JPH08272345A - Image processor - Google Patents

Abstract

PURPOSE: To reduce a sense of incongruity on a display screen by transforming an input color image to a certain extent to an image of a similar color tone to an original one. CONSTITUTION: A color space dividing part 2 equally divides a color space of an input color image into divided color spaces of a palette register number. Computing part 3 for frequency of use determines the frequency of use for each color data used for the input color image. A divided space managing part 4 manages the number of color data in the divided color spaces and the frequency of use in the divided color space for every divided color space. Redividing part 5 selects the most frequently used color space in the case where there are some divided color space where no color data exists, repeats redividing processing of the selected color space on the basis of the color data number in the divided color space, and re-sets the selected color space to the divided color space of the palette register number. A representative color deciding part 6 determines a representative color data for each of the finally divided color spaces and register this decided color data in a palette 27.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention refers to all color pixels forming an input color image, selects a number of color data items that can be simultaneously output and is registered in a palette, and uses the selected color data items. The present invention relates to an image processing device that performs conversion processing of all color pixels of the input color image.

[0002]

2. Description of the Related Art Conventionally, a computer which outputs a color display on a CRT display usually has a look-up table (palette) built therein, and this palette is used for VRAM.
The information is converted into the color information that appears on the information monitor. Here, in order to display and output the full-color (about 16.77 million colors) input color image as it is on the CRT display, the VRAM needs to be configured with 24 bits per pixel. That is, the VRAM has information for each color element R, G, B, and each color element is represented by 8 bits.

However, when a full-color input color image is displayed and output as it is on a CRT display, since the information amount per pixel is 24 bits, the information amount per input color image is large, and as a result, This leads to problems such as an increase in the size of the storage device and an increase in the transfer time of image data.

Therefore, in general, the number of color data is limited to 256 colors and the amount of information per pixel is set to 8 bits to reduce the amount of information per one input color image.

However, it is impossible to faithfully reproduce the original input color image only by this method.

Therefore, in order to display and output an image that is more faithful to the original input color image, the frequency of use of the color data appearing in the input color image is calculated, and the color data of a predetermined number, for example 256 colors, is calculated in descending order of frequency of use. To the palette,
Each pixel of the input color image is converted into color data registered in the palette.

However, in this case, the color data that is not frequently used is not registered in the palette and is ignored, so that, for example, small dots in the input color image cannot be reproduced, and the reproducibility of the image is poor. there were.

In order to solve this problem, for example, in Japanese Patent Laid-Open No. 3-171374, a full color color space is divided into 256 divided color spaces in advance, and color data does not exist in the divided color space. Deletes the divided color space that does not have color data, divides the most frequently used divided color space into two equal parts, and makes all divided color spaces have color data.
A color conversion method is described in which six divided color spaces are generated and representative color data is registered in a palette for each of the generated divided color spaces.

According to this color conversion method, the color data of the color pixels that are rarely used is also registered in the palette, so that a display similar to the color tone of the input color image can be performed to some extent, and the discomfort on the display screen is reduced. be able to.

[0010]

However, in Japanese Patent Laid-Open No. 3-171374, the divided color space that is frequently used is simply divided into two equal parts without considering the number of color data in the divided color space. Since a new divided color space is generated by the above, there may be an extreme difference in the number of color data between the divided color spaces that are finally divided, and an image that faithfully reproduces the input color image may appear. There was a problem that it could not be converted.

Therefore, the present invention eliminates such problems,
When generating the divided color space for determining the color data to be registered in the palette, divide the input color image into divided color spaces that consider both the frequency of use of color pixels in the divided color space and the number of color data, An object of the present invention is to provide an image processing apparatus that can convert an input color image into an image similar in color tone to some extent and reduce the discomfort on the display screen.

