JP2002218500A - Subtractive color processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a subtractive color processor that can generate palette data without the use of a histogram memory, reduce the arithmetic amount and display a natural image. SOLUTION: The subtractive color processor is provided with an RGB color component histogram generating section 11 that converts a YUV color component obtained from a low frequency component extract section into an RGB component and counts its frequency distribution, a palette storage section 7 that stores an image pallet after the subtractive color processing, a palette generating section 6 that extracts the color component having much frequency distribution from the RGB color component histogram generating section 11 with priority and registers it to the palette storage section 7, a subtractive color processing section 8 that finds out the closest color from the palette storage section 7, converts it into a color code given to a selected palette color and stores the converted code into a subtractive color processing image frame memory, and the subtractive color processing image frame memory 9 that stores relation showing which color registered in the palette storage section 7 is closest to each pixel after the subtractive color processing while making the relation correspond to the pixel position of an original image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a color reduction process used in a display of a portable terminal device or the like having a limited calculation resource.

[0002]

2. Description of the Related Art In recent years, terminals for displaying images have rapidly become widespread. Such terminals limit the number of colors that can be displayed to about 256 colors in order to keep the cost of the device low,
Often saves memory for display. Despite these limitations, the display capability is sufficient for applications such as documents and tables. However, in order to display natural images such as photographs, images represented by about 24 bits are represented by 8 bits. For this purpose, it is necessary to perform a color reduction process, and the image quality is greatly affected by the color reduction method. Therefore, there is an increasing demand for realizing high image quality by reducing the amount of calculation and the memory used as much as possible.

[0003] As one of effective means of color reduction, there is an adaptive pallet system in which necessary 256 colors are selected from tens of thousands of colors and displayed. Hereinafter, a conventional color reduction processing device will be described.

FIG. 1 is a block diagram showing a configuration of a conventional color reduction processing apparatus.

[0005] As shown in FIG.
Decoder 2 for decoding compressed data 1 using DCT
The original image frame memory 3 as an image memory for recording full color data (color image data) expressed in full color output from the decoder 2, and the frequency of what color is referred to the original image frame memory A histogram generation unit 4 for counting whether or not the image is used in the histogram generation unit, a histogram memory 5 for recording the histogram generated by the histogram generation unit 4, and a color image display based on the histogram generated on the histogram memory 5. A palette generation unit 6 for selecting a palette color to be used and generating palette data p; a palette storage unit 7 for recording the correspondence between the palette data p output by the palette generation unit 6 and the corresponding color code; Full color stored in the original image frame memory 3 with the palette color stored in the storage unit 7 A color reduction unit 8 for approximating the data with the above-mentioned palette color and converting it into a color code represented by an 8-bit code, and a color reduction image frame memory 9 for storing the color code output from the color reduction unit 4. It is configured.

The operation of the conventional color reduction processing apparatus having the above-described components will be described with reference to the flowchart of FIG.

Step 2-1 First, the decoder 2 calculates color data of each pixel by variable length decoding, inverse quantization and inverse DCT, and stores the obtained color image data in the original image frame memory 3. save.

Step 2-2 Next, the histogram generator 4 generates a color histogram based on the color data stored in the original image frame memory 3, and stores the result in the histogram memory 5.

Step 2-3 Next, the palette data creating section 6 selects a palette color to be used for displaying a color image based on the histogram read out from the histogram memory 5, and sets the palette data p.
Is created and written in the pallet storage unit 7. That is, based on information on which colors are used and in what frequency in a given image, a palette that appropriately covers the color distribution in the given image is selected. Existing algorithms such as the “dynamic partitioning method” are often used as the palette selection algorithm.

Step 2-4 Next, the color reduction unit 8 retrieves the closest color from the palette data in the palette storage unit 7 using an existing algorithm, and approximates (quantizes) the color of each pixel. The approximated color is converted into a symbol (color code) and written into the reduced color image frame memory 9.

[0011]

As described above, since the conventional color reduction processing apparatus requires a histogram memory, the cost is increased due to the enormous memory capacity and the power consumption due to the calculation for generating the histogram is increased. Had the problem of occurring.

SUMMARY OF THE INVENTION An object of the present invention is to provide a color reduction processing apparatus which can reduce the amount of calculation and the power consumption while paying attention to the conventional problems.

[0013]

To achieve the above object, the present invention provides a decoder for expanding compressed data using DCT, and a DCT obtained when expanding compressed data from the decoder. A low-frequency component extracting unit for extracting a DC component and a low-order AC component of a coefficient; an original image frame memory for storing an expanded image output by the decoder; and a YUV obtained from the low-frequency component extracting unit. An RGB color component histogram generation unit that converts a color component into an RGB component and counts it as a frequency distribution, a palette storage unit for storing a palette of images after color reduction, and a frequency count based on the output of the RGB color component histogram generation unit. A palette generation unit that preferentially extracts a large number of color components and registers the extracted color components in the palette storage unit, and registers a color component of each pixel in the original image frame memory in the palette storage unit. A color reduction unit that compares the selected color component with the color components, finds the closest color from the palette storage unit, converts the selected color to a given color code, and records the color code in the color reduction image frame memory. A color reduction processing apparatus having a color reduction image frame memory for storing which color of each subsequent pixel is closest to which color registered in the palette storage unit in association with the pixel position of the original image.

