 2 is that a binary random number R is generated (S403) instead of inputting the pixel value B and the gradation value bg of the binary pattern image (S203). In addition, the binary random number generator 602 outputs the gradation value rg of the binary random number pattern together with the generation of the random number R.

Quantization unit 103
FIG. 17 is a flowchart for explaining the processing of the quantization unit 103.

Quantization unit 103, a value I ₁ of the pixel of interest, compares the tone value rg of the binary random pattern (S1401). The added value I ₁ ′ may be compared with the gradation value rg of the binary random number pattern. Subsequently, it is determined whether or not the binary random number R indicates an on dot (S1402, S1407).

Then, when I ₁ > rg and R = on dot, the quantization unit 103 sets the output pixel value O ₁ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S1406). .
When I ₁ > rg and R = 255, O ₁ = 255… (32)

In addition, when I ₁ > rg and R = off dot, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (for example, 128) (S1403), as shown in Expression (33) The output pixel value O ₁ is determined (S1404, S1405).
I ₁ > rg and R = 0
And if I ₁ '<th, O ₁ = 0… (33)
And if I ₁ '≥ th, O ₁ = 255

Further, when I ₁ ≦ rg and R = off dot, the quantization unit 103 sets the output pixel value O ₁ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₁ ′ (S1411) .
When I ₁ ≤ rg and R = 0, O ₁ = 0… (33)

Further, when I ₁ ≦ rg and R = on dot, the quantization unit 103 compares the added value I ₁ ′ with the quantization threshold th (S1408), and outputs an output pixel value O as shown in Expression (34). ₁ is determined (S1409, S1410).
I ₁ ≤ rg and R = 255
And if I ₁ '<th, O ₁ = 0… (34)
And if I ₁ '≥ th, O ₁ = 255

● Second channel quantization The flow of the second channel quantization process is also shown in FIG.

Quantization unit 111
FIG. 18 is a flowchart for explaining the processing of the quantization unit 111.

The quantization unit 111 compares the value I _{2 of the} pixel of interest with the gradation value rg of the binary random number pattern (S1501). The added value I ₂ ′ may be compared with the gradation value rg of the binary random number pattern. Subsequently, it is determined whether or not the binary random number R indicates an off dot (S1502, S1507).

Then, when I ₂ > rg and R = off dot, the quantization unit 111 sets the output pixel value O ₂ = 255 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S1506) .
When I ₂ > rg and R = 0, O ₂ = 255 (35)

In addition, when I ₂ > rg and R = on dot, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (for example, 128) (S1503), as shown in Expression (36) The output pixel value O ₂ is determined (S1504, S1505).
I ₂ > rg and R = 255
And if I ₂ '<th, O ₂ = 0… (36)
And if I ₂ '≥ th, O ₂ = 255

Further, when I ₂ ≦ rg and R = on dot, the quantization unit 111 sets the output pixel value O ₂ = 0 regardless of the magnitude relationship between the quantization threshold th and the added value I ₂ ′ (S1511). .
When I ₂ ≦ rg and R = 255, O ₂ = 0… (37)

In addition, when I ₂ ≦ rg and R = off dot, the quantization unit 111 compares the added value I ₂ ′ with the quantization threshold th (S1508), and outputs an output pixel value O as shown in Expression (38). ₂ is determined (S1509, S1510).
I ₂ ≤ rg and R = 0
And if I ₂ '<th, O ₂ = 0… (38)
And if I ₂ '≥ th, O ₂ = 255

  The image processing according to the sixth embodiment of the present invention will be described below. Note that the same reference numerals in the sixth embodiment denote the same parts as in the first to fifth embodiments, and a detailed description thereof will be omitted.

[Device configuration]
FIG. 19 is a block diagram illustrating a configuration example of an image processing apparatus according to the sixth embodiment.

  The input terminals 1100 and 1103 are connected to a storage device such as a memory for storing input image data, or an input device such as a scanner or an imaging device. From these devices, image data of the first channel is input to the input terminal 1100 in the raster scan order, and image data of the second channel is input to the input terminal 1103 in the raster scan order. Note that the image data of both channels are input to input terminals 1100 and 1103 in synchronization.