[0012]

The present invention refers to all color pixels forming an input color image, selects a number of color data items that can be simultaneously output and is registered in a palette, and selects the selected color data items. In an image processing device that uses the conversion processing of all color pixels of the input color image, a color space dividing unit that divides the color space of the input color image into the number of divided color spaces that can be simultaneously output, and Calculating means for obtaining the frequency of use for each color data; managing means for managing the number of color data in the divided color space and the frequency of use of color pixels for each of the divided color spaces; and a divided color space managed by the managing means. When there is a divided color space in which color data does not exist, a divided color space that is used most frequently is selected, and the selected divided color space is redivided based on the number of color data in the divided color space. Repeat, color data exists Re-dividing means for resetting to the simultaneous output possible number of divided color spaces and representative color data for each of the simultaneous output possible number of divided color spaces reset by the re-dividing means, and the decided color data And a registration means for registering in the pallet.

[0013]

In the present invention, first, the color space dividing means divides the color space of the input color image into the number of divided color spaces which can be simultaneously output. The calculation means obtains the frequency of use for each color data used for the input color image. Then, the management unit manages the number of color data in the divided color space divided by the color space dividing unit and the frequency of use in the divided color space for each divided color space. When there is a divided color space in which color data does not exist in the divided color space managed by the management unit, the re-dividing unit selects the divided color space with the highest frequency of use, and selects the selected divided color space as the divided color space. The process of re-division based on the number of color data in is repeated, and the color data is reset to the number of divided color spaces in which simultaneous output is possible. Then, the registration means determines representative color data for each final divided color space reset by the re-division means,
The color data thus determined is registered in the palette. And
The input color image is converted into color data registered in the palette.

As a result, the finally generated divided color space takes into consideration the number of color data which is used only for the frequency of use.

[0015]

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the arrangement of an image processing apparatus according to an embodiment of the present invention. In FIG. 1, the image processing apparatus includes a control unit 1, a scanner 10 via a scanner controller C1, a magnetic disk 11 via a magnetic disk controller C2, a keyboard 12 including a mouse 13,
Display 1 via display controller C3
4, and an optical disk 15 via an optical disk controller C4 is connected to the system bus B.

The control unit 1 uses the CP as an actual physical means.
The color space division unit 2, the usage frequency calculation unit 3, the division space management unit 4, the subdivision unit 5, the representative color determination unit 6, the conversion unit 7, the table group 8 and And input color image data 9. Each of these functions is achieved by various programs stored in the magnetic disk 11 in advance, and is read when the image processing apparatus is started and various processing is executed.

The input color image data 9 is transferred to the scanner 10
Is the full-color input color image read in or the data of the full-color input color image stored in the optical disc 15 and is the target data for the image conversion processing.

The color space dividing unit 2 sets the color space of the input color image temporarily stored as the input color image data 9 to the number that can be registered in the palette in the display controller C3, that is, the number that the display 14 can simultaneously output. Division is performed, and the division result is stored in the division color space management table 8a in the table group 8.

The use frequency calculation unit 3 determines all color pixels used in the input color image data 9 as color data (R, G,
The usage frequency is calculated for each image code indicated by B), and the calculation result is stored in the usage frequency management table 8b in the table group 8.

The divided space management unit 4 generates a divided space usage management table 8c for managing the number of color data and usage frequency for each divided color space based on the divided color space management table 8a and the usage frequency management table 8b. To do.

The subdivision unit 5 deletes the divided color space in which the number of color data is "0" in the divided space usage management table 8c,
The divided color space having the highest frequency of use is divided into two based on the number of color data as described later, thereby resetting the divided color space to the number of simultaneously outputtable divided color spaces. With the resetting of the divided color space, the divided color space management table 8a and the divided space usage management table 8c are updated.

The representative color determination unit 6 determines representative color data for each of the divided color spaces indicated by the finally updated divided space usage management table 8c, and the determined color data is used as a palette in the display controller C3. Sequentially register at 27.

The conversion unit 7 refers to the divided color space management table 8a for the color data of all the color pixels in the input color image data 9, determines the corresponding divided color space, and corresponds to this divided color space. It is converted into a divided space identifier (corresponding to a palette number) and transferred to the image memory 21.

The display controller C3 sequentially reads the divided space identifiers in the image data stored in the image memory, converts them into corresponding color data in the palette, and outputs them to the display 14.

Here, the display 14 can output color data of 256 colors at the same time, and the palette 27 has 25 colors.
Color data of 6 colors can be registered. On the other hand, as described above, the input color image is full color, that is, about 16
It is expressed using color data of 770,000 colors.

