KR20080004392A - Scalable video encoding/decoding method and apparatus thereof - Google Patents

Abstract

A scalable video encoding/decoding method and an apparatus thereof are provided to override an existing weight given in a slice unit to a corresponding block of an enhancement layer of a reference frame with a greater weight than the existing weight in forming the reference block of the enhancement layer when the video data of a macro block of a basic layer is in a skip mode, thereby improving efficiency of SVC(Scalable Video Coding). A scalable video encoding method comprises the following steps of: determining whether the block of a basic layer corresponding to the block of an enhancement layer of a current frame to encode is in a skip mode(S210); overriding an existing weight set in the block of the enhancement layer of a reference frame corresponding to the block of the enhancement layer of the current frame with a new weight when the block of the basic layer is in the skip mode(S220); and generating a reference block for the block of the enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the basic layer of the current frame by using the new weight(S230).

Description

Scalable video encoding / decoding method and apparatus thereof

The present invention relates to a scalable video encoding / decoding method and an apparatus thereof, and more particularly, to a weight of a macroblock of an enhancement layer when the macroblock mode of the base layer is a skip (SKIP) block in AR-FGS. The present invention relates to a scalable video encoding / decoding method and apparatus for improving coding efficiency by generating a reference block by overriding a skip mode weight higher than a weighting factor of.

Adaptive Reference Fine Gradunality Scalability (AR-FGS) is a Signal to Noise Ratio (SNR) in Scalable Video Coding (SVC), which is currently being internationally standardized by JVT (Joint Video Team) in MPEG and ITU-T. A technique for improving coding efficiency by performing temporal prediction in a FGS coding method of scalability.

Signal to noise ratio (SNR) scalable is a technology that improves image quality in proportion to the received bit rate according to a variable network situation. Among these SNR scalable technologies, Fine Gradunality Scalability (GFS) is a technique for receiving a bitstream arbitrarily truncated according to a network condition and improving image quality in proportion to the amount of the bitstream received. However, because the FGS technique does not know the bit rate to be transmitted, it cannot have a temporal prediction structure that leads to a high coding efficiency improvement in the video codec. If the temporal prediction structure is applied without any consideration, a drift occurs due to a mismatch between the reference image for motion compensation of the encoder and decoder stages, resulting in a sharp performance degradation in terms of the reproduced image and coding efficiency.

Adaptive Reference-Fine Grain Scalability (AR-FGS) technology effectively controls drift and takes advantage of the performance improvements of temporal prediction structures. This AR-FGS technique generates a reference block (or macroblock) for motion compensation by a weighted sum of reference blocks obtained from a partially decoded upper layer and a lower layer. The AR-FGS method implemented in this way can improve FGS coding performance and control drift.

1 is a conceptual diagram of reference block generation in Adaptive Reference-Fine Grain Scalability (AR-FGS) according to the prior art.

은 FGS 계층(향상 계층)에서 코딩될 블록의 신호이다.

는 기본 계층과 향상 계층의 가중 합의 조합으로 만 들어지는 움직임 보상 참조 블록의 신호를 의미한다. 향상 계층의 참조 블록의 신호는

로 기본 계층의 양자화된 계수는

으로 표기하고 변환은

로 표기하였다. 그리고 기본 계층의 양자화된 변환 계수는

로 표기하였다.Referring to Figure 1, the size of the block is MxN

Is the signal of the block to be coded in the FGS layer (enhancement layer).

Denotes a signal of a motion compensation reference block made by a combination of a weighted sum of a base layer and an enhancement layer. The signal of the reference block of the enhancement layer is

The quantized coefficients of the base layer

And the conversion is

Indicated as. And the quantized transform coefficients of the base layer

Indicated as.

Generation of a reference block in AR-FGS is performed in two ways.

를 향상 계층의 가중치로,

를 기본 계층의 가중치로 이용하여 기본 계층의 대응 블록과 향상 계층의 대응 블록의 가중 합으로 참조 블록을 형성한다.1.If the quantized coefficients in the base layer are all zeros, the equation (1)

