Video decoding method, video decoding device and storage medium

Abstract

To provide a video decoding method.SOLUTION: When a determination is made that multiple sub partitions are multiple vertical sub partitions, each of the sub partitions having a width less than or equal to 2, intra prediction is executed for the sub partitions of a coded block in intra sub partition (ISP) mode. Executing intra prediction for the sub partitions of the coded block includes merging at least two of the multiple sub partitions into one prediction region of the intra prediction, and generating prediction samples in the prediction region of the intra prediction on the basis of multiple reference samples adjacent to the prediction region so that a sample reconstructed from one of the at least two sub partitions may not be used for executing the intra prediction for any other sub partitions of the at least two sub partitions.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] FIELD OF THE DISCLOSURE This disclosure relates generally to video encoding and compression. In particular, this disclosure relates to a system and method for performing video encoding using an intra sub-partition coding mode. More particularly, this disclosure relates to a video decoding method, a video decoding device, and a storage medium. [Background technology]

[0002] This section provides background information related to the present disclosure. Information contained in this section should not necessarily be construed as prior art.

[0003] Any of a variety of video encoding techniques can be used to compress the video data. The video encoding can be performed in accordance with one or more video encoding standards. Some exemplary video encoding standards include versatile video coding (VVC), joint exploration test model (JEM) coding, H.265 / HEVC (high-efficiency video coding), H.264 / AVC (advanced video coding), and moving picture experts group (MPEG) coding.

[0004] Video coding generally utilizes prediction methods (e.g., inter-prediction, intra-prediction, etc.) that exploit redundancy inherent in video images or sequences. One goal of video coding techniques is to compress video data into a format that uses a lower bit rate while avoiding or minimizing degradation of video quality.

[0005] The first version of the HEVC standard, completed in October 2013, offers approximately 50% bitrate savings or equivalent perceptual quality compared to previous-generation video coding standards (H.264 / MPEG AVC). While the HEVC standard offers significant coding improvements over previous technologies, there is evidence that better coding efficiency than HEVC can be achieved by using additional coding tools. Based on this evidence, both the Video Coding Experts Group (VCEG) and the Moving Picture Experts Group (MPEG) have initiated exploratory research to develop new coding techniques for future video coding standards. The Joint Video Exploration Team (JVET), established by ITU-T VECG and ISO / IEC MPEG in October 2015, has initiated significant research into advanced technologies that could enable significant improvements in coding efficiency. One reference software model, called the Joint Exploration Model (JEM), has been maintained by JVET by integrating several additional coding tools on top of the HEVC Test Model (HM).

[0006] In October 2017, ITU-T and ISO / IEC published a Joint Proposal (CfP) for video compression with capabilities beyond HEVC. In April 2018, the 10th JVET received and evaluated 23 CfP responses. These responses demonstrated a compression efficiency gain of approximately 40% over the HEVC standard. Based on these evaluation results, JVET launched a new project to develop a next-generation video coding standard, "Versatile Video Coding (VVC)." Also in April 2018, a reference software code base called the VVC Test Model (VTM) was established to demonstrate a reference implementation of the VVC standard.

[0007] This section provides a general overview of the disclosure and is not an exhaustive disclosure of its entire scope or all of its features. Summary of the Invention [Means for solving the problem]

[0008] According to a first aspect of the present disclosure, a video encoding method is performed on a computing device having one or more processors and a memory storing a plurality of programs executed by the one or more processors. The method includes independently generating a respective intra prediction for each of a plurality of corresponding sub-divisions, each intra prediction being generated using a plurality of reference samples from a current coding block.

[0009] According to a second aspect of the present disclosure, a video encoding method is performed on a computing device having one or more processors and a memory storing a plurality of programs executed by the one or more processors. The method includes generating, for a luma component of an intra sub-partition (ISP) coding block, a respective intra prediction for each of a plurality of corresponding sub-partitions using only N modes out of M possible intra-prediction modes, where M and N are positive integers and N is less than M.

[0010] According to a third aspect of the present disclosure, a video encoding method is performed on a computing device having one or more processors and a memory storing a plurality of programs executed by the one or more processors. The method includes generating intra predictions for a chroma component of an intra sub-partition (ISP) coding block using only N modes out of M possible intra prediction modes, where M and N are positive integers and N is less than M.

[0011] According to a fourth aspect of the present disclosure, a video encoding method is performed on a computing device having one or more processors and a memory storing a plurality of programs executed by the one or more processors, the method generating a respective luma intra prediction for each of a plurality of corresponding sub-partitions of an entire intra sub-partition (ISP) coding block for a luma component and for a chroma component, and generating a chroma intra prediction for the entire intra sub-partition (ISP) coding block.

[0012] According to a fifth aspect of the present disclosure, a video encoding method is executed on a computing device having one or more processors and a memory storing a plurality of programs executed by the one or more processors, the method comprising: generating a first prediction using an intra sub-partitioning mode, generating a second prediction using an inter prediction mode, and combining the first prediction and the second prediction to generate a final prediction by applying a weighted average to the first prediction and the second prediction. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a block diagram illustrating an exemplary encoder that can be used in conjunction with many video coding standards. [Figure 2] FIG. 1 is a block diagram illustrating an exemplary decoder that can be used in conjunction with many video coding standards. [Figure 3] Five types of exemplary block divisions for a multi-type tree structure are shown. [Figure 4] 1 illustrates an exemplary set of intra modes for use with the VVC standard. [Figure 5] FIG. 1 illustrates a set of multiple reference lines for performing intra prediction. [Figure 6A] 10 shows a first set of reference samples and angular directions used to perform intra prediction of a first rectangular block. [Figure 6B]10 shows a second set of reference samples and angular directions used to perform intra prediction of a second rectangular block. [Figure 6C] 10 shows a third set of reference samples and angular directions used to perform intra prediction of square blocks. [Figure 7] 1 illustrates an exemplary set of positions for neighboring reconstructed samples used for position-dependent intra-prediction combining (PDPC) of one coding block. [Figure 8A] 1 illustrates an exemplary set of short-distance intra-prediction partitions (SDIPs) for an 8x4 block. [Figure 8B] 1 illustrates an exemplary set of short-distance intra-prediction partitions (SDIPs) for a 4x8 block. [Figure 8C] 1 illustrates an exemplary set of short-distance intra-prediction partitions (SDIPs) for blocks of any size. [Figure 9A] A plot of chroma values ​​as a function of luminance values, the plot being used to derive a set of linear model parameters. [Figure 9B] 9B shows the sample locations used to derive the linear model parameters of FIG. 9A. [Figure 10] 10 illustrates the generation of reference samples for intra prediction for all sub-partitions, using only reference samples outside the current coding block. [Figure 11] 11 illustrates the combination of inter and intra predictor samples for the first sub-partition of FIG. 10. [Figure 12] 11 illustrates the combination of inter predictor samples and intra predicted samples for the second sub-partition of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0014] A set of exemplary, non-limiting embodiments of the present disclosure will be described below in conjunction with the accompanying drawings. Structural, method, or functional variations may be implemented by those skilled in the relevant art based on the examples presented herein, and all such variations are within the scope of the present disclosure. Where no contradiction exists, the teachings of different embodiments may, but need not, be combined with each other.

