JPH1074248A - Color reduction device and color reduction method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform the color reduction processing of pictures while leaving a colorful part by referring to a lookup table and allocating the color of respective picture elements for constituting an input picture to one of limited colors. SOLUTION: A classification means 101 uses a classification rule 102 for classifying the respective picture elements of the input picture corresponding to the value of the color and classifies the respective picture elements of the input picture by one of the plural attribute values. A color number counting means 103 classifies the respective picture elements for constituting the input picture by the attribute value in the classification means 101 and calculates a color number to be allocated to a picture element group constituted of the picture elements of the same attribute value for the respective attribute values based on the classified result. Also, a lookup table calculation means 104 inputs the picture, calculates color conversion information and stores the calculated result in the lookup table 105. Then, a color reduction means 106 refers to the lookup table 105 and allocates the color of the respective picture elements for constituting the input picture to one of the limited colors.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is intended to reduce the number of colors of a full-color image output by an image output means such as a display means or a printing means, and to maintain the texture of an original image while eliminating realism. The present invention relates to a color reduction device for creating an illustration-like color image and a color reduction method thereof.

[0002]

2. Description of the Related Art Generally, an apparatus for displaying an image expresses pixel data by a combination of red (R), green (G), and blue (B) and displays the pixel data. In particular, in a device that displays an image called a full-color image, generally, each pixel of red (R), green (G), and blue (B) has 8 bits (= 256) of grayscale image data for each pixel. Since the frame buffer has a memory capacity for storing image data, a high-precision full-color image can be displayed.

However, in most workstations, personal computers, and the like, the frame buffer does not have a memory capacity enough to simultaneously display a full-color image. Therefore, an input full-color image is simulated using colors that can be displayed by display means. A method of displaying information is used. For that purpose, it is necessary to convert the color of each pixel of the full-color image into a color that can be displayed by the display means. The method of the processing is large, a method using a dither matrix, and a statistical method of color. There are two ways to use the distribution. Hereinafter, the above two methods will be briefly described.

[0004] 1) Method using dither matrix [0004] This is because noise of a color that cannot be expressed in an input image is simply made inconspicuous in order to make false contour lines generated when the number of gradations is reduced inconspicuous. When the methods using the dither matrix are classified according to the processing algorithm, there are the following two methods.

[0005] A method of determining the color of a pixel in consideration of the color of the pixel surrounding the pixel, etc. [0005] Each pixel is prepared by superimposing a previously prepared dither matrix on the input image without considering the color of the pixel in the vicinity of the pixel. 2) Method of reducing colors using statistical distribution of colors Calculates the statistical value of the color distribution of the input image, and N colors (N: can be reproduced by display means) that are optimal for reproducing the input image Calculate any number less than the maximum value of the color) and create a look-up table. Then, for each pixel of the input image, the full-color image data is converted by converting to a look-up table, if there is the same color, to that color, and if not, to the most similar color in the look-up table. It is converted into a color that can be displayed by the display means and displayed.

The method 1) is described, for example, in Japanese Patent Application Laid-Open
22273, and the method 2) is described in, for example, JP-A-5-89972 and JP-A-5-21.
No. 5260, “Image Analysis Handbook”, pp.5
05-516, The University of Tokyo Press, 1991 ".

[0007]

SUMMARY OF THE INVENTION It is an object of the present invention to input a full-color image and to create an illustration-like color image which retains the texture of the original image while removing the realism from the input image. . For example, in animations and illustrations, it is better to express a person's skin with solid colors without gradations than with gradations using color gradations.
The texture of the illustration appears more strongly. Therefore, when a full-color image is input, it is desirable that the gradation expression of the full-color image be repainted with a single color as much as possible by the color reduction processing. For example, when a full-color image of a human face is input, it is desirable that the color of the skin color can be reduced so that it can be expressed by replacing it with two or three representative skin colors. In addition, the boundaries between different colors are not expressed in a gradation style using pseudo gradations, but when the boundaries are clearly expressed, the texture of animation and illustration is enhanced.

Considering the effectiveness of the conventional color reduction processing from such a viewpoint, first, the method using the dither matrix of the above 1) is a method for accurately reproducing an input full-color image as much as possible using limited colors. Thus, it is clear that the purpose of the present invention is not met.

