JPH08340554A - Color image coder - Google Patents

Abstract

PURPOSE: To obtain a color image coder in which deterioration in image quantity is reduced by selecting a block whose color quality deterioration of a color signal is easily recognized through setting a simple threshold level so as to decrease the quantization width of the selected block. CONSTITUTION: An absolute value of a difference between a mean value of picture element values in each block and an achromatic color level of a color difference signal is calculated. Whether or not a maximum value of the absolute value is larger than a prescribed threshold level is discriminated by a color difference signal check device 15 and a quantization controller 16 decides the quantization width by using the discrimination result and a code string length counted by a generated code amount counter 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image coding device, and more particularly to a color image coding device capable of reducing visual deterioration of a color image.

[0002]

2. Description of the Related Art In recent years, MPEG has been used as an image coding method.
As represented by the (Moving Picture Experts Group) method, discrete cosine transform (DCT, Discrete Cosine Tr)
orthogonal transform such as ansform) and Huffman code (Huffman code)
A method using a variable length code such as Coding) is widely known.

FIG. 8 shows an example of a conventional color image coding apparatus when orthogonal transform and variable length coding are used.
Here, an encoding method in the MPEG system will be described with reference to FIG. In FIG. 8, 11 is a block converter for converting the digital video signal 18 into a block of 8 × 8 pixels for each of the Y signal, the RY signal, and the BY signal, and 12 is a blocked digital signal. Image signal 2
0 is orthogonally transformed for each block and the orthogonal transformation coefficient 12
An orthogonal transformer that generates 8 and a quantizer 13 that quantizes the above orthogonal transform coefficient. This quantization is performed by an 8 × 8 quantization matrix (not shown) corresponding to each frequency after orthogonal transform. No.) and the quantization width 168 determined by the quantization width determiner 16 to perform division on the orthogonal transform coefficient. 14 performs variable-length coding on the quantized orthogonal transform coefficient 138 in macroblock units,
The variable length encoder 17 for generating 9 is a generated code amount counter for receiving the code amount signal 148 of the variable length encoder 14 and cumulatively counting the code amount to detect the generated code amount.
Further, 16 is a count output 1 of the generated code quantity counter 17.
78 is a quantizer width determiner for determining a quantizer width 168 when the orthogonal transform coefficient from 78 is quantized by the quantizer 13.

Next, the operation of the conventional color image coding apparatus will be described with reference to FIGS. 8, 2, 9 and 10. The digital image signal 18 is first converted into the block converter 1
In FIG. 1, first, as shown in FIG. 9, after being divided into 16 × 16 pixel macro blocks MB, 8 × 8 pixels are further divided for each of the Y signal, RY signal, and BY signal. The digital image signal 20 is converted into a unit signal to be a blocked digital image signal 20. FIG. 2 shows a block converter 1
Showing each component of the image signal block-converted by 1
Y, RY, which are included in the 16 × 16 pixel range of the Y signal,
Each block Y1, Cr1, Cb1 of the BY signal is composed of a set of 2 × 2 = 4 blocks, 2 × 1 = 2 blocks and 2 × 1 = 2 blocks in units of 8 × 8 pixels. A group of each block Y1, Cr1, Cb1 is called a macro block MB.

Blocked digital image signal 20
Is orthogonally transformed for each block by the orthogonal transformer 12 and becomes an orthogonal transformation coefficient 128 (see FIG. 9). Then, the orthogonal transform coefficient 128 is quantized by the quantizer 13 (see FIG. 9). At this time, the orthogonal transform coefficient 12 is obtained by multiplying the 8 × 8 quantization matrix corresponding to each frequency by the quantization width 168 determined by the quantization width determiner 16.
Quantization is performed by dividing by 8. Although different quantization matrices can be used for the luminance signal and the color difference signal, the quantization width 168 uses the same value for all blocks in the macroblock. The quantized orthogonal transform coefficient 138 is subjected to a variable length code by the variable length encoder 14 to form a code string 19. Further, the generated code amount counter 17 is the variable length encoder 1
The code amount 148 generated in 4 is counted, for example, in the case of MPEG, it is counted for each macroblock in one frame, and the cumulative amount (BN for the macroblock number N) in the one frame is obtained. Then, the quantization width determiner 16 uses the cumulative value 178 of the generated code amount obtained by the generated code amount counter 17 to set the quantization width 168 (BN to the quantization width 168 so that the generated code amount becomes a predetermined amount. On the other hand, Q _{N + 1} ) is determined (see FIGS. 10 (a) and 10 (b)). As a unit for keeping the generated code amount constant, one frame, or one frame to several frames is selected.

