JP2000115557A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To faithfully reproduce the feature of an inputted picture to be satisfactory to see. SOLUTION: A feature variable calculating circuit 41 executes orthogonal transforming arithmetic in the block of the prescribed number of picture elements of an inputted picture signal Si to calculate feature values α1 to α3 showing the spatial features of the inputted picture in the block from the low frequency component and a high frequency component of a spatial frequency in the block. At the same time, the circuit 41 executes histogram arithmetic in a specific color space area concerning the inputted picture signal in the block to calculate feature values β1 and β2 showing the features of the color distribution of the inputted picture in the block. A discrimination circuit 43 compares the feature values α1 to α3 and β1 and β2 with a threshold to decide feature values x1 and x2. Plural halftone generating matrix patterns different from each other are provided in a halftone generation circuit 51. A selection circuit 53 selects optimum one of the plural halftone generating matrix patterns according to the combination of the feature values x1 and x2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus capable of always outputting a visually pleasing image by performing optimal image processing according to the type of an input color image.

[0002]

2. Description of the Related Art When considering the image quality of output images such as full-color copies and prints, color reproducibility, gradation reproducibility,
Sharpness and granularity are mentioned as important characteristics,
Until now, many image processing techniques for faithfully reproducing a document have been devised. Recently, with the widespread use of network computers and desktop publishing, the types of image information input to full-color copy systems and print systems have also been diverse, and as a result, optimal image quality reproduction according to the type of image information has been achieved. Are required to be performed at the same time.
Therefore, recently, attention has been paid to an adaptive image processing technique for performing optimal image processing according to the type of an input color image.

[0003] As a typical example thereof, JP-A-64-43
No. 69 discloses a technique for processing a pattern and a character by separating and processing the pattern and the character. This is done by determining whether the input image signal is a continuous tone signal including a halftone such as a picture or a photograph, or a binary or nearly binary signal such as a black character or a color character. A method of performing image processing,
Even if an image signal in which picture information and character information are mixed is input, it is determined and an optimum image processing is performed for each of them.

[0004] In particular, in Japanese Patent Application Laid-Open No. Hei 4-212283, it is possible to accurately separate a pattern from a line drawing by determining whether the input image is a halftone image or a line drawing based on a high frequency component of the input image. Thus, the input image can be processed under optimal image processing conditions.

However, any of the above-mentioned methods can be applied to an extreme case of a continuous tone signal or a character signal. For example, the method can be applied to a subtle image change in a picture in the same image information. Can not respond. In addition, these methods focus on separating pattern information and character information with high accuracy,
If there is a pattern similar to a character in the picture, the processing for the letter may be erroneously performed.If a pattern similar to the letter exists in the picture, this is properly reproduced. You may not be able to.

As another example, a method has been considered in which the characteristics of a color space region of an input image signal are calculated and the color correction processing is changed. This is because, for example, in the vicinity of achromatic colors, the amount of under color removal is increased to enhance the reproducibility of gray, and in the vicinity of the color gamut limit, when an image signal that cannot be reproduced is input, a color that is as inconspicuous as possible is minimized. Correction.

As an example of improving sharpness, Japanese Patent Laid-Open No. 5-130404 discloses that processing of a spatial frequency is changed in accordance with the range of the density level of an input image signal. However, this method is for compensating for poor reproducibility of the highlight portion and the shadow portion of the image output device, and does not perform optimal image reproduction according to the characteristics of the input image.

Further, JP-A-8-298596 discloses that
It is shown that the spatial frequency components of an input image signal are compared to perform optimal spatial frequency processing. As an example of a compression technique, Japanese Patent Application Laid-Open No. 5-95484 discloses that the spatial frequency components of an input image signal are compared. Then, it is shown that an optimum quantization method is selected.

[0009]

In many image forming apparatuses, halftones are reproduced by an area modulation method. In such an image forming apparatus, after performing various image processes,
Finally, since the image signal is converted into a binary or numerical pattern and output, color reproducibility and sharpness are greatly influenced by the area modulation method. Therefore, in the area modulation method, it is necessary to perform image processing in consideration of the characteristics of the input image. In other words, according to the characteristics of the input image, the halftoning method, which is one of the factors that most influence the quality of image quality. It is necessary to choose.

