JPH06243244A - Color image display method - Google Patents

Abstract

PURPOSE:To improve the reproducibility for such a color that has a large local change, to improve the impression received from an entire image, and to shorten processing time in regard of a color image display method. CONSTITUTION:The image data are inputted (S1) and then divided into partial areas (S2). Then the number of representative colors is decided at every partial area (S3), and a representative color having the number of colors is selected from each area (S4). Furthermore all picture elements are compared with a part or all of representative colors, and the nearest representative color is decided as the display color of the picture element (S5). Thus a look-up table LUT is produced (S6). Then the input image data are converted into the RGB signals based on the table LUT (S7). These RGB signals are converted into the analog signals (S8) and then displayed on a display (S9).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image display method using a look-up table (also called a color map, hereinafter referred to as LUT).

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents. Reference: Supervised by Mikio Takagi, Yoshihisa Shimoda "Image Analysis Handbook", first edition (1991-1-17), University of Tokyo Press, P.
505-516 As described in the above literature, image processing technology has been used in many fields in recent years, and technology for displaying a color image on a personal computer, a workstation or the like has become important. A schematic diagram of a general color image display system is shown in FIGS.

FIG. 2A shows a full-color color image display system, and FIG. 2B shows a LUT color image display system. A normal natural image has tens of thousands to tens of millions of colors, and to display all these colors,
A full-color system as shown in FIG. 2A is used.
In this full color system, each pixel has a frame buffer (or refresh memory) 1R, 1G, 1B of about 8 bits of red (R), green (G), and blue (B), and a digital buffer is provided on the output side thereof. A CRT3, which is a display device, is connected via an analog converter (hereinafter referred to as a D / A converter) 2R, 2G, and 2B, and approximately 16 million colors (=
2 ²⁴ = 2 ⁸ × 2 ⁸ × 2 ⁸ ) can be displayed at the same time. This full-color method is suitable for displaying a normal natural image in color, but in order to display all the many colors of the natural image, many frame buffers 1R,
Requires 1G and 1B. Therefore, if the full color system shown in FIG. 2A is used for a device such as a personal computer or a workstation, the scale of the device becomes large and the cost becomes high. Therefore, in a general workstation or the like, the number of colors that can be simultaneously displayed is 8 to 256 due to the limitation of the memory capacity of the frame buffer memories 1R, 1G, and 1B.
By limiting the number of colors (hereinafter, referred to as N colors) and using the LUT method as shown in FIG. 2B, a degree of freedom is given to the selection of colors (representative colors) to be displayed at the same time.
In the LUT method of FIG. 2B, the color number corresponding to each pixel value is stored in the memory of the frame buffer 5, and according to the LUT 6 in which the correspondence between the original color of the original image and the representative color is registered, D / CRT8 via A converters 7R, 7G, 7B
To display the image. At this time, what kind of representative color corresponds to the original color is the key to color tone reproduction. This method (algorithm) is called the limited color display method (limited color display algorithm). Can be classified.

Limited Color Display Algorithm (1) Representative Color Selection Step (a) Uniform Quantization Method (b) Subdivision Quantization Method-Frequency Method (Popularity Algorithm, Popularity A)
lgorithm) etc. (2) Display color selection step-Nearest neighbor search method (3) (1), (2) steps-Median Cut Algorithm
hm) method The limited color display algorithm can be considered by dividing it into two steps of (1) representative color selection step and (2) display color selection step. (1) The representative color selection step is
This is a step of calculating the statistic of the color distribution of the original image and selecting the optimum representative color of N colors for reproducing the image. On the other hand, the (2) display color selection step is a step of selecting and displaying a color closest to each pixel value of the original image from N colors. There are various conventional methods for this limited color display algorithm.

(1) The representative color selection step includes (a) a uniform quantization method and (b) a subdivision quantization method. An example of the (b) subdivision quantization method is a frequency method. is there. This frequency method is an algorithm for selecting representative colors in order from the color with the most color distribution statistics (upper N colors). (2) As an example of the display color selection step, there is a nearest neighbor search method, in which each pixel value is compared with all the representative colors, and the one having the smallest difference is used as the representative color. The median cut algorithm method is an algorithm that discusses two steps (1) and (2) at the same time. A space in which the three primary colors of R, G, and B are three-dimensional orthogonal coordinates is called an RGB color space. In the median cut algorithm method, first, a rectangular parallelepiped circumscribing the distribution in the color space of the original image is considered, and this rectangular parallelepiped is divided into two equal parts so that the number of distribution pixels becomes equal. When the number of the divided rectangular parallelepipeds reaches N, the division is terminated, the average value of the pixels of each rectangular parallelepiped is set as each representative color, and all the pixels in the rectangular parallelepiped are displayed in the representative color.

FIG. 3 is a block diagram showing an example of a color image display system according to the conventional limited color display algorithm. Elements common to those in FIG. 2B are designated by common reference numerals. . In this image display system, as shown in FIG.
In the color image display system of (b), limited color display means 1
0 is connected. The limited color display means 10 converts the input image data Pin of 8 bits for each of R, G, and B, that is, 24 bits in total into color number data C of n bits (2 ⁿ = N) and supplies it to the frame buffer 5. . In addition, the limited color display means 10 includes the color number data C and the display colors (R, G,
B) corresponding LUT data M is created and supplied to the LUT 6. Then, each image data Pin is converted into an 8-bit R, G, B signal by the LUT 6, converted into an analog signal by the D / A converters 7R, 7G, 7B, and then supplied to the CRT 8 for correspondence. The color to be displayed is displayed.

FIG. 4 is a functional block diagram showing a configuration example of the limited color display means 10 in FIG. The limited color display means 10 selects a representative color from the 24-bit image data Pin and outputs the 24-bit representative color data C2, and the 24-bit image data Pin and 24.
Select a display color from the bit representative color data C2 and select LUT
The display color selection means 12 outputs data M and n-bit color number data C. Image data Pi
When n is input to the representative color selecting means 11 and the display color selecting means 12, the representative color selecting means 11 produces image data Pi.
The representative color of N colors is selected from n, and the representative color data C2 is supplied to the display color selection means 12. The display color selection means 12 creates and outputs n-bit color number data C and LUT data M from the image data Pin and the representative color data C2.

[0008]

However, the conventional color image display method using the limited color display algorithm has the following problems. (A) In the conventional limited color display algorithm, since the limited representative color is selected from the color distribution statistics of the entire image, the color tone reproduction as a whole is quite good, but the color with low appearance frequency is a color that is considerably different from the original image. Since it is often displayed in, the color tones in the details are not reproduced well. In general, the image quality deterioration due to this is not so noticeable, but if the appearance frequency is low in the image but a characteristic color is present, even if other parts are well reproduced, The impression it receives is quite low. For example, in the case of an image including a person's face, lips and the like have a small statistical frequency, and thus the conventional method rarely assigns an optimum color. (B) In many cases, the background occupying a large area in the image changes gently like the sky. However, in the conventional method, many colors are simply assigned to frequently appearing colors. Although the color reproducibility of these backgrounds is good, the color reproducibility of main images other than the background is poor. In particular, there are many cases where the color that is to be watched is subject to drastic changes in color. Therefore, if a small number of colors are assigned to these areas, sharp edges cannot be reproduced, and the subjective impression of the entire image deteriorates. (C) In the conventional method, the processing time of the display color selection step is considerably long. Particularly, in the method of comparing each pixel with all the representative colors like the nearest neighbor search method, the color reproducibility is improved, but the execution time is too long, which is not so realistic.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the problems of the prior art and to provide the following color image display method. (1) By dividing an image into partial areas and selecting a representative color for each area, a characteristic color that has a low appearance frequency but appears in a part of the image is selected as a representative color, and those colors are reproduced. An object of the present invention is to provide a color image display method in which the subjective evaluation of the entire image is enhanced by enhancing the property. (2) A color image display method in which the color reproducibility of the main part of the image is improved by suppressing the selection of representative colors from the part where the background changes gently and selecting a large number of representative colors from the part that changes drastically. The purpose is to provide. (3) An object of the present invention is to provide a color image display method in which the processing time is shortened by comparing pixel values only with the representative colors of the area including the target pixel and the surrounding area in the display color selection step. To do.

