KR20230029717A - Method for decoding a video partitioned block - Google Patents

Abstract

In accordance with one embodiment, provided is a method for decoding a video-split block. The method includes the following steps of: splitting a coding unit block in a slice into four first subblocks; splitting at least one of the split first subblocks into two second subblocks, wherein a prediction mode for the second subblocks has an intra prediction mode or an inter prediction mode; splitting first subblocks, which are not split into the second subblocks, of the split first subblocks into four third subblocks, wherein a prediction mode for the third subblocks has an intra prediction mode or an inter-prediction mode; performing prediction on the second subblocks according to the prediction mode for each of the second subblocks, and performing prediction on the third subblocks according to the prediction mode for each of the third subblocks; and converting at least one of the second subblocks and at least one of the third subblocks, wherein the second or third subblocks are individually converted by conversion kernels of variable sizes, wherein the conversion kernels of variable sizes are selected based on sizes of subblocks corresponding to the kernels, respectively, wherein the selected conversion kernels of variable sizes are equal to or smaller than the sizes of the subblocks corresponding to the kernels, respectively, wherein the second subblocks are rectangular blocks. Therefore, the present invention is capable of increasing encoding efficiency by increasing the flexibility of encoding mode selection.

Description

Method for decoding a video partitioned block}

The present invention relates to a method of dividing an input picture into blocks in video encoding and decoding and encoding and decoding the divided blocks by simultaneously using intra- and inter-prediction encoding, and particularly improves encoding efficiency. Inter-screen and screen-to-screen video encoding, which can be improved, and the encoding efficiency can be further improved by encoding block image signals using squared transformation or non-squared transformation according to the size of the partition block. It relates to a split block encoding and decoding method through intraprediction.

ISO/IEC 14496-10 (MPEG-4 Part 10 Advanced Video Coding) or H.264 and ISO/IEC 14496-10 Amendment, a technology for video compression coding jointly standardized by ISO/IEC and ITU-T 3 (MPEG-4 Scalable Video Coding) standard, SMPTE standard VC-1, and other Audio Video coding Standard (AVS) standards have made great progress in video data compression efficiency.

There can be many factors to improve video compression efficiency, but in particular, existing video encoding standards (MPEG-1 Video, MPEG-2 Video, MPEG-4 Part 2 Visual, H.261, H.263, etc.) Unlike predictive encoding after dividing the size of the screen to be encoded into macroblocks (16×16 pixels), macroblocks are divided into 16×16, 16×8, 8×16, 8×8, 8×4, It is further subdivided into 4×8 and 4×4 units, predictive encoding is performed on these sub-blocks, and a block that generates cost minimization in terms of rate-distortion cost is encoded as an optimal block mode.

In this way, it is possible to more effectively perform minute motion or motion prediction of complex images, and significantly improve compression efficiency by greatly reducing generated residual signals.

1 is a diagram showing split block types of a 16×16 macroblock unit block for encoding in a conventional H.264|AVC encoder, and shows seven types of motion prediction block split types used in H.264.

As shown in FIG. 1, block-based predictive coding typically encodes an input image by dividing it into 16×16 macroblock units. In particular, in the ISO/IEC 14496-10 (MPEG-4 Advanced Video Coding) or H.264 standard, a macroblock is divided into 7 sub-blocks as shown in FIG. 1, and a block that minimizes rate-distortion cost is finally selected and predicted. perform encoding.

When intra-coding is performed on the sub-blocks divided into 16×16 macroblocks to be encoded, intra-prediction encoding is performed on the macroblocks in units of one 16×16 pixel unit, or sub-blocks are divided into 4 sub-blocks. Intra-prediction encoding of an 8×8 block or intra-prediction encoding of 16 4×4 blocks is performed.

In general, in low-resolution video encoding, this intra-prediction encoding technique has coding efficiency in terms of reducing the number of block mode cases, but it is not suitable for high-definition (HD) or ultra-high definition (UHD) video encoding. there is a problem with That is, in the case of a super macroblock having a size of 32×32 or larger by extending a 16×16 macroblock, which is a coding unit block, the modes of all divided blocks within the super macroblock are set to 16×16, 8 as in the conventional method. Applying the same intra prediction based on a ×8 or 4 × 4 block results in a decrease in coding efficiency.