As the weight of the enhancement layer,

Is used as a weight of the base layer to form a reference block by a weighted sum of the corresponding blocks of the base layer and the corresponding blocks of the enhancement layer.

````(1)

`` '' (1)

를, 향상 계층에 대응되는 변환 계수에는

를 곱하고 이들의 가중 합으로 변환 계수를 얻는다. 변환 계수 영역에서 대응되는 기본 계층 위치의 변환 계수가 0이 아니라면 (3)의 식, 즉 기본 계층 신호를 그대로 이용한다. 이렇게 획득한 변환 계수를 역변환을 통하여 참조 블록을 형성한다.2. In other cases (one or more non-zero quantized coefficients exist in the base layer), a reference block is formed in the transform coefficient region. If the transform signal of the base layer at the corresponding position is 0, the transform coefficient corresponding to the base layer in the equation (2), i.

The transform coefficient corresponding to the enhancement layer

Multiply by the weighted sum of them If the transform coefficient of the corresponding base layer position in the transform coefficient area is not 0, the equation (3), that is, the base layer signal is used as it is. The transform coefficient thus obtained is formed through a inverse transform to form a reference block.

```(2)

`` `(2)

```(3)

`` `(3)

) 와, 각 기본 계층의 블록에서 잔여 영상의 값이 모두 '0'이 아닐 경우 DCT 도메인(domain)으로 변환하여 모든 변환 계수의 값이 '0'이 아닌 계수가 존재할 경우의 가중치(

)를 따로 전송을 한다. 여기서 가중치(

,

)는 상위 계층의 가중치로 0 과 1 사이의 값을 가지며, 하위 계층의 가중치는 1 - 상위계층의 가중치, 즉, (1-

또는 1-

)를 갖는다.The weights are given in slice units, and the weights when the residual image values of all pixels in the blocks of each base layer are all '0' (

) And if the residual image values in the blocks of each base layer are not all '0', the weight is converted to the DCT domain so that the coefficients of all transform coefficients having non-zero coefficients (

Send separately). Where weights (

,

) Is a weight of the upper layer and has a value between 0 and 1, and the weight of the lower layer is 1-the weight of the upper layer, that is, (1-

Or 1-

Has

FGS coding is performed using the advantage of temporal prediction structure with reference blocks formed in the above manner. The method shows a relatively much higher performance than conventional FGS coding in video coding requiring real time, and an improved performance than conventional FGS coding in general video coding.

Video coding techniques, such as the MPEG4 and H.264 standards, use several prediction techniques to improve coding efficiency. The skip (SKIP) mode is a mode in which the block data of the base layer does not exist and the data of the reference picture is taken and used as it is, and there is no change in data in the time direction. Therefore, if the data is imported and used as it is expected that there will be no change of data in the enhancement layer, the performance can be improved, and even if it is not transmitted, the probability of drift through wrong reference in the skip block will be low.

An object of the present invention is to provide a scalable coding method and apparatus for improving coding performance and reducing the probability of drift when image data of a macroblock of a base layer is a skip mode.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

According to another aspect of the present invention, there is provided a scalable video encoding method comprising: determining whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be encoded is a skip mode; If the block of the base layer is the skip mode, overriding an existing weight set in the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame with a new weight; And generating a reference block for the block of the enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight.

According to another aspect of the present invention, there is provided a scalable video decoding method comprising: determining whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be decoded is in a skip mode; If the block of the base layer is the skip mode, overriding an existing weight set in the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame with a new weight; And generating a reference block for the block of the enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight.

According to an aspect of the present invention, there is provided a scalable video encoding apparatus comprising: a mode determination unit determining whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be encoded is a skip mode; A weight override unit for overriding an existing weight set in a block of an enhancement layer of a reference frame corresponding to a block of an enhancement layer of the current frame with a new weight when the block of the base layer of the current frame is a skip mode; And a reference block generator configured to generate a reference block for a block of an enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame using the new weight. Can be.

According to an aspect of the present invention, there is provided a scalable video decoding apparatus comprising: a mode determination unit determining whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be decoded is a skip mode; A weight overriding unit configured to override an existing weight set in a block of an enhancement layer of a reference frame corresponding to a block of an enhancement layer of the current frame with a new weight when the block of the base layer is a skip mode; And a reference block generator configured to generate a reference block for a block of an enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame using the new weight. Can be.

In order to achieve the above technical problem, it is characterized by providing a computer-readable recording medium having recorded thereon a program for executing the scalable video encoding method and the scalable video decoding method in a computer.

According to the present invention, when the image data of the block (macro block) of the base layer is in the skip mode, when the reference block of the enhancement layer is formed, the existing weight given in slice units to the corresponding block of the enhancement layer of the reference frame is higher. By overriding the weight of the SKIP mode, the SVC coding efficiency can be improved.