[0010]

In order to solve the above-mentioned problems, a first invention is a color image display method of creating an LUT storing color data and displaying a color image by referring to the LUT. Area division processing for dividing the original image into partial areas, and color distribution and size of the plurality of partial areas divided by the area division processing are referred to determine the number of representative colors selected from each of the partial areas. The representative color number determining process and the representative color selecting process for selecting the representative color number from each of the partial areas are performed. Then, the display color of all the pixels is selected by the display color selection process from the representative color selected by the representative color selection process to create the LUT. In the second invention, the display color selection processing of the first invention selects the display color only from the representative colors of the area including the target image and the peripheral several areas. In the third invention, the area dividing process of the first or second invention determines the number of areas by using the limited color number N as a parameter, and after determining the size of the image and the size of each area from the number of areas, The original image is divided into partial areas.

In the fourth aspect of the invention, in the area dividing process of the first or second aspect, the areas in which the difference in pixel value between adjacent pixels is smaller than the threshold value are divided as one area. In the fifth invention, the area division processing of the first or second invention is
First, when the partial area is divided into several large areas and the distribution of the color distribution in the area is large, the partial area is 1 / i of the limited color number N (however,
i; threshold value). And
In the representative color number determination process, i representative color numbers are assigned to each area. In the sixth invention, in the representative color number determination process of the first or second invention, first, the representative color number is provisionally determined in each area, and then the total of the provisional representative color number of all areas and the limited color are defined. Number N
And the true representative color number of each area is determined. In the seventh invention, the representative color number determination process of the sixth invention reduces the number of representative colors when the color distribution in the area is dense, and increases the number of representative colors when they are dispersed.

In the eighth invention, in the representative color number determination processing of the seventh invention, when the density of the color distribution is higher than the threshold value and the difference in the average pixel value from the peripheral area is small, the area is determined. The representative color is not selected. In the ninth invention, the area division processing of the seventh or eighth invention divides into M × M (where M is the size of the image block) square rectangular blocks. Then, in the representative color number determination process, discrete Conin transform (hereinafter referred to as DCT) is performed on the image data of each block, the AC component of the transform coefficient is used as a parameter, and when the transform coefficient is large, many representative colors are assigned, When the number is small, the number of representative colors is determined by assigning a small number.
In the tenth invention, in the representative color number determination process of the ninth invention, when the representative color number is determined using the transform coefficient of the DCT, the number of low frequency components (m Xm, M ≧ m) only, and the number of representative colors is determined using only the coefficient of the low frequency component as a parameter.
In the eleventh invention, in the representative color number determination process of the ninth or tenth invention, when the representative color number is determined by using the transform coefficient of the DCT, the transform coefficients having a similar frequency are grouped together in several groups. Then, the number of representative colors in the area is calculated using the sum of the conversion coefficients of each group as a parameter.

In the twelfth invention, in the representative color selection processing of the first or second invention, pixels whose distance is closer than the threshold value k in the color space are grouped into several groups, and the number of groups is assigned to the area. Regroup by increasing k until the number is less than several times the number of representative colors. Then, the regrouped groups are arranged in descending order of the number of constituent pixels, the group having the representative color number is selected from the top, and the representative color is selected from each group. In the thirteenth invention, the representative color selection processing of the first or second invention excludes a pixel whose distance in a color space is closer than a threshold value from a representative color candidate as compared with a representative color of a peripheral area, A representative color is selected from the remaining pixels. In the fourteenth invention, the display color selection processing of the second invention compares the average pixel value of the area including the target pixel with
All the peripheral areas where the difference is smaller than the threshold value are used as reference areas,
A display color is selected from the representative colors of the reference area.

[0014]

According to the first aspect of the invention, since the color image display method is configured as described above, in the representative color selection step, the original image is divided into partial areas by the area division processing and the representative color number determination processing is performed. Sent. In the representative color number determination process, the number of representative colors to be selected from the region is determined by referring to the color distribution of each region and the size of the region, and the determination result is sent to the representative color selection process. In the representative color selection process, a representative color is selected from each area by the number of colors, and the selection result is sent to the representative color selection process in the display color selection step. In the representative color selection process, a display color is selected from the selected representative colors to create an LUT. Then, a color image is displayed with reference to the created LUT. According to the second invention, in the display color selection process in the display color selection step of the first invention, only the representative color of the region including the target pixel and the peripheral region, for example, the pixel value is compared, and the display color is displayed. Is selected. According to the third invention, in the first or second area division processing, the number of areas to be divided is determined using the limited color number N as a parameter, and each area is determined from the determined number of areas and the size of the image. After the size is determined, the area division is performed.

According to the fourth invention, in the area dividing process of the first or second invention, when the difference between the pixel values of two adjacent pixels is smaller than a certain threshold value, the two pixels are placed in the same area. I include it. Then, the same processing is performed on all the pixels, and the pixels are grouped into some areas and the areas are divided. According to the fifth invention, in the area dividing process of the first or second invention, first, when the original image is divided into several large areas, and in each area, the variance of the color distribution in the area is large, Further, the area is divided into smaller areas. The number of areas is 1 / i of the limited number of colors N
The above process is repeated until the whole area is divided. Further, when determining the number of representative colors, i
A representative color number is assigned to each. According to the sixth invention, in the representative color number determination process of the first or second invention,
When determining the number of representative colors, first, the number of representative colors is provisionally determined using the color distribution and image size of each area as parameters. Next, the total number of provisional representative colors of all areas is compared with the limited number of colors N, and the number of representative colors of each area is corrected to determine the true number of representative colors. According to the seventh invention, in the representative color number determination process of the sixth invention, when the representative color number is determined, if the color distribution in the area is dense (if the pixel value is gently changed). If the number of representative colors is reduced and the color distribution is dispersed (if the pixel value changes drastically), the number of representative colors is increased.

In the eighth invention, in the representative color number determination process of the seventh invention, when the representative color number is determined, the density of the color distribution is higher than the threshold value (the pixel value hardly changes). When the difference in the average pixel value between the area and the peripheral area is small (the color is almost the same as that of the peripheral area), the representative color number is set so that no representative color is selected from that area (representative color number 0). A decision is made. According to the ninth aspect, in the seventh or eighth area division process, when the original image is divided into partial areas, the original image is divided into M × M square rectangular blocks. Then, when determining the number of representative colors, DCT is performed on the image data of each block, and the AC component of the conversion coefficient is used as a parameter. When the coefficient is large, a large number of representative colors are assigned, and when it is small, The number of representative colors is determined by assigning a small number. According to the tenth invention, in the representative color number determination process of the ninth invention, when determining the representative color number, only the coefficient of the number of low frequency components of the DCT is irrespective of the size M of the image block. The number of calculated representative colors is determined by using only the coefficient of the low frequency component as a parameter.
According to the eleventh aspect, in the representative color number determination process of the ninth or tenth aspect, when the representative color number is determined, the DCT
The conversion coefficients of (1) are grouped into those having similar frequencies, and the sum of the coefficients of each group is used as a parameter to calculate the number of representative colors in that area. For example, when the coefficient of the low frequency component is large, the pixel value changes gently in the area, so it is not necessary to allocate too many colors, but when the coefficient of the high frequency component is large, it changes drastically. Therefore, many colors are assigned and the optimum number of colors is selected.