In other words, what is noteworthy in the conventional split block-based predictive encoding method is that all of the split blocks are coded only through intra- or inter-prediction encoding. That is, both intra-prediction encoding and inter-prediction encoding of a split block screen are not applied, and only one method is selected and applied. This is due to the simplification of the syntax expressing the block coding mode by applying only one of intra- or inter-screen coding, so there may be a gain in coding efficiency in image or video compression of HD resolution or lower, but the coding unit is a macroblock or larger superblock. In the case of a macro block, it may act as a factor that reduces encoding efficiency.

The present invention has been devised to solve the above-described problems, and intra- and inter-prediction coding is performed by extending an intra- or inter-prediction coding selection method for sub-blocks of a block divided during video encoding of the present invention. The purpose is to provide a more effective predictive encoding method by extending all modes to be selectable and performing encoding by selectively applying a square or non-square transformation kernel according to the block size to the residual signal after motion compensation of the divided block. .

Another object of the present invention is to provide a computer-readable recording medium in which the above method is implemented.

In order to achieve the above object, the present invention divides an input picture into coding unit blocks, divides the coding unit blocks into sub-blocks, and selects one of intra-prediction coding and inter-prediction coding for each sub-division block. to encode

An embodiment comprises receiving a video data sequence of encoded video data, wherein the video data sequence comprises a slice, and the slice comprises coding unit blocks, wherein the coding blocks comprise a first sub-coding unit blocks, at least one of the first sub coding unit blocks is divided into second sub coding unit blocks, and either the first sub coding unit blocks or the second sub coding unit blocks is a square block; decoding a coding mode corresponding to any one of the first sub coding unit blocks and the second sub coding unit blocks; performing intra prediction or inter prediction on the sub coding unit block corresponding to the coding mode based on the decoded coding mode; and transforming residual data of the sub coding unit block corresponding to the coding mode, wherein one of transform kernels having variable sizes is applied to the sub coding unit block corresponding to the coding mode. and here, the applied transform kernel is selected according to the size of the sub coding unit block corresponding to the coding mode.

In the present invention, in encoding a coding unit block pixel value in video encoding, at least one intra-prediction encoding and inter-prediction encoding is applied to a divided sub-block or its sub-division blocks, and encoding is performed so that the divided sub-blocks are displayed on the screen. By enabling encoding in intra- or inter-prediction coding mode, coding unit blocks or subdivision blocks thereof are predictive-encoded by using both intra- and inter-prediction, thereby increasing flexibility in selection of coding modes, thereby increasing coding efficiency. there is

In addition, in the present invention, in the split block-based prediction encoding, each split block can apply both intra prediction and inter prediction to sub-division blocks, and selectively variable transform according to the size of the split block. Encoding is performed by applying the size of a variable block-size transform kernel (KERN), thereby greatly improving the encoding efficiency.

1 is a diagram showing partition block types of a 16×16 macroblock unit block for encoding in a conventional H.264|AVC encoder;
2 is a diagram illustrating super macroblock unit blocks and split block types for intra- or inter-prediction encoding in an encoder according to an embodiment of the present invention;
3 is a flowchart illustrating a partition block encoding method in video encoding according to an embodiment of the present invention.
4 is a flowchart illustrating a method of decoding a bitstream encoded by a video division block method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, in adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible, even if they are displayed on different drawings. And, in describing the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted.

2 is a diagram illustrating super macroblock unit blocks and split block types for intra or inter prediction encoding in an encoder according to an embodiment of the present invention. 2 is a diagram showing an example of a super macroblock having a macroblock size of 16×16 pixels or larger and a block division type applied in an embodiment of the present invention.