On the other hand, in the method of color reduction using the statistical distribution of colors in the above 2), the distance is reduced in an RGB color space including three lightness axes of red (R), green (G), and blue (B). This is a mechanism for reducing the number of colors by defining the colors that are close to each other into one representative color. That is, in the conventional method, when the number of display colors (M is assumed here) is designated, the entire input full-color image is scanned,
Calculates the optimal distance in the RGB space for separating into M colors, calculates M representative colors, and converts each pixel of the input image into M representative colors according to the distance. In order to determine the representative M color by scanning the entire input full-color image, if the value of M is set to a small value, the distance becomes large. Therefore, the gradation portion where similar colors continuously change is somewhat a single color. Alternatively, it can be replaced by a plurality of colors.

By the way, for example, the whole is dull,
Consider an image that has only a few colorful parts.
In such an image, the colorful portion is remarkably impressive to human eyes, but in the conventional method, the color that is used less frequently in the input image is replaced with any of the representative colors. Because of the absorption, the color of the vivid portion will be replaced by one of the calculated representative colors, and the impression received from the reduced color image obtained as a result of the processing will be the input image Is likely to be different from the impression received from

Therefore, a problem to be solved is an apparatus and a method capable of performing color reduction processing of an image while inputting an image having only a little colorful portion while leaving the colorful portion. It is to provide.

[0012]

In order to solve the above-mentioned problems, a color reduction apparatus according to the present invention uses a look-up table for storing color conversion information for expressing an image with a limited color, and a color conversion device according to a color value. A classification rule for classifying each pixel of the input image, a classification unit for classifying each pixel of the input image with any one of a plurality of attribute values using the classification rule, and The pixels are classified by the attribute value, and based on the result of the classification, the number of colors to be assigned to a pixel group composed of pixels having the same attribute value is calculated for each attribute value. A look-up table calculating means for inputting, calculating color conversion information and storing the calculation result in the look-up table, and referring to the look-up table to determine the color of each pixel constituting the input image. Characterized by comprising a color reduction means for assigning to any of Teiiro, as main components.

Further, in the color reduction method according to the present invention, a first step of classifying each pixel of the input image by any one of a plurality of attribute values using the classification rule, and a processing result in the first step A second step of calculating, for each attribute value, the number of colors to be assigned to a pixel group composed of pixels having the same attribute value, inputting an image, calculating color conversion information, and calculating the calculation result. A third step of storing in a look-up table, a fourth step of referencing the look-up table and assigning the color of each pixel constituting the input image to one of the limited colors, and a filter using a square filter. And a fifth step of performing processing.

According to the present invention, first, each pixel of the input image is classified by one of a plurality of attribute values by referring to the classification rule by the classifying means.
Referring to the classification result, the number of colors to be assigned to a pixel group having the same attribute value is calculated for each attribute value.

The look-up table calculating means creates a look-up table corresponding to each attribute value with reference to the processing result of the number-of-colors calculating means. Color conversion of each pixel of the input image is performed with reference to the table to obtain a desired color-reduced image.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a color reduction apparatus according to claim 1 of the present invention.

The color reduction apparatus according to the present invention includes, as main components, a lookup table (105) for storing color conversion information for expressing an image with a limited color, and stores each pixel of an input image in accordance with a color value. Classification rule (102) for classification, classification means (101) for classifying each pixel of the input image by any one of a plurality of attribute values using the classification rule (102), and input by the classification means (101) Color number calculation for classifying each pixel constituting an image based on the attribute value and calculating, for each attribute value, the number of colors to be assigned to a pixel group composed of pixels having the same attribute value based on the result of the classification Means (103), input an image, look-up table calculating means (104) for calculating the color conversion information and storing the calculation result in a look-up table (105), refer to the look-up table (105),
A color reduction means (106) for assigning the color of each pixel constituting the input image to any of the limited colors, further comprising an input means (108) for inputting an image, and a square filter. Filter means for filtering (1
07) and display means (109) for displaying an image.