According to the twelfth invention, when the representative color is selected in the representative color selection processing of the first or second invention,
Pixels whose distance in the color space is closer than the threshold value k (pixels of almost the same color) are grouped together, and k is increased until the number of groups is less than several times the number of representative colors assigned to the area. Be regrouped.
Then, the regrouped groups are arranged in descending order of the number of constituent pixels, the group having the representative color number is selected from the top, and the representative color is selected from each group. According to the thirteenth invention, in the representative color selection processing of the first or second invention, when the representative color is selected, the peripheral region (first
The invention is compared with several neighboring areas, and the second invention is compared with a representative color of a peripheral pixel referred to when selecting a display color, and a pixel whose distance in the color space is closer than a threshold value (already in the peripheral area). Pixels having almost the same color as the representative color are excluded from the representative color candidates, and the representative color is selected from the remaining pixels in the area. According to the fourteenth invention, in the display color selection process of the second invention, when the display color is selected, the average pixel value of the area including the target pixel is compared, and the difference is smaller than the threshold value. All areas (peripheral areas of almost the same color) are used as reference areas, and display colors are selected from the representative colors of these reference pixels. Therefore, the above problem can be solved.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A color image display method according to an embodiment of the present invention will be described with reference to the drawings. Color Image Display System FIG. 1 is a schematic configuration diagram of a color image display method of an embodiment of the present invention and a color image display system for implementing the method. The color image display system used in this embodiment is
Similar to FIG. 3 of the related art, the input R, G, B image data Pin is converted into n-bit color number data C, and LUT data M corresponding to each color number and display color (RGB) is created. The limited color display unit 100 is provided, and the frame buffer 200 and the LUT 300 are connected to the output side thereof. The frame buffer 200 is a memory that stores a color number corresponding to each pixel value, and its output side is an LU.
It is connected to T300. The LUT 300 is a circuit that inputs the LUT data M and the output of the frame buffer 200 and converts each pixel data into R, G, B signals, and has an R, G, B D / A converter 400 on its output side. A display (for example, a CRT) 500 is connected via the. The R, G, B D / A converter 400 has a function of converting the R, G, B signals into analog signals and then displaying them on the CRT 500. This color image display system is configured by using hardware such as an integrated circuit or software such as a program-controlled processor.

An outline of the color image display method of this embodiment using such a color image display system will be described. First, in the limited color display means 100, in the process S1, for example, 2
4-bit image data Pin is input, and in step S2, the input image data Pin is divided into partial areas. Then, in step S3, the number of representative colors is determined for each of the divided areas, and in step S4, the number of representative colors is selected from each area. After the selection of the representative color, in step S5, for all pixels, the representative color closest to some or all of the representative colors is set as the display color of the pixel, and in step S6, the LUT data M is created. Color number data C is given to LUT300.
Is output and given to the frame buffer 200. LUT3
At 00, in step S7, the input image data Pin is converted into R, G, B signals according to the created LUT data M. The converted R, G, B signals are D / A of R, G, B.
After being converted into analog signals in the process S8 in the converter 400, in the process S9 in the CRT 500,
Appears on the display.

FIG. 5 shows the limited color display means 100 shown in FIG.
It is a functional block diagram of. The limited color display means 100 performs a process corresponding to the first aspect of the invention, and is composed of hardware using an integrated circuit or the like, or software using a program-controlled processor or the like. The limited color display means 100 has area dividing means 110 for performing area dividing processing of S1 and S2 of FIG. 1 and outputting 8-bit area data A1, and on the output side thereof, a representative color number determining means 120. The representative color selection means 130 is connected via. The representative color number determination unit 120 uses the original image data Pin.
1 has a function of performing the representative color number determination process of S3 of FIG. 1 based on the area data A1 and supplying the representative color number data C1 as the processing result to the representative color selection means 130. The representative color selection means 130 performs the representative color selection process of S4 of FIG. 1 based on the image data Pin and the representative color number data C1.
The 24-bit representative color data C2 is displayed in the display color selection means 140.
It has a function to give to. The display color selection means 140 is
The display color selection process of S5 of FIG. 1 is performed based on the 24-bit image data Pin and the 24-bit representative color data C2, and the LUT data M and the n-bit color number data C are output. .

Next, each processing content in the limited color display means 100 of FIG. 5 will be described below with reference to FIGS. FIG. 6 is a block division example, FIG. 7 is a line division example, and FIG. 8 is a correspondence table example. In the limited color display means 100 of FIG. 5, 24-bit image data Pin
Is input, the area dividing unit 110 divides the image data Pin into partial areas, and the divided area data A
1 is sent to the representative color number determination means 120. In this area division processing, for example, as shown in FIG. 6, the area is divided into blocks each having a square or rectangular shape of 8 × 8 pixels or the like, or
As shown in FIG. 7, the image data Pi
Divide n. The representative color number determining unit 120 inputs the 24-bit image data Pin and the area data A1, determines the number of representative colors to be selected from each divided area by referring to the size and the number of colors of the area, The representative color number data C1 is sent to the representative color selecting means 130. In this representative color number determination processing, for example, the threshold value is set to 10 and the number of colors (types) in the area is set.
If 25 is 3 colors, if the number of colors is 7, it is 1 color.
This can be achieved by preparing and applying a simple correspondence table (FIG. 8) using division. Further, in the area division processing, when the areas are divided so as to have different sizes, the number of colors can be determined in proportion to the size.

Here, if the number of colors is arbitrarily determined in each area, the total number of colors in all areas does not match the specified limited number of colors. Therefore, a limited number is provided in each area in advance. It may not be exceeded, or the conventional limited color method may be applied again after the representative color number determination processing is finished to perform fine adjustment. In the representative color selection means 130,
The 24-bit image data Pin and the representative color number data C1 are input, the representative colors are selected from the respective regions by the assigned number of colors, and the 24-bit representative color data C2 is sent to the display color selecting means 140. In this representative color selection process, for example, a representative color is selected from colors that frequently appear in the area, as in the frequency method. The display color selection means 140 selects a color for displaying each pixel from the 24-bit image data Pin and the 24-bit representative color data C2, and the n-bit color number data C is selected.
And LUT data M are output. For example, it can be realized by applying a nearest neighbor search method or the like, in which all the representative colors selected from each area are compared with the pixels to be selected. As mentioned above,
In the limited color display unit 100 shown in FIG. 5, the original image is divided into partial areas, and the characteristic color from each area is used as the representative color. Therefore, the appearance frequency that appears in a part where the reproducibility is poor is low. However, the reproducibility is improved for characteristic colors. That is, the reproducibility is improved for a color that locally changes greatly. Therefore, the impression received from the entire image is improved.

Embodiment of limited color display means Each means 110 constituting the limited color display means 100 of FIG.
FIG. 9 shows the embodiments (1) to (4) of 120, 130, 140.
The following is a description with reference to FIG. (1) Area dividing means 110 (FIG. 9) FIG. 9 is a functional block diagram of the area dividing means 110 shown in FIG. The area dividing means 110 performs the processing corresponding to the third aspect of the invention, and the 24-bit image data Pi
24, an area number determining means 111 for performing the area number determining process based on n and the limited color number N and outputting the area number data A2.
The small area dividing unit 112 performs a small area dividing process based on the bit image data Pin and the area number data A2 and outputs the area data A1.