In the conventional color image coding apparatus as shown in FIG. 8, the quantization width of the block to be coded is determined on the basis of the previously generated code amount. It was done regardless of the visual characteristics. Therefore, while human vision is sensitive to, for example, red, the above-described conventional device has the same quantization distortion for all colors, and thus the human vision has red portions as described above. Since it is sensitive, the deterioration due to the quantization distortion is more noticeable in the red portion than in other colors.

In order to solve such a problem, Japanese Patent Laid-Open No. 6-225340 discloses a method of prohibiting or reducing the increase of the quantization width when the color contains a saturated state or a color close to the saturated state. Is proposed. That is, when the color includes a saturated state or a color close to the saturated state as described above, a little deterioration is noticeable. Therefore, the deterioration can be prevented by reducing the quantization width and finely quantizing. It is a thing. However, this method has a problem in that it is not possible to prevent deterioration of cyan, yellow, and the like, which are in opposite phases to red and blue, for example.

The above publication also proposes a method of reducing the quantization width when a saturation level is detected for a specific color, for example, purple. However, in this conventional method, the thresholds for the RY and BY signals for detecting the specific color, for example, the color such as purple, are determined in order to determine whether the specific color is at a saturation level. It is necessary to set various values, that is, it is necessary to set a fine threshold value, and the setting of the threshold value is very complicated.

[0009]

As described above, in the conventional color image coding apparatus, the quantization is performed regardless of the visual characteristics of human beings. Therefore, deterioration due to quantization distortion is more noticeable in the red part than other colors, and in order to solve such a problem, when the color contains a saturated state or a color close to the saturated state, , There is also a method of prohibiting or reducing the increase of the quantization width, but this method cannot prevent the deterioration of cyan, yellow, etc., which are in antiphase with red and blue, for example. . Further, a method of reducing the quantization width when a saturation level is detected for a specific color, for example, purple has been proposed. However, in this conventional method, in order to detect the specific color, There is a problem that it is necessary to set a fine threshold value for the color difference signal, which makes the setting of the threshold value very complicated.

The present invention has been made to solve the above-mentioned conventional problems. In a color image coding apparatus using orthogonal transform and variable length coding, the quantization width can be set by a simple threshold setting. An object of the present invention is to provide a color image coding apparatus capable of reducing visual deterioration of image quality by selecting a block to be reduced and reducing a quantization width of the selected block.

[0011]

In order to achieve the above object, a color image coding apparatus according to the first aspect of the invention uses a Y signal which is a luminance signal and two color difference signals RY and BY signals. In a color image encoding device for encoding a digital color image signal that is configured, a block converter that divides the digital color image signal into blocks, and a digital color image signal that is divided for each block , An orthogonal transformer that performs an orthogonal transformation for each block, a quantizer that quantizes the orthogonal transform coefficient obtained by the orthogonal transformer, and a variable that performs variable-length coding on the quantized orthogonal transform coefficient Within the set of the long encoder and the block whose code amount by the variable length encoder should be constant,
A generated code amount counter that cumulatively counts code string lengths that are sequentially encoded by the variable length encoder, and an average of pixel values in each block for at least one color difference signal of the two color difference signals. The absolute value of the difference between the value and the colorless level value of the color difference signal is calculated, and the maximum absolute value of the difference in the pixel set unit in which the quantization width in the quantizer is to be changed is a predetermined value. A color difference signal inspector that determines whether the difference is greater than a threshold value,
When it is determined that the maximum value is less than or equal to the threshold value in the color difference signal inspector, the quantization width used in the quantizer is obtained based on the code string length counted by the generated code amount counter. When the maximum value is determined to be larger than the threshold value in the color difference signal inspector, the quantization width used in the quantizer is set to the code string length counted by the generated code amount counter. And a quantizer width determiner for setting the value to be smaller than the quantizer width obtained based on the above.

A color image coding apparatus according to a second aspect of the present invention is the color image coding apparatus according to the first aspect,
The threshold value used in the color difference signal inspecting device is assumed to be a fixed value.

A color image coding apparatus according to a third aspect of the present invention is the color image coding apparatus according to the first aspect.
The threshold used in the color difference signal inspector is a variable value.

A color image coding apparatus according to a fourth aspect of the present invention is the color image coding apparatus according to the third aspect, wherein
The threshold value used in the color difference signal tester is an average value of the maximum absolute values of the differences in the frame encoded immediately before.

A color image coding apparatus according to a fifth aspect of the present invention is the color image coding apparatus according to any one of the first to fourth aspects, wherein the actual quantization width is 1/2 of the temporary quantization width. It is supposed to be set to ⅓.

A color image coding apparatus according to a sixth aspect of the present invention is the color image coding apparatus according to any one of the first to fifth aspects, wherein the provisional quantization width is Qi, the actual quantization width is Qr, and the absolute value of the difference. Where X is a maximum value of T and T is a threshold value, the setting of the actual quantization width with respect to the temporary quantization width is Qr = Qi / {2+ (X−T) / (112−T)} (X> T ) Is used.

