JPH11298895A - Image coder and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image coder where quantization distortion is much more reduced and an image is compression-coded with higher quality. SOLUTION: An AC component MAX value detection section 31 detects a maximum value of an AC component for each macro block of discrete cosine transform DCT and a class number selection section 32 compares the maximum value with a prescribed threshold value to select a class number. On the other hand, an AC component arithmetic section 34 calculates a total sum of the AC components, an AC component data analysis section 35 compares the total sum with a predetermined prescribed threshold value and outputs a signal indicating correction of the class number to a class number discrimination section 36 when the total sum of the AC components is larger. The class number discrimination section 36 makes correction of decreasing the selected class number by one or the like to decide the class number when the correction is instructed and outputs the result to a quantization number estimate section 22.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus for compressing and encoding image data, and more particularly to an image encoding apparatus capable of reducing quantization distortion and encoding an image with higher quality. Regarding the method.

[0002]

2. Description of the Related Art As a method of compressing and encoding an image, for example, a transform such as DCT is performed, the obtained transform coefficient is adaptively quantized so that the total data amount is equal to or less than a predetermined value, and a variable length code is further encoded. Is generally used as an efficient encoding method. When encoding is performed by such a method, the quantization step is determined, for example, by detecting the maximum value of AC data after DCT, comparing the maximum value with a predetermined threshold, and determining the class number. The combination is determined by selecting a combination from the predetermined combination of the class number and the Q number such that the data amount after quantization is within a predetermined value.

[0003]

However, in the conventional method of determining the class number based only on the maximum value of the AC component of the DCT block and determining the quantization step, quantization distortion is generated. Easy to do,
There is a problem that image quality is likely to be distorted during reproduction.

Accordingly, it is an object of the present invention to provide an image encoding apparatus capable of reducing the quantization distortion and compressing and encoding an image with higher quality. It is another object of the present invention to provide an image encoding method capable of reducing the quantization distortion and compressing and encoding an image with higher quality.

[0005]

Means for Solving the Problems To solve the above-mentioned problems, a method in which the amount of AC component in one DCT unit is analyzed, and the class number is corrected based on the result, whereby only the maximum value of the AC component is referred to. Quantization distortion can be suppressed as compared with.

Therefore, the image encoding apparatus of the present invention
DCT that performs discrete cosine transform (DCT) on an input digital video signal for each predetermined DCT block
Means, class selection means for selecting a quantization class based on the maximum value of the AC component of the coefficient obtained by the DCT, and the selected class based on the sum of the AC components of the coefficient obtained by the DCT Class correction means for correcting the coefficient, a quantization step determining means for determining a quantization step based on the corrected class, and a coefficient obtained by the DCT based on the determined quantization step. And a quantizing means for transforming the data.

Further, another image coding apparatus of the present invention is a DCT means for performing a discrete cosine transform (DCT) for each predetermined DCT block on an input digital video signal, and obtained by the DCT. First class selecting means for selecting a first quantization class based on the maximum value of the AC component of the coefficient, and selecting a second quantization class based on the sum of the AC components of the coefficient obtained by the DCT Second
Classifying means, and quantizing the coefficient obtained by the DCT by selecting the one of the selected first quantization class and the selected second quantization class that has a smaller quantization step. Class determining means for determining as a class to be provided, based on the determined class,
Quantization step determining means for determining a quantization step;
Based on the determined quantization step, the DCT
And a quantization means for quantizing the coefficient obtained by the above.

Further, in the image coding method according to the present invention, the input digital video signal is subjected to discrete cosine transform (DCT) for each predetermined DCT block, and the DCT is performed.
A quantization class is selected based on the maximum value of the AC component of the coefficient obtained by
Correcting the selected class based on the sum of the C components, determining a quantization step based on the corrected class, based on the determined quantization step,
The coefficients obtained by the DCT are quantized.

Further, another image encoding method according to the present invention comprises:
The input digital video signal is subjected to a discrete cosine transform (DCT) for each predetermined DCT block, and a first quantization class is selected based on the maximum value of the AC component of the coefficient obtained by the DCT. , Selecting a second quantization class based on the sum of the AC components of the coefficients obtained by the DCT, among the selected first quantization class and the selected second quantization class, A class having a smaller quantization step is determined as a class to be used for quantization of the coefficient obtained by the DCT, and a quantization step is determined based on the determined class. , The coefficient obtained by the DCT is quantized.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment A first embodiment of the present invention will be described with reference to FIGS. In the first embodiment, the present invention will be described by exemplifying an image encoding apparatus that compresses and encodes an image using a discrete cosine transform (DCT) in order to record image data on a desired recording medium. .