In addition, compared to the case where the temporal prediction structure is applied without considering whether the mode is a skip mode, a probability of occurrence of drift due to a mismatch between a reference picture for motion compensation between the encoder and the decoder may be reduced.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the same elements among the drawings are denoted by the same reference numerals and symbols as much as possible even though they are shown in different drawings. In the following description of the present invention, detailed descriptions of well-known functions or configurations will be omitted when it is determined that the detailed description may unnecessarily obscure the subject matter of the present invention.

In the present invention, the terms "picture" and "frame" are used interchangeably to represent image data in a video sequence.

2 is a flowchart illustrating a scalable encoding method according to an embodiment of the present invention.

Referring to FIG. 2, the scalable video encoding apparatus determines whether a block mode of a base layer corresponding to a block of an enhancement layer of a current frame to be encoded is a skip (SKIP) mode (S210). The skip mode is a mode in which the block of the base layer of the current frame does not transmit additional data of the base layer and directly imports the block data of the base layer of the reference frame. Therefore, the encoding apparatus may compare the blocks of the base layer of the current frame and the reference frame, for example, and determine whether the skip mode is the same whether the block data of the current frame is the same as the block data of the reference frame in the time direction.

If the block of the base layer of the current frame is the skip mode, the encoding apparatus overrides a new weight (hereinafter, referred to as a skip mode weight) to an existing weight set in the block of the enhancement layer of the reference frame (S220). ). The skip mode weight may be set to be larger than the existing weight set in units of slices to increase coding efficiency by increasing the data usage ratio of the enhancement layer. The skip mode weight may be coded and transmitted in the slice header along with the existing weight. As a result, the decoder may check the mode for each block, and then use the skip mode weight for only the skip mode block, and the other blocks may generate the reference block using the existing weight.

The encoding apparatus configures a reference block for the block of the enhancement layer of the current frame to be encoded using the weighted sum (S230). If the block mode of the base layer of the current frame is the skip mode, the encoding apparatus performs a weighted sum of the corresponding block of the enhancement layer of the reference frame to which the skip mode weight is applied and the corresponding block of the base layer of the current layer to which the weight is calculated based on the skip mode weight. Generates a reference block. If the block mode of the base layer of the current frame is not the skip mode, the encoding apparatus generates a reference block by weighted sum of the corresponding block of the enhancement layer of the reference frame and the corresponding block of the base layer of the current frame using existing weights. do.

AR-FGS block encoding is performed on the block of the enhancement layer of the current frame based on the generated reference block (S240).

3 is a flowchart illustrating a scalable video decoding method according to an embodiment of the present invention.

Referring to FIG. 3, the scalable video decoding apparatus receives the encoded bitstream from the encoding apparatus (S310). The received bitstream may include a block encoded in a skip mode, skip mode information, and skip mode weight information for generating a reference block.

Next, the decoding apparatus determines whether the block mode of the base layer corresponding to the block of the enhancement layer of the current frame to be decoded in the bitstream is the SKIP mode (S320). The skip mode determination indicates that skip mode information included in the bitstream, for example, information indicating that a corresponding block has no data, a corresponding syntax element such as a specific syntax element such as a skip flag, is a skip mode. Can be inferred from the information.

When the block mode of the base layer of the current frame is the skip mode, the decoding apparatus overrides a skip (SKIP) mode weight to an existing weight set in the block of the enhancement layer of the reference frame (S330). If the block mode of the base layer is the skip mode, the decoding apparatus extracts a skip mode weight included in the received bitstream, and overrides the skip mode weight with an existing weight of the block of the enhancement layer of the reference frame. The skip mode weight may be extracted from a slice header in the bitstream.

The decoding apparatus configures a reference block for the block of the enhancement layer of the current frame to be decoded using the weighted sum (S340). When the block of the base layer of the current frame is determined to be the skip mode, the decoding apparatus weights the corresponding block of the enhancement layer of the reference frame to which the skip mode weight is applied and the corresponding block of the base layer of the current layer to which the weight calculated from the skip mode weight is applied. Generate a reference block by sum. If it is determined that the block mode of the base layer of the current frame is not the skip mode, the decoding apparatus generates a reference block by weighted sum of the corresponding block of the enhancement layer of the reference frame and the block of the base layer of the current frame using the existing weights. do.