Next, the processing contents of the area dividing means 110 will be described. First, when the 24-bit image data Pin and the limited color number N are input, the area number determination unit 111 determines an appropriate number of areas to be divided from the image data Pin and the limited color number N, and the number of areas is determined. The data A2 is sent to the small area dividing unit 112. This area number determination processing can be easily realized by setting the value obtained by dividing the limited color number N by 2 or 3 as the area number. For example, when the limited color number N is 256, 128 or 64 is set as the number of areas. If you can't divide it, just cut it down. The small area dividing unit 112 determines the area size from the 24-bit image data Pin and the area number data A2, divides the area into small areas, and outputs the area data A1. This small area division processing is performed by, for example, the area number determination unit 1
The number of regions supplied from 11 is 64, and the size of the image is 6
When 40 × 480, 640 and 480 with the square root of 64
When divided by, the size of one region becomes 80 × 60, which can be realized by dividing into this size. If you can't divide it, just cut it down. As described above, in the area dividing unit 110 of FIG. 9, each processing can be performed using the same threshold value in each area regardless of the size of the image and the number of limited colors.

(2) Representative color number determining means 120 (see FIG. 1)
0) FIG. 10 is a functional block diagram of the representative color number determination unit 120 shown in FIG. The representative color number determining means 120 is the sixth
The process corresponding to the present invention is performed. The representative color number tentative determination unit 121 that performs the representative color number tentative determination process based on the 24-bit image data Pin and the area data A1 and outputs the representative color data C3, and the limited color Number N and representative color number data C3
The color number comparing means 122 for performing the color number comparing processing based on the above and outputting the representative color number data C1 and the weight calculating means 123 for performing the weight calculating processing based on the output of the color number comparing means 122 and outputting the weight data W1. And the color number adjusting means 124 for inputting the weight data W1 and performing the color number adjusting process.

Next, the processing contents of the representative color number determining means 120 will be described with reference to FIGS. 8 and 11. FIG. 11 is a diagram showing an example of a proportional function of inclination. First, the representative color number tentative determination means 121 tentatively determines the representative color number in each area based on the 24-bit image data Pin and the area data A1, and the tentatively determined representative color number data C3 is used in the color number comparison means. Send to 122. This representative color number tentative determination process is, for example,
As described with reference to FIG. 5, when the threshold value is 10, the number of colors (types) in the area is 25, the number of colors is 3, and when the number of colors is 7, the number of colors is 1, so that the simple correspondence of FIG. It can be realized by preparing a table and applying it. In the color number comparison means 122, the total of the representative color number data C3 supplied from the representative color number tentative determination means 121 is compared with the limited number of colors N. If the number of representative colors is equal to N, the input representative color The numerical data C3 is output as it is as the representative color number data C1. In other cases, the representative color number data C1 is supplied to the weight calculating means 123. In this color number comparison processing, for example, the limited color number N is 256 and the total number of temporary representative colors (C3) is 40.
In the case of 0, it can be determined that fine adjustment to reduce 144 is necessary.

The weight calculating means 123 calculates the weight of the color of the area with reference to the representative color number data C1 of each area, and supplies the weight data W1 to the color number adjusting means 124. This weight calculation processing is performed by, for example, the limited number of colors N as in the above example.
Is 256, and when the adjustment range is large as in the case where the total number of temporary representative colors (C3) is 400, the inclination 25 is set as shown in FIG.
This can be realized by using a proportional function of 6/400 and calculating a weighting function with an adjusted representative color number being an integer that does not fall below that value. When the adjustment range is small, the number of colors is subtracted one by one from the area with the largest number of colors, and fine adjustment is performed. The color number adjusting unit 124 adjusts the number of colors by making the weight data W1 supplied from the weight calculating unit 123 correspond to the representative color number data C3, and the representative color number data C3 is compared with the representative color number data C3.
Supply to 2. As described above, in the representative color number determination means 120 of FIG. 10, in the representative color number determination processing in each area,
Since it is not necessary to handle the information of the entire image at all, the processing is simplified, and the memory amount can be reduced even when it is realized by hardware.

(3) Representative Color Selection Means 130 (FIG. 12) FIG. 12 is a functional block diagram of the representative color selection means 130 shown in FIG. The representative color selecting means 130 corresponds to the twelfth invention, and is the 24-bit image data Pin.
The neighboring pixel merging unit 131 that merges (merges) the neighboring images by the neighboring pixel merging process to output group data GP, and the neighboring pixel merging unit 131 that compares the representative color number data C1 and the group data GP. Group number determination means 132 for performing group number determination processing for requesting (Q) regrouping, and pixel group sorting means for performing pixel group sorting processing for rearranging the image groups obtained by the neighboring pixel merging means 131. 133
And 2 based on the output of the pixel group sorting means 133.
The representative color is calculated from the 4-bit image data Pin and calculated as 24
The representative color calculation unit 134 that performs the representative color calculation process that outputs the representative color data C2 of bits is configured.

Next, the representative color selection processing of the representative color selection means 130 will be described with reference to FIG. Figure 13
It is a figure which shows the example of merging of neighboring pixel X, Y, and Z. First,
In the neighboring pixel merging unit 131, all pixels in the area are compared based on the input 24-bit image data Pin to merge the pixels whose distance is closer than the threshold value k in the color space, and the merged group. The data GP is sent to the group number determination means 132. This neighborhood pixel merging process is performed by, for example, as shown in FIG.
When Y and Z are present (the color space is represented by a plane for simplicity), if the threshold value k is 20, the pixel Y is merged into the group of pixels X and the pixel Z is not merged. After that, the pixel value of this group of pixels X is provisionally X.
The pixel value is the same as the pixel value of, and comparison and merging with other pixels are performed. In the group number determination means 132, the group data (number of groups) merged by the neighboring pixel merge means 131
GP and representative color number data C1 assigned to the area
And the re-grouping is requested (Q) to the neighboring pixel merging means 131. In the group number determination process, for example, when the number of groups is twice the number of representative colors or less, regrouping is requested. Neighborhood pixel merging means 131
Then, when regrouping is requested, the threshold value k is increased,
Group a wider range of pixels to reduce the number of groups. For example, in the case of FIG. 13, when the threshold value k is increased to 30, the pixel Z is also merged into the same group, so that the number of groups decreases.

On the other hand, the pixel group sorting means 133 arranges the pixel groups obtained by the neighboring pixel merging means 131 in descending order of the number of constituent pixels, selects a representative color number group from the top, and supplies it to the representative color calculating means 134. . The representative color calculation unit 134 calculates 2 from each supplied group.
An appropriate color in the 4-bit image data Pin is selected as a representative color, and 24-bit representative color data C2 is output. In this representative color calculation processing, for example, the average of all pixel values in the group is set as the representative color, or the color having the largest number of constituent pixels is set as the representative color. As described above, the representative color selecting unit 130 of FIG. 12 can select colors having a certain degree of difference as the representative colors, so that the representative colors are not concentrated at a position close to each other in the color space and the color reproducibility is improved. improves.