According to the present invention, pallet data can be created without using a memory for storing a histogram, and the amount of calculation can be reduced by not performing processing for creating a histogram of the entire image. And a color reduction processing device capable of reducing power consumption can be obtained.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The invention according to claim 1 of the present invention provides a decoder for expanding compressed data using DCT,
A low-frequency component extraction unit for extracting a DC component and a low-order AC component of a DCT coefficient obtained when the compressed data is expanded from the decoder, and an original storing an expanded image output from the decoder. An image frame memory, an RGB color component histogram generation unit that converts the YUV color component obtained from the low frequency component extraction unit into an RGB component and counts it as a frequency distribution, and calculates a frequency from the output of the RGB color component histogram generation unit. A palette generation unit that preferentially extracts many color components and registers them in a palette storage unit, a palette storage unit that stores a palette of images after color reduction, and a color component of each pixel in the original image frame memory. , Comparing each color component registered in the palette storage unit, searching for the closest color from the palette storage unit, and assigning the color to the selected palette color. A color reduction unit that converts the color code into a color code and stores the color code in the color reduction image frame memory, and which color each pixel after color reduction is closest to the color registered in the palette storage unit is stored in association with the pixel position of the original image. A color reduction processing apparatus having a color reduction image frame memory, which selects palette colors before decompressing compressed image data (compressed data), creates palette data, and sequentially approximates the palette colors to directly reduce color images. Since the data can be written in the frame memory, the histogram memory for storing the histogram for creating the pallet data can be deleted. In addition, by selecting a color for display from a DC component already output from the low-frequency component extraction unit and a color generated by the low-order AC component, it is possible to greatly reduce the calculation for selecting the color. This has the effect of reducing power consumption, which is extremely important in mobile terminal devices and the like.

According to a second aspect of the present invention, there is provided a color reduction processing apparatus having a configuration in which an AC component order designation unit is added to the color reduction processing apparatus according to the first aspect, wherein a user can freely set the order of the AC component. Depending on how many colors you want to reduce, for example, "up to first order for 256 colors" or "up to third order for 4096 colors"
There is an effect that more accurate color reduction processing can be performed depending on the target number of colors.

According to a third aspect of the present invention, there is provided a color reduction processing apparatus having a configuration in which a "brightness-darkness frequency linearity correction unit" is added to the color reduction processing apparatus according to the second aspect, wherein human eyes are dark. However, utilizing the property of being sensitive to colors in brighter places, the higher the frequency, the higher the frequency, and the higher the frequency is counted, so that the bright colors are preferentially registered in the palette and have the effect of improving the quality of the apparently reduced color image. .

According to a fourth aspect of the present invention, there is provided a color reduction processing apparatus having a configuration in which a YUV color component histogram generation section is provided instead of the RGB color component histogram generation section in the color reduction processing apparatus according to the first aspect. Yes, human eyes are UV
Since the resolution of the Y component is higher than that of the UV component using the property of being more sensitive to the Y component than the component, it is possible to “register even a slight difference in the Y component as a different color in the palette”. This has the effect of improving the quality of the color-reduced image in human vision.

According to a fifth aspect of the present invention, there is provided a color reduction processing apparatus having a configuration in which a low frequency component sequential extraction unit is provided instead of the low frequency component extraction unit in the color reduction processing apparatus according to the first aspect. If the palette does not fill the expected number of colors, the low-frequency sequential extraction unit can extract the data of the next higher-order AC component, so that the color distribution in the extracted palette is gradually enhanced. And a high-quality color-reduced image can be obtained.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1) FIG. 3 is a block diagram showing a configuration of a color reduction processing apparatus according to Embodiment 1 of the present invention.
The same components as those in the conventional example are denoted by the same reference numerals.

The color reduction processing device according to the first embodiment of the present invention
As shown in FIG. 3, a decoder 2 for decoding compressed data 1 using DCT, and full-color data (color image data) output from the decoder 2 and displayed in full color are recorded. Referring to the original image frame memory 3 as an image memory and the decoder 2, DC
Low frequency component extraction unit 1 for extracting low frequency components and low order AC components
The RGB color component histogram generator 11 and the RGB color component histogram generator 11 that count the frequency of each appearing color over the entire image with reference to 0 and the output of the low-frequency component extractor 10 are referred to as high. A palette generation unit 6 that preferentially selects a color used frequently as a palette color and creates palette data p, and a correspondence relationship between the palette data p output by the palette generation unit 6 and the corresponding color code. A palette storage unit 7 to be recorded and full color data stored in the original image frame memory 3 with the palette colors stored in the palette storage unit 7 are approximated by the above-described palette colors, and are represented by 8-bit codes. The image processing apparatus is provided with a color reduction section 8 for converting to a code, and a color reduction image frame memory 9 for storing a color code output from the color reduction section 4.

As described above, the color reduction processing apparatus according to the first embodiment of the present invention refers to a low-frequency component extracting unit 10 for extracting a DC component and a low-order AC component, and The entire system is characterized by having an RGB color component histogram generation unit 11 for counting the frequency of each color that appears, and its function will be described below.

The compressed data 1 is obtained by subjecting the original image to DCT conversion or the like. In JPEG, which is a standard method for compressing still images, this DCT conversion is performed for each 8 × 8 pixel block, and the DCT conversion is performed. The image data is compressed by performing quantization centering on the high frequency component obtained by the above and further performing variable length coding on the data.

The decoder 2 performs variable length decoding and inverse quantization on the compressed data 1 to extract DCT coefficients. Further, inverse DCT is performed, and DCT is performed in block units.
Actual color image data (full color data) from coefficients
Generate f. In the first embodiment, a description will be given assuming that color reduction processing is performed using the primary AC component. It is assumed that the number of colors after the color reduction is 256 colors.

The low-frequency component extraction unit 10 determines a DC component (DC component) D in the DCT coefficient recorded at the head of the block data and a vertical direction adjacent to the DC component in the block data. AC components (lower-order AC components) Av1 and Ah1 each representing a gradual change in the horizontal direction are sequentially extracted, and each color component is converted into RGB and passed to an RGB color component histogram generation unit 11. At this time, for example, when the target number of colors is 256, R is 4 bits and G is 4 bits.
Bits and B are quantized by 4 bits, and also have a function of preventing colors that are too close from being registered in the palette.

The palette generation unit 6 receives the output of the RGB color component histogram generation unit 11 and selects 256 different colors from those having the highest frequency by priority and selects the palette data p.
Is written in the pallet storage unit 7.