Hereinafter, each component constituting the present invention will be described. FIG. 2 to FIG. 4 are conceptual diagrams showing an example of the configuration of a look-up table (105) required in the overall processing in the present invention. First, the lookup table (105) shown in FIG. 2 is a specific example in the case where the brightness of each of red (R), green (G), and blue (B) has a gradation of 8 bits (= 256). . In this example, the RGB values are set in the order of 8
A numerical value represented by a total of 3 bytes by padding each bit is set as a table number (201, corresponding to an element number in the array) of the lookup table (105), and the array element (202) includes:
The configuration is such that data of a color to be actually replaced is stored.

For example, in the example of FIG. 2, the value of (R, G, B) is (0
x01,0x01,0x01), if these are expressed together in 3 bytes, it becomes 0x010101, and if 0x010101 is converted to decimal, it is the data of the 65793th array element, and the data stored there is Since it is 0x000000, it can be read from the lookup table (105) in FIG. 2 that if there is the data of (0x01, 0x01, 0x01), this data is replaced with (0x00, 0x00, 0x00).

However, as described above, red (R),
The brightness of each of green (G) and blue (B) is 8 bits (= 25
When the lookup table (105) having the gradation of 6) is created, there is a problem that the memory capacity for storing the lookup table (105) becomes large. For example, in the example of FIG. 2, the data for storing the data after the color conversion is 1
Since 3 bytes are required for each color, and the data array size is equal to the number of colors, that is, 16777216, it is not suitable for mounting in terms of memory capacity and processing speed.

Therefore, in practice, for example, a look-up table (105) is used in which the brightness of each of red (R), green (G), and blue (B) is reduced to a 4- or 5-bit gradation. FIG. 3 shows an example of the look-up table (105) when the brightness of each of red (R), green (G), and blue (B) is 4 bits. In this case, the size of the look-up table (105) is 4096. It only needs an individual.

FIG. 4 shows that the number of colors after color reduction is 256 (=
If it is less than 1 byte, it is a lookup table (105) configured to further reduce the required memory capacity. In this example, instead of directly writing the color after the color reduction in the conventional look-up table (105), the corresponding table (in particular, the color palette and Call) (203). In this case, the size of the place for writing the color after the color reduction of the lookup table (105) is only 1 byte. Although it is necessary to prepare a separate color palette, it is clear from the consideration of the total required memory that a smaller memory capacity is required if a separate color palette is created.

The classification means (101) according to the present invention classifies each pixel of the input image by one of a plurality of attribute values using the classification rule (102) according to the present invention. 102) is a rule for classifying each pixel of the input image according to the color value of the pixel for a certain pixel,
For example, it is as shown in the following example. Here, the saturation is represented by s and the lightness is represented by v, and both s and v are 0 to 2
A value in the range of 55 is assumed.

The processing in the classification means (101) corresponds to the processing in the first step in the color reduction method according to claim 8 of the present invention.

Example of classification rule: "For a pixel, calculate the value of s, v of the pixel, and calculate (s is 128 or more and v is 144 or more) or (v is 224 or more). If so, the attribute value of the pixel is A, otherwise the attribute value is B. "

The classification rule (102) used in the above example
Can also be used as an embodiment of the high brightness chroma pixel detection rule according to the third aspect of the present invention. Hereinafter, the above example is referred to as a classification rule (1).

Next, the flow of processing in the classification means (101) will be described with reference to FIG. FIG. 5 shows a classification means (10
3 is a flowchart showing the flow of the process in 1).

(301a) The lookup table (105) is initialized with a value other than the value used as the attribute value.

(302a) An arbitrary pixel whose attribute value has not yet been determined is extracted from the input image, and the attribute value of the pixel is determined by the classification rule (102).

(303a) The determined attribute value is written in the lookup table (105) corresponding to the color of the pixel.

(304a) For all pixels of the input image,
If the above processes (302a) to (303a) are completed,
The process at (101) ends. If it is not finished,
Returning to (302a), the processing of (302a) to (303a) is repeated.

FIG. 6 shows an example of a specific processing flow when the classification rule (1) is applied in FIG. The details of the processing flow are the same as those in FIG.

The number-of-colors calculating means (103) according to the present invention comprises a pixel group consisting of pixels having the same attribute value based on the result of classification of each pixel constituting the input image by the classification means (101). Is calculated for each attribute value. FIG. 7 is a flowchart showing an example of the flow of processing in the number-of-colors calculating means (103) according to the present invention. An example of processing in the number-of-colors calculating means (103) will be described with reference to FIG. The processing in the color number calculation means (103) corresponds to the processing in the second step in the color reduction method according to claim 8 of the present invention.