Color space division by the color space division unit 2 will be described below with reference to FIGS. 2 and 3.

The color space division unit 2 divides the color space of the input color image shown in FIG. 2 (a) into 256 divided color spaces shown in FIG. 2 (b). That is, the color space of the input color image shown in FIG. 2A has two color elements of R, G, and B, respectively.
Since it is expressed by the 8th power of 256, that is, 256, 256
X256 x 256 (about 16.77 million colors), but the color space division unit 2 sets R, G, and B color elements to 8, 8 and
It is divided into four regions, and as a result, divided into a plurality of divided color spaces of 8 × 8 × 4 = 256. The divided color spaces are all rectangular parallelepipeds of the same shape.

Here, FIG. 3 shows the data structure of the divided color space management table 8a, and the divided color spaces divided by the color space dividing unit 2 are defined. That is, the color element R is divided into eight equal parts, and the range of each divided color space becomes 32,
The color element G is also divided into 8 equal parts, and the range of each divided color space becomes 32. The color element B is divided into 4 equal parts and the range of each divided color space becomes 64.
Becomes Therefore, the divided color space to which each color data of the input color image belongs can be obtained by referring to the divided color space management table 8a. Each divided color space is identified by a divided space identifier.

Next, referring to FIG. 4, the usage frequency calculation unit 3
The data structure of the usage frequency management table 8b, which is the calculation result of, will be described. The use frequency management table 8b stores R, G, and B image codes used in the input color image, that is, the use frequency for each color data. The frequency of use for each color data is represented by the number of each color data that appears in the input color image, based on the color data of all color pixels used in the input color image data 9 by the frequency-of-use calculating unit 3. It For example, in FIG. 4, color data (R, G,
B) is (0,0,0), color data (0,0,0)
0) indicates that 10 images have appeared in the input color image.

Next, referring to FIG. 5, the division space management unit 4
The data structure of the divided space usage management table 8c generated by the above will be described.

The divided space usage management table 8c is shown in FIG.
For each divided color space, using the divided space identifier of the divided color space management table 8a, the number of color data indicating the type of color data appearing in the divided color space, and the color pixel appearing in the divided color space The number (the degree of use of the divided space) is stored. The number of color pixels in each divided color space is a cumulative value of the usage frequency of each color data existing in the divided color space. For example, FIG. 5 shows that there are three types of color data in the divided color space of the divided space identifier “1”, and the number of color pixels appearing in this divided color space is 44. .

An image conversion processing procedure in the image processing apparatus shown in FIG. 1 will be described with reference to FIG.

In FIG. 6, first, the color space division unit 2 divides the full-color color space into the number of palette registrations, that is, the number of divided color spaces that can be simultaneously output from the display 14,
The divided color space management table 8a is created (step 101). Next, the control unit 1 reads the input color image data into the control unit 1 (step 102). Here, the input color image may be a document image read by the scanner 10 as described above, or an image stored in the optical disc 15. Then, the usage frequency calculation unit 3
Calculates the number of color pixels appearing for each color data in the read input color image data 9 as the usage frequency (step 103). Further, the divided space management unit 4 determines which divided color space of the divided color spaces defined by the divided color space management table 8a the color data of the usage frequency management table 8b belongs to, thereby obtaining the input color image. For each color pixel, the number of color data for each divided color space and the usage of the color pixel for each divided color space are calculated, and the divided space usage management table 8c is generated (step 10).
4). Then, the divided space management unit 4 refers to the divided space usage management table 8c, and when there is a divided color space in which the number of color data is “0”, deletes the divided color space and also divides the divided color space. The management data of (step 1
05). Further, the divided space management unit 4 also deletes the divided color space of the divided color space management table 8a.

After that, the subdivision unit 5 refers to the division space usage management table 8c and determines that the number of division color spaces is 255.
It is determined whether or not the following (step 106). That is, it is determined whether the number of divided color spaces is less than the number that can be registered in the palette. As a result, when it is determined that the number of divided color spaces is equal to or less than 255, the divided color space with the highest usage of the color pixel for each divided color space in the divided space usage management table 8c is selected (step 10
7). For example, the divided space usage management table 8c in FIG.
From the value of the degree of use shown in, the divided color space of the divided space identifier “3” having the degree of use of “100” is selected.