First, the overall configuration of the image encoding apparatus will be briefly described. FIG. 1 is a block diagram showing the configuration of the image encoding device. The image encoding device 10 includes:
Analog / digital conversion unit 11, blocking unit 12,
Shuffling unit 13, DCT unit 14, quantization unit 15,
It has a variable length coding unit 16, a framing unit 17, a deshuffling unit 18, and an error correction coding unit 19.

In the image coding apparatus 10 having such a configuration, the input video data is converted into a digital signal by an analog / digital conversion unit 11, and each pixel data of the video data is converted into a macroblock by a blocking unit 12. It divides and performs shuffling in a macroblock unit in order to average the code amount in the shuffling unit 13,
The DCT unit performs discrete cosine transform (DCT) for each 8 × 8 DCT block. The converted DCT coefficients are read by zigzag scan, and are classified in the quantization unit 15 according to the DCT code amount for each predetermined area, and quantization is performed using a quantization table. And
The variable-length coding section 16 modulates and codes the data, thereby reducing the number of effective bits, and sends the result to an error correction coding section 19 via a framing section 17 and a deshuffling section 18. The signal to which the error correction code is added is recorded on a recording medium, for example, after being further modulated.

Next, the quantization section 15 of the image encoding apparatus 10 according to the present invention will be described with reference to FIG. FIG. 2 is a block diagram illustrating a configuration of the quantization unit 15. The quantization unit 15 includes a class number determination unit 21, a quantization number estimation unit 22, and a quantization unit 23.

The class number determining unit 21 is a DCT unit 14
A class number for determining a quantization step is determined for the image data of each 8 × 8 DCT block that has been DCT-transformed in step (1), and is output to the quantization number estimation unit 22. The configuration of the class number determination unit 21 will be described later in detail.

The quantization number estimating unit 22 selects a quantization step based on the class number determined by the class number determining unit 21 so that the data can be compressed to, for example, data having a length of 3040 bits per block.

The quantization unit 23 includes a quantization number estimation unit 2
Based on the quantization step determined in step 2, the DCT unit 1
The image data for each 8 × 8 DCT block that has been DCT-transformed in step 4 is sequentially quantized, and the quantized image data is sequentially output to the variable-length encoding unit 16.

Next, the configuration of the class number determining unit 21 will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of the class number determination unit 21. The class number determining unit 21 includes an AC component MAX value detecting unit 31, a class number selecting unit 32, a class number correcting unit 33, and a clath number determining unit 36. The class number correction unit 33 further includes an AC component calculation unit 34 and an AC component data analysis unit 35.

The AC component MAX value detecting section 31 converts the DCT image data input from the DCT section 14 into 8 × 8
For each DCT block, the AC of that DCT block
The maximum value of the component is detected and output to the class number selection unit 32. Note that the class number determination unit 21 has a DC
It is assumed that the image data of each 8 × 8 DCT block that has been DCT-transformed by the T unit 14 is sequentially zigzag-scanned from a low-frequency component to a high-frequency component and input as shown in FIG.

The class number selecting section 32 outputs the AC component MA
AC of each DCT block input from X value detecting section 31
The class number is selected based on the maximum value of the component, and is output to the clath number determination unit 36. In the class number selection unit 32, as shown in FIG.
If the number is 11 or less, 0 is set as the class number and AC
If the maximum value of the component is 12 or more and 23 or less, 1 is set as the class number. If the maximum value of the AC component is 24 or more and 35 or less, 2 is set as the class number. If the maximum value of the AC component is 36 or more, Selects 3 as the class number.

The AC component calculation unit 34 of the class number correction unit 33 converts the DCT image data input from the DCT unit 14 into DCT blocks for each 8 × 8 DCT block.
The sum of the AC components of the T block is calculated and output to the AC component data analysis unit 35.

The AC component data analyzer 35 compares the sum of the AC components of each DCT block input from the AC component calculator 34 with a predetermined threshold value, and
If the sum of the C components is larger, this is indicated, and a signal instructing the correction of the class number is output to the clath number judging unit 36. When the total sum of the input AC components is equal to or less than the threshold value, no instruction to correct the class number is issued.

The class number determination unit 36 receives a signal from the AC component data analysis unit 35 of the class number correction unit 33 to instruct the class number input from the class number selection unit 32 to correct the class number. If so, the correction is performed by, for example, decreasing the class number by one, and the resulting class number is output to the quantization number estimation unit 22. When the signal instructing the correction of the class number is not input from the AC component data analysis unit 35, the clan number determination unit 36 sends the class number input from the class number selection unit 32 to the quantization number estimation unit 22 as it is. Output.