As shown in FIG. 2, the super macro block is divided into sub-blocks, and intra- or inter-prediction encoding processes are performed on each divided block, so that the super macro block uses both intra- and inter-prediction encoding modes in the final encoding mode. By enabling encoding to be performed, the encoding efficiency of the image can be increased very effectively. In actual encoding, as shown in Equation 1, it is possible to select and encode a block mode that minimizes rate-distortion cost.

는 모드에 의존적인 라그랑지안 승수이고, MODE는 분할 블록 모드 타입을 나타낸다.Here, J _MODE is the rate-distortion function for the block coding mode, s is the original block pixel input to be encoded, r is the reference image pixel input, QP is the quantization parameter,

is a mode-dependent Lagrangian multiplier, and MODE represents a partition block mode type.

In addition, when transform coding is applied to the residual signal of the super macroblock whose size is increased, a 16x16 or larger square transform kernel larger than the existing 4x4 and 8x8, or a 16x8 for non-square transform, or Encoding efficiency can be increased by selectively applying an 8x16 or larger non-square transformation kernel according to the size of the partition block.

When a 16x16 or larger square transformation kernel is applied to the super macro block, it can be calculated as shown in Equation 2.

Here, X represents an NxN input video signal matrix, A represents an NxN square transform kernel matrix, and Y represents a transform coefficient matrix. If the divided sub-block is a non-square block, transformation is performed as shown in Equation 3.

Here, when the input video signal X is an Mx(M/2) matrix, A ₁ represents an MxM square transform kernel matrix, A ₂ represents a (M/2)x(M/2) square transform kernel matrix, and Y represents a transform represents the coefficient matrix.

In an embodiment of the present invention, when a square or non-square kernel transformation is applied, it is preferable to perform transform encoding by applying a kernel having the same or smaller size as compared to the smaller number of horizontal or vertical pixels of a partition block. .

3 is a flowchart illustrating a partition block encoding method in video encoding according to an embodiment of the present invention.

Referring to FIG. 3, first, the order of screen i for encoding is initialized (i=0) (S101). Then, the screen i is input in order for encoding (S102).

Next, the input screen i is divided into coding unit blocks (S103). In an embodiment of the present invention, a coding unit block may be a macro block or a super macro block.

Next, it is checked whether the current screen (i) is to be inter-prediction-encoded (S104). If the current picture (i) is not inter-prediction, intra-prediction encoding is performed (S105). Otherwise, if the current picture (i) is inter-prediction, a coding unit that is encoded within one picture (i) The order of block (j) is initialized (j = 0) (S106).

Thereafter, the unit block (j) to be encoded is divided into sub-blocks (S107). Then, the order of the lower block mode (k) is initialized (k=0) (S108). One block mode is selected from the lower block modes (k) (S109).

It is checked whether intra- or inter-prediction is performed for the sub-block mode encoded in the coding unit block (S110). In the case of performing inter- and intra-prediction, intra- and inter-picture encoding is performed (S111), and otherwise, only inter-prediction encoding is performed (S112). Then, the prediction encoding result and the rate-distortion cost value, which are encoding results, are stored (S113).

It is checked whether the lower block mode (k) is the final block mode (S114). If the lower block mode (k) is not the final block mode, steps S109 to S113 are repeated for the next block mode. On the other hand, if the lower block mode (k) is the final block mode, the optimal split block mode is determined and the corresponding encoding result is finally selected (S115).

Then, it is determined whether the current coding unit block (j) is the last block in the current screen (i) (S116). If the current coding unit block (j) is not the final block, the next coding unit block is input and steps S107 to S115 are repeated.

If the coding unit block (j) is the final block in the current screen (i) in step S116, it is checked whether the current screen (i) is the final screen (S117). If the current screen (i) is the final screen, the algorithm is terminated. Otherwise, the process returns to step S102, inputs the next screen, and repeats steps S102 to S116.

4 is a flowchart illustrating a method of decoding a bitstream encoded by a video division block method according to an embodiment of the present invention.

Referring to FIG. 4, first, the order of screen i for decoding is initialized (i=0) (S201). Then, the screen encoding bitstream i is input in order for decoding (S202).