  The output terminals 1102 and 1110 are connected to a storage device or an output device such as a display device or a printing device. The output terminal 1102 outputs binary image data of the first channel, and the output terminal 1110 outputs binary image data of the second channel. Note that the image data of both channels are output from the output terminals 1102 and 1110 after being synchronized by the buffer 1111.

The quantization processing unit 1101 outputs image data O _{1 obtained} by binarizing the image data I ₁ of the first channel by normal error diffusion processing. Note that the pseudo halftone processing method of the quantization processing unit 1101 is not limited and may be a dither method.

On the other hand, the quantization processing unit of the second channel refers to the image data I ₁ and O ₁ of the _first channel and outputs the image data O ₂ binarized from the image data I ₂ of the second channel To do.

  The accumulated error adding unit 1104 adds the error diffusion value stored in the accumulated error memory 1109 to the image data input from the input terminal 1103 on a pixel basis. Although details will be described later, the quantization unit 1105 determines the quantization value of the pixel of interest of the second channel based on the quantization threshold stored in the threshold memory 1106 and the like.

  The error calculation unit 1107 calculates a quantization error from the input / output image data of the quantization unit 1105. The error diffusion unit 1108 updates the error diffusion value stored in the cumulative error memory 1109 so that the quantization error is diffused to a predetermined pixel before quantization.

[Device operation]
FIG. 20 is a flowchart for explaining the operation of the quantization processor of the second channel.

First inputs of the pixel of interest, the input image data I ₁ of the first channel (S1201), the output image data O ₁ of the first channel type (S1202), the second channel input image data I ₂ Is input (S1203).

The cumulative error adding unit 1104 calculates the target pixel obtained by adding the error diffusion value corresponding to the pixel position of the cumulative error memory 1109 to the input pixel data (the target pixel value before quantization) I _{2 according} to Expression (9). The value (added value) I ₂ 'is output (S1204). The error diffusion values stored in the cumulative error memory 1109 are the same as those in FIG.

Next, the quantization unit 1105 determines the output pixel value O ₂ by comparing the pixel values I ₁ , O ₁ , I ₂ , I ₂ ′ of the pixel of interest and the quantization threshold th2 stored in the threshold memory 1106. (S1205). Details of the quantization will be described later.

Next, the error calculation unit 1107 calculates a quantization error E ₂ that is a difference between the added value I ₂ ′ and the output pixel value O ₂ using Equation (10) (S1206).

  Next, the error diffusion unit 1108 performs error diffusion processing, and updates the error diffusion value stored in the cumulative error memory 1109 by using equation (11) (S1207).

  This completes the binarization process for one pixel of the second channel. Then, when the binarization process is performed on all the pixels of the image of the second channel, the process ends, and when there is an unprocessed pixel, the process returns to step S1201, and the above operation is repeated (S1208).

Quantization unit 1105
FIG. 21 is a flowchart for explaining the processing of the quantization unit 1105.

The quantization unit 1105 compares the pixel value I ₁ + I ₂ with the maximum gradation value (255 in this example) of the image data (S2501), and determines whether or not the output value O ₁ of the first channel is off-dot. Is determined (S2502, S2507).

Then, when I ₁ + I ₂ > 255 and O ₁ = off dot, the quantization unit 1105 sets the output pixel value O ₂ = 255 regardless of the magnitude relationship between the quantization threshold th2 and the added value I ₂ ′. (S2508).
When I ₁ + I ₂ > 255 and O ₁ = 0, O ₂ = 255… (39)

In addition, when I ₁ + I ₂ > 255 g and O ₁ = on dot, the quantization unit 1105 compares the added value I ₂ ′ with the quantization threshold th2 (for example, 128) (S2509), and the expression (40) The output pixel value O ₂ is determined as shown in (S2510, S2511).
I ₁ + I ₂ > 255 and O ₁ = 255
And if I ₂ '<th2, O ₂ = 0… (40)
And if I ₂ '≥ th2, O ₂ = 255