[0015] The terms used in this disclosure are intended to describe particular examples rather than to limit the disclosure. As used in this disclosure and the appended claims, the singular forms "a," "an," and the like also refer to the plural unless the context clearly dictates otherwise. As used herein, the term "and / or" should be understood to refer to any and all possible combinations of one or more of the associated listed items.

[0016] Terms such as "first," "second," and "third" may be used herein to describe various pieces of information, but these terms should not be used to limit the information. These terms are used only to distinguish one category of information from another. For example, first information may be referred to as second information, and similarly, second information may be referred to as first information, without departing from the scope of this disclosure. As used herein, the term "if" means "when," "upon," or "in response to," depending on the context.

[0017] Throughout this specification, references to "one embodiment," "an embodiment," "another embodiment," etc. mean that one or more particular features, structures, or characteristics described in connection with one embodiment are included in at least one embodiment of the present disclosure; thus, the appearances of the phrases "in one embodiment," "in an embodiment," "in another embodiment," etc. in various places throughout this specification do not necessarily all refer to the same embodiment. Furthermore, particular features, structures, or characteristics in one or more embodiments may be combined in any suitable manner.

[0018] Conceptually, many video coding standards are similar, as noted above. For example, virtually all video coding standards use block-based processing and share similar video coding block diagrams to achieve video compression. Like HEVC, the VVC standard is built on a block-based hybrid video coding framework. There are.

[0019] 1 shows a block diagram of an exemplary encoder 100 that may be used in conjunction with many video coding standards. In encoder 100, a video frame is divided into multiple video blocks for processing. For each given video block, a prediction is formed based on either an inter-prediction or an intra-prediction approach. In inter-prediction, one or more predictors are formed by motion estimation and motion compensation based on pixels from a previously reconstructed frame. In intra-prediction, a predictor is formed based on reconstructed pixels in a current frame. A mode decision allows the best predictor to be selected for predicting the current block.

[0020] A prediction residual, which represents the difference between the current video block and its predictor, is sent to transform circuit 102. The transform coefficients are then sent from transform circuit 102 to quantization circuit 104 for entropy reduction. The quantized coefficients are then provided to entropy coding circuit 106 to generate a compressed video bitstream. As shown in FIG. 1, prediction-related information 110 from inter-prediction and / or intra-prediction circuit 112, such as video block partition information, motion vectors, reference picture indices, and intra-prediction modes, is also provided through entropy coding circuit 106 and stored in compressed video bitstream 114.

[0021] The encoder 100 also requires decoder-related circuitry to reconstruct pixels for prediction purposes. First, a prediction residual is reconstructed through an inverse quantization circuit 116 and an inverse transform circuit 118. This reconstructed prediction residual is combined with a block predictor 120 to generate unfiltered reconstructed pixels for the current video block.

[0022] To improve coding efficiency and visual quality, an in-loop filter 115 is commonly used. For example, a deblocking filter is available in the current versions of AVC, HEVC, and VVC. In HEVC, an additional in-loop filter called Sample Adaptive Offset (SAO) is defined to further improve coding efficiency. In the current VVC standard, yet another in-loop filter called Adaptive Loop Filter (ALF) is under active consideration and is likely to be included in the final standard.

[0023] These in-loop filter operations are optional. Performing these operations helps improve coding efficiency and visual quality. They can also be turned off as a decision by the encoder 100 to save computational complexity.

[0024] Note that intra prediction is typically based on unfiltered reconstructed pixels, while inter prediction is based on filtered reconstructed pixels if these filter options are turned on by the encoder 100.

[0025] Figure 2 is a block diagram illustrating an exemplary decoder 200 that can be used with many video coding standards. This decoder 200 is similar to the reconstruction-related portions present in the encoder 100 of Figure 1. In the decoder 200 (Figure 2), an input video bitstream 201 is first decoded through an entropy decoding circuit 202 to derive quantized coefficient levels and prediction-related information. The quantized coefficient levels are processed through an inverse quantization circuit 204 and an inverse transform circuit 206 to obtain a reconstructed prediction residual. A block predictor implemented in an intra / inter mode selector 212 is configured to perform either an intra prediction procedure 208 or a motion compensation procedure 210 based on the decoded prediction information. A filtered prediction output is generated by summing the reconstructed prediction residual from the inverse transform circuit 206 with the prediction output generated by the block predictor using an adder 214. The reconstructed video in the picture buffer 213 can then be sent to drive a display device and can also be used to predict future video blocks. In situations where the in-loop filter 209 is turned on, a filtering operation is performed on these reconstructed pixels to derive the final reconstructed video output 222.

[0026] Returning to FIG. 1, the input video signal to the encoder 100 is processed block by block. Each block is called a coding unit (CU). In VTM-1.0, a CU may be up to 128x128 pixels. In High Efficiency Video Encoding (HEVC), Joint Exploration Test Model (JEM), and Versatile Video Encoding (VVC), the basic unit of compression is called a coding tree unit (CTU). However, in contrast to the HEVC standard, which divides blocks based only on a quadtree, in the VVC standard, one CTU is divided into CUs based on a quadtree / binarytree / ternarytree structure to adapt to changing local characteristics. Furthermore, the concept of multiple division unit types in the HEVC standard does not exist in the VVC standard. That is, the separation of CUs, prediction units (PUs), and transform units (TUs) does not exist in the VVC standard. Instead, each CU is always used as a basic unit for both prediction and transformation without further division. The maximum CTU size for HEVC and JEM is defined as two blocks of up to 64x64 luma pixels and, for 4:2:0 chroma format, 32x32 chroma pixels. The maximum allowed size of a luma block within a CTU is specified as 128x128 (although the maximum size of a luma transform block is 64x64).