The decoding apparatus performs AR-FGS block decoding on the block of the enhancement layer of the current frame based on the generated reference block (S350).

4 is a flowchart illustrating a scalable video encoding method according to another preferred embodiment of the present invention. Detailed description of the contents overlapping with FIG. 2 will be omitted.

Referring to FIG. 4, the encoding apparatus determines whether to set a flag indicating execution of SKIP mode weight overriding (S410).

When the overriding performance flag is set to '1', the encoding apparatus determines whether the block mode of the base layer of the current frame is the SKIP mode (S420).

If the block of the base layer is the skip mode, the encoding apparatus overrides the skip mode weight to the existing weight set in the block of the enhancement layer of the reference frame (S430).

The encoding apparatus configures a reference block for the block of the enhancement layer of the current frame by using the weighted sum (S440). If the overriding execution flag is set to '1' and the block of the base layer is determined to be the skip mode, the block of the enhancement layer of the skipped frame weighted reference frame and the base layer of the base layer of the current frame to which the weight is calculated are calculated from the skipped mode weight. The reference block is generated by the weighted sum of the blocks. If the encoding apparatus does not set the overriding performance flag to '1' or the block mode of the base layer of the current frame is not the skip mode, the corresponding layer of the enhancement layer of the reference frame and the base layer of the current frame are used by using the existing weight. A reference block is generated by the weighted sum of the corresponding blocks.

The encoding apparatus performs AR-FGS block encoding on the enhancement layer of the current frame based on the generated reference block (S450).

5 is a flowchart illustrating a scalable video decoding method according to another preferred embodiment of the present invention. Detailed description of the content overlapping with FIG. 3 will be omitted.

Referring to FIG. 5, the scalable video decoding apparatus of the present invention receives a bitstream including a block encoded in a skip mode from the encoding apparatus (S510).

The decoding apparatus determines whether or not a flag indicating execution of SKIP mode weight overriding is set (S520). The received bitstream may include a block encoded in a skip mode, whether a skip mode is performed, skip mode information, and a skip mode weight for generating a reference block.

If the overriding performance flag is set to '1', the decoding apparatus determines whether the mode of the block of the base layer of the current frame is the SKIP mode (S530).

If the block of the base layer is the skip mode, the decoding apparatus overrides the SKIP mode weight with the existing weight set in the block of the enhancement layer of the reference frame (S540).

The decoding apparatus configures a reference block for the block of the enhancement layer of the current frame to be decoded using the weighted sum (S550). If the overriding performance flag is set to '1' and the block of the base layer is the skip mode, the decoding apparatus determines the weight of the current frame to which the weight is calculated from the corresponding block of the enhancement layer of the reference frame to which the skip mode weight is applied and the skip mode weight. The reference block is generated by the weighted sum of the corresponding blocks of the base layer. If the overriding performance flag is not set to '1' or the block mode of the base layer of the current frame is not the skip mode, the decoding apparatus uses the existing weight to base the corresponding block of the enhancement layer of the reference frame and the current frame. A reference block is generated by the weighted sum of the corresponding blocks of the layer.

The decoding apparatus performs AR-FGS block decoding on the enhancement layer of the current frame based on the generated reference block (S560).

In FIG. 2 to FIG. 5, the case where the block mode of the base layer is the SKIP mode has been described as an example, but the block is predicted from reference pictures (blocks on the left, left diagonal, upper part of the block in H.264). It will be apparent to those skilled in the art that new weights can be overridden for the block even when within a range of values.

The SKIP mode weight used for overriding may be coded in the slice header, and fixed length or variable length coding of N bits may be used.

6 is a diagram of an embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

Referring to FIG. 6, in an embodiment of coding in slice header in scalable extension syntax of actual scalable video coding, flag information “override_max_diff_ref_scala_for_zero_base_block_flag”, which is flag information on whether to use SKIP mode weight overriding, is coded. If 1, the coded "max_diff_ref_scale_for_skipped_base_block", which is SKIP mode weight overriding information, is coded with 2 bits. If the flag is 0, "max_diff_ref_scale_for_skipped_base_block" is not coded. The value of "max_diff_ref_scale_for_skipped_base_block" has a value from 0 to 3. 0 gives the weight of the enhancement layer 32/32, 1 gives 31/32, 2 gives 30/32, and 3 gives 29/32. When coding a block of an enhancement layer in which the block mode of the base layer is the SKIP mode, the SKIP mode weight "max_diff_ref_scale_for_skipped_base_block" is overridden to "max_diff_ref_scale_for_zero_base_block" when "override_max_diff_ref_scala_for_zero_base_block_flag" is 1.

7 is a diagram of another embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