(401) The number of colors N after color reduction is input. Also,
By referring to the look-up table (105), the number of attribute values classified by the classification means (101) is checked. Here, for convenience of explanation, the number is assumed to be k, and i
The (1 ≦ i ≦ k) -th attribute value is Ai.

(402) For i = 1,..., K, the number of samples Hi of the pixel group whose attribute value is Ai and the degree Vi of the scattering of the pixel group whose attribute value is Ai are calculated.

There are various methods for calculating Vi. For example, a pixel is decomposed into components of red (R), green (G), and blue (B), and red (R), Green (G),
If the dispersion of pixel groups in the color space can be reflected, for example, the value of Vi is calculated by the sum of the variances calculated in the parameter space for each of blue (B), V
The calculation method of i may be any.

(403) For i = 1,..., K, the number of colors Ni to be assigned to the pixel group whose attribute value is Ai is calculated from the value of Hi and the value of Vi.

Here, there are various methods for calculating Ni. The basic policy is as follows: When the frequency Hi is large, the value of Ni is increased, and when the frequency Hi is small, the value of Ni is decreased.

If the degree of dispersion Vi is large, the value of Ni is increased, and if Vi is small, the value of Ni is decreased.

In this case, it is considered that the number of colors can be appropriately assigned. Based on the above policy, for example, Ni is calculated by the following equation (1), Ni = Sqrt (Hi) × Vi Equation (1) (where Sqrt (x) is a function that returns the square root of x)
Then, it is considered that a method of normalizing Ni to obtain Ｎｉ (i = 1,..., K) Ni = N to obtain a target Ni is effective.

In the above description, a method of calculating the number of colors Ni based on the number of samples Hi of the pixel group whose attribute value is Ai and the degree of dispersion Vi of the pixel group whose attribute value is Ai has been described. If the processing cost is to be as small as possible with a device having a low processing speed, i = 1,
.., K, the sample number H of the pixel group whose attribute value is Ai
A method of observing only i and determining the value of Ni by the square root ratio may be used.

That is, for example, Ni is calculated by the following equation (2), Ni = Sqrt (Hi) Equation (2) (where Sqrt (x) is a function that returns the square root of x)
Subsequently, it is considered that a method of normalizing Ni to obtain 目的 (i = 1,..., K) Ni = N to obtain a target Ni is effective.

In the calculation of Ni, the value of Hi affects Ni regardless of either of the above formulas (1) and (2).
Giving i a value of 1 or more is meaningless. In that case,
The i-th attribute value may be replaced by any other attribute value by some means. For example, a classification rule
In the situation where the two attributes should be divided according to (102), the total number of pixels classified into one of the attributes is less than, for example, 0.3% of the total number of pixels classified into the other attribute. In such a case, all pixels have the same attribute value, which is an example of the above replacement.If additional processing like the example is necessary according to the classification result,
The condition that requires additional processing and the content of the additional processing may be described together in the classification rule (102) (see the following example).

Example in which the processing of the classification rule (1) is added:
"For a pixel, calculate the value of s, v for that pixel, (s
Is 128 or more and the value of v is 144 or more) or (the value of v is 224 or more), the attribute value of the pixel is A, otherwise the attribute value is B. "

After the classification process, if necessary, the following 2
Execute the process. If the ratio of the pixels having the attribute value A is less than 0.3% of the whole, the attribute values of the pixels having the attribute value A are all changed to B.

If the proportion of the pixels having the attribute value B is less than 0.3% of the whole, the attribute values of the pixels having the attribute value B are all changed to A.

The lookup table calculating means (104) according to the present invention calculates a representative color of a pixel group having the same attribute value based on the number of colors assigned to the pixel group having the same attribute value. Calculate and store the calculation result in the lookup table (105).

FIG. 8 is a flow chart showing an example of the flow of processing in the lookup table calculation means (104) according to the present invention.
The processing example in (104) will be described. The processing in the lookup table calculation means (104) corresponds to the processing in the third step in the color reduction method according to claim 8 of the present invention.