[0027] 3 shows five types of exemplary block divisions for a multi-type tree structure. The five types of exemplary block divisions include 4-way division 301, horizontal 2-way division 302, vertical 2-way division 303, horizontal 3-way division 304, and vertical 3-way division 305. In a situation where a multi-type tree structure is used, one CTU is first divided by a quadtree structure. Then, each quadtree leaf node can be further divided by a binary tree structure and a ternary tree structure.

[0028] Using one or more of the example block partitions 301, 302, 303, 304, or 305 of Figure 3, spatial prediction and / or temporal prediction can be performed using the configuration shown in Figure 1. Spatial prediction (or "intra prediction") uses pixels from samples of already coded neighboring blocks (called reference samples) within the same video picture / slice to predict a current video block. Spatial prediction reduces spatial redundancy inherent in video signals.

[0029] Temporal prediction (also called "inter-prediction" or "motion-compensated prediction") predicts a current video block using reconstructed pixels from an already-encoded video picture. Temporal prediction reduces the temporal redundancy inherent in video signals. The temporal prediction signal for a given CU is typically signaled by one or more motion vectors (MVs), which indicate the amount and direction of motion between the current CU and its temporal reference. Additionally, if multiple reference pictures are supported, a reference picture index is also transmitted, which is used to identify which reference picture in the reference picture store the temporal prediction signal comes from.

[0030] After spatial and / or temporal prediction is performed, intra / inter mode decision circuit 121 within encoder 100 selects the best prediction mode, for example, based on a rate-distortion optimization method. Block predictor 120 is then subtracted from the current video block, and the resulting prediction residual is non-correlatedly quantized using transform circuit 102 and quantization circuit 104. The resulting quantized residual coefficients are inverse quantized by an inverse quantization circuit 116 and inverse transformed by an inverse transform circuit 118 to form a reconstructed residual, which is then fed back into the prediction block to form a reconstructed signal for the CU. Further in-loop filters 115, such as a deblocking filter, a sample adaptive offset (SAO), and / or an adaptive in-loop filter (ALF), can be applied to the reconstructed CU before it is placed into a reference picture store in a picture buffer 117 and used to encode future video blocks. To form an output video bitstream 114, the coding mode (inter or intra), prediction mode information, motion information, and quantized residual coefficients are all sent to an entropy coding unit 106 for further compression and packing to form a bitstream.

[0031] The basic intra prediction scheme applied in the VVC standard remains largely the same as that in the HEVC standard, except that some modules are further extended and / or improved in the VVC standard, such as intra sub-partition (ISP) coding mode, extended intra prediction with intra wide-angle direction, position-dependent intra prediction combining (PDPC), and 4-tap intra interpolation. One broad aspect of the present disclosure is directed to improving the existing ISP design in the VVC standard. Furthermore, other coding tools (e.g., tools in intra prediction and transform coding processes) included in the VVC standard and closely related to the techniques proposed in this disclosure will be described in more detail below.

[0032] [Intra prediction mode with wide intra angle direction] As in the HEVC standard, the VVC standard uses a set of previously decoded samples adjacent to one current CU (i.e., above or to the left) to predict samples for that CU. However, to capture the finer edge orientations present in natural video (especially for video content with high resolution, e.g., 4K), the amount of intra-angle modes is expanded from 33 modes in the HEVC standard to 93 modes in the VVC standard. In addition to angle orientation, both the HEVC standard and the VVC standard provide for a planar mode (assuming a slowly varying surface with horizontal and vertical slopes derived from the boundaries) and a DC mode (assuming a flat surface).

[0033] Figure 4 illustrates a set of exemplary intra modes 400 for use with the VVC standard, and Figure 5 illustrates a set of reference lines for performing intra prediction. Referring to Figure 4, the set of exemplary intra modes 400 includes modes 0, 1, -14, -12, -10, -8, -6, -4, -2, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, and 80. Mode 0 corresponds to planar mode, and mode 1 corresponds to DC mode. Similar to the intra prediction process of the HEVC standard, all of the defined intra modes (i.e., planar, DC, and angular) in the VVC standard utilize a set of neighboring reconstructed samples above and to the left of the predicted block as references for intra prediction. However, unlike the HEVC standard, in which only the nearest row / column of the reconstructed samples (row 0, 501 in FIG. 5) is used as a reference, a multi-reference line (MRL) is introduced in VVC, in which two additional rows / columns (i.e., row 1, 503 and row 3, 505 in FIG. 5) are used for the intra prediction process. The index of the selected reference row / column is signaled from the encoder 100 (FIG. 1) to the decoder 200 (FIG. 2). If a row / column that is not the nearest in FIG. 5, such as row 1, 503 or row 3, 505, is selected, the planar and DC modes of FIG. 4 are excluded from the set of intra modes that can be used to predict the current block.

[0034] FIG. 6A illustrates a first set of reference samples and angular directions 602, 604 used for intra-prediction of rectangular blocks (width W divided by height H equals 2). The first set of reference samples includes a first sample 601 and a second sample 603. FIG. 6B illustrates a second set of reference samples and angular directions 606, 608 used for intra-prediction of vertically long rectangular blocks (W / H=1 / 2). The second set of reference samples includes a third sample 605 and a fourth sample 607. FIG. 6C illustrates a third set of reference samples and angular directions 610, 612 used for intra-prediction of square blocks (W=H). The third set of reference samples includes a fifth sample 609 and a sixth sample 610. Assuming nearest neighbors are used, FIG. 6C illustrates the location of a third reference sample that can be used in the VVC standard to derive prediction samples for an intra-block. As shown in FIG. 6C, in addition to square coding blocks, there are also rectangular coding blocks in the intra prediction procedure in the context of the VVC standard, since a quadtree / binary / ternary tree partitioning structure is applied.

[0035] Due to the unequal widths and heights of a given block, various sets of angular directions are selected for different block shapes, which is also called wide-angle intra prediction. Specifically, in addition to planar and DC modes, 65 of the 93 angular directions are also supported for each block shape for both square and rectangular coding blocks, as shown in Table 1. This design not only efficiently captures the directional structure typically present in video (by adaptively selecting angular directions based on block shape), but also ensures that a total of 67 intra modes (i.e., planar, DC, and 65 angular directions) are enabled for each coding block. This can achieve good intra mode efficiency while providing a consistent design across different block sizes.

[0036] [Table 1] Table 1: Angular directions selected for intra prediction of different block shapes in VVC.