Further, it is determined whether or not the divided color space selected in step 107 can be divided. Here, the case where the color data is not dividable is the case where the number of color data is “1”. For example, when the divided color space of the divided space identifier “3” shown in the divided space usage management table 8c of FIG. 5 is selected, the number of color data in this divided color space is “1”,
Judged as indivisible. If it is determined in step 108 that the divided color space can be divided, the divided color space is divided into two parts based on the number of color data, and the divided color space management table 8a and the divided space usage management table 8c are set. Update (step 109) and move to step 106.
Then, the above division color space division processing is repeated until the number of division color spaces reaches 256. The newly divided color space is also the selection target in step 107.

On the other hand, if it is determined in step 108 that the division is not possible, the division color space selected in step 107 is not divided and the division color space with the next highest degree of use is selected. Moving to 108, the divided color space with the next highest degree of use is sequentially selected until the division processing becomes possible. For example, in step 107, when the divided color space of the divided space identifier “3” in the divided space use degree management table 8c of FIG. 5 is selected, the divided space identifier “255” whose degree of use is “38” is changed to It is selected as a color space that is highly used.

Here, the dividing process of the divided color space in step 109 will be described. This dividing process does not divide the selected divided color space into a range simply indicated by the distance, but the selected divided color space. Divide the number of color data in space so that it can be divided into two. For example, as shown in FIG. 2 (c), four color elements R
When the selected divided color space having 1 to R4 is divided, the selected division is performed with the coordinate value P1 of the color element R so as to be divided into the color elements R1 and R2 and the color elements R3 and R4. Divide the color space. By performing the division of the divided color space based on the number of color data as described above, the color data is evenly or almost equally arranged in the two newly divided color spaces.

On the other hand, when the number of divided color spaces is not less than 255 in step 106, that is, when the number of divided color spaces is 256, the process proceeds to step 111, and the representative color determination unit 6 determines each divided color space. Then, one representative color data is determined. The representative color data may be determined by the representative color determination unit 6 by selecting the color data closest to the average value of the color data in consideration of the degree of use in the divided color space. Color data close to the average value of the color data may be selected. It is also possible to newly determine color data indicating the average value of the color data in consideration of the degree of use in the divided color space, and use this determined color data as the representative color data.

Thereafter, the representative color determination section 6 determines the determined 25
The six color data are registered in the palette 27 (step 1
12). Further, the conversion unit 7 obtains the color space of each color pixel of the input color image data, based on the color space management table 8a, into a color space to which the color pixel belongs, and a color space identifier ( Index),
The image data is transferred to the image memory 21 (step 113). Then, the display controller C3 refers to the index at a predetermined timing, converts the image data stored in the image memory 21 into color data registered in the palette 27, and outputs the color data displayed on the display 14 (step 114), and this image conversion processing ends.

Next, the detailed configuration of the display controller C3 will be described with reference to FIGS.

In FIG. 7, the display controller C3 has an image memory 21, a palette circuit 22, a D / A converter 23, a driver circuit 24 and a timing generation circuit 25.
2 and the timing generation circuit 25 are connected to the system bus B.

The image memory 21 stores the image data converted by the converter 7 via the system bus B.

The palette circuit 22 is connected to the palette 27 as shown in FIG. 8, and the image data from the image memory 21 via the system bus B is stored in the palette 27 based on the image data. Select, D / A
It has a selector circuit 26 for outputting to the converter 23. The palette 27 stores color data for each divided color space determined by the representative color determination unit 6 at an index position corresponding to the divided color space.

The D / A converter 23 is used in the pallet circuit 2
2 performs digital / analog conversion for each color element R, G, B of the color data output from
The analog signal for each color element converted by the / D converter 23 is converted into an input signal format in the display 13, and the converted signal is output to the display 13.

The timing generation circuit 25 controls the read control timing of the image data.