(4) Display Color Selection Means 140 FIG. 14 is a functional block diagram of the display color selection means 140 shown in FIG. The display color selection means 140 corresponds to the second aspect of the present invention, and includes a reference area selection means 141 for selecting a reference area in a reference area selection process based on 24-bit image data Pin and outputting reference area data R1. , The image data Pin in the reference area display color selection processing based on the reference area data R1 and the 24-bit representative color data C2.
From the LUT data M and the n-bit color number data C to select the display color in the reference area and the display color selecting means 142 in the reference area.

Next, the display color selection processing by the display color selection means 140 will be described with reference to FIGS. 15 and 16. 15 and 16 are diagrams showing examples of reference areas.
First, the reference area selection means 141 inputs 24-bit image data Pin, selects an area including a representative color to be referred from the peripheral area of the area including the target pixel for display color selection, and then the reference area. The data R1 is sent to the reference area display color selection means 142. In the reference area selection process, for example, when the blocks are divided as shown in FIG. 15, 9 areas (including the target area) adjacent to the target area are used as the reference areas, or 25 neighboring areas are used as the reference areas. To do.
In the case of line division as shown in FIG. 16, three adjacent lines or five adjacent lines are selected as the reference area. In the reference area display color selecting means 142, the display color is selected by comparing the pixel value only with the representative color in the reference area selected by the reference area selecting means 141, and the LUT data M and the n-bit color number data C are selected. Is output. In this reference area display color selection processing, for example, the difference between the pixel values of all the representative colors in the reference area and the target pixel is calculated, and the representative color with the smallest difference is used as the display color of that pixel. As described above, the display color selection means 140 of FIG. 14 limits the representative colors to be referred to and compares only a part of the representative colors, so that the processing time is significantly shortened, and when it is realized by hardware. Can also reduce the memory capacity.

Other Embodiments of Limited Color Display Means Each of the means 110 to 140 in the limited color display means 100 of FIG.
Another embodiment of the present invention will be described with reference to FIGS. (1-1) Area dividing means 110 (FIG. 17) FIG. 17 is another functional block diagram of the area dividing means 110 shown in FIG. The area dividing means 110 corresponds to the fourth aspect of the invention, and determines the pixel value in the pixel value determination processing based on the 24-bit image data Pin and outputs the area data A1 or the pixel data PE1. It is composed of a means 113 and a pixel merging means 114 for carrying out a pixel merging process of the pixel data PE1. Next, the area division processing of the area division means 110 will be described with reference to FIG. FIG. 18 is a diagram illustrating an example of pixel merging. First, in the pixel value determination means 113, 24-bit image data Pin is input, pixel values of a plurality of adjacent pixels are compared, and it is determined whether to merge into the same region or different regions, and to merge. Supplies the pixel data PE1 to the pixel merging unit 114, compares the pixel data PE1 with another adjacent pixel when not merging, and outputs the area data A1 when there are no pixels to be merged. Pixel merging means 11
In 4, the pixel data PE1 supplied from the pixel value determination means 113 is merged, and the pixel value determination means 113 is merged with P.
Supply E1 pixel data. In this pixel merging process, for example, when pixels are arranged as shown in FIG. 18, the threshold value is set to 5
As a result, pixels having a difference of 5 or less as compared with adjacent pixels on the left, right, top and bottom are merged into one area. In the case of FIG. 18, 4
It is divided into two areas. As described above, since the area dividing unit 110 in FIG. 17 includes pixels having similar colors in the same area, it does not need to perform processing such as merging reference areas in the representative color selection and the display color selection. , You can choose a suitable color.

(2-1) Representative Color Number Determination Means 120 (FIG. 19) FIG. 19 is another functional block diagram of the representative color number determination means 120 shown in FIG. The representative color number determining means 120
Since it corresponds to the seventh invention, a color dispersion calculating unit 125a that calculates the color dispersion data CD1 by the color dispersion calculation processing based on the area data A1 and the 24-bit image data Pin.
A representative color number calculating means 125b for calculating the representative color number data C1 by the representative color number calculating process based on the color dispersion data CD1 and the image data Pin. Next, the representative color number determination processing of the representative color number determination means 120 will be described with reference to FIG.
Will be described with reference to. FIG. 20 is a diagram showing an example of the conversion function. First, in the color variance calculating means 125a, the area data A1 and the 24-bit image data Pin are input, and, for example, the variance (the sum of the square of the difference between the average and the pixel value) is used to calculate the color statistics in the area. And calculating the color dispersion data CD1 as the representative color number calculating means 125.
supply to b. The 24-bit image data Pin is input to the representative color number calculating means 125b, and the color variance calculating means 12 is inputted.
By referring to the color dispersion data CD1 supplied from 5a, a large number of colors is allocated to a large dispersion area, a small number of colors is allocated to a small dispersion area, and the like.
The assigned color number, that is, the calculated color number is output as the representative color number of the area in the form of representative color number data C1. In this representative color number calculation process, for example, the variance value is set to 1 as shown in FIG.
This can be realized by preparing a simple conversion function that is simply divided by 00 and calculating the number of colors. As described above, FIG.
The representative color number determination unit 120 of 9 reduces the number of colors in the gently changing portion and allocates many colors to the rapidly changing portion to improve the color reproducibility of the notable portion. , The impression received from the entire image is also improved.

(2-2) Representative Color Number Determining Unit 120 (FIG. 21) FIG. 21 is another functional block diagram of the representative color number determining unit 120 shown in FIG. The representative color number determining means 120
This is equivalent to the eighth invention, and color dispersion calculation means 125a for performing the same color dispersion calculation processing as in FIG. 19, and color dispersion data CD1, 24 output from the color dispersion calculation means 125a.
Mean value comparison means 126a for comparing the average values in the average value comparison processing based on the bit image data Pin and the reference area data R1 and outputting the minimum difference data MD, the color dispersion data CD1, the image data Pin and the minimum difference. And a representative color number calculating means 126b for calculating the representative color number data C1 based on the data MD. Next, the representative color number determination processing of the representative color number determination means 120 will be described. First, in the color variance calculating means 125a,
Similar to FIG. 19, the statistical dispersion of colors in the area is calculated from the area data A1 and the 24-bit image data Pin, and the calculated color dispersion data CD1 is used as the average value comparison means 126a and the representative color number calculation means 126b. Supply to. In the average value comparing means 126a, the reference area selecting means 1 of FIG.
The average value of the area of the reference area data R1 output from 41 is calculated, compared with the average value of the target area, and the minimum difference data MD is supplied to the representative color number calculating means 126b. In this average value comparison process, for example, when the reference area is adjacent 9 blocks as shown in FIG. 15, the difference between the average values of the target area and the other 8 blocks is calculated, and the minimum difference data MD having the smallest difference is calculated. Output. In the representative color number calculating means 126b, as shown in FIG.
9 has the same function as the representative color number calculation unit 125b, the color distribution supplied from the color dispersion calculation unit 125a is small, and the minimum difference data MD supplied from the average value comparison unit 126a is smaller than the threshold value. Number of representative colors is 0
Also, it has a function of outputting in the form of the representative color number data C1. Therefore, in this representative color number calculation process, for example,
When the color dispersion threshold is 20 and the average value difference threshold is 30, the number of representative colors is set to 0 for the region where the color dispersion is 20 or less and the average value difference is 30 or less. However, when the number of representative colors in the reference area becomes all 0, there is no representative color to be compared when selecting the display color, so the number of representative colors in the adjacent area is all 0.
In the case of, the number of representative colors is exceptionally set to 1 and output. As described above, the representative color number determining unit 120 of FIG. 21 further improves the color reproducibility of other portions by further reducing the number of colors assigned to the gently changing portions.