(501) For i = 1,..., K, the value of the sample number Hi of the pixel group whose attribute value is Ai and the attribute value of A
From the number of colors Ni assigned to the pixel group i, a distance Di for determining whether or not the pixel group having the attribute value Ai is a similar color in the color space is calculated.

Various methods for calculating Di have been disclosed. For example, the distance defined by the cubic close-packed method disclosed in pp.513-514 of the above-mentioned document "" Image Analysis Handbook "" seems to be one of the distance definitions suitable for this method. .

(502) For i = 1,..., K, Ni representative colors for a pixel group whose attribute value is Ai are calculated based on the distance Di. In the method based on the legislative close-packed arrangement disclosed in the above-mentioned document, an algorithm for selecting the representative color is also disclosed. However, the algorithm can be applied to the selection of the representative color in the present processing.
In the above algorithm, a method of determining a more appropriate representative color by repeating a predetermined calculation is disclosed.However, in a case where it is desired to reduce the processing cost as much as possible with a device having a low processing speed, the above-described method is used. It is possible to obtain a result with a certain level of quality by performing the calculation only once.

(503) For i = 1,..., K, which representative color of each pixel Ni whose attribute value is Ai is to be replaced is calculated based on the distance in the color space, and the result is looked up. The result is written in the table (105), and the processing in the lookup table calculation means (104) is completed.

The color reduction means (106) according to the present invention refers to the look-up table (105) and performs a process of assigning the color of each pixel constituting the input image to one of the limited colors. FIG. 9 is a flowchart simply showing the flow of processing in the color reduction means (106) according to the present invention. The flow of processing in the color reduction means (106) will be briefly described with reference to FIG. The processing in the color reduction means (106) is performed according to claim 8 of the present invention.
This corresponds to the processing of the fourth step in the color reduction method described in (1).

(601) An arbitrary pixel which has not been subjected to color conversion in the input image is extracted, and a look-up table is obtained.
The color of the pixel corresponding to the pixel of (105) is referred to, and the color of the pixel is rewritten with the converted color written therein.

(602) For all pixels of the input image,
When the processing of (601) is completed, the processing by the color reduction means (106) is completed. If the processing has not been completed, the processing of (601) is repeated.

Next, the overall flow of the color reduction processing by the color reduction apparatus and the color reduction method according to the present invention will be described with reference to the flowchart of FIG. FIG. 11 shows how the contents of the lookup table (105) are rewritten as the processing proceeds. In this example, the brightness of each of red (R), green (G), and blue (B) of the color data of each pixel of the input image has a gradation of 4 bits (= 16), and the classification rule
(102) is the case where the above-described classification rule (1) is used.

(701) An image is input by the input means (108), and the lookup table (105) is initialized. FIG. 11A shows a state in which the lookup table (105) has been initialized.

(702) The classification means (101) uses the classification rule
Referring to (102), each pixel of the input image is classified by one of a plurality of attribute values, and the result is overwritten on the lookup table (105) (FIG. 11 (b)). This example is a case where the above-described classification rule (1) is used as an example of the classification rule (102).

(703) The color number calculating means (103) refers to the look-up table (105), and the classification means (101)
It is checked whether or not the number of attribute values classified by is plural. Let this number be k. If k = 1, the following processing
(704) to (705), if k> 1, the following processing (706) to (708)
Execute According to the classification rule (1), pixels are A and B
Since the rule is classified into the two types of attribute values, k = 2.

Since the lookup table (105) in the following processing (705) is a known technique, the lookup table (105) is reconstructed and its contents are rewritten in the following. How to go (707)
Description will be made with reference to FIG.

(704) Look-up table calculating means (104)
Thus, the representative colors for the number of colors after the color reduction processing specified in advance are calculated, and the lookup table (105) is completed. If k = 1, look-up table (105)
Is a look-up table (105) in the usual format, as depicted in FIG.

(705) The color reduction means (106) refers to the look-up table (105) and rewrites the color of each pixel constituting the input image with the color corresponding to the limited color.

(706) Based on the result of classification of each pixel constituting the input image by the color number calculation means (103) by the classification means (101), pixels having the same attribute value The number of colors to be assigned to the group is calculated for each attribute value. Here, it is assumed that the number of colors assigned to A is Na, and the number of colors assigned to B is Nb.