The color reduction unit 8 refers to the original image frame memory and determines the color of the pixel of the color image (full color image) f in the palette data p stored in the palette storage unit 7 using the existing algorithm. Find a close color,
The color of the color image f is sequentially expressed by a color code indicating the color, and is written to the reduced color image frame memory 9.

The reduced color image frame memory 9 has a capacity enough to store all data (color code data) representing the pixels of the entire (multiple blocks) color image in a color code.

The operation of the color reduction processing apparatus according to the first embodiment in which each component functions as described above will be described with reference to FIG. FIG. 4 is a flowchart illustrating the operation of the color reduction processing apparatus according to the first embodiment. Here, the target color reduction color number is assumed to be 256 colors, and the display capability of the display device is assumed to be 256 colors out of 4096 colors (4 bits for each of R, G, and B). Also, the input image size is 256 × 256.

Step 4-1 First, the decoder 2 to which the compressed data 1 has been input in units of blocks performs variable length decoding and inverse quantization on the compressed data 1, and converts the decompressed image data into the original image. It is stored in the frame memory 3. At this time, the decoder 2 stores the DC component D of the DCT coefficient and the primary AC components Av1, Ah1.

Step 4-2 The low frequency component extraction unit 10 extracts the low frequency components (here, the DC component D of the DCT coefficient and the primary AC components Av1, Ah1) stored in the decoder 2.

Step 4-3 The RGB color component histogram generation unit 11 generates the DC component D and the primary AC component Av extracted by the low frequency component extraction unit 10.
1, D, D + Av1, D−Av1, D + Ah1, DA using Ah1
After converting h1 to RGB, it counts which colors are used and how often for all blocks of the original image. Now, the input image size is 256 ×
Since it is 256, the number of blocks is 16 × 16 = 256 blocks. The above five sets of RGB for each block
Is obtained, the color of the palette candidate is 256 × 5
= 1280 colors. However, considering the display ability, R,
When quantization is performed with 4 bits for each of G and B, the obtained 256 colors are degenerated (close colors are assigned to the same RGB value), and thus the number of colors usually decreases. Here, it is assumed that the color is reduced to 270 colors.

Step 4-4 The pallet generator 6 generates the RGB color component histogram.
Referring to the result of No. 1, the color used frequently is written into the palette storage unit 7 with priority. As a result, of the 270 colors obtained in step 4-3, 256 colors that are used frequently are selected.

Step 4-5 The color reduction unit 4 searches for the closest color among the palette colors stored in the palette storage unit 7 and quantizes (approximates) the color data of the color image f. At this time, the quantized color is converted into a color code, and the color code is written into the reduced color image frame memory 9.

As described above, according to the first embodiment, D
DC component D and primary AC components Av1, A of CT coefficients
By using h1 to select a color (palette color) to be used for the palette from among D, D + Av1, D-Av1, D + Ah1, and D-Ah1, a histogram for creating palette data is referred to a vast original image. It is not necessary to create the histogram, and it is possible to greatly reduce the amount of calculation and power consumption, and at the same time, to greatly reduce the amount of memory required for the histogram.

In the first embodiment, the method is described in which the full-color image is temporarily expanded in the original image frame memory and then the color reduction processing is performed. However, the compressed data is expanded for each block, and the low-frequency The components are extracted, RGB color component histograms are created for all blocks, a palette is generated from the results, and expansion processing is performed again for each block, and color reduction processing is performed each time one block is expanded. If you go, the original image frame memory is 1
Blocks are sufficient, and the memory can be significantly reduced.
Even if such a method is used, the gist of the present invention is not changed at all.

(Embodiment 2) FIG. 5 is a block diagram showing a configuration of a color reduction processing apparatus according to Embodiment 2 of the present invention. In FIG. 5, reference numeral 2 denotes a decoder for decoding compressed data 1 using DCT, and reference numeral 3 denotes full-color data (color image data) output from the decoder 2 and displayed in full color. An original image frame memory 10 as an image memory refers to the decoder 2 and D
A low-frequency component extraction unit 11 for extracting the C component and the low-order AC component refers to an output of the low-frequency component extraction unit 10 and counts the frequency of each appearing color over the entire image with reference to RGB.
The color component histogram generation unit 6 refers to the RGB color component histogram generation unit 11, preferentially selects a color that is used frequently as a palette color, and creates a palette data p. A palette storage unit for recording the correspondence between the color data output by the generation unit and the corresponding color code. The palette storage unit 8 stores the full color data stored in the original image frame memory with the palette colors stored in the palette storage unit. A color reduction unit that approximates a palette color and converts it to a color code represented by an 8-bit code;
Reference numeral 9 denotes a color-reduction image frame memory for storing the color code output from the color-reduction unit 4, and these components are configured in the same manner as in the first embodiment.

The feature of the color reduction processing apparatus according to the second embodiment is as follows.
The configuration of the first embodiment is provided with an AC component order designation unit 12 for designating “the order of the AC component to be extracted”.

The functions and the like of the color reduction processing apparatus according to the second embodiment composed of the above components will be described.

The compressed data 1 is obtained by subjecting the original image to DCT conversion or the like. In JPEG, which is a standard method for compressing still images, this DCT conversion is performed for each 8 × 8 pixel block. The image data is compressed by performing quantization centering on the high frequency component obtained by the above and further performing variable length coding on the data.

The decoder 2 performs variable length decoding and inverse quantization on the compressed data 1 to extract DCT coefficients. Further, inverse DCT is performed, and DCT is performed in block units.
Actual color image data (full color data) from coefficients
Generate f.

The AC component order designation unit 12 designates the order of the AC component to be used for pallet generation in consideration of the number of colors to be reduced by the user.