(2-3) Representative Color Number Determination Means 120 (FIG. 22) FIG. 22 is another functional block diagram of the representative color number determination means 120 shown in FIG. The representative color number determining means 120
This corresponds to the ninth aspect of the invention and includes area data A1 and 24.
DCT calculation means 1 for performing the DCT calculation process based on the bit image data Pin and outputting the conversion coefficient data CC1
27a, the conversion coefficient data CC1 and the image data Pi
Representative color number data C is calculated by performing a representative color number calculation process based on n.
And a representative color number calculating means 127b that outputs 1. Next, the representative color number determination processing of the representative color number determination means 120 will be described. First, the DCT calculation means 127a uses the area data A1 and the image data Pi of the square block.
n is input, DCT is performed on the image data Pin to calculate the conversion coefficient, and the conversion coefficient data CC1 is supplied to the representative color number calculation means 127b. Where DCT is
Although it may be performed for all the data (R, G, B, etc.) of the image data Pin, a good result can be obtained by performing the calculation only for the luminance data. Representative color number calculation means 12
In 7b, the conversion coefficient data CC1 and the 24-bit image data Pin are input, and the number of colors is assigned with reference to the conversion coefficient data CC1. That is, when the AC component of the conversion coefficient is large, the number of colors is large because the color variation is large, and when the AC component is small, the number of colors is small.
The number of colors thus assigned is output as the representative color number data C1. This representative color number calculation processing is realized, for example, by summing all the conversion coefficients of the AC components and dividing the sum by a threshold value to obtain the number of colors. As described above, FIG.
In the representative color number determining unit 120, the change of the pixel value in the area can be obtained in more detail than the variance, so that more appropriate processing can be performed.

(2-4) Representative Color Number Determination Unit 120 (FIG. 23) FIG. 23 is another functional block diagram of the representative color number determination unit 120 shown in FIG. The representative color number determining means 120
This is equivalent to the tenth invention and includes area data A1 and area data A2.
Simplified DCT that performs a simplified DCT calculation process based on 4-bit image data Pin and outputs transform coefficient data CC2
The calculation unit 128 and the same representative color number calculation unit 1 as in FIG.
And 27b. Next, the representative color number determination processing of the representative color number determination means 120 will be described with reference to FIG. FIG. 24 is a diagram showing an example of simplified DCT calculation. First, in the simplified DCT calculation means 128, the area data A1 and the 24-bit image data Pin are input, only a small size conversion coefficient is calculated regardless of the block size of the image, and the calculated conversion coefficient data CC2 is represented. It is sent to the color number calculation means 127b. The representative color is a color that is representative of the color in the region, but a color that changes considerably finely as a source of the high frequency component of the DCT coefficient is insensitive to the human eye, and thus has low reproducibility. good. Therefore, it is possible to calculate the number of colors only by the coefficient of the low frequency component. For example, an 8 × 8 conversion coefficient is normally obtained from the 8 × 8 image data Pin. However, as shown in FIG. 24, if there is a 4 × 4 coefficient (inside the thick frame in the figure) for color number calculation, it is sufficient for color number determination. Therefore, by calculating only this part,
A simple DCT calculation process is performed. Representative color number calculation means 1
In 27b, the representative color number data C1 is output in the same manner as in FIG. As described above, the representative color number determining unit 1 in FIG.
In 20, the processing time is shortened and the scale of the hardware configuration is reduced by calculating only a part of the coefficients.

(2-5) Representative Color Number Determination Means 120 (FIG. 25) FIG. 25 is another functional block diagram of the representative color number determination means 120 shown in FIG. The representative color number determining means 120
The DCT calculating means 129a, which corresponds to the eleventh invention and has the same function as the DCT calculating means 127a in FIG. 22 or the simplified DCT calculating means 128 in FIG. 23, and the DCT calculating means 129a.
Transform coefficient data CC3 output from the calculation means 129a
A conversion coefficient merging unit 129b that performs conversion coefficient merging processing for the above, and a representative color number calculation process that performs representative color number calculation processing based on the output of the conversion coefficient merging unit 129b and the 24-bit image data Pin The color number calculation means 129c is included. Next, the representative color number determination processing of the representative color number determination means 120 will be described with reference to FIGS. 26 and 27. FIG. 26 is a diagram showing an example of conversion coefficient merging, and FIG. 27 is a diagram showing an example of representative color number calculation. First, in the DCT calculation means 129a, the area data A1 and the 24-bit image data Pin are input, and the DC in FIG.
T calculation means 127a or simplified DCT calculation means 1 of FIG.
The DCT is calculated in the same manner as 28, and the calculated conversion coefficient data CC3 is supplied to the conversion coefficient merging means 129b. The transform coefficient merging unit 129b inputs the transform coefficient data CC3, adds the transform coefficients having similar frequencies, and groups them into three groups as shown in FIG. 26, for example. When this DCT calculation means 129a is configured by the simplified DCT calculation means 128 of FIG. 23, it is only in the thick frame and can be realized by setting 10 coefficients to the third group. Further, when the DCT calculating means 129a is configured by the DCT calculating means 127a in FIG. 22, it can be realized by setting 58 coefficients in the third group. Representative color number calculation means 129
In c, the output of the conversion coefficient merging unit 129b and the 24-bit image data Pin are input, the number of colors is calculated using the total coefficient of each group as a parameter, and the representative number of colors is represented in the form of representative color number data C1. Output. In this representative color number calculation process, for example, the coefficients in each group are summed, and the number of colors is calculated by dividing by a threshold value set for each group or compared, and the total number of colors in all groups is represented. The number of representative colors C1 is determined and output as the color.

Here, the coefficients of the first group in FIG.
It represents the magnitude of the ultra-low frequency component. If this value is large, the image changes greatly globally, so it is sufficient to assign a certain number of colors. The coefficient of the third group represents the magnitude of the high frequency component, and if this value is large, the image changes drastically. Therefore, when the coefficient of this group is large, many colors may be assigned. A concrete example thereof is shown in FIG. In FIG. 27, the thresholds of the respective groups are set to 80, 10, and 100, the number of colors is set to 1 in the first group and the second group when the threshold is exceeded, and the result of dividing the third group by the threshold is set as the number of colors. There is. In addition, since the total number of colors may be 0, the number of representative colors is 1 added to this total. As described above, the representative color number determining unit 120 of FIG. 25 can handle global changes and local changes of pixels by grouping for each frequency, and the optimum number of colors for images of various characteristics. Can be assigned.

(2-6) Area dividing means 110 / representative color number determining means 120 (FIG. 28) FIG. 28 is a functional block diagram in which the area dividing means 110 and the representative color number determining means 120 shown in FIG. 5 are combined. . The area dividing means 110 and the representative color number determining means 120
Corresponding to the invention of 24 bits of 24-bit image data Pi
initial area dividing means 115 for outputting area data A3 by the initial area dividing processing based on n, and the area data A3
And the number-of-regions determination unit 116 that performs the number-of-regions determination process based on the limited number of colors N and outputs the region data A1 and the representative number-of-colors data C1, and the color variance calculation based on the output of the region number determination unit 116 and the image data Pin. Functions equivalent to the color dispersion calculation means 117 that performs processing and outputs the color dispersion data CD2, the division area selection means 118 that inputs the color dispersion data CD2 and performs division area selection processing, and the initial area division means 115. The image data Pin and the output of the divided area selection unit 118 are re-divided to obtain area data A.
And a re-dividing means 119 for outputting 3.