[0037] [Position-dependent Intra Prediction Combination] As mentioned earlier, intra prediction samples are generated from either unfiltered or filtered sets of neighboring reference samples, which may introduce discontinuities along block boundaries between the current coding block and its neighbors. To solve such problems, boundary filtering is applied in the HEVC standard by combining the first row / column of prediction samples for DC, horizontal (i.e., mode 18 in FIG. 4), and vertical (i.e., mode 50) prediction modes with unfiltered reference samples, utilizing a two-tap filter (for DC mode) or a gradient-based smoothing filter (for horizontal and vertical prediction modes).

[0038] The Position-Dependent Intra-Prediction Combining (PDPC) tool in the VVC standard extends the above concept by using a weighted combination of intra-predicted samples and unfiltered reference samples. In the current VVC working draft, PDPC is enabled for the following intra-modes without signaling: planar, DC, horizontal (i.e., mode 18), vertical (i.e., mode 50), angular directions close to the lower-left diagonal (i.e., modes 2, 3, 4, ..., 10), and angular directions close to the upper-right diagonal (i.e., modes 58, 59, 60, ..., 66). Assuming a predicted sample located as coordinates (x, y) is pred(x, y), its corresponding value after PDPC is calculated as follows: pred(x,y)=(wL×R -1,y +wT×R x-1 -wTL×R -1,-1 +(64-wL-wT+wTL)×pred(x,y)+32)>>6 … (1) where R x-1 , R -1,yrepresent the reference samples above and to the left of the current sample (x, y), respectively, and R -1,-1 represents the reference sample in the top-left corner of the current block.

[0039] 7 shows an example set of positions for neighboring reconstructed samples used for position-dependent intra-prediction combining (PDPC) of one coding block. x-1 ) represents the reference sample located above the current predicted sample (x, y). -1,y ) represents the reference sample located to the left of the current predicted sample (x, y). The third reference sample 705 (R -1,-1 ) represents the reference sample located at the top left corner of the current predicted sample (x, y).

[0040] The reference samples, including the first, second, and third reference samples 701, 703, and 705, are combined with the current predicted sample (x, y) during the PDPC process. Assuming that the current coding block has a size of W×H, the weights wL, wT, and wTL in equation (1) are adaptively selected according to the prediction mode and sample position, as follows:

[0041] In DC mode, wT=32>>((y<<1)>>shift), wL=32>>((x<<1)>>shift), wTL=(wL>>4)+(wT>>4)…(2)

[0042] In planar mode, wT=32>>(y<<1)>>shift), wL=32>>((x<<1)>>shift), wTL=0…(3)

[0043] In horizontal mode, wT=32>>((y<<1)>>shift), wL=32>>((x<<1)>>shift), wTL=wT …(4)

[0044] In vertical mode, wT=32>>((y<<1)>>shift), wL=32>>((x<<1)>>shift), wTL=wL…(5)

[0045] In the lower left diagonal, wT=16>>((y<<1)>>shift), wL=16>>((x<<1)>>s hift), wTL=0…(6)

[0046] In the upper right diagonal direction, wT=16>>((y<<1)>>shift), wL=16>>((x<<1)>>shift), wTL=0…(7)

[0047] Here, shift=(log2(W)-2+log2(H)-2+2)>>2

[0048] [Multiple transformation selection and shape adaptive transformation selection]

[0049] In addition to the DCT-II transform used in the HEVC standard, the introduction of additional core transforms, DCT-VIII and DST-VII, enables a Multiple Transform Selection (MTS) tool in the VVC standard. In the VVC standard, adaptive selection of transforms is enabled at the coding block level by signaling one MTS flag in the bitstream. In particular, if the MTS flag is equal to 0 for a block, a pair of fixed transforms (e.g., DCT-II) are applied in the horizontal and vertical directions. Otherwise (if the MTS flag is equal to 1), two additional flags are further signaled for the block to indicate the transform type (either DCT-VIII or DST-VII) in each direction.

[0050] On the other hand, due to the introduction of quadtree / binary / ternary tree-based block partitioning structures in the VVC standard, the distribution of intra prediction residuals is strongly correlated with block shape. Therefore, when MTS is disabled (i.e., the MTS flag is equal to 0 for a coding block), one shape-adaptive transform selection method is applied to all intra-coded blocks, where DCT-II and DST-VII transforms are implicitly enabled based on the width and height of the current block. More specifically, for each rectangular block, this method uses the DST-VII transform in the direction related to the short side of the block and the DCT-II transform in the direction related to the long side of the block. For each square block, the DST-VII is applied in both directions. Furthermore, to avoid introducing new transforms for different block sizes, the DST-VII transform is enabled only if the short side of an intra-coded block is 16 or less. Otherwise, the DCT-II transform is always applied.

[0051] Table 2 shows the enabled horizontal and vertical transforms for intra-coded blocks based on the shape adaptive transform selection method in VVC.

[0052] [Table 2] [Table 2: Shape adaptive transformation selection within a VVC block]

[0053] [Intra sub-division coding mode] Conventional intra-mode uses only the reconstructed samples adjacent to a coding block to generate intra-predicted samples for that block. The spatial correlation between the predicted sample and the reference sample based on this method is roughly proportional to the distance between the predicted sample and the reference sample. Therefore, samples in the inner part (especially those located in the lower right corner of the block) usually have lower prediction accuracy than samples near the block boundary. To further improve intra prediction efficiency, short-distance intra prediction (SDIP) has been proposed and extensively studied. This method divides an intra-coded block into multiple sub-blocks horizontally or vertically for prediction. Typically, a square block is divided into four sub-blocks. For example, an 8x8 block can be divided into four 2x8 or four 8x2 sub-blocks. One extreme case of such sub-block-based intra prediction is so-called line-based prediction, in which a block is divided into one-dimensional rows / columns for prediction. For example, a WxH (width x height) block can be divided into H sub-blocks of size Wx1 or W sub-blocks of size 1xH for intra prediction. Each of the resulting rows / columns is coded in the same way as a normal two-dimensional (2D) block (as shown in Figures 6A, 6B, 6C, and 7), i.e., predicted by one of the available intra-modes, and the prediction error is decorrelated based on transform and quantization and sent to decoder 200 (Figure 2) for reconstruction. As a result, the reconstructed samples in one sub-block (e.g., row / column) can be used as a reference to predict samples in the next sub-block. The above process is repeated until all sub-blocks in the current block have been predicted and coded. Furthermore, to reduce signaling overhead, all sub-blocks in one coding block share the same intra-mode.