The low-frequency component extraction unit 10 determines the DC component (DC component) D of the DCT coefficient recorded at the head of the block data among the DCT coefficients and the vertical direction adjacent to the DC component in the block data. AC components (lower-order AC components) representing gradual changes in the horizontal direction Av1, Ah1, Av2,
Ah2, ... are sequentially extracted, each color component is converted to RGB, and
It is passed to the GB color component histogram generator 11. At this time,
For example, if the target number of colors is 256, R is 4
Bit, G is 4 bits, B is 4 bits,
It also has a function to prevent colors that are too close from being registered in the palette.

The color reduction unit 8 refers to the original image frame memory and determines the color of the pixel of the color image (full color image) f in the palette data p stored in the palette storage unit 7 using the existing algorithm. Find a close color,
The color of the color image f is sequentially expressed by a color code indicating the color, and is written to the reduced color image frame memory 9.

The reduced-color image frame memory 9 has a capacity to store all data (color code data) representing the pixels of the entire (multiple blocks) color image by color codes.

The operation of the color reduction processing apparatus according to the second embodiment, in which the components function as described above, will be described with reference to FIG. FIG. 6 is a flowchart illustrating the operation of the color reduction processing apparatus according to the second embodiment of the present invention. here,
The target number of color reductions is 256 colors, and the display capability of the display device is 4
It is assumed that there are 256 colors out of the 096 colors (4 bits for each of R, G, and B). Also, the input image size is 256 × 256.

Step 6-1 First, the decoder 2 to which the compressed data 1 has been input in units of blocks performs variable length decoding and inverse quantization on the compressed data 1 and converts the decompressed image data into the original image data. It is stored in the frame memory 3. At this time, the decoder 2 stores the DC component D and the AC components Av1, Ah1, Av2, Ah2,... Of the DCT coefficients.

Step 6-2: The user inputs “the A-th order A”
Whether up to the C component is to be extracted? " Here, it is assumed that 3 is specified.

Step 6-3: The low-frequency component extracting section 10 stores the low-frequency components stored in the decoder 2 (here, the DC components D and AC of the DCT coefficients).
Components Av1, Ah1, Av2, Ah2, Av3, Ah3) are extracted.

Step 6-4: The RGB color component histogram generating section 11 extracts the DC component D and the AC component (D, D) extracted by the low frequency component extracting section 10.
−Av1, D + Av1, D−Ah1, D + Ah1, D−Av2, D + Av2, D−
Ah2, D + Ah2, D-Av3, D + Av3, D-Ah3, D + Ah3)
After the conversion to GB, it counts which colors are used and how often for all blocks of the original image. Now, since the input image is 256 × 256, the number of blocks is 16 × 16 = 256, and since 13 sets of RGB are obtained from each block, 256 × 1
Palette candidate data of 3 = 3328 colors is obtained. However, when quantization is performed with R, G, and B bits each taking into account the display capability, the obtained 256 colors are degenerated (close colors are assigned to the same RGB values), so the number of colors usually decreases. I do. Here, it is assumed that the image has been reduced to 546 colors.

Step 6-5 The pallet generator 6 is the RGB color component histogram generator 1
Referring to the result of No. 1, the color used frequently is written into the palette storage unit 7 with priority. Here, of the 546 colors obtained in step 6-5, 25 frequently used colors are used.
Six colors are selected.

Step 6-6 The color reduction unit 4 finds the closest color among the palette colors stored in the palette storage unit 7 and quantizes (approximates) the color data of the color image f. At this time, the quantized color is converted into a color code, and the color code is written into the reduced color image frame memory 9.

As described above, the color reduction processing according to the second embodiment allows the user to freely specify the order of the AC component.
For example, "2
Up to first order for 56 colors ”or“ 3 for 4096 colors ”
It is possible to perform more accurate color reduction processing in accordance with the target number of colors, such as "until next".

(Embodiment 3) FIG. 7 is a block diagram showing a configuration of a color reduction processing apparatus according to Embodiment 3 of the present invention.
In FIG. 7, reference numeral 2 denotes a decoder for decoding the compressed data 1 using DCT, and reference numeral 3 denotes full-color data (color image data) output from the decoder 2 and displayed in full color. An original image frame memory 10 as an image memory refers to the decoder 2,
A low-frequency component extraction unit 11 for extracting a DC component and a low-order AC component refers to an output of the low-frequency component extraction unit 10 and counts the frequency of each appearing color over the entire image with reference to RGB
A color component histogram generation unit, 6 is a palette generation unit, 7 is a palette storage unit that records the correspondence between the color data output by the palette generation unit 6 and the corresponding color code, 8
A color reduction unit 9 for approximating the full color data stored in the original image frame memory with the palette colors stored in the palette storage unit 7 by the above-described palette colors and converting them into a color code represented by an 8-bit code; Is a reduced color image frame memory for storing the color code output from the color reduction unit 4, and these components are configured in the same manner as in the second embodiment.

The feature of the color reduction processing apparatus according to the third embodiment is as follows.
The configuration of the second embodiment is characterized in that a light-dark power linear correction unit 13 for increasing the power for a lighter color is provided, and the pallet generation unit 6 performs the light-dark power linear correction unit 1.
3, the color used frequently is preferentially selected as the palette color, and the palette data p is created.

The functions and the like of the color reduction processing apparatus according to the third embodiment composed of the above components will be described.

The compressed data 1 is obtained by subjecting the original image to DCT conversion or the like. In JPPEG, which is a standard method for compressing a still image, this compressed data 1 is subjected to this DCT conversion for each 8 × 8 pixel block. The image data is compressed by performing quantization centering on the high frequency component obtained by the above and further performing variable length coding on the data.

The decoder 2 performs variable length decoding and inverse quantization on the compressed data 1 to extract DCT coefficients. Further, inverse DCT is performed, and DCT is performed in block units.
Actual color image data (full color data) from coefficients
Generate f. In the second embodiment, a description will be given assuming that color reduction processing is performed using the primary AC component. It is assumed that the number of colors after the color reduction is 256 colors.