(707) Look-up table calculating means (104)
Thus, a lookup table (105) corresponding to each attribute value is created, and at the same time, representative colors for the number of colors after color reduction are calculated, thereby completing the lookup table (105).
In the example of FIG. 11C, two lookup tables (105) corresponding to the attribute values A and B are created. In addition, a color palette (FIG. 11 (d)) whose size (size of the array) is Na or Nb is simultaneously created. Then, a representative color for Na is calculated for the attribute value A, and a representative color for Nb is calculated for the attribute value B, and the corresponding lookup table (105) is calculated.
To complete.

(708) The color reduction means (106) determines the color of each pixel constituting the input image by referring to the look-up table (105) corresponding to the attribute values A and B of the pixel. Rewrite with the corresponding color of.

(709) If necessary, the image after the color reduction processing is subjected to filter processing such as noise removal. (710) The reduced color image is displayed on the display means (109), and the color reduction processing ends.

FIG. 12 is a block diagram showing a configuration example of the color reduction apparatus according to the seventh aspect of the present invention. In the color reduction device according to the seventh aspect of the present invention, as shown in FIG. 12, the color reduction device according to the first aspect of the present invention (FIG. 1)
Further, an edge extracting means (110) for extracting an edge component of the input image by using a filtering means (107), and a superimposing means (111) for comparing brightness values of pixels to synthesize two images.
, The edge image can be superimposed on the color-reduced image. In this case, the superimposing means (111) calculates the color after superimposing the images by comparing the lightness of points at the same coordinate position of the two input images and determining the pixel having the lower lightness. The color of
A method of setting the color of the pixel after superposition may be used. By the superimposition processing, an image in which an animation-like effect has been applied to an input image can be automatically created.

The processing in the filter means (107) is the same as the processing in the fifth step in the color reduction method according to the present invention, and the processing in the superimposing means (111) is the same as that in the present invention. This corresponds to the processing of the seventh step in the color reduction method according to claim 10.

FIG. 13 is a flowchart showing the flow of processing in a color reduction apparatus according to a seventh aspect of the present invention and a color reduction method according to a tenth aspect of the present invention. The overall processing is to execute the color reduction processing ((701) to (710)) and the edge extraction processing (711) in parallel (either processing may be performed first), and finally (701) the color reduction processing. The image of the processing result and the image of the edge extraction processing result are overlapped, and the processing ends. Regarding the configuration method of the edge extracting means (110) and the configuration method of the square filter used at that time, refer to Japanese Patent Application No. 8-8 / 1996.
1268, the description is omitted here. The processes (701) to (710) in FIG.
0, so the description is omitted. The processing in the edge extracting means (110) corresponds to the processing in the sixth step in the color reduction method according to claim 10 of the present invention.

[0070]

According to the present invention, a classification rule according to a use is prepared, and the color distribution of an input image is analyzed and classified by the classification means using the classification rule. For an image having a slightly colorful portion, it is possible to confirm the degree of the presence of the colorful portion, and it is possible to appropriately assign the colorful portion and the other portion according to the processing result of the classification means. A process for calculating the number of colors by the number-of-colors calculating means is newly incorporated.

In the conventional color reduction processing, there is a problem that the data having a low frequency is absorbed by the data having a high frequency statistically with respect to the whole, so that the data having a low frequency is lost by the color reduction processing. However, the introduction of the classification rule, the classification means, and the number-of-colors calculation means enabled the color reduction processing while leaving a necessary part even if the frequency is low. For example, in the example of the above-described image, it is possible to realize that the color reduction processing is performed independently on the slightly colorful part and the other part, and as a result, the colorful part can be left after the color reduction processing. .

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a color reduction apparatus according to claim 1 of the present invention.

FIG. 2 is a conceptual diagram showing a configuration example of a conventional lookup table.

FIG. 3 is a conceptual diagram showing a configuration example of a conventional lookup table.

FIG. 4 is a conceptual diagram showing a configuration example of a conventional lookup table.

FIG. 5 is a flowchart showing a flow of processing in a classification unit according to the present invention.

FIG. 6 is a flowchart showing the flow of processing in the classification means according to the present invention based on an example of a classification rule;

FIG. 7 is a flowchart showing an example of the flow of processing in the number-of-colors calculating means according to the present invention.