[0054] In SDIP, different subblock partitions may provide different coding efficiencies. Generally, line-based prediction provides the best coding efficiency because it provides the "shortest prediction distance" between different partitions. On the other hand, codec hardware implementations have the worst encoding / decoding throughput. For example, considering a block with 4x4 subblocks and the same block with 4x1 or 1x4 subblocks, the latter case has only one-quarter the throughput of the former case. In HEVC, the minimum intra-prediction block size for luma is 4x4.

[0055] 8A shows an exemplary set of short-distance intra-prediction (SDIP) partitions for an 8x4 block 801, FIG. 8B shows an exemplary set of short-distance intra-prediction (SDIP) partitions for a 4x8 block 803, and FIG. 8C shows an exemplary set of short-distance intra-prediction (SDIP) partitions for a block of any size 805. Recently, a video coding tool called sub-partition prediction (ISP) has been introduced into the VVC standard. Conceptually, ISP is very similar to SDIP. In particular, depending on the block size, ISP divides the current coding block into two or four sub-blocks in either the horizontal or vertical direction, with each sub-block containing at least 16 samples.

[0056] 8A, 8B, and 8C show all possible partitioning cases for different coding block sizes. Furthermore, to handle interactions with other coding tools in the VVC standard, the current ISP design also includes the following main aspects:

[0057] Interaction with wide-angle intra direction: ISP is combined with wide-angle intra direction. In the current design, the block size (i.e., width / height ratio) used to determine whether the regular intra direction or its corresponding wide-angle intra direction should be applied is one of the original coding blocks, i.e., the block before sub-block division.

[0058] Interaction with multiple reference lines: ISP cannot be enabled together with multiple reference lines. In particular, in the current VVC signaling design, the ISP enable / disable flag is signaled after the MRL index. If an intra block has one non-zero MRL index (i.e., it refers to a non-nearest neighboring sample), the ISP enable / disable flag is not signaled and is inferred as 0, i.e., in this case, ISP is automatically disabled for the coding block. will be installed.

[0059] ● Interaction with Most Probable Mode: Similar to normal intra modes, the intra mode to be used for one ISP block is signaled via the Most Probable Mode (MPM) mechanism. However, compared to normal intra modes, the MPM scheme for ISP has the following modifications: 1) A split ISP block enables only the intra modes included in its MPM list and disables all other intra modes not in the MPM list. 2) For a split ISP block, its MPM list excludes DC mode, prioritizes horizontal intra modes for ISP horizontal split, and prioritizes vertical modes for ISP vertical split.

[0060] At least one non-zero coefficient block flag (CBF): In current VVC, a CBF flag is signaled for each transform unit (TU) to identify that the transform block contains one or more transform coefficient levels not equal to 0. Given a particular block that uses ISP, the decoder assumes that at least one of the subdivisions has a non-zero CBF. Thus, if the first n-1 subdivisions produce a CBF of zero, where n is the number of subdivisions, then the CBF of the nth subdivision is inferred to be 1. Therefore, there is no need to transmit and decode it.

[0061] Interaction with Multiple Transform Selection: ISP is applied exclusively with MTS. That is, if a coding block uses ISP, its MTS flag is not signaled but is always inferred to be 0 (disabled). However, instead of always using the DCT-II transform, a fixed set of core transforms (including DST-VII and DCT-II) is implicitly applied to the ISP coding block based on the block size. In particular, assuming W and H are the width and height of one ISP subdivision, its horizontal and vertical transforms are selected according to the following rules, as listed in Table 3:

[0062] [Table 3] [Table 3 Selected horizontal and vertical transformations for ISP blocks]

[0063] [Cross-component linear model prediction] 9A is a plot of chroma values ​​as a function of luma values, where the plot is used to derive a set of linear model parameters. More specifically, a line 901 representing the relationship between chroma values ​​and luma values ​​is used to derive a set of linear model parameters as follows: To reduce cross-component redundancy, a cross-component linear model (CCLM) prediction mode is used in VVC, whereby chroma samples are By using such a linear model, the predicted values ​​are based on the reconstructed luminance samples of the same CU. pred C (i,j)=α·rec L '(i,j)+β …(8) where pred C (i,j) represents the predicted chroma sample in the CU, and rec L '(i,j) represents the downsampled and reconstructed luma sample of the same CU. The linear model parameters α and β are derived from the line 901 representing the relationship between luma and chroma values ​​from the two samples, as illustrated in FIG. 9A. They are the linear model parameters α and β for the minimum luma sample A(X A , Y A ) and the maximum luminance sample B(X B , Y B ) where X A , Y A are the x-coordinate (brightness value) and y-coordinate (saturation value) values ​​of sample A, and X B , Y B are the x and y coordinate values ​​of sample B. The linear model parameters are obtained according to the following equations: α=(y B -y A ) / (x B -x A ) β=y A -αx A This method is also called min-max method. The division in the above equation can be avoided and replaced with multiplication and shift.

[0064] Figure 9B shows the locations of samples used to derive the linear model parameters of Figure 9A. For coding blocks with a square shape, the above two equations for the linear model parameters α and β apply directly. For non-square coding blocks, the adjacent samples on the longer boundary are first subsampled to have the same number of samples as the shorter boundary. Figure 9B shows the locations of the left and top samples involved in CCLM mode, as well as the samples of the current block, which includes an N x N set of chroma samples 903 and a 2N x 2N set of luma samples 905. In addition to using the top and left templates to jointly calculate the linear model coefficients, these templates can also be alternatively used in two other LM modes, called LM_A and LM_L modes.

[0065] In LM_A mode, only pixel samples in the template are used to calculate the linear model coefficients. To obtain more samples, the template is extended to (W+W). In LM_L mode, only pixel samples in the left template are used to calculate the linear model coefficients. To obtain more samples, the left template is extended to (H+H). Note that when the upper reference line is at the CTU boundary, only one luma line (a typical line buffer in intra prediction) is used to create the downsampled luma samples.

[0066] For chroma intra mode coding, a total of eight intra modes are allowed for chroma intra mode coding. These modes include five conventional intra modes and three cross-component linear model modes (CCLM, LM_A, and LM_L). The chroma mode signaling and derivation process is shown in Table 4. Chroma mode coding directly depends on the intra prediction mode of the corresponding luma block. In an I slice, separate block partition structures for luma and chroma components are available, so one chroma block can correspond to multiple luma blocks. Therefore, for chroma DM mode, the intra prediction mode of the corresponding luma block that covers the center position of the current chroma block is directly inherited.