The low-frequency component extraction unit 10 determines a DC component (DC component) D in the DCT coefficient recorded at the head of the block data among the DCT coefficients and a vertical direction adjacent to the DC component in the block data. AC components (lower-order AC components) Av1 and Ah1 each representing a gradual change in the horizontal direction are sequentially extracted, and each color component is converted into RGB and passed to an RGB color component histogram generation unit 11. At this time, for example, when the target number of colors is 256, R is 4 bits, G is 4 bits, and B is 4 bits, and a function of preventing a color that is too close from being registered in the palette is also provided. It shall have both.

The pallet generation unit 6 receives the output of the RGB color component histogram generation unit 11 and selects 256 different colors with higher frequency from the palette data p.
Is written in the pallet storage unit 7.

The color reduction unit 8 refers to the original image frame memory and determines the color of the pixel of the color image (full color image) f in the palette data p stored in the palette storage unit 7 using the existing algorithm. Find a close color,
The color of the color image f is sequentially expressed by a color code indicating the color, and is written to the reduced color image frame memory 9. The reduced-color image frame memory 9 has a capacity to store all data (color code data) representing the pixels of the entire (multiple blocks) color image by color codes.

The light-dark frequency linear correction unit 13 newly sets the frequency obtained by multiplying the appearance frequency of each color by the value of the Y component. That is, the frequency increases as the color becomes brighter.

The operation of the color reduction processing apparatus according to the second embodiment, in which the components function as described above, will be described with reference to FIG. FIG. 8 is a flowchart illustrating the operation of the color reduction processing apparatus according to the third embodiment. Here, the target color reduction color number is assumed to be 256 colors, and the display capability of the display device is assumed to be 256 colors out of 4096 colors (4 bits for each of R, G, and B). Also, the input image size is 256 × 256.

Step 8-1 First, the decoder 2 to which the compressed data 1 has been input in units of blocks performs variable length decoding and inverse quantization on the compressed data 1, and converts the decompressed image data into the original image data. It is stored in the frame memory 3. At this time, the decoder 2 stores the DC component D of the DCT coefficient and the primary AC components Av1 and Ah1.

Step 8-2 The low frequency component extracting section 10 extracts the low frequency components (here, the DC component D of the DCT coefficient and the primary AC components Av1, Ah1) stored in the decoder 2.

Step 8-3: The RGB color component histogram generator 11 converts the DC component D and the primary AC component Av1,
Using Ah1, D, D + Av1, D−Av1, D + Ah1, D−Ah1
Is converted to RGB, then, for each block of the original image, which color is used and how often are counted. Now, the input image has a size of 256 × 256, and the number of blocks is 16 × 16. Since five sets of RGB components have been obtained for each block in step 8-3, candidates for palette data of 256 × 5 = 1280 colors have been obtained. However, when quantization is performed with R, G, and B bits each taking into account the display capability, the obtained 256 colors are degenerated (close colors are assigned to the same RGB values), so the number of colors usually decreases. I do. Here, it is assumed that the color is reduced to 270 colors.

Step 8-4 The light-dark frequency linear correction unit 13 newly sets a value obtained by multiplying the appearance frequency of each color by the value of the Y component. That is, the frequency increases as the color becomes brighter.

Step 8-5: The palette generating section 6 refers to the result of the light-dark frequency linearity correcting section 13 and stores the colors used frequently in the palette storing section 7.
To write preferentially. Thereby, Step 8-
From the 270 colors obtained in step 4 and the inflated frequency corresponding to the colors, 256 colors that are frequently used are selected.

Step 8-6 The color reduction unit 4 finds the closest color among the palette colors stored in the palette storage unit 7 and quantizes (approximates) the color data of the color image f.

At this time, the quantized color is converted into a color code, and the color code is written into the reduced color image frame memory 9.

As described above, the color reduction processing apparatus according to the third embodiment utilizes the property that the human eye is more sensitive to colors in brighter places than in dark places. Bright colors are preferentially registered in the palette, and the quality of the apparently reduced color image can be improved.

(Embodiment 4) FIG. 9 is a block diagram showing a configuration of a color reduction processing apparatus according to Embodiment 4 of the present invention.
In FIG. 9, reference numeral 2 denotes a decoder for decoding compressed data 1 using DCT, and reference numeral 3 denotes full-color data (color image data) output from the decoder 2 and displayed in full color. An original image frame memory 10 as an image memory refers to the decoder 2,
A low-frequency component extraction unit 14 for extracting a DC component and a low-order AC component is referred to as an output of the low-frequency component extraction unit 10 and counts the frequency of each appearing color over the entire image with reference to the YUV.
A color component histogram generation unit, 6 is a palette generation unit, 7 is a palette storage unit that records the correspondence between color data output by the palette generation unit and corresponding color codes, and 8 is a palette color stored in the palette storage unit. A color reduction unit that approximates the full-color data stored in the original image frame memory with the above-mentioned palette color and converts it into a color code represented by an 8-bit code, and 9 is a color code output from the color reduction unit 8. A color-reduced image frame memory for storing
These components are configured in the same manner as in the first embodiment.

The feature of the color reduction processing apparatus according to the fourth embodiment is as follows.
A YUV color component histogram generator that counts the frequency of each color appearing over the entire image by referring to the output of the low frequency component extractor 10 instead of the RGB component histogram generator 11 in the configuration of the first embodiment. The palette generation unit 6 refers to the YUV color component histogram generation unit 14 and preferentially selects a color that is frequently used as a palette color to generate palette data p. Has become.

The functions and the like of the color reduction processing apparatus according to the fourth embodiment composed of the above components will be described.

The compressed data 1 is obtained by subjecting an original image to DCT conversion or the like. In JPEG, which is a standard method for compressing still images, this DCT conversion is performed for each 8 × 8 pixel block. The image data is compressed by performing quantization centering on the high frequency component obtained by the above and further performing variable length coding on the data.