Next, the area division processing and the representative color number determination processing of FIG. 28 will be described with reference to FIG. FIG. 29 shows
It is a figure which shows the example of initial area division. First, the initial area dividing unit 115 inputs the 24-bit image data Pin,
The image is initially divided into several areas, and the divided area data A3 is supplied to the area number determination means 116.
For example, in the division example of FIG. 29, the original image is divided into four. In the area number determination means 116, the area data A3 and the limited color number N are input, and the number of divided areas and the limited color number N are input.
And (for example, the number of regions is compared with N / i (where i is a threshold value)) to determine whether or not to further divide. In the case of further division, the area data A1 is supplied to the color variance calculating means 117. If there is no need to divide, the area data A1 is output, and the threshold value i is further set to the representative color number data C1.
Output as. Thereby, the representative color number determining means can be omitted. In this area number determination processing, for example, when the limited color number N is 256 and the threshold value i is 4, the number of areas is 64.
In the following cases, it is determined that further division is necessary. The color variance calculation means 117 inputs the area data A1 and the 24-bit image data Pin from the area number determination means 116, calculates the distribution of the color distribution of the already divided areas, and calculates the calculated color dispersion. The data CD2 is supplied to the divided area selection means 118. For example, the color dispersion of each area is obtained from the sum of squares of the difference between the average value and each pixel value. The division area selection means 118 inputs the color dispersion data CD2, selects an area having a large color dispersion, and supplies it to the re-division means 119. The subdivision means 119 is a division area selection means 118.
It has a function of re-dividing the area selected in step 1 into smaller areas. For example, the area having the largest chromatic dispersion is divided into four as shown in FIG. Here, the initial area dividing unit 115 and the re-dividing unit 119 have the same functions, so they may be combined into one. As described above, the means of FIG.
The process of determining the number of representative colors can be omitted, which simplifies the process.

(3-1) Representative color selection means 130 (see FIG. 3)
0) FIG. 30 is another functional block diagram of the representative color selection unit 130 shown in FIG. The representative color selection means 130 is the first
Which corresponds to the third aspect of the invention, and has 24-bit image data P
image deletion means 135 that performs image deletion processing based on in and reference area data R1 and outputs image data P; and representative color calculation processing based on the image data P and representative color number data C1 The representative color calculation unit 136 that outputs the data C2. next,
The representative color selection processing of the representative color selection means 130 is shown in FIG.
Will be described with reference to. FIG. 31 is a diagram showing a reference area. First, the pixel deleting means 135 inputs the 24-bit image data Pin and the reference area data R1 and deletes a pixel having a color close to the representative color of the reference area from candidates for selecting the representative color of the area. The image data P of the above is supplied to the representative color calculating unit 136. In this pixel deletion processing, for example, when adjacent 9 blocks are used as reference areas as shown in FIG. 15, the representative color of 8 blocks other than the target area is compared with the target pixel. Pixels having a difference smaller than the threshold value (pixels of almost the same color) have already become the representative color in the reference area even if they are not newly selected as the representative color, so this target pixel is deleted from the representative color candidates. However, in the pixel deletion processing, the representative color may not yet be selected in the reference area depending on the order of processing, and at that time, only the already selected representative color is compared. For example, when processing is performed in the raster direction, as shown in FIG.
Since the representative colors have been selected in only four areas, only the representative colors in these four areas are compared. The representative color calculating unit 136 inputs the image data P and the representative color number data C1 and calculates the representative color by outputting the representative color in order from the color with the most appearance pixels, and outputs the representative color data C2. As described above, the representative color selection unit 130 of FIG. 30 avoids selecting a similar color as the representative color, so that the representative color is not concentrated and the color reproducibility is improved.

(4-1) Display color selecting means 140 (see FIG. 3)
2) FIG. 32 is another functional block diagram of the display color selection means 140 shown in FIG. This display color selection means 140 is the first
The invention corresponds to the fourth aspect of the invention, and includes a reference area merging unit 143 that performs reference area merging processing based on the area data A4 and 24-bit image data Pin to output pixel data PE2 and reference area data R2, and the pixel data PE2.
Pixel average comparison means 144 for performing a pixel average comparison process based on the above and outputting a merge request MQ, and the reference area data R
2. Display color calculation means 145 that performs display color calculation processing based on image data Pin and representative color data C2 and outputs LUT data M and n-bit color number data C. Next, the display color selection processing of the display color selection means 140 will be described with reference to FIG. FIG. 33 is a diagram showing an example of reference area merging. First, the reference area merging unit 143 inputs the area data A4 and 24-bit image data Pin, merges all peripheral pixels to be referred to when selecting a display color as a reference area, and averages the image data PE2 of the neighboring area with the pixel average. It is supplied to the comparison means 144, and when the merge request MQ comes back, that area is merged. In this reference area merging process, it is determined whether or not to merge all areas adjacent to the reference area. The pixel average comparison unit 144 inputs the pixel data PE2, compares the averages of the pixel values of the supplied two regions, and when the difference is smaller than the threshold value, sends a merge request MQ to the reference region merge unit 143. For example, when regions having an average value are distributed as shown in FIG. 33, the average values are compared with the upper, lower, left, and right regions (threshold value 10 in the figure), and the regions are merged to form a reference region. However, depending on the processing direction, the region in the lower right direction has not yet been selected as a representative color, so these regions are not referred to. In the display color calculation means 145,
The reference area data R2 from the reference area merging means 143,
24-bit image data Pin and representative color data C2
, The pixel is compared with all the representative colors in the reference area, the closest color is used as the display color, and the LUT data M and the n-bit color number data C are output. As described above, the display color selection means 140 of FIG. 32 can refer to only the representative colors of the close colors without waste, so that the optimum display color can be efficiently selected.

[0044]

As described in detail above, according to the first invention, the original image is divided into partial areas, and the characteristic color from each area is used as the representative color. Reproducibility is improved for a characteristic color that has a low appearance frequency that appears in a bad part. That is, the reproducibility is improved with respect to a color having a large local change, thereby improving the impression received from the entire image. According to the second aspect of the invention, the representative color to be referred to is limited and only a part of the representative colors are compared, so that the processing time can be greatly shortened and the memory capacity can be reduced even when it is realized by hardware. According to the third aspect of the present invention, each processing can be performed using the same threshold value in each area regardless of the size of the image and the number of limited colors. According to the fourth aspect, by including pixels having similar colors in the same area, it is possible to select an optimum color without performing processing such as merging reference areas in selecting representative colors and display colors. it can. According to the fifth aspect, the process of determining the number of representative colors can be omitted, which simplifies the process. According to the sixth invention, in the representative color number determination process in each area, it is not necessary to handle the information of the entire image at all, so that the process is simplified and the memory amount is reduced even when it is realized by hardware. it can. 7th
According to the invention, by reducing the number of colors in the part that changes smoothly and allocating a large number of colors to the part that changes drastically, the color reproducibility of the part to be noticed is improved, and The impression you receive will also improve.

According to the eighth aspect of the invention, the color reproducibility of the other portions is improved by further reducing the number of colors assigned to the gently changing portions. According to the ninth aspect, the state of the change in the pixel value in the area can be obtained in more detail than the variance, and thus more appropriate processing can be performed.
According to the tenth invention, by calculating only some of the coefficients, the processing time is shortened and the scale of the hardware configuration is reduced. According to the eleventh aspect, it is possible to cope with a global change and a local change of pixels by grouping them by frequency, and it is possible to assign an optimum number of colors to an image having various properties. According to the twelfth invention, it is possible to select colors having a certain degree of difference as representative colors.
The representative color is not concentrated in a short distance in the color space, and the color reproducibility is improved. According to the thirteenth aspect, since it is avoided to select a similar color as the representative color, the representative color is not concentrated and the color reproducibility is improved. According to the fourteenth invention, since it is possible to refer to only the representative colors of close colors without waste,
You can efficiently select the optimum display color.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a color image display method and a color image display system according to an embodiment of the present invention.