[0067] [Table 4] Table 4: Deriving chroma prediction mode from luma mode when CCLM_ is enabled

[0068] Although ISP tools in VVC can improve intra-prediction efficiency, there is room for further improvement in VVC performance. Meanwhile, some parts of existing ISPs would benefit from further simplification to provide more efficient codec hardware implementations and / or to provide improved coding efficiency. In this disclosure, several methods are proposed to further improve ISP coding efficiency, simplify existing ISP designs, and / or facilitate improved hardware implementations.

[0069] [Independent (or parallel) subdivision predictor generation for ISP] 10 illustrates the generation of reference samples for intra prediction 1007 for all subdivisions using only reference samples outside the current coding block 1000. The current coding block 1000 includes a first subdivision 1, 1001, a second subdivision 2, 1002, a third subdivision 3, 1003, and a fourth subdivision 4, 1004. This disclosure proposes generating intra predictions independently for each subdivision 1001, 1002, 1003, and 1004. In other words, all predictors for subdivisions 1001, 1002, 1003, and 1004 can be generated in parallel. In one embodiment, predictors for all subdivisions are generated using the same technique as used in conventional non-subdivision intra mode. In particular, the reconstructed samples of one sub-division are not used to generate intra-prediction samples for any other sub-division within the same coding unit, and all predictors for each sub-division 1001, 1002, 1003, 1004 are generated using the reference samples of the current coding block 1000, as shown in Figure 10.

[0070] In the ISP mode of the VVC standard, the width of each subdivision can be 2 or less. One detailed example is as follows: According to the ISP mode of the VVC standard, the dependency of 2×N (width × height) sub-block prediction on the reconstructed values ​​of previously decoded 2×N sub-blocks of the coding block is not allowed, so that the minimum width for the prediction of a sub-block is 4 samples. For example, an 8×8 coding block coded using ISP with vertical division is divided into four prediction regions, each of size 2×8, and the left two 2×8 prediction regions are merged into the first 4×8 prediction region to perform intra prediction. A conversion circuit 102 (FIG. 1) is applied to each 2×8 division.

[0071] According to this example, the two right 2x8 prediction regions are merged into a second 4x8 prediction region to perform intra prediction. A transform circuit 102 is applied to each 2x8 division. The first 4x8 prediction region generates an intra predictor using neighboring pixels of the current coding block. Note that the second 4x8 region uses reconstructed pixels from the first 4x8 region (located to the left of the second 4x8 region) or neighboring pixels from the current coding block (located above the second 4x8).

[0072] It should be noted that in another embodiment, the intra prediction of the horizontal sub-division can be formed using only horizontal (HOR) prediction modes (mode 18 as shown in FIG. 4 ) and only prediction modes with mode indices less than 18 (as shown in FIG. 4 ), and the intra prediction of the vertical sub-division can be formed using only vertical (VER) prediction modes (i.e., mode 50 as shown in FIG. 4 ) and prediction modes with mode indices greater than 50 (as shown in FIG. 4 ). As a result, in the HOR prediction mode (shown as mode 18 in FIG. 4 ) and all of the angular prediction modes with modes less than 18, the intra prediction for each horizontal sub-division can be performed independently and in parallel. Similarly, in the VER prediction mode (mode 50 in FIG. 4 ) and all of the angular prediction modes with modes greater than 50, the intra prediction for each vertical sub-division can be performed independently and in parallel.

[0073] [Intra-Prediction Mode Coding for Luminance Component for ISP] This disclosure proposes allowing only N modes of all possible intra-prediction modes for the luma component for an ISP-coded block (N is a positive integer). In one embodiment, only one mode is allowed for the luma component of the ISP-coded block. For example, this single-allowed mode may be planar mode. In another example, this single-allowed mode may be DC mode. In a third example, this single-allowed mode may be one of HOR prediction mode, VER prediction mode, or diagonal (DIA) mode (mode 34 in FIG. 4) intra-prediction modes.

[0074] In another embodiment, only one mode is allowed for the luma component of an ISP coded block, and this mode may differ depending on the direction of the subdivision, i.e., whether it is horizontal or vertical subdivision. For example, only HOR prediction mode is allowed for horizontal subdivision, while only VER prediction mode is allowed for vertical subdivision. In yet another example, only VER prediction mode is allowed for horizontal subdivision, while only HOR prediction mode is allowed for vertical subdivision.

[0075] In yet another embodiment, only two modes are allowed for the luma component of an ISP coded block, each of which may be selected in response to the direction of the corresponding subdivision, i.e., whether the subdivision is horizontal or vertical. For example, only planar and HOR prediction modes are allowed for horizontal subdivision, while only planar and VER prediction modes are allowed for vertical subdivision.

[0076] The conventional Most Probable Mode (MPM) mechanism is not utilized to signal the N modes allowed for the luma component of an ISP-coded block. Instead, we propose to signal intra modes for an ISP-coded block using a binary codeword determined for each intra mode. The codewords can be generated using any of a variety of different processes, including a truncated binary (TB) binarization process, a fixed-length binarization process, a truncated Rice (TR) binarization process, a k-th order Exp-Golomb binarization process, and a restricted EGk binarization process. These binary codeword generation processes are clearly defined in the HEVC specification. Truncated Rice, with the Rice parameter equal to zero, is also known as truncated unary binarization. Example sets of codewords using different binarization methods are shown in Table 5.

[0077] [Table 5] Table 5. Binary code words generated using various binarization methods.

[0078] [Independent (or parallel) subdivision predictor generation for ISP] In yet another embodiment, the MPM derivation process for normal Intra mode is directly reused for ISP mode, and the signaling method of MPM flag and MPM index is kept the same as the existing ISP design.

[0079] [Intra-prediction mode coding for chroma components for ISP] In this disclosure, the ISP coding blocks (N c is a positive integer), N from all possible chroma intra prediction modes c It is proposed to allow only one mode. In one embodiment, only one mode is allowed for the chroma components of an ISP coded block. For example, this single allowed mode may be a direct mode (DM). In another example, this single allowed mode may be an LM mode. In a third example, this single allowed mode may be one of an HOR prediction mode or a VER prediction mode. The direct mode (DM) is configured to apply the same intra prediction mode used by the corresponding luma block to the chroma block.

[0080] In another embodiment, only two modes are allowed for the chroma components of an ISP coded block: In one example, only DM and LM are allowed for the chroma components of an ISP coded block.

[0081] In yet another embodiment, only four modes are allowed for the chroma components of an ISP coded block: In one example, only DM, LM, LM_L, and LM_A are allowed for the chroma components of an ISP coded block.