The decoder 2 performs variable length decoding and inverse quantization on the compressed data 1 to extract DCT coefficients. Further, inverse DCT is performed, and DCT is performed in block units.
Actual color image data (full color data) from coefficients
Generate f. In the fourth embodiment, a description will be given assuming that the color reduction processing is performed using the primary AC component. It is assumed that the number of colors after the color reduction is 256 colors.

The low-frequency component extraction unit 10 determines a DC component (DC component) D of the DCT coefficient recorded at the head of the block data and a vertical direction adjacent to the DC component in the block data. AC components (lower-order AC components) Av1 and Ah1 each representing a gradual change in the horizontal direction are sequentially extracted, and each color component is converted into RGB and passed to the YUV color component histogram generation unit 14. At this time, for example, when the target number of colors is 256, R is 4 bits, G is 4 bits, and B is 4 bits, so that a color that is not so close is not registered in the palette. Shall also have.

The pallet generation unit 6 receives the output of the YUV color component histogram generation unit 14, selects 256 different colors from those having the highest frequency, and selects the pallet data p.
Is written in the pallet storage unit 7.

The color reduction unit 8 refers to the original image frame memory and determines the color of the pixel of the color image (full color image) f in the palette data p stored in the palette storage unit 7 using the existing algorithm. Find a close color,
The color of the color image f is sequentially expressed by a color code indicating the color, and is written to the reduced color image frame memory 9.

The color-reduced image frame memory 9 has a capacity enough to store data (color code data) representing the pixels of the entire (multiple blocks) color image in color code.

The YUV color component histogram generation unit 14
For example, when the display capability of the display device has only 4096 colors, that is, when the display device is represented by each of 4 bits of RGB, the histogram is once generated by quantizing Y5 bits and UV to 4 bits. By doing so, the sensitivity for the Y component is improved, and even a slight difference in the Y component is registered as another color.

The operation of the color reduction processing apparatus according to the fourth embodiment, in which each component functions as described above, will be described with reference to FIG. FIG. 10 is a flowchart illustrating the operation of the color reduction processing apparatus according to the fourth embodiment. Here, the target color reduction color number is 256 colors, and the display capability of the display device is 409.
256 colors out of 6 colors (4 bits each for R, G, B). Also, assume that the input image size is 256 × 256.

Step 10-1 First, the decoder 2 to which the compressed data 1 is input in units of blocks performs variable length decoding and inverse quantization on the compressed data 1, and converts the decompressed image data into the original image. It is stored in the frame memory 3. At this time, the decoder 2 stores the DC component D of the DCT coefficient and the primary AC components Av1 and Ah1.

Step 10-2 The low frequency component extraction unit 10 extracts the low frequency components (here, the DC component D of the DCT coefficient and the primary AC components Av1, Ah1) stored in the decoder 2.

Step 10-3: The YUV color component histogram generator 14 extracts the DC component D and the primary AC component Av1, extracted by the low frequency component extractor 10.
Using Ah1, D, D + Av1, D−Av1, D + Ah1, D−Ah1
And count which colors are used and how often for all blocks in the original image. Now, the input image has a size of 256 × 256.
The number of blocks is 16 × 16. Since five sets of YUV components are provided for each block, 256 × 5 = 12
80 color YUV palette data candidates are obtained. However, in consideration of the display capability, since quantization is performed with each of 4 bits of R, G, and B, quantization is performed with 4 bits of each of 5 bits of Y here. Allocating 5 bits only to Y is because human eyes are sensitive to Y. In this way 12
The 80 colors are degenerated (close colors are assigned the same RGB values) and the number of colors is usually reduced. Here, it is assumed that the color is reduced to 270 colors.

Step 10-4: The pallet generator 6 generates the YUV color component histogram generator 1
4, the color used frequently is preferentially written in the palette storage unit 7. As a result, among the 270 colors obtained in step 10-3, 256 with a high use frequency are selected.

Step 10-5 The color reduction unit 4 finds the closest color among the palette colors stored in the palette storage unit 7 and quantizes (approximates) the color data of the color image f. At this time, the quantized color is converted into a color code, and the color code is written into the reduced color image frame memory 9.

As described above, the color reduction processing apparatus according to the fourth embodiment utilizes the property that the human eye is sensitive to the UV component, and makes the resolution of the Y component higher than that of the UV component.
“Even a slight difference in the Y component can be registered in the palette as another color”, so that the quality of a color-reduced image can be improved visually for humans.

(Embodiment 5) FIG. 11 is a block diagram showing a configuration of a color reduction processing apparatus according to Embodiment 5 of the present invention.

In FIG. 11, reference numeral 2 denotes a decoder for decoding the compressed data 1 using DCT, and reference numeral 3 denotes full-color data (color image data) output from the decoder 2 and displayed in full color. Original image frame memory as an image memory for recording
A color component histogram generation unit 6 refers to the RGB color component histogram generation unit 11, preferentially selects a color used frequently with a palette color, and generates a palette data p. A pallet storage unit for recording the correspondence between the color data output by the unit 6 and the corresponding color codes. A palette storage unit 8 stores the full color data stored in the original image frame memory in the pallet colors stored in the pallet storage unit 7 as described above. A color reduction unit 9 for approximating the color of the pallet and converting it into a color code represented by an 8-bit code; and 9, a color reduction image frame memory for storing the color code output from the color reduction unit 8. Is configured in the same manner as in the first embodiment.

The feature of the color reduction processing apparatus according to the fifth embodiment is as follows.
Low frequency component extraction unit 10 in the configuration of the first embodiment
, The DC component and the lower order A
Because the low-frequency component sequential extraction unit 15 that extracts the C component is provided, the RGB color component histogram generation unit 11 refers to the output of the low-frequency component sequential extraction unit 15, and outputs each color that appears over the entire image. The frequency is counted.

The functions and the like of the color reduction processing apparatus according to the fourth embodiment composed of the above components will be described.