Referring to FIG. 7, the SKIP mode weight overriding information “max_diff_ref_scale_for_skipped_base_block” is coded with 5 bits as an embodiment of coding in the slice header in scalable extension syntax of actual scalable video coding.

8 is a diagram of another embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

Referring to FIG. 8, an embodiment of coding SKIP mode weight overriding information “max_diff_ref_scale_for_skipped_base_block” as a variable length code as an embodiment of coding in a slice header in scalable extension syntax of actual scalable video coding is used in H.264. Here is an example of coding using Exp-Golomb code.

Pseudo Code is an example of the actual application to the scalable video coding standardization currently under international standardization.

9A to 9C are diagrams of an embodiment of a scalable header in scalabe extension syntax including syntax for representing a scalable coding method of image data according to the present invention.

Referring to FIGS. 9A to 9C, as the syntax applied to the pseudo code in the SVC international standard, the semantics of the variables used in FIGS. 9A to 9C are as follows.

The following are embodiments of the decoding process of the pseudocode, including a scaling process of a differential inter prediction sample for a 4x4 luma block, a scaling process of a differential inter prediction sample for an 8x8 luma block, and Represents a scaling process of differential inter prediction samples for chroma blocks.

For 4x4 luminance blocks Difference Inter  Scaling process of predictive samples Scaling  process for differential Inter prediction samples of  4x4 luma blocks )

For 8x8 luma blocks Difference Inter  Scaling process of predictive samples Scaling  process for differential Inter prediction samples of  8x8 luma blocks )

saturation( chroma ) For blocks Difference Inter  Scaling process of predictive samples Scaling process for differential Inter predictin samples of chroma blocks )

10 is a block diagram illustrating a schematic internal configuration of a scalable video encoding apparatus according to an embodiment of the present invention. Detailed descriptions of contents overlapping with the above description will be omitted.

Referring to FIG. 10, the encoding apparatus of the present invention includes a mode determiner 1010, a weight overriding unit 1020, a reference block generator 1030, and an encoder 1040.

The mode determination unit 1010 determines whether a block of the base layer of the current frame corresponding to the block of the enhancement layer of the current frame to be encoded is a skip mode. In addition, the mode determination unit 1010 determines whether to set a flag for instructing skip mode weight overriding. When the overriding flag is set to '1', the mode determining unit 1010 determines whether the block of the base layer is the skip mode. If the overriding flag is not set, the mode determination unit 1010 determines whether the block of the base layer is the skip mode. Do not judge.

When the block of the base layer of the current frame is the skip mode, the weight overriding unit 1020 has a larger value than the existing weight set in the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame. Override with skip mode weight.

The reference block generator 1030 generates a reference block based on a weight set in the block of the enhancement layer of the reference frame. The reference block generator 1030 sets the overriding flag to '1' in the mode determining unit 1010 and determines that the block of the base layer of the current frame is the skip mode, the enhancement layer of the reference frame to which the new weight is applied. The reference block is generated by a weighted sum of corresponding blocks of and corresponding blocks of the base layer of the current frame to which weights calculated from the new weights are applied. When the reference block generator 1030 determines that the block of the base layer of the current frame is not the skip mode, the mode determiner 1010 uses the existing weight to correspond to the corresponding block of the enhancement layer of the reference frame and the base layer of the current frame. Generate a reference block based on the corresponding block in. If the reference block generation unit 1030 does not set the overriding flag to '1' in the mode determination unit 1010, the corresponding block of the enhancement layer of the reference frame and the base layer of the current frame using existing weights are used. A reference block is generated by the weighted sum of the blocks.

The encoding unit 1040 performs AR-FGS block encoding on the block of the enhancement layer of the current frame by using the generated reference block and generates a bitstream.

11 is a block diagram illustrating a schematic internal configuration of a scalable video decoding apparatus according to an exemplary embodiment of the present invention. Detailed descriptions of contents overlapping with the above description will be omitted.

Referring to FIG. 11, the decoding apparatus of the present invention includes a receiver 1110, a mode determiner 1120, a weight overriding unit 1130, a reference block generator 1140, and a decoder 1150.

The receiver 1110 receives a bitstream including a block encoded in a skip mode.

The mode determiner 1120 determines whether the block of the base layer corresponding to the block of the enhancement layer of the current frame to be decoded is the skip mode. In addition, the mode determination unit 1120 determines whether a flag indicating skip mode weight overriding is set in the received bitstream. When the mode determining unit 1120 determines that the overriding execution flag is set to '1', the mode determination unit 1120 determines whether the corresponding block of the base layer is the skip mode, and if the overriding execution flag is not set to '1', It is not determined whether the corresponding block of the layer is the skip mode.