A color image coding apparatus according to a seventh aspect of the present invention is the color image coding apparatus according to any one of the first to sixth aspects, wherein the orthogonal transformer is the average value of the pixel values of the color difference signal blocks. The DC component of the block of the color difference signal obtained by the above is used.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1 . First, the first embodiment of the present invention will be described with reference to FIG. Here, a case will be described in which the input digital image signal is in the 4: 2: 2 format and the quantization width is changed in units of macroblocks in which a plurality of blocks are gathered as in the MPEG system.

In FIG. 1, numeral 11 converts the digital video signal 18 into blocks of 8 × 8 pixels for each signal,
A block converter that outputs a blocked digital image signal 20. An orthogonal converter that performs an orthogonal transform on the blocked digital image signal 20 for each block to generate an orthogonal transform coefficient 121. It is a quantizer for quantizing the orthogonal transformation coefficient, and this quantization is performed by the quantization matrix of 8 × 8 corresponding to each frequency after orthogonal transformation and the quantization determined by the quantization width determiner 16. The width is used to divide the orthogonal transform coefficient. 14
Is a variable-length encoder that applies a variable-length code to the quantized orthogonal transform coefficient in macroblock units to generate a code string 19, and 17 is a code amount signal 1 of the variable-length encoder 14.
41 is an input, and is a generated code amount counter for cumulatively counting the code amount and detecting the generated code amount. Reference numeral 15 denotes a color difference signal inspecting device for inspecting a color difference signal from the output 21 of the block converter 11, which is a pixel value in each block for at least one of the two color difference signals. The absolute value of the difference between the average value and the value of the colorless level of the color difference signal is calculated and the quantization width in the quantizer 13 is to be changed in pixel set units, here the absolute value of the difference in the macroblock. Is compared with a predetermined threshold value. 16 is an output 22 of the color difference signal inspector 15
And a count output 171 of the generated code amount counter 17, which is a quantization width determiner for determining a quantization width when the orthogonal transform coefficient is quantized by the quantizer 13.
This is because when the color difference signal inspector 15 determines that the maximum value is less than or equal to the threshold value, the quantization width 161 used by the quantizer 13 is counted by the generated code amount counter 17. When the color difference signal inspector 15 determines that the maximum value is larger than the threshold value, the quantization width 161 used in the quantizer 13 is set to the generated code amount. The quantization width is smaller than the quantization width obtained based on the code string length 171 counted by the counter 17, and the quantization width 161 which is the output is given to the quantizer 13.

Next, the operation will be described. As described with reference to FIG. 8 above, the digital video signal 18 is first transmitted to the block converter 11 as shown in FIG.
After being divided into macroblocks of × 16 pixels, Y
8 × 8 for each signal, RY signal and BY signal
It is converted into a signal in units of pixels and becomes a blocked digital image signal 20. That is, as shown in FIG. 2, the blocks Y1 and C of the image signals Y, RY, and BY signals that have been block-converted by the block converter 11.
r1 and Cb1 are 2 × in units of 8 × 8 pixels, respectively.
It is composed of a set of 2 = 4 blocks, 2 × 1 = 2 blocks, and 2 × 1 = 2 blocks, and each of these blocks Y1, Cr
A group of 1 and Cb1 is called a macro block MB.

Further, the block-shaped digital image signal 20 is subjected to orthogonal transformation for each block by the orthogonal transformation unit 12 and becomes an orthogonal transformation coefficient 121 (see FIG. 9). As the orthogonal transform method, for example, in the MPEG method, discrete cosine transform is used. The orthogonal transform coefficient 121 is quantized by the quantizer 13 (see FIG. 9). Here, the orthogonal transform coefficient is divided by a value obtained by multiplying the 8 × 8 quantization matrix corresponding to each frequency of the orthogonal transform coefficient by the quantization width 161 determined by the quantization width determiner 16. Quantize. The orthogonal transform coefficient 131 quantized by the quantizer 13 is variable-length coded by the variable-length encoder 14 to form a code string 19. In addition, the generated code amount counter 17 has a code string 19 output from the variable length encoder 14.
Of each macroblock in one frame, and the accumulated amount in one frame (macroblock number
BN) is calculated for N (see FIG. 11).