The functions and the like of the color reduction processing apparatus according to the fifth embodiment composed of the above components will be described.

The compressed data 1 is obtained by subjecting the original image to DCT conversion or the like. In JPEG, which is a standard method for compressing still images, this DCT conversion is performed for each 8 × 8 pixel block. Quantization is performed centering on the high-frequency components obtained by the above, and the data is further subjected to variable-length encoding to realize compression of image data.

The decoder 2 performs variable length decoding and inverse quantization on the compressed data 1 to extract DCT coefficients. Further, inverse DCT is performed, and DCT is performed in block units.
Actual color image data (full color data) from coefficients
Generate f. In the fifth embodiment, a description will be given assuming that color reduction processing is performed using the primary AC component. It is assumed that the number of colors after the color reduction is 256 colors.

The low frequency component sequential extraction section 15 uses the DCT
DC recorded at the beginning of block data among coefficients
A DC component (DC component) D in the T coefficient and an AC component (lower-order AC component) Av1, Ah1, which represents a gradual change in the vertical and horizontal directions adjacent to the DC component in the block data.
.. Are sequentially extracted, each color component is converted to RGB, and is passed to the RGB color component histogram generation unit 11. At this time, for example, if the target number of colors is 256 colors, R
Is quantized by 4 bits, G is 4 bits, and B is 4 bits, and has a function of preventing a color that is too close from being registered in the palette.

The palette generation unit 6 receives the output of the RGB color component histogram generation unit 11 and selects 256 different colors from those with the highest frequency by priority and selects the palette data p
Is written in the pallet storage unit 7.

The color reduction unit 8 refers to the original image frame memory and determines the color of the pixel of the color image (full color image) f in the palette data p stored in the palette storage unit 7 using the existing algorithm. Find a close color,
The color of the color image f is sequentially expressed by a color code indicating the color, and is written to the reduced color image frame memory 9.

The color-reduced image frame memory 9 has a capacity enough to store data (color code data) representing the pixels of the entire (multiple blocks) color image in color code.

The operation of the color reduction processing apparatus according to the fifth embodiment, in which the components function as described above, will be described with reference to FIG. FIG. 12 is a flowchart illustrating the operation of the color reduction processing apparatus according to the fifth embodiment. Here, the target color reduction color number is 256 colors, and the display capability of the display device is 409.
256 colors out of 6 colors (4 bits each for R, G, B). Also, assume that the input image size is 256 × 256.

Step 12-1 First, the decoder 2 to which the compressed data 1 has been input in units of blocks performs variable length decoding and inverse quantization on the compressed data 1, and converts the decompressed image data into the original image data. It is stored in the frame memory 3. At this time, the decoder 2 stores the DC component D of the DCT coefficient and the primary AC components Av1, Ah1, Av2, Ah2,.

First, the decoder 2 instructs the low frequency component sequential extraction unit 15 to extract a DC component, that is, a zero-order AC component.

Step 12-2 The low frequency component sequential extraction unit 15 extracts the low frequency components (here, the DC component D of the DCT coefficient and the primary AC components Av1, Ah1) stored in the decoder 2.

Step 12-3 The RGB color component histogram generation unit 11 converts the DC component D extracted by the low frequency component sequential extraction unit 15 into RGB, and then, for each block of the original image, Count what is used in Now, the input image has a size of 256 × 256, and the number of blocks is 16 × 16. One set of R for each block
Since the GB component is obtained, 256 colors are conveniently obtained. However, when quantization is performed on each of R, G, and B bits in consideration of the display capability, the obtained 256 colors are degenerated (close colors are the same R colors).
(Assigned to GB values), the number of colors usually decreases. Here, it is assumed that the color has been reduced to 120 colors. That is, there are 136 remaining “empties” in the target color reduction 256 colors.

Step 12-4: The pallet generator 6 generates the RGB color component histogram generator 1.
Referring to the result of No. 1, the color used frequently is written into the palette storage unit 7 with priority. In this case,
Since the number of obtained colors is less than the available space in the palette, all 120 colors are adopted regardless of the frequency.

Step 12-5: If there is still room in the pallet, the next higher order is instructed to the low-frequency component sequential extracting section 15, and the process proceeds to step 12-2. Otherwise, go to step 12-6. Since 0 was previously designated as the order, the description is continued here assuming that 1 is designated.

Step 12-2 Since the order 1 is designated, D + Av1, D−Av1, and D−Av1, D−Av1, are calculated using the DC component D and the primary AC components Av1 and Ah1.
D + Ah1. Extract D-Ah1.

Step 12-3 After converting the color data extracted in the previous step into RGB, it is counted which color is used and how often in all blocks of the original image. In this case, since four sets of colors are obtained for each block, 256 blocks × 4 colors are newly set as candidates. Here, it is assumed that each of the R, G, and B components is reduced to 1024 colors by quantizing with 4 bits.

Step 12-4: The pallet generator 6 is the RGB color component histogram generator 1.
With reference to the result of No. 1, the color used frequently is written in the palette storage unit 7. Since there are 136 available pallets, 136 colors that are frequently used are selected from the obtained 270 colors.

Step 12-5: If there is still a vacancy in the palette, the next higher order is instructed to the low frequency component sequential extraction section 15, and the process proceeds to step 12-2. Otherwise, proceed to 12-6. The explanation is continued assuming that 1 has been designated earlier and the pallet has no empty space. In this case, the process proceeds to step 12-6.

Step 12-6 The color reduction unit 8 searches for the closest color among the palette colors stored in the palette storage unit 7 and quantizes (approximates) the color data of the color image f. At this time, the quantized color is converted into a color code, and the color code is written into the reduced color image frame memory 9.

As described above, in the color reduction processing apparatus according to the fifth embodiment, when the pallet is not filled with the expected number of colors, the low-frequency component sequential extraction unit 15 outputs the data of the AC component having the next higher order. Can be extracted, the color distribution in the extracted palette can be gradually enhanced, and a high-quality color-reduced image can be obtained.