FIG. 8 is a flowchart showing an example of the flow of processing in a lookup table calculation unit according to the present invention.

FIG. 9 is a flowchart showing the flow of processing in the color reduction means according to the present invention.

FIG. 10 is a flowchart showing an overall flow of a color reduction process in the color reduction apparatus and the color reduction method according to the present invention.

FIG. 11 is a conceptual diagram showing a configuration example of a lookup table according to the present invention.

FIG. 12 is a block diagram showing a configuration example of a color reduction apparatus according to claim 7 of the present invention.

FIG. 13 is a flowchart showing a processing flow in a color reduction apparatus according to a seventh aspect of the present invention and a color reduction method according to the tenth aspect;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 Classification means 102 Classification rule 103 Number of colors calculation means 104 Lookup table calculation means 105 Lookup table 106 Color reduction means 107 Filter means 108 Input means 109 Display means 110 Edge extraction means 111 Superposition means

Claims (10)

[Claims] 1. A lookup table for storing color conversion information for expressing an image in a limited color, and a lookup table for inputting an image, calculating color conversion information, and storing the calculation result in the lookup table. A color reduction device comprising: a calculation unit; and a color reduction unit that refers to a look-up table and assigns the color of each pixel constituting the input image to any of the limited colors. A color reduction apparatus, comprising: a rule for classifying each pixel of an image, wherein the classification rule classifies each pixel of the input image with one of a plurality of attribute values using the classification rule. 2. Classification means classifies each pixel constituting an input image by the attribute value, and, based on the result of the classification,
2. The color reduction apparatus according to claim 1, further comprising a color number calculation unit that calculates the number of colors assigned to a pixel group having the same attribute value for each attribute value. 3. The classification rule includes a high-brightness / saturation pixel detection rule that inputs at least one of a saturation value and a lightness value among elements constituting a color and outputs one of attribute values. The color reduction apparatus according to claim 1 or 2, wherein 4. The apparatus according to claim 1, further comprising a filter means for performing a filtering process using a square filter.
A color reduction apparatus according to any one of the above. 5. A two-dimensional square filter 1 is a filter that inputs an image to the filter and uses the most frequently input value as an output signal of the filter. The color reduction apparatus according to any one of claims 1 to 4, wherein filter processing is performed by filter means using the square filter 1. 6. A two-dimensional square filter 2, which is an edge extraction filter configured to obtain an output signal of the square filter for an input image as edge information of the input image, wherein the two-dimensional square filter 2 converts the input image into at least two or more floors. Edge extracting means for converting the converted result into a black-and-white gray-scale image, applying a filtering process to the conversion result by the filter means using the square filter 2, and extracting edge components from the input image. The color reduction apparatus according to any one of claims 1 to 4, wherein: 7. The two images are input, and a color after superimposition is calculated according to a result of comparing colors of points at the same coordinate position of the two images, and a color of the two images is calculated. 7. The color reduction apparatus according to claim 6, further comprising a superimposing unit that overwrites the pixel at the coordinate position of any one of the images with the calculated color. 8. The look-up table stores color conversion information for expressing an image with a limited color, and the classification rule is a rule for classifying each pixel of an input image according to a color value. A first step of classifying each pixel of the input image with one of a plurality of attribute values using a rule;
A second step of calculating, for each attribute value, the number of colors to be assigned to a pixel group having the same attribute value based on the processing result in the first step; A third step of calculating information and storing the calculation result in the lookup table; and a fourth step of referring to the lookup table and assigning a color of each pixel constituting the input image to one of the limited colors. And a fifth step of performing a filtering process using a square filter. 9. The color reduction according to claim 8, wherein the processing result in the fourth step is subjected to filtering processing in the fifth step using the square filter 1 according to claim 5. Method. 10. The color filter method according to claim 8, further comprising: converting the input image into a black-and-white shading image having at least two or more tones; In a fifth step, a filtering process is performed in a fifth step, and an edge component is extracted from the input image, and two images are input.
According to the result of comparing the colors of the points at the same coordinate position of the two images, the color after superimposition is calculated, and the pixel at the coordinate position of any one of the two images is calculated. And a seventh step of overwriting the calculated color.