In addition, when I ₁ + I ₂ ≦ 255 and O ₁ = on-dot, the quantization unit 1105 sets the output pixel value O ₂ = 0 regardless of the magnitude relationship between the quantization threshold th2 and the added value I ₂ ′. (S2506).
When I ₁ + I ₂ ≤ 255 and O ₁ = 255, O ₂ = 0… (41)

In addition, when I ₁ + I ₂ ≦ 255 and O ₁ = off dot, the quantization unit 1105 compares the added value I ₂ ′ with the quantization threshold th2 (S2503), as shown in Equation (42) The output pixel value O ₂ is determined (S2504, S2505).
I ₁ + I ₂ ≤ 255 and O ₁ = 0
And if I ₂ '<th2, O ₂ = 0… (42)
And if I ₂ '≥ th2, O ₂ = 255

  Thus, when the sum of the pixel values of the first and second channels at the same position is 255 or less, the second channel can form an on dot only when the first channel is an off dot. Can not. Therefore, when the dot recording rates of both the first and second channels are low, it is possible to prevent the on-dots of the two channels from overlapping.

  Hereinafter, image processing according to the seventh embodiment of the present invention will be described. Note that the same reference numerals in the seventh embodiment denote the same parts as in the first to sixth embodiments, and a detailed description thereof will be omitted.

  In the sixth embodiment, the case where the input image has two channels has been described. It is possible to prevent dots from being superimposed on an input image exceeding 3 channels.

[Device configuration]
FIG. 22 is a block diagram illustrating a configuration example of the image processing apparatus according to the seventh embodiment.

  The input terminal 1203 is connected to a storage device such as a memory for storing input image data, or an input device such as a scanner or an imaging apparatus. The image data of the third channel is input to the input terminal 1203 from these devices in the raster scanning order. Note that the image data of all channels are synchronized and input to the input terminals 1100, 1103, and 1203.

  The output terminal 1210 is connected to a storage device or an output device such as a display device or a printing device. The output terminal 1210 outputs binary image data of the third channel. Note that the image data of all channels are output from the output terminals 1102, 1110, and 1210 after being synchronized by the buffers 1111 and 1112.

The quantization processing unit of the third channel refers to the image data I ₁ and O _{1 of the first} channel and the image data I ₂ and O ₂ of the _second channel, and the image data of the third channel Image data O _{3 obtained} by binarizing I ₃ is output.

  The accumulated error adding unit 1204 adds the error diffusion value stored in the accumulated error memory 1209 to the image data input from the input terminal 1203 for each pixel. Although details will be described later, the quantization unit 1205 determines the quantization value of the target pixel of the third channel based on the quantization threshold stored in the threshold memory 1206 and the like.

  The error calculation unit 1207 calculates a quantization error from the input / output image data of the quantization unit 1205. The error diffusion unit 1208 updates the error diffusion value stored in the accumulated error memory 1209 so that the quantization error is diffused to a predetermined pixel before quantization.

[Device operation]
FIG. 23 is a flowchart for explaining the operation of the quantization processing unit of the third channel. The first and second channel quantization processing operations are the same as those in the sixth embodiment.

First, the target pixel, the input image data I ₁ of the first channel type (S1211), receives the output image data O ₁ of the first channel (S1212). Furthermore, by entering the input image data I ₂ of the second channel (S1213), receives the output image data O ₂ of the second channel (S1214). Then, enter the input image data I ₃ of the third channel (S1215).

The cumulative error adding unit 1204 calculates the target pixel obtained by adding the error diffusion value corresponding to the pixel position of the cumulative error memory 1209 to the input pixel data (the target pixel value before quantization) I _{3 according} to Expression (9). The value (added value) I ₃ 'is output (S1216). The error diffusion values stored in the cumulative error memory 1209 are the same as those in FIG.

Next, the quantization unit 1205 compares the pixel values I ₁ , O ₁ , I ₂ , O ₂ , I ₃ , I ₃ ′ of the pixel of interest and the quantization threshold th3 stored in the threshold memory 1206 and outputs the result determining the pixel value O ₃ (S1217). Details of the quantization will be described later.