[0114]

As is apparent from the above description, the color reduction processing apparatus according to the present invention selects palette colors before decompressing compressed image data (compressed data) and creates palette data. Image data for storing color image data and a histogram memory for storing a histogram for creating palette data, since full color data of each color can be sequentially approximated to a palette color and written directly to the display memory without being developed on the image memory. Has the advantageous effect of reducing In addition, the calculation for display can be significantly reduced, and an advantageous effect that power consumption, which is extremely important in a portable terminal device or the like, can be reduced can be obtained.

Further, the color reduction processing apparatus according to the present invention allows the user to freely specify the order of the AC component. Therefore, depending on "what color do you want to reduce?" .., Etc., the advantageous effect of being able to perform highly accurate color reduction processing according to the target number of colors is obtained.

Further, the color reduction processing apparatus according to the present invention utilizes the property that human eyes are more sensitive to colors where they are brighter than where they are dark. Is registered in the pallet,
This has the advantageous effect that the quality of the apparently reduced color image can be improved.

Further, the color reduction processing apparatus according to the present invention makes use of the property that the human eye is more sensitive to the Y component than the UV component, and makes the resolution of the Y component higher than that of the UV component.
Since “a slight difference in the Y component can be registered as another color in the palette”, the effect of improving the quality of the color-reduced image in human vision can be obtained.

In the color reduction processing apparatus according to the present invention, when the palette does not fill the expected number of colors, the low frequency component sequential extraction unit can extract the data of the AC component having the next higher order. Since the color distribution in the selected palette can be gradually enhanced, an advantageous effect that a high-quality color-reduced image can be obtained can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a conventional color reduction processing device.

FIG. 2 is a flowchart showing the operation of the color reduction processing apparatus.

FIG. 3 is a block diagram illustrating a configuration of a color reduction processing device according to the first embodiment of the present invention;

FIG. 4 is a flowchart showing the operation of the color reduction processing apparatus.

FIG. 5 is a block diagram illustrating a configuration of a color reduction processing apparatus according to a second embodiment of the present invention.

FIG. 6 is a flowchart showing the operation of the color reduction processing apparatus.

FIG. 7 is a block diagram illustrating a configuration of a color reduction processing apparatus according to a third embodiment of the present invention.

FIG. 8 is a flowchart showing the operation of the color reduction processing apparatus.

FIG. 9 is a block diagram illustrating a configuration of a color reduction processing apparatus according to a fourth embodiment of the present invention.

FIG. 10 is a flowchart showing the operation of the color reduction processing apparatus.

FIG. 11 is a block diagram illustrating a configuration of a color reduction processing apparatus according to a fifth embodiment of the present invention.

FIG. 12 is a flowchart showing the operation of the color reduction processing apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Compressed data 2 Decoder 3 Original image frame memory 6 Palette generation unit 7 Palette storage unit 8 Color reduction unit 9 Color reduction image frame memory 10 Low frequency component extraction unit 11 RGB component histogram generation unit 12 AC component order designation unit 13 Light-dark Frequency linearity correction unit 14 YUV component histogram generation unit 15 Low frequency component sequential extraction unit

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 1/60 H04N 1/40 D 1/46 1/46 Z F term (Reference) 5B057 BA26 CA01 CA08 CA12 CA16 CB01 CB08 CB12 CB16 CC01 CE18 CG05 DC23 5C057 AA06 AA11 CA01 DA00 EA01 EA02 EA07 EC01 EL01 EM01 GE08 GF00 GG01 5C077 LL18 LL20 MP08 PP32 PP34 PP48 PQ08 PQ19 PQ22 RR21 NA05 MA02 NA01 LA01 BA12 BA29 BA34 BA35 BA36 BB16 BB25 BB42 BB44 DA53 DA63 DA73 DA89 MM02 MM04 MM07

Claims (5)

[Claims] 1. A decoder for decompressing compressed data using DCT, and a low-frequency component for extracting DC components and low-order AC components of DCT coefficients obtained when decompressing the compressed data from the decoder. An extraction unit, an original image frame memory for storing the decompressed image output from the decoder, and RGB for converting the YUV color component obtained from the low frequency component extraction unit into an RGB component and counting it as a frequency distribution A color component histogram generation unit, a palette storage unit for storing a palette of images after color reduction, and a color component having a high frequency is preferentially extracted from the output of the RGB color component histogram generation unit and registered in the palette storage unit. The palette generation unit compares the color components of each pixel in the original image frame memory with each color component registered in the palette storage unit, and determines the closest color to the palette. Find out the storage unit, a color reduction process unit to be recorded in the reduced color image frame memory is converted into the color code given to the selected palette colors,
A color-reduction processing apparatus comprising: a color-reduction image frame memory for storing which color of each pixel after color reduction is closest to a color registered in a palette storage unit in association with a pixel position of an original image. 2. A color reduction processing apparatus according to claim 1, further comprising an AC component order designation unit, which allows the user to designate the order of the AC components to be used for pallet generation. Processing equipment. 3. A color reduction process according to claim 1, further comprising adding a light-dark frequency linear correction unit to the color reduction device, wherein a color brighter than a dark color is preferentially registered in a palette when a palette is generated. apparatus. 4. A Y-color component in a YUV color component obtained from a low-frequency component extraction unit, in which a large number of bits are compared with the UV component instead of the RGB color component histogram generation unit in the color reduction processing device according to claim 1. A color reduction processing apparatus characterized by adding a YUV color component histogram generation unit that performs quantization by assigning. 5. A low-frequency component extracting unit according to claim 1, wherein a low-frequency component extracting unit is provided instead of the low-frequency component extracting unit. Requests the low frequency component sequential extraction unit to extract higher order AC components, and increases the AC component order by one until there is no more space on the pallet.
A color reduction processing apparatus characterized in that colors registered in a palette are enhanced as much as possible.