[0010] In this respect, the conventional methods described above only switch the half-toning method depending on whether the image is a pattern or a character, and perform halftone reproduction in accordance with the characteristics of the pattern of the input image. The halftoning method is not selected, and the conventional method cannot realize halftone reproduction according to the characteristics of the pattern of the input image.

In view of the above, the present invention is intended to reproduce the characteristics of an input image more faithfully and more visually.

[0012]

According to the present invention, a feature value calculating means for calculating a feature value of an input color image signal in a plurality of pixel blocks, and a feature value calculated by the feature value calculating means are set to a predetermined threshold value. Comparing, determining means for determining the type of the input color image signal in the block, a plurality of halftone generating means for generating mutually different halftone patterns from the input color image signal, and according to the determination result of the determining means Selecting means for selecting one of the plurality of halftone generating means, and converting the input color image signal in the block into an output color image signal by the halftone generating means selected by the selecting means. It shall be.

[0013]

According to the image processing apparatus of the present invention, the input color image signal in the plurality of pixel blocks has a large number of low frequency components, a large number of high frequency components, or a highlight. Gradation reproducibility, sharpness, etc., depending on the characteristics of the input color image signal in the multi-pixel block, such as whether it is an area, a shadow area, a warm hue, or a cool hue. Different halftone patterns, such as character or graininess, are selected. Therefore, the features of the input image are reproduced more faithfully and more visually.

[0014]

FIG. 1 shows a basic configuration of an image output system provided with an image processing apparatus according to the present invention. This image output system includes an image input device 10, an image output device 2
0, and the image processing device 30.

The image input device 10 optically reads a color image, such as a digital camera or an image scanner, and separates the color image into three colors of BGR (blue, green, red). Each color component is 400 spi, 8 bits. It converts the input image signal Si into digital data.

The image output device 20 uses an output image signal So obtained from the image processing device 30 as described later,
An image is formed on a recording medium such as paper by reproducing halftones by an area modulation method. For example, an electrophotographic method that forms an image using YMCK (yellow, magenta, cyan, and black) toners is used. is there.

FIG. 1 shows a basic configuration of the image processing apparatus 30. The image processing apparatus 30 analyzes a characteristic of an input image signal Si from the image input apparatus 10, and analyzes the characteristic by the characteristic analyzing section 40. An optimum image processing unit that performs optimal image processing on the input image signal Si in accordance with the characteristics of the input image signal Si and generates an output image signal So.

The feature analyzing section 40 includes a feature amount calculating circuit 41 and a discriminating circuit 43.
Then, an orthogonal transformation operation is performed within a block of a predetermined number of pixels including the target pixel of the input image signal Si and its surrounding pixels,
From the low-frequency component and the high-frequency component of the spatial frequency in the block, feature amounts α1, α2, α3 indicating the spatial characteristics of the input image in the block are calculated.

At the same time, the feature value calculation circuit 41 performs a histogram operation on the input image signal in the block in a specific color space area, and obtains the feature values β1 and β2 indicating the characteristics of the color distribution of the input image in the block. Is calculated.

The discriminating circuit 43 of the feature analyzing unit 40 calculates the feature amounts α1, α2, α calculated by the feature amount calculating circuit 41.
3 and β1, β2, the characteristics of the input image in the block are determined, and the characteristic amounts x1 and x2 of the determination result are output.

The optimum processing unit 50 includes a halftone generation circuit 51 and a selection circuit 53.
A plurality of halftone generation matrix patterns for generating a halftone pattern corresponding to a combination of the characteristic amounts x1 and x2 of the determination result of the determination circuit 43 are provided in advance. For example, two types of halftone generation matrix patterns are prepared so that two types of output image signals So with different sharpness can be obtained.

The selection circuit 53 of the optimum processing section 50 selects one of a plurality of halftone generation matrix patterns provided in the halftone generation circuit 51 in accordance with the combination of the characteristic quantities x1 and x2 of the determination result of the determination circuit 43. Choose the best one. Then, the halftone generation circuit 51 generates a halftone using the selected halftone generation matrix pattern and forms an output image signal So.

For example, as a feature in a certain pattern, when the low frequency component of the spatial frequency is large and the high frequency component is small, the halftone generation matrix pattern with low sharpness is selected, so that the output image signal So is obtained. When a low sharpness is formed and the high frequency component is large and the low frequency component is small, a halftone generation matrix pattern with high sharpness is selected, so that the output image signal So has high sharpness. Things are formed.