Quantization width 161 used in the quantizer 13
(Q _{N + 1} or Q _{N + 1} ′ for BN) is based on the output 22 of the color difference signal inspector 15 and the output 171 of the generated code amount counter 17 by the quantization width determiner 16. It is determined. This determination method will be described below. FIG. 3 shows the configuration of the color difference signal inspector 15a according to the first embodiment. The color difference signal inspector 15a includes a block converter 11 and an R
The block 21 of the -Y signal is input. For the RY signal 21 input to the color difference signal inspector 15a, the average value calculator 31 first obtains the average value of pixel values in each block, that is, the DC component. The absolute difference calculator 32 calculates the absolute value of the difference between the average value of the pixel values in each block obtained by the average value calculator 31 and the colorless level value of the color difference signal. For example, ITU-R REC. Which is a standard for digital video signals. At 601, each pixel is represented with an accuracy of 8 bits, and the colorless level 35 of the color difference signal is represented.
a is 128, and the RY signal has a pixel value of 16
It can take a value of ~ 240.

Therefore, in such a case, when the average value of the pixel values obtained by the average value calculator 31 is 100, the output 32a of the absolute difference calculator 32 is 28. The output 32a of the absolute difference calculator 32 is input to the maximum value detector 33, where two blocks of RY signals in the macroblock (in the case of 4: 2: 2 system) are compared with a colorless level. Find the maximum absolute difference. For example, as shown in Figure 4 (a),
When the average value of the pixel values of the block Cr1 of the two RY signals 21 is 150 and 60, respectively, the absolute values of the difference from the colorless level are 22 and 68, respectively, and the maximum absolute value of the difference is It becomes 68. The maximum value 33a obtained by the maximum value detector 33 is input to the discriminator 34, and is compared with a predetermined threshold value 36a by this, and it is determined as the determination result 22a whether it is larger than the threshold value 36a. Is output.

Next, the quantization width determiner 16 first obtains the cumulative value 171 of the generated code amount obtained so far from the generated code amount counter 17, for example, in the case of MPEG, the accumulated value up to that point in the frame. From the cumulative value of all generated code amounts (BN for macro block number N), the provisional quantization width Qi (Q for BN is set so that the generated code amount by the encoding becomes a predetermined amount).
_{N + 1} ) is determined. Next, the quantization width determiner 16
Based on the determination result of the color difference signal inspector 15, that is, the output 22a of the discriminator 34, the temporary quantization width Qi is changed to determine the actual quantization width Qr (see FIGS. 11 (a) (b)). ).

That is, when the discriminator 34 determines that the maximum absolute value 33a of the difference is less than or equal to the threshold value 36a, the provisional quantization width Qi is set to the actual quantization width Qr.
And When the discriminator 34 determines that the maximum absolute difference value 33a is larger than the threshold value 36a, the actual quantization width Qr is made smaller than the temporary quantization width Qi (FIG. 11 (a)). Then, Q _{N + 1} and Q _{N + 4} are changed to Q _{N + 1} ′ and Q _{N + 4} ′). For example, consider the case where the threshold value 36a is set to 30 for the macroblock MB in FIG. In FIG. 4 (a), the maximum absolute value of the difference is 68 as shown in FIG. 4 (c), which is larger than the threshold value 30. Therefore, the actual quantization width Qr is calculated from the provisional quantization width Qi. Also make it smaller. FIG.
In (b), since the maximum absolute value of the difference is 18, the actual quantization width Qr is set to the same value as the temporary quantization width Qi (FIG. 11).
In (a), Q _N , Q _{N + 2} , and Q _{N + 3} are directly changed to Q _N , Q _{N + 2} ,
Q _{N + 3} ).

As a method for making the actual quantization width Qr smaller than the tentative quantization width Qi, consider a case where the actual quantization width Qr is made smaller as the maximum absolute value of the difference becomes larger.
Here, the pixel value that each block of the RY signal can take is 1
6 to 240, and the value of the colorless level of the color difference signal is 128.
, The maximum possible absolute value of the difference is 11
It is 2. Therefore, when the actual quantization width Qr is set to be between 1/3 and 1/2 times the temporary quantization width Qi, the following expression (1) may be used. Here, the temporary quantization width is Qi, the actual quantization width is Qr, the maximum absolute value of the difference is X, and the threshold value is T. Qr = Qi / {2+ (X−T) / (112−T)} (X> T) (1) Here, the actual quantization width Qr is made to be 1/3 or more of the temporary quantization width Qi. Is smaller than 1/3 of the temporary quantization width, too much code amount is allocated to the block whose absolute value of difference is larger than the threshold value, and the code amount cannot be allocated to other macroblocks. This is because the deterioration of those macroblocks becomes large. On the other hand, the reason why the provisional quantization width is made 1/2 or less is that if it is made larger than 1/2, the effect of suppressing image quality deterioration cannot be obtained so much. Quantization is performed by the quantizer 13 using the actual quantization width Qr thus obtained by the quantization width determiner 16 and a predetermined quantization matrix.