Next, the error calculation unit 1207 calculates a quantization error E ₃ that is a difference between the added value I ₃ ′ and the output pixel value O ₃ by Expression (10) (S1218).

  Next, the error diffusion unit 1208 performs error diffusion processing, and updates the error diffusion value stored in the cumulative error memory 1209 by using equation (11) (S1219).

  This completes the binarization processing for one pixel of the third channel. If the binarization process is performed on all the pixels of the third channel image, the process ends. If there is an unprocessed pixel, the process returns to step S1211 to repeat the above operation (S1220).

Quantization unit 1205
FIG. 24 is a flowchart illustrating the processing of the quantization unit 1205.

The quantization unit 1205 compares the value I ₁ + I ₂ + I ₃ of the target pixel with the maximum gradation value (255 in this example) of the image data (S901), and outputs the output values O of the first and second channels. _It is determined whether ₁ and O ₂ are off dots (S902, S907).

Then, when I ₁ + I ₂ + I ₃ > 255 and O ₁ = O ₂ = off dot, the quantization unit 1205 outputs the output pixel regardless of the magnitude relationship between the quantization threshold th3 and the added value I ₃ ′. The value O ₃ is set to 255 (S908).
When I ₁ + I ₂ + I ₃ > 255 and O ₁ = O ₂ = 0, O ₃ = 255… (43)

Further, when I ₁ + I ₂ + I ₃ > 255 g and O ₁ = O ₂ = on dot, the quantization unit 1205 compares the added value I ₃ ′ with the quantization threshold th3 (for example, 128) (S2509). ), The output pixel value O ₃ is determined as shown in the equation (44) (S910, S911).
I ₁ + I ₂ + I ₃ > 255 and O ₁ = O ₂ = 255
And if I ₃ '<th3, O ₃ = 0… (44)
And if I ₃ '≥ th3, O ₃ = 255

In addition, when I ₁ + I ₂ + I ₃ ≦ 255 and O ₁ = O ₂ = on-dot, the quantization unit 1205 outputs the output pixel regardless of the magnitude relationship between the quantization threshold th3 and the added value I ₃ ′. The value O ₃ is set to 0 (S906).
When I ₁ + I ₂ + I ₃ ≤ 255 and O ₁ = O ₂ = 255, O ₃ = 0… (45)

In addition, when I ₁ + I ₂ + I ₃ ≦ 255 and O ₁ = O ₂ = off dot, the quantization unit 1205 compares the added value I ₃ ′ with the quantization threshold th3 (S903), and formula ( As shown in 46), the output pixel value O ₃ is determined (S904, S905).
I ₁ + I ₂ + I ₃ ≤ 255 and O ₁ = O ₂ = 0
And if I ₃ '<th3, O ₃ = 0… (46)
And if I ₃ '≥ th3, O ₃ = 255

  Thus, when the sum of the pixel values of the first, second, and third channels at the same position is 255 or less, the second channel forms on dots only when the first channel is off dots. Can not do it. Furthermore, the third channel can form on dots only when the first and second channels are off dots. Therefore, when all the dot recording rates of the first, second, and third channels are low, it is possible to prevent the on-dots of the three channels from overlapping.

[Other embodiments]
Note that the present invention can be applied to a system composed of a plurality of devices (for example, a computer, an interface device, a reader, a printer, etc.), or a device (for example, a copier, a facsimile machine, a control device) composed of a single device. Etc.).

  Another object of the present invention is to supply a storage medium storing a computer program for realizing the functions of the above-described embodiments to a system or apparatus, and the computer (CPU or MPU) of the system or apparatus executes the computer program. But it is achieved. In this case, the software read from the storage medium itself realizes the functions of the above embodiments, and the computer program and the computer-readable storage medium storing the computer program constitute the present invention. .

  Further, the above functions are not only realized by the execution of the computer program. That is, it includes a case where an operating system (OS) or the like running on the computer performs part or all of the actual processing according to an instruction of the computer program, thereby realizing the above functions.