As described above, by selecting the halftone generation matrix pattern in accordance with the spatial characteristics in a predetermined block of the input image, whatever the input image is, the input image is output as the output image signal So. That has good sharpness faithful to the above.

FIG. 2 shows a specific example of the image processing device 30. The image processing device 30 of this example includes a gradation conversion unit 60 that performs gradation conversion of an input image signal Si of BGR data from the image input device 10, and an image signal ENL * B after the gradation conversion.
The color conversion unit 70 performs color conversion so as to separate the GR into a lightness signal L * and chromaticity signals a * and b *, and the feature analysis unit 40 and the optimal processing unit 50 shown in FIG. The asterisk * should be superscripted, but not in the specification.

The feature amount calculation circuit 41 of the feature analysis unit 40
The characteristic amounts α1, α2, α3 and β1, β2 are calculated from the lightness / chromaticity signals L * a * b * from the color conversion unit 70, and the determination circuit 43 determines the characteristic amounts α1, α2, α3 and β1,
From β2, the feature quantities x1 and x2 of the discrimination result are output.

In this example, the optimum processing unit 50 includes a subtractive color mixture conversion circuit 55 and a UCR circuit 57 in addition to the halftone generation circuit 51 and the selection circuit 53 shown in FIG. Reference numeral 55 denotes a lightness / chromaticity signal L * a * b * from the color conversion unit 70 when a characteristic is selected by the selection circuit 53 in accordance with a combination of the characteristic amounts x1 and x2 of the determination result of the determination circuit 43. Performs subtractive color conversion according to the features in the block of the input image signal Si,
The UCR circuit 57 similarly removes the subtractive color mixture image signal ENL * YMC from the subtraction color mixture conversion circuit 55 according to the characteristics of the block of the input image signal Si according to the characteristics selected by the selection circuit 53. I do.

The halftone generation circuit 51 selects the halftone generation matrix pattern by the selection circuit 53 in accordance with the combination of the characteristic quantities x1 and x2 of the determination result of the determination circuit 43. Is converted into the coverage image signal Y in accordance with the features of the input image signal Si in the block.
Convert to MCK. This coverage image signal YMCK
Is sent to the image output device 20 as an output image signal So.

Specifically, the gradation conversion section 60 is constituted by a table for converting the input image signal Si of BGR data into equivalent brightness. In this example, the color conversion unit 70 converts the image signal ENL * BGR converted to the equivalent brightness into a 3 × 1
The matrix is converted to a CIEL * a * b * color space, which is one of the uniform color spaces, by a matrix operation of 0.

The feature value calculating circuit 41 calculates one or both of the spatial frequency feature values α1, α2, α3 and the color distribution feature values β1, β2, and calculates the spatial frequency feature values α1, α2, α3. For example, two-dimensional Fourier transform is used for the calculation, and for example, histogram calculation is used for calculating the color distribution feature amounts β1 and β2.

FIG. 3 shows an example of a portion of the feature value calculation circuit 41 for calculating the spatial frequency feature values α1, α2, α3.
First, the brightness signal L * of a block of 64 × 64 pixels including the target pixel and its surrounding pixels is stored in the two-dimensional buffer memory 4.
11 and then stores the stored brightness signal L * of the 64 × 64 pixel block in the two-dimensional Fourier transform circuit 41.
2 to perform a two-dimensional Fourier transform. In this case, an original that cannot be subjected to two-dimensional Fourier transform (input image for one page)
The processing for 32 pixels in the peripheral portion is based on the calculation result of the inner block adjacent to the 32 pixels.

Next, from the signal after the two-dimensional Fourier transform,
A first spatial frequency band masking circuit 413 extracts a power spectrum component in a low / middle frequency range, and a second spatial frequency band masking circuit 414 extracts a power spectrum component in a high frequency range.

Then, the first spatial frequency band masking circuit 4
The average value α1 of the power spectrum components of the low and middle frequency components from the signal 13 is calculated, and the average value α1 is sent to the spatial frequency feature amount discrimination circuit 431.

On the other hand, the power spectrum component of the high frequency component from the second spatial frequency band masking circuit 414 is separated into a regular component and an irregular component by a regular component The average value α2 of the rule component is calculated by the value calculation circuit 417.
The power spectrum irregular component average value calculation circuit 418
The average value α3 of the irregular component is calculated by
The respective average values α2 and α3 are sent to the spatial frequency feature value discrimination circuit 431.