Finally, the operation of the quantization unit 15 will be described. The image data DCT-transformed by the DCT unit 14 is sequentially zigzag scanned from a low-frequency component to a high-frequency component and input to the class number determination unit 21, where the AC component M
The AX value detector 31 detects the maximum value of the AC component for each DCT block, and the class number selector 32
The maximum value of the component is compared with a predetermined threshold to select a class number. On the other hand, in the AC component calculation unit 34 of the class number correction unit 33, the DCT
The sum of the AC components of the block is calculated, and the sum of the AC components is compared with a predetermined threshold by the AC component data analysis unit. If the sum of the input AC components is larger, an instruction to correct the class number is issued. The signal to be changed is determined by the clath number judging unit 36.
Output to

When a signal instructing this correction is input, the class number determining unit 36 corrects the class number selected by the class number selecting unit 32 by, for example, decreasing the class number by one. The number is output to the quantization number estimating unit 22. The quantization number estimating unit 23 determines a quantization step based on the detected class number so that the code amount per DCT block is equal to or less than a predetermined amount, and uses the quantization step to perform the quantization step. At 24, quantization is performed.

As described above, in the image coding apparatus 10 according to the present embodiment, in the class number determination unit 21 of the quantization unit 15, the AC component calculation unit 34 calculates the total sum of the AC components of the image data for each DCT block. And the AC component data analyzer 35 analyzes it to correct the class number. Therefore, compared with the case where the maximum value of the AC component is simply used, the quantization distortion can be reduced, and the image can be encoded with high quality.

Second Embodiment A second embodiment of the present invention will be described. Although the second embodiment of the present invention has the same configuration as the first embodiment, the method of correcting the class number performed in the class number determination unit 21 of the quantization unit 15 is the same as the first embodiment described above. An example different from the embodiment will be described. Image encoding device 1
Since the configurations of 0 and the quantization unit 15 are exactly the same as those of the first embodiment described above, a description thereof will be omitted, and only the class number determination unit 21 of the quantization unit 15 will be described.

The AC component MAX value detection unit 31 of the image encoding apparatus 10 according to the second embodiment converts the DCT-inputted image data of each 8 × 8 DCT block input from the DCT unit 14 into Four areas as shown in FIG. 5 are defined, and the maximum value of the AC component is detected for each area,
Output to the class number selection unit 32.

The class number selecting section 32 outputs the AC component MA
Based on the maximum value of the AC component of each area of each DCT block input from the X value detection unit 31, referring to a predetermined threshold value determined in the same manner as in the first embodiment,
The class number is selected for each area, and the selected class number is output to the clath number determination unit 36.

On the other hand, the AC component calculation unit 34 of the class number correction unit 33 calculates the area of each of the 8 × 8 DCT blocks of the DCT image data input from the DCT unit 14 as shown in FIG. Then, the sum of the AC components is calculated,
Output to the AC component data analyzer 35.

The AC component data analysis unit 35, based on the sum of the AC components for each area of each DCT block input from the AC component calculation unit 34,
With reference to a predetermined threshold value determined in the same manner as in 2, the provisional class number for each area is determined and output to the clath number determination unit 36.

The class number determining unit 36 includes a class number input from the class number selecting unit 32 and an AC component data analyzing unit 35 for each area of each DCT block.
By comparing the input class number with the provisional class number, a class number having a smaller quantization step is determined as a normal class number, and is output to the quantization number estimating unit 22.

As a result, the quantization number estimating unit 22
With reference to the class number determined by the class number determination unit 21, for example, with reference to a table as shown in FIG. 6, the quantum or step is determined and output to the quantization unit 23.

As described above, the class numbers are respectively obtained based on the maximum value of the AC component and the sum total of the AC components, and by selecting the preferred one, the class number substantially determined only by the maximum value of the AC component is determined. The number can be corrected, and the present invention may have such a configuration. Further, even in the image encoding apparatus according to the second embodiment, compared with the case where the maximum value of the AC component is simply used,
Quantization distortion can be reduced, and an image can be encoded with high quality.

[0034]

According to the image encoding apparatus and method of the present invention, quantization distortion can be reduced and an image can be compression-encoded in a high quality state.