  The computer program may be written in a function expansion card connected to the computer or a memory of the unit. That is, the case where the above functions are realized by the CPU or the CPU of the unit performing part or all of the actual processing in accordance with the instruction of the computer program.

  When the present invention is applied to the storage medium, the storage medium stores a computer program corresponding to or related to the flowchart described above.

Block diagram showing a configuration example of the image processing apparatus of Example 1, A flowchart for explaining the quantization process; Diagram showing scanning of input image, The figure explaining the error diffusion value which a cumulative error memory stores, A flowchart for explaining the processing of the quantization unit; The figure explaining the error diffusion value which a cumulative error memory stores, A flowchart for explaining the processing of the quantization unit; Block diagram showing a configuration example of an image processing apparatus of Example 2, Block diagram showing a configuration example of an image processing apparatus of Example 3, A flowchart for explaining the processing of the quantization unit; A flowchart for explaining the processing of the quantization unit; The figure explaining the error diffusion value which a cumulative error memory stores, A flowchart for explaining the processing of the quantization unit; Block diagram showing a configuration example of an image processing apparatus of Example 4, Block diagram showing a configuration example of an image processing apparatus of Example 5, A flowchart for explaining the operation of the image processing apparatus; A flowchart for explaining the processing of the quantization unit; A flowchart for explaining the processing of the quantization unit; Block diagram showing a configuration example of an image processing apparatus of Example 6, A flowchart for explaining the operation of the quantization processor of the second channel; A flowchart for explaining the processing of the quantization unit 1105; Block diagram showing a configuration example of an image processing device of Example 7, A flowchart for explaining the operation of the quantization processing unit of the third channel; 10 is a flowchart for describing processing of a quantization unit 1205.

Claims (6)

An image processing apparatus having a holding unit that holds a binary pattern image having a predetermined average gradation value and a blue noise characteristic, and a binarization unit that binarizes image data of two channels, respectively. The binarization means includes
An error memory for storing error diffusion values;
Adding means for outputting an addition value obtained by adding the error diffusion value and the pixel value of the target pixel;
A computing means for computing an error between the added value and the binarized value of the target pixel;
Diffusion means for updating an error diffusion value stored in the error memory so as to diffuse the error to pixels before binarization;
Based on the comparison between the pixel value of the target pixel and the average gradation value of the pattern image, the pixel value of the pattern image corresponding to the position of the target pixel, and the comparison of the addition value and the quantization threshold Quantizing means for determining the binarized value of
The pattern image indicates a determination condition of a binarization value indicating dot arrangement in the two channels , and the binarization means refers to the pattern image and the determination conditions of the two channels are opposite to each other . the image processing apparatus according to claim that you control as.   2. The image processing apparatus according to claim 1, wherein the average gradation value is an intermediate gradation value at which texture is likely to occur due to error diffusion. It said holding means, the image processing apparatus according to claim 1 or claim 2, characterized in that to hold the pattern image of the exclusive on-dot each other as a pattern image corresponding to each of the two channels.   4. The image processing apparatus according to claim 1, wherein the average gradation value is proportional to an on-dot appearance rate of the pattern image. Holding means for holding a binary pattern image having a predetermined average gradation value and blue noise characteristics, an error memory for storing an error diffusion value, and a binarizing means for binarizing the image data of two channels respectively An image processing method for an image processing apparatus having the binarization means,
Calculating an added value obtained by adding the error diffusion value and the pixel value of the target pixel;
Calculating an error between the added value and the binarized value of the target pixel;
Update the error diffusion value stored in the error memory so that the error is diffused to the pixel before binarization,
Based on the comparison between the pixel value of the target pixel and the average gradation value of the pattern image, the pixel value of the pattern image corresponding to the position of the target pixel, and the comparison of the addition value and the quantization threshold Process to determine the binarized value of
The pattern image indicates a determination condition of a binarization value indicating dot arrangement in the two channels , and the binarization means refers to the pattern image and the determination conditions of the two channels are opposite to each other . an image processing method characterized that you control as. A program for causing a computer to function as each unit of the image processing apparatus according to any one of claims 1 to 4.

Images