FIG. 2 is a configuration block diagram of a conventional general color image display system.

FIG. 3 is a configuration block diagram of a color image display system according to a conventional limited color display algorithm.

FIG. 4 is a functional block diagram of limited color display means in FIG.

5 is a functional block diagram of limited color display means in FIG.

FIG. 6 is a diagram showing an example of block division.

FIG. 7 is a diagram showing an example of line division.

FIG. 8 is a diagram showing an example of a correspondence table.

9 is a functional block diagram of area dividing means in FIG.

10 is a functional block diagram of a representative color number determining unit in FIG.

FIG. 11 is a diagram showing an example of a proportional function of inclination.

12 is a functional block diagram of a representative color selecting unit in FIG.

FIG. 13 is a diagram illustrating an example of neighboring pixel merging.

FIG. 14 is a functional block diagram of display color selection means in FIG.

FIG. 15 is a diagram showing an example of a reference area.

FIG. 16 is a diagram showing an example of a reference area.

17 is another functional block diagram of the area dividing means in FIG.

FIG. 18 is a diagram illustrating an example of pixel merging.

FIG. 19 is another functional block diagram of the representative color number determining means in FIG.

FIG. 20 is a diagram showing an example of a conversion function.

FIG. 21 is another functional block diagram of the representative color number determining unit in FIG.

FIG. 22 is another functional block diagram of the representative color number determining unit in FIG.

FIG. 23 is another functional block diagram of the representative color number determination means in FIG.

FIG. 24 is a diagram showing an example of simplified DCT calculation.

FIG. 25 is another functional block diagram of the representative color number determination means in FIG.

FIG. 26 is a diagram showing an example of transform coefficient merging.

FIG. 27 is a diagram illustrating a representative color number calculation example.

28 is another functional block diagram of the area dividing unit / representative color number determining unit in FIG.

FIG. 29 is a diagram showing an example of initial area division.

FIG. 30 is another functional block diagram of the representative color selecting means in FIG.

FIG. 31 is a diagram showing an example of a reference area.

FIG. 32 is another functional block diagram of the display color selecting means in FIG.

FIG. 33 is a diagram showing an example of reference area merging.

[Explanation of reference numerals] 100 limited color display means 110 area dividing means 111 area number determining means 112 small area dividing means 113 pixel value judging means 114 pixel merging means 115 initial area dividing means 116 area number judging means 117 color variance calculating means 118 division Region selecting means 119 Re-dividing means 120 Representative color number determining means 121 Representative color number tentative determining means 122 Color number comparing means 123 Weight calculating means 124 Color number adjusting means 125a Color variance calculating means 125b Representative color number calculating means 126a Average value comparing means 126b Representative color number calculation means 127a DCT calculation means 127b Representative color number calculation means 128 Simplified DCT calculation means 129a DCT calculation means 129b Conversion coefficient merge means 129c Representative color number calculation means 130 Representative color selection means 131 Neighborhood pixel merge means 132 Group number judgment Setting means 133 Pixel group sorting means 134 Representative color calculating means 135 Pixel deleting means 136 Representative color calculating means 140 Display color selecting means 141 Reference area selecting means 142 Reference area display color selecting means 143 Reference area merging means 144 Pixel average comparing means 145 Display color calculation means 200 Frame buffer 300 Look-up table (LUT) 400 D / A converter 500 CRT

Claims (14)

[Claims] 1. A color image display method of creating a look-up table storing color data and displaying a color image by referring to the look-up table; area division processing for dividing an original image into partial areas; A representative color number determination process of determining a representative color number to be selected from each of the partial regions by referring to the color distribution and the size of the regions in the plurality of partial regions divided by the region dividing process, and from each of the partial regions A representative color selection process of selecting representative colors by the number of representative colors is executed, and display colors of all pixels are selected from the representative colors selected in the representative color selection process by the display color selection process to create the lookup table. A color image display method characterized by creating. 2. The color image display method according to claim 1, wherein the display color selection processing selects a display color only from the representative colors of the area including the target pixel and the peripheral several areas. 3. The area dividing process determines the number of areas using the limited number of colors N as a parameter, determines the size of the image and the size of each area from the number of areas, and then divides the original image into partial areas. The color image display method according to claim 1 or 2, characterized in that. 4. The color image display method according to claim 1, wherein in the area division processing, areas in which a difference in pixel value between adjacent pixels is smaller than a threshold value are combined and divided into one area. 5. In the area dividing process, first, the area is divided into a plurality of large areas, and when the color distribution in the area is large,
It is characterized in that the partial region is further divided into 1 / i (where i is a threshold value) of the limited number of colors N, and in the representative color number determination process, i number of representative colors is assigned to each region. The color image display method according to claim 1 or 2. 6. In the representative color number determination process, first, a representative color number is provisionally determined in each area, and then the total of the provisional representative color numbers of all areas is compared with a limited color number N. 3. The color image display method according to claim 1, wherein the true representative color number of the area is determined. 7. The color according to claim 6, wherein in the representative color number determination processing, the number of representative colors is reduced if the color distribution in the area is dense, and the representative color is increased if the color distribution is dispersed. Image display method. 8. In the representative color number determination process, in a region where the density of color distribution is higher than a threshold value and the difference in the average pixel value from the surrounding region is small, the representative color is not selected from that region. 8. The color image display method according to claim 7, wherein. 9. In the area division processing, M × M (however,
M; the size of the image block), and in the representative color number determination process, discrete cosine transform is performed on the image data of each block, and the AC component of the transform coefficient is used as a parameter. 9. The color image display method according to claim 7, wherein the number of representative colors is determined by allocating a large number of representative colors when the value is large and allocating a small number when the value is small. 10. The number of low frequency components (m × m) regardless of the size M of the image block when the number of representative colors is determined by using the transform coefficient of the discrete cosine transform in the number of representative colors determination process. , M ≧ m) is calculated, and the number of representative colors is determined by using only the coefficient of the low frequency component as a parameter. 11. In the representative color number determination process, when the representative color number is determined by using the transform coefficient of the discrete cosine transform, those having similar frequencies are grouped into several groups, and 11. The color image display method according to claim 9, wherein the number of representative colors in the area is calculated using the total of the conversion coefficients as a parameter. 12. In the representative color selection processing, pixels whose distances are closer than a threshold value k in a color space are grouped into several groups until the number of groups becomes less than several times the number of representative colors assigned to the area. 3. The method according to claim 1, wherein k is increased, regrouping is performed, the groups are arranged in descending order of the number of constituent pixels, a group having a representative color number is selected from the top, and a representative color is selected from each group. The color image display method described. 13. The representative color selection process excludes pixels having a color space distance closer than a threshold value from a representative color candidate as compared with the representative color of a peripheral area, and selects the representative color from the remaining pixels in the area. 3. The color image display method according to claim 1, wherein the color image display method is selected. 14. In the display color selection processing, all peripheral areas having a difference smaller than a threshold value compared with an average pixel value of an area including a target pixel are set as reference areas, and displayed from a representative color of the reference areas. The color image display method according to claim 2, wherein a color is selected.