[0082] N for the chroma component of the ISP coding block cThe conventional MPM mechanism is not utilized to signal the mode. Instead, a fixed binary codeword is used to indicate the selected chroma mode in the bitstream. For example, the chroma intra-prediction mode of an ISP-coded block can be signaled using a determined binary codeword. The codeword can be generated using different processes, including a truncated binary (TB) binarization process, a fixed-length binarization process, a truncated Rice (TR) binarization process, a k-th order Exp-Golomb binarization process, a limited EGk binarization process, etc.

[0083] [Saturation component coding block size of ISP coding block] This disclosure allows subdivision coding for chroma components for ISP coded blocks. It is proposed that there is no subdivision for ISP coded blocks. Instead, regular full-block-based intra prediction is used for the chroma components of ISP coded blocks. In other words, for ISP coded blocks, subdivision is only performed for their luma components that do not have subdivision for their chroma components.

[0084] [Combination of ISP and inter prediction] Figure 11 shows the combination of inter and intra predictor samples for the first sub-division 1001 of Figure 10. Similarly, Figure 12 shows the combination of inter and intra prediction samples for the second sub-division 1002 of Figure 10. More specifically, to further improve coding efficiency, a new prediction mode is provided in which prediction is generated as a weighted combination (e.g., weighted averaging) of ISP mode and inter prediction mode. Intra predictor generation for ISP mode is the same as that described above in connection with Figure 10. The inter predictor may be generated by a merge mode or inter mode process.

[0085] In the illustrative example of Figure 11, inter predictor samples 1101 of the current block (including all sub-divisions) are generated by performing motion compensation using merge candidates indicated by the merge index. Intra predictor samples 1103 are generated by performing intra prediction using the signaled intra mode. Note that this process can also generate intra predictor samples for sub-divisions other than the first (i.e., second sub-division 1002) using reconstructed samples of previous sub-divisions, as shown in Figure 12. After the inter and intra predictor samples are generated, they are weighted-averaged to generate the final predictor samples for the sub-division. The combined mode can be treated as an intra mode. Alternatively, the combined mode may be treated as an inter mode or a merge mode instead of an intra mode.

[0086] [CBF Signaling for ISP Coding Blocks] To simplify the design of the ISP, this disclosure proposes to always signal the CBF for the last subdivision.

[0087] In another embodiment of the present disclosure, it is proposed to not signal the CBF of the last subdivision but to infer its value at the decoder side. For example, the value of the CBF of the last subdivision is always inferred as 1. In another example, the value of the CBF for the last subdivision is always inferred as zero.

[0088] According to one embodiment of the present disclosure, a method of video encoding includes independently generating a respective intra prediction for each of a plurality of corresponding sub-divisions, each intra prediction being generated using a plurality of reference samples from a current coding block.

[0089] In some examples, reconstructed samples from a first sub-division of the plurality of corresponding sub-divisions are not used to generate a respective intra prediction for any other sub-division of the plurality of corresponding sub-divisions.

[0090] In some examples, each of the plurality of corresponding subdivisions has a width of 2 or less.

[0091] In some examples, the plurality of corresponding sub-divisions include a plurality of vertical sub-divisions and a plurality of horizontal sub-divisions, and the method further comprises generating a first set of intra predictions for the plurality of horizontal sub-divisions using only a horizontal prediction mode, and generating a first set of intra predictions for the plurality of vertical sub-divisions using only a vertical prediction mode. generating a second set of intra predictions for the partition.

[0092] In some examples, horizontal prediction modes are performed using mode indices less than 18.

[0093] In some examples, the vertical prediction mode is implemented using a mode index greater than 50.

[0094] In some examples, the horizontal prediction mode is performed independently and in parallel for each of the multiple horizontal subdivisions.

[0095] In some examples, the vertical prediction mode is performed independently and in parallel for each of the multiple vertical subdivisions.

[0096] In some examples, the plurality of corresponding sub-divisions includes a last sub-division, and the method further comprises signaling a coefficient block flag (CBF) value of the last sub-division.

[0097] In some examples, the plurality of corresponding sub-divisions includes a last sub-division, and the method further comprises inferring, at the decoder, a coefficient block flag (CBF) value of the last sub-division.

[0098] In some examples, the coefficient block flag (CBF) value is always inferred as one.

[0099] In some examples, the coefficient block flag (CBF) value is always assumed as zero.

[0100] According to another embodiment of the present disclosure, a method of video encoding includes, for a luma component of an intra sub-partition (ISP) coding block, generating a respective intra prediction for each of a plurality of corresponding sub-partitions using only N modes out of M possible intra prediction modes, where M and N are positive integers and N is less than M.

[0101] In some examples, each of the plurality of corresponding subdivisions has a width of 2 or less.

[0102] In some instances, N is equal to 1, so that only a single mode is allowed for the luminance component.

[0103] In some instances, the single mode is a planar mode.

[0104] In some examples, the single mode is a DC mode.

[0105] In some examples, the single mode is one of a horizontal (HOR) prediction mode, a vertical (VER) prediction mode, or a diagonal (DIA) prediction mode.

[0106] In some examples, the single mode is selected in response to the direction of the subdivision, the horizontal (HOR) prediction mode is selected in response to the direction of the subdivision being horizontal, and the vertical (VER) prediction mode is selected in response to the direction of the subdivision being vertical.

[0107] In some examples, the single mode is selected in response to the direction of the subdivision, the horizontal (HOR) prediction mode is selected in response to the direction of the subdivision being vertical, and the vertical (VER) prediction mode is selected in response to the direction of the subdivision being horizontal.

[0108] In some examples, N is equal to 2, and two modes are allowed for the luminance component.

[0109] In some examples, a first set of two modes is selected in response to the direction of the first subdivision, and a second set of two modes is selected in response to the direction of the second subdivision.

[0110] In some examples, the first set of two modes includes a planar mode and a horizontal (HOR) prediction mode, and the first sub-division direction includes horizontal sub-division, and the second set of modes includes a planar mode and a vertical (VER) prediction mode, and the second sub-division direction includes vertical sub-division.

[0111] In some examples, each of the N modes is signaled using a corresponding binary codeword from a predetermined set of binary codewords.

[0112] In some examples, the set of predetermined binary code words is generated using at least one of a truncated binary (TB) binarization process, a fixed length binarization process, a truncated Ricean (TR) binarization process, a truncated unary binarization process, a k-th order Exp-Golumb binarization process, or a limited EGk binarization process.