As described above, in the first embodiment of the present invention, a simple threshold value calculation is performed on the RY signal.
That is, the absolute value of the difference between the average value of the pixel values in the block and the colorless level of the color difference signal is calculated, and the maximum value in the macroblock, which is the pixel set unit in which the quantization width of the absolute value is to be changed, is predetermined. Is larger than the threshold value of, and the quantization width for the block larger than the predetermined threshold value is
From the quantization width obtained based on the generated code amount obtained by cumulatively counting the code string lengths sequentially coded by the variable length encoder in the set of blocks where the code amount by the variable length encoder should be constant Since the image quality is reduced, the visual deterioration of the image quality can be reduced.

Embodiment 2 . Next, a second embodiment of the present invention will be described with reference to FIGS. 1 and 5. The second embodiment differs from the first embodiment in that the color difference signal inspector is
Checking not only the block of -Y signals, but also the block of BY signals.

FIG. 5 is a block diagram of the color difference signal inspecting device according to the second embodiment, in which 51 is an R-
A Y signal average value calculator, 52 is a BY signal average value calculator. Others 32, 33 and 34 are an absolute difference calculator, a maximum value detector and a discriminator, as in FIG.

Next, the operation of the color image coding apparatus according to the second embodiment will be described. The digital image signal 18 is supplied to the block converter 11 by the Y signal, the RY signal,
After each BY signal is converted into a block, R
Blocks Cr1 and Cb1 of the −Y signal and the BY signal are input to the color difference signal inspector 15b. Color difference signal inspector 15b
The RY signal block Cr1 input to the RY signal average value calculator 51 is input to the RY signal block Cb1.
Is input to the BY signal average value calculator 52. RY
Signal average value calculator 51 and BY signal average value calculator 52
In the same way as the average value calculator 31 of the first embodiment, the average value 51b, 52b of the pixel values in each block of each RY signal and BY signal is obtained.

The average value 5 of the pixel values obtained by the RY signal average value calculator 51 and the BY signal average value calculator 52, respectively.
1b and 52b are input to the absolute difference calculator 32. Then, for each average value 51b, 52b, the absolute value 32b of the difference from the value (128) of the colorless level 35a of the color difference signal is obtained. In this way, the absolute value 32b of the difference is obtained for all blocks of the RY signal and the BY signal in the macroblock MB, and then the maximum value detector 33 is used.
Then, all color difference signal blocks Cr in the macro block MB
The maximum absolute value 33b of the difference is calculated for each block of 8 × 8 pixels 1 and Cb1. Then, the discriminator 34 compares the obtained maximum value 33b with the threshold value 36a, and outputs the determination result 22b.

For example, in the macro block MB of FIG. 4A, two RY signal and BY signal blocks Cr1,
The average pixel value of each block of Cb1 is 15
0, 60, 50, 170. When the absolute value of the difference between the average value and the value (128) of the colorless level 35a is obtained, as shown in FIG.
It becomes 8,78,42. Therefore, the maximum value is 78, and a comparison is made between this value 78 and the value 30 of the threshold value 36a. The method of determining the actual quantization width by the quantization width determiner 16 using the determination result 22b of the discriminator 34 is the same as in the first embodiment.

With the above operation, the second embodiment
Then, a simple threshold value calculation is performed on the RY signal and the BY signal, that is, the absolute value of the difference between the average value of the pixel values in the block and the colorless level of the color difference signal is obtained, and the absolute value thereof is calculated. It is determined whether the maximum value in the macroblock, which is the unit of pixel set whose value quantization width is to be changed, is greater than a predetermined threshold value, and the quantization width for blocks larger than the predetermined threshold value is determined. Is reduced, it is possible to reduce the visual deterioration in image quality.

Third Embodiment Next, in the first embodiment, the threshold value 36a is set to a predetermined fixed value, but it can be set to a variable value. FIG. 6 shows an example of the configuration of the color difference signal inspecting device 15c according to the third embodiment of the present invention having such a configuration.

In FIG. 6, reference numerals 31, 32, 33 and 34 denote an average value calculator, an absolute difference calculator, a maximum value detector and a discriminator, as in FIG. 3, and 61 denotes the maximum value detector. The maximum value 33c, which is the output of 33, is input, and an average value calculator for calculating the average value, 36c is the output of the average value calculator 61, and is input to the discriminator 34 as a threshold value. .

In the third embodiment, the average value calculator 6 calculates the maximum value 33c of the absolute value of the difference obtained by the maximum value detector 33 in the macroblock whose quantization width is to be changed.
Enter 1 as well. Then, the average value calculator 61 obtains, for example, the average value of the maximum value for one frame and the average value is used by the discriminator 34 as the threshold value 36c in the next frame.