When the corresponding block of the base layer of the current frame is the skip mode, the weight overriding unit 1130 extracts a skip mode weight from the bitstream, and the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame. Override the existing weight set in the skip mode weight.

The reference block generator 1140 generates a reference block based on a weight set in the block of the enhancement layer of the reference frame. When the reference block generation unit 1140 determines that the overriding execution flag is set to '1' by the mode determination unit 1120 and that the block of the base layer of the current frame is the skip mode, the reference frame enhancement is performed. The reference block is generated by a weighted sum of corresponding blocks of the layer and corresponding blocks of the base layer of the current frame to which the weight calculated from the new weights is applied. When the reference block generator 1140 determines that the corresponding block of the base layer of the current frame is not the skip mode, the mode determiner 1120 determines that the corresponding block of the enhancement layer of the reference frame and the current frame are based on the existing weight. A reference block is generated based on the corresponding block of the base layer. When the reference block generation unit 1140 determines that the overriding execution flag is not set to '1' in the mode determination unit 1120, the reference block generation unit 1140 uses the existing weights to form the base of the corresponding block in the enhancement layer of the reference frame and the current frame. A reference block is generated by the weighted sum of the corresponding blocks of the layer.

The decoding unit 1150 performs AR-FGS block decoding on the block of the enhancement layer of the current frame by using the generated reference block and reconstructs the block.

12 to 15 are comparisons of peak signal-to-noise ratio (PSNR) between the results of coding in JSVM 5.10, a codec of scalable video coding standardization, and the results of implementing embodiments of the coding method according to the present invention in JSVM 5.10. Graphs.

Coded using the syntax (Syntax) applied to the SVC International Standard (FIGS. 9A to 9C), Semanics and Decoding Process, the existing weight parameter "max_diff_ref_scale_for_zero_base_coeff" is fixed to the weight of the upper layer. Was 18/32. If the base layer is in SKIP mode, the existing weight of the parent layer, "max_diff_ref_scale_for_zero_base_block", is applied as the weight of the parent layer of 28/32 (circle graph), 16/32 (triangle graph), and 8/32 (diamond graph) It was.

Referring to FIG. 12, the Foreman CIF 15 Hz sequence is coded as "IPPP ..." and the FGS layer is a graph comparing performance to the second layer. When "max_diff_ref_scale_for_zero_base_block" is 28/32, "max_diff_ref_scale_for_skipped_base_block" is 30/32. 32, and in the other cases 32/32. In the case of "max_diff_ref_scale_for_zero_base_block" is 28/32, the performance is improved up to 0.15dB, in 16/32, the performance is improved by 1dB, and in 8/32, the performance is improved by 1.35dB. .

Referring to FIG. 13, a bus CIF 15Hz sequence is coded as "IPPP ...", and a performance comparison of the FGS layer up to two layers. When "max_diff_ref_scale_for_zero_base_block" is 28/32, "max_diff_ref_scale_for_skipped_base_block" is 30/32. In the other cases, it was set to 32/32. In the case of "max_diff_ref_scale_for_zero_base_block" is 28/32, the performance is improved up to 0.1dB, in the case of 16/32, the performance is improved to 0.43dB, and in 8/32, the performance is improved to 0.73dB. Can be.

Referring to FIG. 14, the Mobile CIF 15Hz sequence is coded as "IPPP ...", and the FGS layer is a graph comparing the performance up to two layers. When "max_diff_ref_scale_for_zero_base_block" is 28/32, "max_diff_ref_scale_for_skipped_base_block" is 29/32, and in other cases, it is 32/32. When "max_diff_ref_scale_for_zero_base_block" is 28/32, the performance is improved up to 0.09dB, and when it is 16/32, the performance is improved up to 0.84dB, and 8/32 is 2.07dB. have.

Referring to FIG. 15, a football CIF 15 Hz sequence is coded as "IPPP ...", and the FGS layer is a graph comparing performances up to two layers. When "max_diff_ref_scale_for_zero_base_block" is 28/32, "max_diff_ref_scale_for_skipped_base_block" is 29 It was set at / 32, and in the other cases, 32/32. When "max_diff_ref_scale_for_zero_base_block" is 28/32, the performance is improved up to 0.01dB, 16/32 is 0.39dB, and 8/32 is 0.47dB. have.

Therefore, the SVC coding efficiency can be improved by the encoding / decoding method and the apparatus to which the reference block generation method according to the present invention is applied.

In describing the present invention, a scalable video encoding / decoding method performed in macroblocks or block units has been described. However, the present invention can be applied to a scalable video encoding / decoding method performed in slice units or frame units. Those skilled in the art will be able to predict sufficiently.

Also, for convenience of description, the case where the FGS layer is one layer has been described as an example, but a person having ordinary skill in the art can sufficiently predict that the present invention can be applied even when the FGS layer is two or more layers. will be.

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

So far, the present invention has been described with reference to preferred embodiments. Although specific terms have been used herein, they are used only for the purpose of describing the present invention and are not used to limit the scope of the present invention as defined in the meaning or claims.

Therefore, it will be understood by those skilled in the art that the present invention may be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a conceptual diagram of a reference block generation method of Adaptive Reference-Fine Grain Scalability (AR-FGS) according to the prior art.