[Brief description of the drawings]

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

FIG. 2 is a block diagram illustrating a configuration of a quantization unit of the image encoding device illustrated in FIG. 1;

FIG. 3 is a block diagram illustrating a configuration of a class number determining unit of the quantization unit illustrated in FIG. 2;

FIG. 4 is a diagram for explaining a zigzag scan.

FIG. 5 is used in a second embodiment of the present invention;
FIG. 3 is a diagram for describing area division in a DCT block.

FIG. 6 is a diagram illustrating an example of a method of determining a quantization step performed by a quantization number estimating unit of the quantization unit illustrated in FIG. 2;

[Explanation of symbols]

10 image encoding device, 11 analog / digital conversion unit, 12 blocking unit, 13 shuffling unit,
14 DCT unit, 15 quantization unit, 16 variable length coding unit, 17 framing unit, 18 deshuffling unit, 19 error correction coding unit, 21 class number determination unit, 22 quantization number Estimating unit, 23 ... Quantizing unit, 3
1 AC component MAX value detector, 32 class number selector, 33 class number corrector, 34 AC component calculator, 35 AC component data analyzer, 36 clans number determiner

Claims (10)

[Claims] An input digital video signal is subjected to discrete cosine transform (DCT) for each predetermined DCT block.
DCT means for performing quantization; class selection means for selecting a quantization class based on the maximum value of AC components of the coefficients obtained by the DCT; and the selection based on the sum of AC components of the coefficients obtained by the DCT. Class correction means for correcting the corrected class, quantization step determination means for determining a quantization step based on the corrected class, and DCT based on the determined quantization step
And a quantizing means for quantizing the coefficient obtained by the above. 2. The method according to claim 1, wherein the class selection unit selects the quantization class for each of the plurality of regions obtained by dividing the DCT block into a plurality of predetermined regions, and the class correction unit selects the quantization class for each of the regions. The image encoding device according to claim 1, wherein a class is corrected, and the quantization step determination unit determines the quantization step for each of the regions. 3. The image coding apparatus according to claim 2, further comprising a variable length coding unit for performing variable length coding on the quantized coefficient. 4. A discrete cosine transform (DCT) for an input digital video signal for each predetermined DCT block
And selecting a quantization class based on the maximum value of the AC component of the coefficient obtained by the DCT, and correcting the selected class based on the sum of the AC components of the coefficient obtained by the DCT, A quantization step is determined based on the corrected class, and the DCT is determined based on the determined quantization step.
Image coding method for quantizing the coefficients obtained by the above. 5. The correction of the class is performed for each DCT block, and the selection of the class and the determination of the quantization step include:
5. The image encoding method according to claim 4, wherein the method is performed for each of a plurality of predetermined areas obtained by dividing the DCT block. 6. A discrete cosine transform (DCT) for each predetermined DCT block with respect to an input digital video signal.
DCT means for performing the following; first class selecting means for selecting a first quantization class based on the maximum value of the AC component of the coefficient obtained by the DCT; and DCT means of the AC component of the coefficient obtained by the DCT. A second class selecting means for selecting a second quantization class based on the summation, wherein the quantization step is smaller among the selected first quantization class and the selected second quantization class. Class determining means for determining a class to be used as a class to be subjected to quantization of the coefficient obtained by the DCT; quantization step determining means for determining a quantization step based on the determined class; The DCT based on the quantized step
And a quantizing means for quantizing the coefficient obtained by the above. 7. The first class selecting means and the second class selecting means.
Class selecting means selects the first and second quantization classes for each area obtained by dividing the DCT block into a plurality of predetermined areas, and the class determining means selects the class for each of the areas. The image encoding apparatus according to claim 6, wherein the quantization step determination unit determines the quantization step for each of the regions. 8. The image coding apparatus according to claim 7, further comprising a variable length coding unit for performing variable length coding on the quantized coefficient. 9. A discrete cosine transform (DCT) for an input digital video signal for each predetermined DCT block
And selecting a first quantization class based on the maximum value of the AC component of the coefficient obtained by the DCT, and selecting a second quantization class based on the sum of the AC components of the coefficient obtained by the DCT. And among the selected first quantization class and the selected second quantization class, a class having a smaller quantization step is provided for quantization of the coefficient obtained by the DCT. A quantization step based on the determined class; and a DCT based on the determined quantization step.
Image coding method for quantizing the coefficients obtained by the above. 10. The method according to claim 1, wherein the selection of the first class, the selection of the second class, the determination of the class, and the determination of the quantization step are performed for each of a plurality of predetermined regions obtained by dividing the DCT block. 10. The image encoding method according to 9.