[0113] According to another embodiment of the present disclosure, a method of video encoding includes generating intra predictions for a chroma component of an intra sub-partition (ISP) coded block using only N modes out of M possible intra prediction modes, where M and N are positive integers and N is less than M.

[0114] In some instances, N is equal to 1, so that only a single mode is allowed for the chroma component.

[0115] In some examples, the single mode is a direct mode (DM), a linear model (LM) mode, a horizontal (HOR) prediction mode, or a vertical (VER) prediction mode.

[0116] In some examples, N is equal to 2, so that two modes are allowed for the chroma component.

[0117] In some examples, N is equal to 4, so that four modes are allowed for the chroma component.

[0118] In some examples, each of the N modes is signaled using a corresponding binary codeword from a predetermined set of binary codewords.

[0119] In some examples, the set of predetermined binary code words is generated using at least one of a truncated binary (TB) binarization process, a fixed length binarization process, a truncated Ricean (TR) binarization process, a truncated unary binarization process, a k-th order Exp-Golumb binarization process, or a limited EGk binarization process.

[0120] According to another embodiment of the present disclosure, a method of video encoding comprises generating a respective luma intra prediction for each of a plurality of corresponding sub-partitions of an entire intra sub-partition (ISP) coded block for the luma component and for the chroma component, and generating a chroma intra prediction for the entire intra sub-partition (ISP) coded block.

[0121] In some examples, each of the plurality of corresponding subdivisions has a width of 2 or less.

[0122] According to another embodiment of the present disclosure, a method of video encoding generates a first prediction using an intra sub-partitioning mode, generates a second prediction using an inter prediction mode, and combines the first and second predictions to generate a final prediction by applying a weighted average to the first and second predictions.

[0123] In some examples, the second prediction is generated using at least one of a merge mode or an inter mode.

[0124] In one or more examples, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored on or transmitted via a computer-readable medium as one or more instructions or code and executed by a hardware-based processing unit. Computer-readable media may include computer-readable storage media, which correspond to tangible media, such as data storage media, or communication media, including any medium that facilitates transfer of a computer program from one place to another, for example, according to a communication protocol. In this manner, computer-readable media generally can correspond to (1) tangible computer-readable storage media that is non-transitory, or (2) communication media, such as a signal or carrier wave. Data storage media may be any available medium accessible by one or more computers or one or more processors to retrieve instructions, code, and / or data structures for implementation as described herein. A computer program product may include computer-readable media.

[0125] Furthermore, the above methods may be implemented using an apparatus including one or more circuits, including an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components. The apparatus may use the circuits in combination with other hardware or software components to perform the above-described methods. Each module, sub-module, unit, or sub-unit disclosed above may be implemented at least in part using one or more circuits.

[0126] Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention which follow its general principles, including departures from the present disclosure which come within known or customary practice in the art. The specification and embodiments are intended to be exemplary only, with the true scope and spirit of the invention being indicated by the following claims.

[0127] It will be understood that the present invention is not limited to the precise examples described above and shown in the accompanying drawings, and that various modifications and changes may be made thereto without departing from the scope of the present invention, which is intended to be limited only by the appended claims.

[0128] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Patent Application No. 62 / 801,214, filed February 5, 2019. The entire disclosure of the foregoing application is incorporated herein by reference in its entirety.

Claims

1. Obtaining a bitstream including the encoded coding blocks; if it is determined that the plurality of sub-divisions are a plurality of vertical sub-divisions, each of the sub-divisions having a width less than or equal to two, performing intra prediction on the plurality of sub-divisions of the coding block in an intra sub-division (ISP) mode; the coding block is associated with a predetermined division scheme, and the predetermined division scheme includes four-way division, horizontal three-way division, vertical three-way division, horizontal two-way division, and vertical two-way division; performing intra prediction on the plurality of sub-partitions of the coding block, merging at least two subdivisions of the plurality of subdivisions into one prediction region of the intra prediction; generating a prediction sample for the prediction region of the intra prediction based on a plurality of reference samples adjacent to the prediction region; Including, Video decoding methods.

2. samples reconstructed from one subdivision of the at least two subdivisions are not used to perform the intra prediction of any other subdivision of the at least two subdivisions; 10. The video decoding method of claim 1.

3. generating a set of prediction samples for the plurality of vertical sub-divisions using a vertical prediction mode with a mode index greater than 50.

10. The video decoding method of claim 1.

4. performing the intra prediction independently and in parallel for a plurality of prediction regions; 10. The video decoding method of claim 1.

5. the plurality of subdivisions includes a final subdivision; signaling a coefficient block flag (CBF) value for the last sub-division.

10. The video decoding method of claim 1.

6. the plurality of subdivisions includes a final subdivision; At the decoder, further comprising inferring a coefficient block flag (CBF) value of the last sub-partition.

6. The video decoding method of claim 5.

7. The coefficient block flag (CBF) value is always inferred as 1 or 0.

7. The video decoding method of claim 6.

8. two or more prediction samples of the prediction region are generated in parallel; 10. The video decoding method of claim 1.

9. one or more processors; a non-transitory computer-readable memory storing instructions executable by the one or more processors; Including, The one or more processors: Obtaining a bitstream including the encoded coding blocks; performing intra prediction on the subdivisions of the coding block in an intra subdivision (ISP) mode if it is determined that the subdivisions are vertical subdivisions and each of the subdivisions has a width less than or equal to 2; It is configured as follows: the coding block is associated with a predetermined division scheme, and the predetermined division scheme includes four-way division, horizontal three-way division, vertical three-way division, horizontal two-way division, and vertical two-way division; performing intra prediction on the plurality of sub-partitions of the coding block, merging at least two subdivisions of the plurality of subdivisions into one prediction region of the intra prediction; generating a prediction sample for the prediction region of the intra prediction based on a plurality of reference samples adjacent to the prediction region; Including, Video decoding device.

10. samples reconstructed from one subdivision of the at least two subdivisions are not used to perform the intra prediction of any other subdivision of the at least two subdivisions; 10. A video decoding apparatus according to claim 9.

11. The one or more processors further generating a set of prediction samples for the plurality of vertical sub-divisions using a vertical prediction mode with a mode index greater than 50; It is configured as follows:

10. A video decoding apparatus according to claim 9.

12. A non-transitory computer-readable storage medium storing a plurality of programs to be executed by a video decoding device having one or more processors, comprising: When executed by the one or more processors, the plurality of programs cause the video decoding device to perform the video decoding method according to any one of claims 1 to 8. A non-transitory computer-readable storage medium.

Images