2 is a flowchart illustrating a scalable encoding method according to an embodiment of the present invention.

3 is a flowchart illustrating a scalable video decoding method according to an embodiment of the present invention.

4 is a flowchart illustrating a scalable video encoding method according to another preferred embodiment of the present invention.

5 is a flowchart illustrating a scalable video decoding method according to another preferred embodiment of the present invention.

6 is a diagram of an embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

7 is a diagram of another embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

8 is a diagram of another embodiment of a syntax for expressing a scalable coding method of image data according to the present invention.

9A to 9C are diagrams illustrating an example of a slice scalable header in scalabe extension syntax including syntax for representing a scalable coding method of image data according to the present invention.

10 is a block diagram illustrating a schematic internal configuration of a scalable video encoding apparatus according to an embodiment of the present invention.

11 is a block diagram illustrating a schematic internal configuration of a scalable video decoding apparatus according to an exemplary embodiment of the present invention.

12 is a graph comparing PSNR performance versus bit rate between an embodiment of a scalable coding method of image data and a method proposed in JSVM 5.10 according to the present invention.

13 is a graph comparing PSNR performance versus bit rate between another embodiment of a scalable coding method of image data and a method proposed in JSVM 5.10 according to the present invention.

FIG. 14 is a graph comparing PSNR performance versus bit rate between another embodiment of a scalable coding method of image data and a method proposed in JSVM 5.10. FIG.

15 is a graph comparing PSNR performance versus bit rate between another embodiment of a scalable coding method of image data according to the present invention and a method proposed in JSVM 5.10.

Claims (37)

(a) determining whether a block of the base layer corresponding to the block of the enhancement layer of the current frame to be encoded is a skip mode; (b) if the block of the base layer is the skip mode, overriding an existing weight set in the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame with a new weight; And (c) generating a reference block for the block of the enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight; A scalable video encoding method. The method of claim 1, And AR-FGS encoding the block of the enhancement layer of the current frame based on the generated reference block. According to claim 1, wherein the step (a), And determining a skip mode when the block data of the base layer of the current frame is the same in the time direction as the block data of the base layer of the reference frame. 2. The method of claim 1, The new weight is set to be larger than the existing weight to increase the block data utilization rate of the enhancement layer of the reference frame. The method of claim 1, wherein step (c) comprises: Generating the reference block by a weighted sum of blocks of an enhancement layer of the reference frame to which the new weight is applied and blocks of the base layer of the current frame to which the weight calculated from the new weight is applied; Scalable video encoding method. The method of claim 1, (d) when the block of the base layer of the current frame is not the skip mode, generating the reference block based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight. And further comprising; a scalable video encoding method. According to claim 1, wherein the step (a), (a1) determining whether to set a flag indicating to perform a new weight overriding; And (a2) when the flag is set, determining whether a block of a base layer of the current frame is a skip mode; The method of claim 7, wherein (e) generating the reference block based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight if the flag is not set; A scalable video encoding method. The method of claim 1, And the new weight is set in a slice header with a fixed length or a variable length of a predetermined bit together with the reference weight. (a) determining whether a block of the base layer corresponding to the block of the enhancement layer of the current frame to be decoded is in a skip mode; (b) if the block of the base layer is the skip mode, overriding an existing weight set in the block of the enhancement layer of the reference frame corresponding to the block of the enhancement layer of the current frame with a new weight; And (c) generating a reference block for the block of the enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight; Scalable video decoding method. The method of claim 10, And AR-FGS decoding the block of the enhancement layer of the current frame based on the generated reference block. The method of claim 10, wherein step (a) comprises: And determining whether a skip mode is based on skip mode information included in the bitstream. The method of claim 10, And the new weight is greater than the existing weight to increase the block data utilization rate of the enhancement layer of the reference frame. The method of claim 10, wherein step (b) comprises: (b1) extracting the new weight included in the bitstream; And (b2) overriding the new weights to blocks of the enhancement layer of the reference frame. The method of claim 14, The new weight is extracted from a slice header in the bitstream. The method of claim 10, wherein step (c) comprises: Generate a reference block for the block of the enhancement layer of the current frame by a weighted sum of the block of the enhancement layer of the reference frame to which