Therefore, by using the color difference signal inspector 15c according to the third embodiment, the threshold value for determining whether or not the color difference signal exceeds the threshold value is set in advance. It is possible to adaptively change it depending on the nature of the image of the frame, and it is possible to adaptively select the macroblock for reducing the quantization width according to the nature of the image.

With the above-described operation, the color image encoding apparatus according to the third embodiment of the present invention can provide RY
By performing a simple threshold value operation on the signal and the BY signal, and by making the threshold value a variable value, a macro block for reducing the quantization width in order to reduce the visual deterioration of the image quality, It can be adaptively selected according to the nature of the image.

The configuration in which the threshold value of the color difference signal discrimination value in the third embodiment is variable is the same as in the second embodiment.
Of course, the same can be applied to the above.

Fourth Embodiment The first embodiment of the present invention
In 3 to 3, the average value of the pixel values of the color difference signal block is obtained by the average value calculator 31, the RY signal average value calculator 51, and the BY signal average value calculator 52. Alternatively, the DC component of the block of the color difference signal obtained by the orthogonal transformer 12 may be used. For example, the configuration in this case in FIGS. 1 and 3 is as shown in FIG. That is, in FIG. 7 (a) and FIG. 7 (b) showing a block diagram of a color image coding apparatus and a block diagram of a color difference signal inspector according to Embodiment 4 of the present invention, the color difference signal inspector 15 is an orthogonal transformer. 12 to DC component 71 of block Cr1 of RY signal
By receiving, it is possible to perform the same operation as in the first and second embodiments of the present invention. This also applies to the other embodiments.

Further, in the above-described first to fourth embodiments, the case where the MPEG system is used as the color image encoding system has been described, but another system may be used.

In the first to fourth embodiments, Y, R-
The case where the same quantization width is used when performing quantization on the Y and B-Y signals has been described, but these use the same quantization width.

Further, in the above-described first to fourth embodiments, the case where the MPEG system is used as the color image encoding system has been described, but other system may be used.

In the first to fourth embodiments, Y, R-
The case where the same quantization width is used when performing quantization on the Y and B-Y signals has been described, but these do not use the same quantization width, but for example, for the Y signal, the conventional method is used. The temporary quantization width which is the same as the method may be used, and the actual quantization width may be used only for the RY and BY signals.

In the above-described first to fourth embodiments, the case where the digital image signal is in the 4: 2: 2 format has been described. However, in the present invention, the digital image signal is in the 4: 2 format.
It can also be applied to the case of 2: 0 or 4: 1: 1 format. That is, the digital image signal 18 is 4:
If the format is 2: 0, each component Y2, Cr2, C of the digital image signal 18 converted into blocks
b2 is as shown in FIG. 12 (a), and if the digital image signal 18 is in the 4: 1: 1 format,
The components Y3, Cr3, Cb3 of the digital image signal 18 converted into blocks are as shown in FIG. 12 (b).

When the input digital signal is in 4: 2: 0 or 4: 1: 1 format, as shown in FIG.
As shown in (b), since the RY signal blocks Cr2 and Cr3 in the macro blocks MB2 and MB3 are one,
The maximum value detector 33 is unnecessary.

Further, in the above-mentioned first to fourth embodiments, the size of the block is 8 × 8 pixels, but this may be another size.

In the first to fourth embodiments, the mathematical expression (1)
Although the case where the actual quantization width is set to 1/2 to 1/3 of the provisional quantization width has been described above, the actual quantization width may be set to a range different from this.

In the first to fourth embodiments, the mathematical expression (1)
Although the actual quantization width is obtained from the temporary quantization width by using as a conversion formula, other methods may be used as long as the actual quantization width is made smaller than the temporary quantization width.

In the first to fourth embodiments, the formula (1) is used as the quantization width conversion formula, but the conversion table recorded in the ROM or the like may be used for the conversion of the quantization width. Good.

Further, in the explanation of FIG. 6, the average value calculator 31 calculates the average value of one frame with respect to the maximum absolute value of the difference. The average value of Alternatively, the average value may be obtained in units other than the frame, such as the average value up to the immediately preceding macroblock.

[0052]

As described above, according to the color image encoding apparatus of the present invention, at the time of encoding, the difference between the average value of the pixel values in each block of the color difference signal and the value of the colorless level is determined. Since the absolute value is calculated and the quantization width is reduced when the absolute value is larger than the predetermined threshold value, a block for which the quantization width is reduced is selected by a simple threshold value setting for the color difference signal. It is possible to reduce the visual deterioration of the image quality.