Then, the spatial frequency characteristic amount discriminating circuit 431
Then, based on the respective average values α1, α2, α3,
As will be described later, a feature amount x1 as a determination result is determined.

FIG. 6A shows an example of the spatial frequency characteristic of the low and medium frequency component extraction mask used in the first spatial frequency band masking circuit 413, and FIG. 6B shows the second spatial frequency band masking circuit. An example of the spatial frequency characteristic of the high frequency component extraction mask used in 414 is shown. In this example,
The low / medium frequency component extraction mask passes a spatial frequency band of 1 to 5 (lp / mm), and the high frequency component extraction mask passes a spatial frequency band of 6 to 8 (lp / mm). Things.

FIG. 4 shows an example of a portion for calculating the color distribution feature values β1 and β2 of the feature value calculation circuit 41 of FIG. First, the lightness signal L * and the chromaticity signals a * and b * of a block of 64 × 64 pixels including the target pixel and its surrounding pixels are
The chromaticity signals a *, b of the stored 64 × 64 pixel block are stored in the two-dimensional buffer memory 421 by the first color space area histogram calculation circuit 423 and the second color space area histogram calculation circuit 424. * And the brightness distribution value β1 of the brightness signal L * in the first color space area and the second color space area set in advance.
And β2 are calculated, and respective distribution values β1 and β2
Is sent to the color distribution feature value discriminating circuit 432.

Then, the color distribution feature value discrimination circuit 432 determines a feature value x2 as a discrimination result based on the respective distribution values β1 and β2, as described later.

The spatial frequency feature value discriminating circuit 431 in FIG. 3 and the color distribution feature value discriminating circuit 432 in FIG. 4 determine the feature values x1 and x2 based on the average values α1, α2, α3 and the distribution values β1, β2, respectively. In order to do so, tables T1 and T2 for discriminating the characteristic amounts as shown in FIG. 7 are used.

The feature amount discriminating tables T1 and T2
Have previously determined mean values α1, α2, α3 and distribution values β1, β2, respectively, with a predetermined threshold s
The values of the feature values x1 and x2 to be determined are described according to the relationship between 1,1, s2, s3 and t1, t2.

Then, the spatial frequency feature value discriminating circuit 431
In the color distribution characteristic amount discriminating circuit 432, the average values α1, α2, α3 calculated by the spatial frequency characteristic amount calculating circuit 410 in FIG. 3 and the distribution values calculated by the color distribution characteristic amount calculating circuit 420 in FIG. β1, β2,
The thresholds s1, s2, s3 and t1, t2 are compared, and based on the result, the feature amounts x1 and x2 are determined with reference to the feature amount determination tables T1 and T2.

For example, α1 <s1, α2 <s2, and α
If 3 <s3, it is determined that x1 = 1. In this case, the low-medium frequency component of the spatial frequency is small, and the high-frequency component is also small in both the regular component and the irregular component, and the block of the input image signal Si is data read from a general photograph or a high-resolution printed matter. Or a picture pattern having a smooth, almost flat spatial feature corresponding thereto.

Further, α1> s1, α2 <s2, and α3
In the case of <s3, it is determined that x1 = 2. in this case,
In the case where the low-medium frequency component of the spatial frequency is large and the high-frequency component is small in both the regular component and the irregular component, the input image signal S
This indicates that the block i is data read from an enlarged photograph or the like, or a corresponding pattern pattern having a spatial characteristic with a density change in a large cycle.

Further, α1 <s1, α2> s2, α3 <
In the case of s3, it is determined that x1 = 3. In this case, when the low-medium frequency component and the high frequency irregular component of the spatial frequency are small and the high frequency regular component is large, the input image signal S
The block i is data read from a general printed matter of about 175 lines or the like, or
This indicates that the pattern pattern is relatively small and regular.

Α1> s1, α2> s2, and α3
In the case of <s3, it is determined that x1 = 4. in this case,
When the low-medium frequency component and the high frequency regular component of the spatial frequency are large and the high frequency irregular component is small, the block of the input image signal Si is data read from a low-quality print or the like, or is equivalent thereto. It is a relatively large regular pattern pattern.

Further, α1> s1, α2> s2, and α3
If> s3, it is determined that x1 = 4. in this case,
In the case where there are many low and medium frequency components of the spatial frequency, and there are also many high frequency components of both regular and irregular components, the input image signal Si
Indicates that the pattern is data read from general xerography or the like, or a corresponding pattern pattern having spatial characteristics with relatively large regular and irregular density fluctuations. .