the new weight is applied and the block of the base layer of the current frame to which the weight calculated from the new weight is applied. Scalable video decoding method comprising a. The method of claim 10, (d) when the block of the base layer of the current frame is not the skip mode, generating the reference block based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight. And further comprising; a scalable video decoding method. The method of claim 10, wherein step (a) comprises: (a1) determining whether a flag indicating new weight overriding is set; And (a2) if the flag is set, determining whether a block of the base layer is a skip mode; The method of claim 18, (f) if the flag is not set, generating the reference block based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight; Scalable video decoding method. A mode determination unit that determines whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be encoded is a skip mode; A weight override unit for overriding an existing weight set in a block of an enhancement layer of a reference frame corresponding to a block of an enhancement layer of the current frame with a new weight when the block of the base layer of the current frame is a skip mode; And And a reference block generator configured to generate a reference block for a block of an enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight. Scalable video decoding apparatus. The method of claim 20, And an encoder configured to perform AR-FGS encoding on a block of an enhancement layer of the current frame based on the generated reference block. The method of claim 20, wherein the mode determination unit, And a skip mode is determined when the block data of the base layer of the current frame is the same as the block data of the base layer of the reference frame in a time direction. The method of claim 20, And the new weight is greater than the existing weight to increase the block data utilization rate of the enhancement layer of the reference frame. The method of claim 20, wherein the reference block generation unit, Scalable video encoding, wherein the reference block is generated by a weighted sum of blocks of the enhancement layer of the reference frame to which the new weight is applied and blocks of the base layer of the current frame to which the weight is calculated from the new weight. Device. The method of claim 20, wherein the reference block generation unit, When the block of the base layer of the current frame is not the skip mode, the reference block is generated based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight. Scalable video encoding device. The method of claim 20, wherein the mode determination unit, And determining whether to set a flag indicating new weight overriding, and if the flag is set, determining whether a block of a base layer of the current frame is a skip mode. The method of claim 26, wherein the reference block generation unit, And when the flag is not set, the reference block is generated based on a block of an enhancement layer of the reference frame and a block of a base layer of the current frame using the existing weight. A mode determination unit that determines whether a block of a base layer corresponding to a block of an enhancement layer of a current frame to be decoded is a skip mode; A weight override unit configured to override an existing weight set in a block of an enhancement layer of a reference frame corresponding to a block of an enhancement layer of the current frame with a new weight when the block of the base layer is in a skip mode; And And a reference block generator configured to generate a reference block for a block of an enhancement layer of the current frame based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the new weight. Scalable video decoding apparatus. The method of claim 28, And a decoding unit to AR-FGS decode a block of an enhancement layer of the current frame based on the generated reference block. The method of claim 28, wherein the mode determination unit, And a skip mode is determined based on skip mode information included in the bitstream. The method of claim 28, The new weight is scalable video decoding apparatus, characterized in that greater than the existing weight to increase the block data utilization rate of the enhancement layer of the reference frame. The method of claim 28, wherein the weight overriding unit, And extracting a new weight included in a bitstream and overriding the new weight to a block of an enhancement layer of the reference frame. The method of claim 28, wherein the reference block generation unit, Generate a reference block for the block of the enhancement layer of the current frame by a weighted sum of the block of the enhancement layer of the reference frame to which the new weight is applied and the block of the base layer of the current frame to which the weight calculated from the new weight is applied. Scalable video decoding apparatus, characterized in that. The method of claim 28, wherein the reference block generation unit, When the block of the base layer of the current frame is not the skip mode, the reference block is generated based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame by using the existing weight. Scalable video decoding apparatus. The method of claim 28, wherein the mode determination unit, And determining whether a flag indicating new weight overriding is set, and if the flag is set, whether a block of the base layer is in a skip mode. The method of claim 35, wherein the mode determination unit, If the flag is not set, the scalable video decoding apparatus generates the reference block based on the block of the enhancement layer of the reference frame and the block of the base layer of the current frame using the existing weight. . A computer-readable recording medium having recorded thereon a program for executing the method of any one of claims 1 to 19 on a computer.

Images