Specifically, when the color pixel data stored in the image memory 21 has the index "0", the timing generation circuit 25 controls the read timing by referring to the position data of the color pixel data. Then, the index “0” of the color pixel data is input to the palette circuit 22. Then, the selector circuit 26 sets the value 0R of the color element R and the value 0G of the color element G corresponding to the index “0”.
And the value 0B of the color element B are selected and output to the D / A converter 23 and converted into an analog signal, and further converted into an analog signal compatible with the input signal format of the display 13 by the driver circuit 24 and displayed on the display 13. Is output.

As described above, in this embodiment, when resetting the divided color space, not only the frequency of use of the divided color space but also the number of color data in the divided color space is taken into account for the division processing. Therefore, the object to be divided is a divided color space that is frequently used, and the number of color data in the divided color space becomes even, and even if the number of color data is limited, the input color image is The input color image can be converted into an image more similar in color tone.

[0049]

As described above in detail, in the present invention, the color space dividing means first divides the color space of the input color image into the number of divided color spaces which can be simultaneously output. The calculation means obtains the frequency of use for each color data used for the input color image. Then, the management unit manages the number of color data in the divided color space divided by the color space dividing unit and the frequency of use in the divided color space for each divided color space. When there is a divided color space in which color data does not exist in the divided color space managed by the management unit, the re-dividing unit selects the divided color space with the highest frequency of use, and selects the selected divided color space as the divided color space. The process of re-division based on the number of color data in is repeated, and the color data is reset to the number of divided color spaces in which simultaneous output is possible. Then, the registration unit determines representative color data for each final divided color space reset by the re-division unit, and registers the determined color data in the palette. Then, the input color image is converted into color data registered in the palette.

Therefore, the finally-generated divided color space has the number of color data equalizing the frequency of use taken into consideration, and the image to be converted and output is similar to the color tone of the input color image to some extent, and the limited color data is limited. There is an advantage that the discomfort caused by the conversion process by is reduced.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of an image processing apparatus that is an embodiment of the present invention.

FIG. 2 is a diagram illustrating a color space division process and a divided color space re-division process.

FIG. 3 is a diagram showing a data configuration of a divided color space management table 8a.

FIG. 4 is a diagram showing a data structure of a usage frequency management table 8b.

FIG. 5 is a diagram showing a data structure of a divided space usage management table 8c.

FIG. 6 is a flowchart showing an image conversion processing procedure.

FIG. 7 is a diagram showing a detailed configuration of a display controller C3.

FIG. 8 is a diagram showing a detailed configuration of a pallet circuit 22.

[Explanation of symbols]

1 ... Control unit 2 ... Color space division unit 3 ... Usage frequency calculation unit
4 ... Division space management section 5 ... Redivision section 6 ... Representative color determination section 7 ... Conversion section 8 ... Table group 8a ... Division color space management table 8b ... Usage frequency management table 8c ... Division space usage management table 9 ... Input color Image data 10 ... Scanner 11 ... Magnetic disk 12 ... Keyboard 13 ... Mouse 14 ... Display 15 ... Optical disk 21 ... Image memory 27 ... Palette C1 ... Scanner controller C2 ... Magnetic disk controller C3 ... Display controller C4 ... Optical disk controller B ... System bus

Claims (1)

[Claims] 1. A reference is made to all color pixels forming an input color image to select a number of color data items that can be simultaneously output and registered in a palette, and the selected color data are used to select all color data items of the input color image. In an image processing device that performs color pixel conversion processing, a color space dividing unit that divides the color space of the input color image into the number of divided color spaces that can be simultaneously output, and the frequency of use for each color data of all the color pixels Calculating means; managing means for managing, for each of the divided color spaces, the number of color data and the frequency of use of color pixels in the divided color space; and a division in which no color data exists in the divided color space managed by the managing means. If there is a color space, the division color space with the highest frequency of use is selected, and the process of re-dividing the selected division color space based on the number of color data in the division color space is repeated,
Re-dividing means for resetting to the simultaneous output possible number of divided color spaces in which color data exists, and representative color data is determined for each simultaneous output possible number of divided color spaces reset by the re-dividing means, and the determination is made. An image processing apparatus comprising: a registration unit that registers the registered color data in the palette.