That is, according to the color image coding apparatus of the first aspect, for a digital color image signal composed of a Y signal which is a luminance signal and two color difference signals RY and BY signals, In a color image encoding device for encoding, a block converter that divides the digital color image signal into blocks, and an orthogonal converter that performs an orthogonal transform for each block on the digital color image signal divided in each of the blocks , A quantizer for quantizing the orthogonal transform coefficient obtained by the orthogonal transformer, a variable length encoder for performing variable length coding on the quantized orthogonal transform coefficient, and the variable length encoder And a generated code amount counter for accumulatively counting code string lengths sequentially coded by the variable length encoder in the set of blocks for which the code amount according to One The absolute value of the difference between the average value of the pixel values in each block and the colorless level value of the color difference signal is calculated for at least one of the color difference signals, and the quantization width in the quantizer is changed. A color difference signal inspecting device that determines whether the maximum value of the absolute value in the pixel set unit is greater than a predetermined threshold value, and the maximum value in the color difference signal inspecting device is less than or equal to the threshold value. If judged,
When the quantization width used in the quantizer is obtained based on the code string length counted by the generated code amount counter, and when it is determined that the maximum value is larger than the threshold value in the color difference signal inspector, A quantizer width determiner for setting the quantizer width used in the quantizer to a value smaller than the quantizer width obtained based on the code string length counted by the generated code quantity counter. For the color difference signal, the difference between the average value of the pixel values in the block and the colorless level of the color difference signal is obtained, and the block in which the absolute value of the difference is larger than a predetermined threshold is the block in which visual deterioration is noticeable. Since the decision is made to reduce the quantization width for that block, the block for which the quantization width is reduced can be selected by a simple threshold value setting for the color difference signal. Picture Effect which can reduce the degradation is obtained.

According to the color image coding apparatus of the second aspect, in the color image coding apparatus of the first aspect, the threshold value used in the color difference signal inspector is:
Since the value is a fixed value, it is possible to select a block in which the quantization width is reduced by simply setting a threshold value for the color difference signal, and it is possible to obtain an effect that visual deterioration in image quality can be reduced.

According to the color image coding apparatus of the third aspect, in the color image coding apparatus of the first aspect, the threshold value used in the color difference signal inspector is:
Since it is assumed to be a variable value, by adaptively varying the threshold value according to the input digital video signal,
It is possible to adaptively reduce visual deterioration of image quality according to the property of the image signal.

According to the color image coding apparatus of the fourth aspect, in the color image coding apparatus of the third aspect, the threshold value used in the color difference signal inspector is:
Since it is the average value of the maximum values of the absolute values in the frame encoded immediately before, it is possible to adaptively reduce the deterioration of the visual image quality according to the property of the image of the immediately preceding frame. .

According to a fifth aspect of the color image encoding apparatus of the present invention, in the color image encoding apparatus according to any one of the first to fourth aspects, the actual quantization width is 1 of the temporary quantization width. Since it is set to / 2 to 1/3, it is possible to effectively reduce the visual deterioration of the image quality.

According to a sixth aspect of the color image encoding apparatus of the present invention, in the color image encoding apparatus of the first to fifth aspects, the temporary quantization width is Qi and the actual quantization width is Qi.
Let r be the maximum absolute value of the difference, X be the threshold value, and T be the threshold, and set the actual quantization width to the temporary quantization width as follows: Qr = Qi / {2+ (X−T) / (112−T )} (X> T), the visual deterioration of the image quality can be effectively reduced.

According to a seventh aspect of the color image encoding device of the present invention, in the color image encoding device of the first to sixth aspects, the orthogonal transformer converts the pixel value of the color difference signal block as an average value. Since the DC component of the pixel block of the obtained color difference signal is used, the average value calculator of the color difference signal can be omitted in the configuration according to any one of claims 1 to 6, and the circuit configuration can be simplified. can do.

[Brief description of drawings]

FIG. 1 is a block diagram of a color image encoding device according to a first embodiment of the present invention.

FIG. 2 is a schematic pattern diagram showing a block relationship between a luminance signal and a color difference signal when a digital video signal has a 4: 2: 2 pattern according to the first embodiment of the present invention.

FIG. 3 is a block diagram of a color difference signal tester in the color image coding device according to the first embodiment of the present invention.

FIG. 4 is a schematic pattern diagram showing an example ((a) and (b)) of average values of pixel values of color difference signal blocks in the color image encoding device according to the first embodiment of the present invention, and respective cases FIG. 3 is a diagram ((c), (d)) showing the absolute value, the maximum value, the real quantization width, and the temporary quantization width of the difference of.

FIG. 5 is a block diagram of a color difference signal tester in a color image coding apparatus according to a second embodiment of the present invention.

FIG. 6 is a block diagram of a color difference signal tester in a color image coding apparatus according to a third embodiment of the present invention.

FIG. 7 is a block diagram (a) of a color image encoding device according to a fourth embodiment of the present invention and a block diagram (b) of a color difference signal inspector.