The average values α1, α2, α3 and the threshold values s1, s
Similarly, the characteristic amount x1 is determined according to the characteristic amount discrimination table T1 in the other three cases with respect to the relationship with 2, s3. The threshold values s1, s2, and s3 are arbitrary constants, and can be appropriately changed according to the environment and conditions to be used.

In this example, the spatial frequency characteristic amount calculating circuit 410 and the spatial frequency characteristic amount determining circuit 431 shown in FIG.
, The spatial frequency band of the input image signal Si is divided into a low-medium frequency region and a high-frequency region, and the high-frequency component is further divided into a regular component and an irregular component.
In this case, the average value α1, α2, α3 is calculated to determine the feature amount x1, but the present invention is not limited to this, and the spatial frequency band of the input image signal Si is further subdivided to obtain more average values. A value may be calculated to determine the feature amount x1.

On the other hand, regarding the determination of the characteristic amount x2 by the characteristic amount determination table T2, the first color space area histogram calculating circuit 423 of the color distribution characteristic amount calculating circuit 420 in FIG.
In order to determine whether the block of the input image signal Si is a warm color hue or a cool color hue, the chromaticity signals a * and b * are subjected to the histogram operation in the specific color space area as described above, and the distribution value is determined. β1 is calculated, and the second color space area histogram calculation circuit 424 determines whether the block of the input image signal Si is a highlight area, a mid area or a shadow area, as described above. Histogram calculation in the spatial domain is performed to obtain the distribution value β
2 is calculated.

Here, the first color space area and the second color space area used in the first color space area histogram calculation circuit 423 and the second color space area histogram calculation circuit 424 are the color reproduction areas of the image output device 20 used. It is set as appropriate according to the characteristics.

Then, for example, β1 <t1 and β2 <
In the case of t2, it is determined that x2 = 1. In this case, it is indicated that the block of the input image signal Si is a picture composed of a warm color highlight area.

When β1> t1 and β2 <t2, it is determined that x2 = 2. In this case, it is indicated that the block of the input image signal Si is a picture composed of a highlight region of a cool color system.

Further, when β1 <t1 and β2> t2, x2 = 3 is determined. In this case, it is indicated that the block of the input image signal Si is a picture composed of a warm color mid area or a shadow area.

If β1> t1 and β2> t2, it is determined that x2 = 4. In this case, it is indicated that the block of the input image signal Si is a picture composed of a cool color mid area or a shadow area.

The threshold values t1 and t2 are also arbitrary constants.
It can be changed as appropriate according to the environment and conditions used.

In the selection circuit 53 shown in FIG.
That is, in the above example, three halftone generations prepared in advance according to the combination of the feature amounts x1 and x2 determined by the spatial frequency feature amount determination circuit 431 of FIG. 3 and the color distribution feature amount determination circuit 432 of FIG. In order to select one of the matrix patterns, FIG.
A pattern selection table T3 as shown in FIG.

In this pattern selection table T3, any one of the first, second and third halftone generation matrix patterns p1, p2 and p3 is selected in advance in accordance with the combination of the feature amounts x1 and x2. Describe what to do.

In this example, when x1 = 1, 2 and x2 = 1, and when x1 = 3, 4 and x2 = 1, the first
Is selected, and x1
= 1 and x2 = 3,4, etc., the second halftone generation matrix pattern p2 is selected, x1 = 3,4 and x2 = 2, x1 = 3,4 and x2 = 4 For example, the third halftone generation matrix pattern p3 is set to be selected. Then, in the selection circuit 53,
The halftone generation matrix pattern is selected according to the pattern selection table T3.

As shown in FIG. 5, the optimum processing unit 50 shown in FIG.
Subtraction color conversion circuits 551 and 55 corresponding to p2 and p3
2, 553, UCR circuits 571, 572, 573, and halftone generation circuits 511, 512, 513, and the brightness chromaticity signal L * a * b * from the color conversion unit 70 is supplied to the selection circuit 53 as described above. The output image signal So output to the image output device 20 is obtained by being processed by a system of a subtractive color mixture conversion circuit, a UCR circuit and a halftone generation circuit according to the halftone generation matrix pattern selected by the above.