Next, it is checked whether the input screen bitstream (i) is inter prediction encoding (S203). If the current screen bitstream (i) is not inter-predictive coding, intra-prediction decoding is performed (S207), otherwise, if the current input screen bitstream (i) is inter-prediction coding, one screen ( The order of slice j to be decoded in i) is initialized (j=0) (S204).

Next, slice information is analyzed for the input screen bitstream (S205). The order of unit decoding blocks j to be decoded in each slice within one screen i is initialized (k=0) (S206). In an embodiment of the present invention, a decoding unit block may be a macro block or a super macro block.

Next, after analyzing each unit coded block information (S208), the order of lower division blocks in the unit coded block is initialized (m=0) (S209). Thereafter, the encoding mode of the subdivision block within the unit coding block is decoded (S210). After checking whether the lower division block is an inter prediction coding block (S211), if it is an inter prediction coding block, inter prediction decoding is performed (S213), and if it is an intra prediction coding block, intra prediction coding is performed if it is a mode (S213). S212).

Subsequently, sub-division block pixel values are restored using the sub-division block decoding result (S214). Then, after checking whether the current sub-division block (m) is the final block (S215), if it is the final sub-division block, unit decoding block pixel values are configured (S216), and if not, step S210 for decoding the next sub-division block. Return to and perform steps S210 to S214.

Thereafter, after checking whether the current unit coding block (k) is the final unit coding block (S217), if it is the final unit coding block, a slice pixel is formed (S218). Otherwise, the process returns to step S208 and steps S208 to S216 are performed. do. Thereafter, after checking whether the current slice (j) is the final slice (S219), screen pixels are configured if the current slice (j) is the final slice (S220), and if not, steps S205 to S218 are performed. Thereafter, after determining whether the current screen (i) is the final screen (S221), if it is the final screen, it is terminated.

In video encoding according to an embodiment of the present invention, an input image is divided into coding unit blocks, the coding unit blocks are further divided into sub-blocks, and then at least one of intra-prediction and inter-prediction is performed in encoding each sub-division block. Encoding is selectively performed using

In this way, the coding mode of the coding unit block can be coded by mixing inter-prediction and intra-prediction sub-division block modes, and at the same time, the coding efficiency can be improved by selectively applying a variable kernel transform according to the division block.

In addition, video decoding according to an embodiment of the present invention can decode a compressed bitstream with improved coding efficiency by performing a reverse process of the coding process.

As another embodiment of the present invention, it is possible to implement the subdivision block encoding method in video encoding described above as computer readable code on a computer readable recording medium. A computer-readable recording medium includes all types of recording devices in which data that can be read by a computer system is stored.

For example, computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, hard disk, floppy disk, removable storage device, and non-volatile memory. , optical data storage devices, and the like, and also includes those implemented in the form of carrier waves (eg, transmission through the Internet).

In addition, the computer-readable recording medium may be distributed to computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

In the video encoding according to the present invention described above, descriptions of preferred embodiments of the intra prediction coding method and/or inter prediction coding method and the decoding method, which is the reverse process thereof, have been described, but these embodiments are exemplary and It is not limiting. Those skilled in the art to which the present invention pertains will understand that various changes and modifications can be made without departing from the spirit of the present invention and the scope of rights set forth in the appended claims.

Claims (1)

receiving a video data sequence of encoded video data;
Here, the video data sequence includes a slice, the slice includes a coding unit block,
Here, the coding unit block is divided into first sub coding unit blocks,
At least one of the first sub coding unit blocks is divided into second sub coding unit blocks, and
either of the first sub coding unit blocks or the second sub coding unit blocks is a square block;
decoding a coding mode corresponding to any one of the first sub-coding unit-blocks and the second sub-coding unit-blocks;
performing intra prediction or inter prediction on the sub coding unit block corresponding to the coding mode based on the decoded coding mode; and
converting residual data of the sub coding unit block corresponding to the coding mode;
Here, one of transform kernels of variable size is applied to the sub-coding unit block corresponding to the coding mode, and
Here, the applied transform kernel is selected according to the size of the sub coding unit block corresponding to the coding mode.

Images