FIG. 8 is a block diagram of a conventional color image encoding device.

9 is a diagram for explaining the operation of the color image coding apparatus according to Embodiment 1 of the present invention in FIG. 1 and the conventional color image coding apparatus in FIG.

FIG. 10 is a diagram (a) showing the cumulative value of the generated bit amount of the macroblock in the frame and the quantization width for explaining the operation of the conventional color image coding device, and quantization from the cumulative value. It is a figure (b) which shows the flow which determines the width.

FIG. 11 is a diagram illustrating an operation of the color image encoding device according to the first embodiment of the present invention, in which a cumulative value of generated code amounts of macroblocks in a frame, a temporary quantization width, and an output of a maximum value detector; Diagram showing discriminator output and actual quantization width (a), and diagram showing flow of determining actual quantization width from the cumulative value (b)
Is.

FIG. 12: Digital video signals are 4: 2: 0 respectively
FIG. 4 is a schematic pattern diagram showing a relationship between blocks of a luminance signal and a color difference signal in the case of a pattern, which is a 4: 1: 1 pattern ((a), (b)).

[Explanation of symbols]

 11 Block Transformer 20 Output of Block Transformer 12 Orthogonal Transformer 121 Output of Orthogonal Transformer 13 Quantizer 131 Quantizer Output 14 Variable Length Encoder 141 Variable Length Encoder Output 15a Color Difference Signal Tester 15b Color difference signal inspector 15c Color difference signal inspector 16 Quantization width determiner 161 Output of quantization width determiner 17 Generated code amount counter 171 Cumulative value of generated code amount counter 31 Average value calculator 31a Output of average value calculator 32 Absolute difference calculator 32a Output of absolute difference calculator 33 Maximum value detector 33a Output of maximum value detector 34 Discriminator 22a Discrimination result 22b Discrimination result 61 Average value calculator 36a Threshold value 36c Output of average value calculator Y1 Y signal block Cr1 RY signal block Cb1 BY signal block Y2 Y signal block Cr2 RY signal block Click Cb2 B-Y signal block Y3 Y signal block Cr3 R-Y signal block Cb3 B-Y signal block

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location H04N 7/32 G06F 15/66 330C H04N 7/137 Z

Claims (7)

[Claims] 1. A color image encoding apparatus for encoding a digital color image signal composed of a Y signal which is a luminance signal and two color difference signals RY and BY signals. A block converter that divides an image signal into blocks, an orthogonal transformer that performs an orthogonal transformation for each block on a digital color image signal that is divided for each block, and an orthogonal transformation coefficient obtained by the orthogonal transformer. , A variable length encoder for performing variable length coding on the quantized orthogonal transform coefficient, and a set of blocks in which the code amount by the variable length encoder should be constant. Then, a generated code amount counter for cumulatively counting the code string lengths sequentially encoded by the variable length encoder, and each block for at least one color difference signal of the two color difference signals The absolute value of the difference between the average value of the pixel values in and the value of the colorless level of the color difference signal is calculated, and the maximum value of the absolute value in the pixel set unit that changes the quantization width in the quantizer is A color difference signal inspecting device that determines whether it is greater than a predetermined threshold value, and if the maximum value is less than or equal to the threshold value in the color difference signal inspecting device, the generated code amount counter counts. The actual quantization width used in the quantizer is obtained based on the coded string length, and used in the quantizer when the color difference signal inspector determines that the maximum value is larger than the threshold value. Quantization width determiner that makes the actual quantization width a value smaller than the temporary quantization width, with the quantization width obtained based on the code string length counted by the generated code amount counter as the temporary quantization width And a color image coding device. 2. The color image coding apparatus according to claim 1, wherein the threshold value used in the color difference signal inspector is a fixed value. 3. The color image coding apparatus according to claim 1, wherein the threshold value used in the color difference signal tester is a variable value. 4. The color image coding apparatus according to claim 3, wherein the threshold value used in the color difference signal inspector is an average value of the maximum absolute values in the immediately preceding coded frame. A color image encoding device characterized by being present. 5. The color image coding apparatus according to claim 1, wherein the actual quantization width is set to 1/2 to 1/3 of the provisional quantization width. Color image encoding device. 6. The color image coding apparatus according to claim 1, wherein the temporary quantization width is Qi, the actual quantization width is Qr, the maximum absolute difference is X, and the threshold is X. Let T be a value, and set the actual quantization width with respect to the temporary quantization width by using the formula: Qr = Qi / {2+ (X−T) / (112−T)} (X> T) A color image encoding device characterized by the above. 7. The color image coding apparatus according to claim 1, wherein a DC component of a block of color difference signals obtained by the orthogonal converter is used as an average value of pixel values of the color difference signal block. A color image encoding device characterized by using.