FIG. 9 shows an example of a halftone generation matrix pattern p1, p2, p3. The simplest example is 4
It shows a matrix pattern of × 4 cells, and the features of each are as follows.

FIG. 9A shows a matrix pattern p1.
Although the sharpness is low as a whole, the gradation is reproduced smoothly, so the input image is sharp and the S / N is good.
In addition, when the reproducibility of the image output device is high, it can be said that the matrix pattern is particularly suitable for reproducing the highlight area.

FIG. 7B shows a matrix pattern p2.
The overall reproducibility of the details is not very high and the details are reproduced firmly.However, the contrast is emphasized and reproduced accordingly, so the input image from the mid area to the shadow area and the image output device It can be said that this is a matrix pattern suitable for reproducing a highlight region when reproducibility is low.

FIG. 9C shows a matrix pattern p3.
And has an intermediate feature between the above two matrix patterns, and is used for reproducing an input image that does not belong to the two matrix patterns or an input image of a joint region where the two matrix patterns are switched.

The above-described example is a case where three types of halftone generation matrix patterns having different sharpness and contrast are used. However, the present invention is not limited to this. Alternatively, the halftone generation matrix pattern in the above-described example may be replaced with another halftone generation matrix pattern having a different image reproduction characteristic.

[0065]

As described above, according to the image processing apparatus of the present invention, the characteristics of the input image can be reproduced more faithfully and more visually.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic configuration of an image processing apparatus according to the present invention.

FIG. 2 is a diagram showing an embodiment of the image processing apparatus of the present invention.

FIG. 3 is a diagram illustrating an example of a spatial frequency feature amount calculation circuit.

FIG. 4 is a diagram illustrating an example of a color distribution feature amount calculation circuit.

FIG. 5 is a diagram illustrating an example of an optimal processing unit.

FIG. 6 is a diagram illustrating an example of the spatial frequency characteristics of a mask for extracting low and middle frequency components and a mask for extracting high frequency components.

FIG. 7 is a diagram illustrating an example of a feature amount determination table.

FIG. 8 is a diagram illustrating an example of a pattern selection table.

FIG. 9 is a diagram illustrating an example of a plurality of halftone generation matrix patterns and an output image based on each of the halftone generation matrix patterns.

[Description of Signs] 10 ... Image input device 20 ... Image output device 30 ... Image processing device 40 ... Feature analysis unit 41 ... Feature amount calculation circuit 43 ... Discrimination circuit 50 ... Optimal processing unit 51 ... Midtone generation circuit 53 ... Selection circuit 60: gradation conversion unit 70: color conversion unit

Continued on the front page F-term (reference) 5B057 CA01 CA08 CA16 CB01 CB07 CB16 CE13 CG05 CH18 DC19 DC25 5C077 LL19 MP08 NN04 NN19 PP32 PP33 PP36 PP43 PP49 PQ19 5C079 LA02 LA31 LC02 LC11 NA01 NA07 5L096 AA23 AA35 FA9 FA9

Claims (6)

[Claims] 1. A feature value calculating means for calculating a feature value of an input color image signal in a plurality of pixel blocks, and a feature value calculated by the feature value calculating means is compared with a predetermined threshold value. Determining means for determining the type of the input color image signal; a plurality of halftone generating means for generating mutually different halftone patterns from the input color image signal; and the plurality of halftone generating means according to the determination result of the determining means Selecting means for selecting one of the blocks, wherein the halftone generating means selected by the selecting means converts an input color image signal in the block into an output color image signal. apparatus. 2. The image processing apparatus according to claim 1, wherein the characteristic amount calculating means calculates the characteristic amount based on a spatial frequency component in at least one spatial frequency band of the input color image signal in the block. An image processing apparatus comprising: 3. The image processing apparatus according to claim 1, wherein the characteristic amount calculating means calculates the characteristic amount based on a frequency of a histogram in at least one color space region of the input color image signal in the block. An image processing apparatus comprising: 4. The image processing apparatus according to claim 1, wherein said plurality of halftone generating means generates at least two types of halftone patterns having different tone reproducibility. Image processing apparatus. 5. The image processing apparatus according to claim 1, wherein said plurality of halftone generating means generates at least two types of halftone patterns having different sharpness. Image processing device. 6. The image processing apparatus according to claim 1, wherein said plurality of halftone generating means generates at least two types of halftone patterns having different graininess. Image processing device.