RU2803532C2 - Method and device for effective encoding of residual blocks - Google Patents

Abstract

FIELD: image coding.

SUBSTANCE: method for decoding video images and a device for decoding video images. In order to efficiently encode the residual blocks obtained from the block-based motion compensation, the video encoder and the video decoder divide the corresponding residual block of the current block into two sub-blocks in the horizontal or vertical direction, and encode only one residual sub-block of the two residual sub-blocks, where the filtering block is made with the ability to set a grid of N samples at equal intervals in the horizontal and vertical directions and perform deblocking filtering on the boundary between two sub-blocks in the current block, which coincides with the grid boundary. The current block is a brightness block, a horizontal transform kernel and a vertical transform kernel for said one sub-block are determined based on the relative position of one sub-block in the current block.

EFFECT: increasing the efficiency of encoding/decoding of the residual blocks.

14 cl, 11 dwg

Description

Technical field

The present invention, described by way of example of some embodiments, relates to encoding and decoding of video images. In particular, the present invention relates to a method and apparatus for efficiently encoding residual blocks.

State of the art

Since video data has a large data volume compared to audio data or still image data, it requires a lot of hardware resources, including memory, to store or transmit the data in raw form before subjecting it to the compression process.

Accordingly, storage or transmission of video data is typically accompanied by compression using an encoder before the decoding unit can receive, decompress, and play back the compressed video data. Well-known video compression technologies include the H.264/AVC standard and the High Efficiency Video Coding (HEVC) standard, which improves H.264/AVC coding efficiency by approximately 40%.

However, the ever-increasing size, resolution, and frame rate of video images, and the resulting increase in the amount of data that needs to be encoded, requires a new and improved compression method with higher encoding efficiency and higher image quality than existing compression methods.

In the video encoding process, the video encoding apparatus generates a residual block by predicting the current block by intra-prediction or inter-prediction, and then subtracting the sample values of the current block from the sample values of the prediction block. The video encoding apparatus divides the residual block into one or more transform blocks, applies a transform to the one or more transform blocks, and thereby transforms the residual values of the transform blocks from the pixel domain to the frequency domain. Depending on the prediction accuracy, some regions of the residual block may have no residual values or not many residual values, and sometimes it is very inefficient to blindly divide the residual block into smaller transform blocks.

Disclosure of the invention

Technical problem solved by the invention

The present invention, described by way of some embodiments, aims to provide a residual block encoding scheme suitable for residual blocks having regions with no residual values or small residual values.

Technical solution

In accordance with at least one aspect of the present invention, a video decoding apparatus is provided, including a decoding unit, a prediction unit, an inverse quantization and inverse transform unit, an adder, and a filtering unit. The decoding unit is configured to decode from the bitstream a flag indicating whether residual signals corresponding to only a partial area of the current block have been encoded, and decode from the bitstream information about transform coefficients for one subblock of two divided subblocks from the current block to recover the transform coefficients. when the flag indicates that residual signals corresponding to only a partial area of the current block have been encoded. The prediction block is configured to predict the current block to generate a prediction block. The inverse quantization and inverse transform block is configured to perform inverse quantization and inverse transform transformation of transform coefficients in one sub-block for which the transform coefficient information has been decoded to generate a residual block for the current block. The adder is configured to add the prediction block and the residual block to restore the current block. The filtering unit is configured to set a grid of N samples at regular intervals in the horizontal and vertical directions and perform deblocking filtering at the boundary between two subblocks in the current block, which coincides with the grid boundary.

According to another aspect of the present invention, a method for decoding video images is provided, the method including the steps of (i) decoding from the bitstream a flag indicating whether residual signals corresponding to only a partial area of the current block have been encoded, (ii) recovering transform coefficients by decoding from the stream transform coefficient information bits for one sub-block of two sub-blocks separated from the current block, when the flag indicates that residual signals corresponding to only a partial area of the current block have been encoded, (iii) predicting the current block to generate a prediction block, (iv) creating a residual block for the current block by performing inverse quantization and inverse transformation of the transform coefficients in one subblock for which the transform coefficient information has been decoded, (v) reconstructing the current block by adding the prediction block and the residual block, and (vi) establishing a grid of N samples through equal spacing in the horizontal and vertical directions, and performing deblocking filtering at the boundary between two subblocks in the current block that coincides with the grid boundary.

Brief description of drawings

In fig. 1 is a diagram illustrating a video encoding apparatus that can implement the methods proposed by the present invention.

In fig. 2 is a diagram explaining a block division method using the QTBTTT structure.

In fig. 3A is a diagram illustrating a variety of intra-prediction modes.

In FIG. 3B is a diagram illustrating a variety of intra-prediction modes, including wide-angle intra-prediction modes.

In FIG. 4 is a diagram illustrating a video decoding apparatus capable of implementing the technologies proposed by the present invention.

In fig. 5A-5D show various types of division of the residual block and subblocks to be encoded.

In fig. 6 is a diagram explaining the case of encoding a single sub-block in a relevant residual block of a current block, illustrating the positions of neighboring blocks to be referred to for quantization parameter (QP) prediction.

In fig. 7 shows a diagram of a chrominance block and at least one luma block, which is combined with the chrominance block.

In fig. 8 is a diagram of reconstructed current blocks, illustrating their relevant residual blocks undergoing vertical splits and transformations in their left subblocks to reveal their interblock separation boundaries along the deblocking boundary.

In fig. 9 is a diagram for explaining deblocking filtering performed at the boundary between residual subblocks at the residual signal level.

In fig. 10 is a flowchart of a method performed by a video encoding apparatus for inter-predictive encoding of a current video block according to at least one embodiment of the present invention.

In fig. 11 is a flowchart of a method performed by a video decoding apparatus for decoding a current inter-prediction encoded block from a coded video bitstream according to at least one embodiment of the present invention.

Carrying out the invention

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings. In the following description, like reference numerals preferably refer to like elements even though the elements are shown in different drawings. In addition, for the sake of clarity and brevity, the following description of certain embodiments has omitted detailed descriptions of related known components and functions insofar as they obscure the subject matter of the present invention.

In fig. 1 is a diagram illustrating a video encoding apparatus that can implement the methods proposed by the present invention. Next, the video encoding device and components of the device will be described with reference to FIG. 1.

The video encoding apparatus may be configured to include an image division unit 110, a prediction unit 120, a subtraction unit 130, a transform unit 140, a quantization unit 145, a reordering unit 150, an entropy encoding unit 155, an inverse quantization unit 160, an inverse transform unit 165 , adder 170, filtering unit 180 and memory 190.

The corresponding components of the video encoding apparatus may be implemented as hardware or software, or a combination of hardware and software. In addition, the function of each component may be implemented in software, and the software function of each component may be implemented to be executed by a microprocessor.

A video image consists of many images. Each image is divided into multiple regions, and coding is performed for each region. For example, one image is divided into one or more fragments and/or sections. Here, one or more fragments may be defined as a group of fragments. Each fragment and/or section is divided into one or more coding tree units (CTUs). And each CTU is divided into one or more coding units (CUs) in a tree structure. Information applied to the corresponding CUs is encoded as CU syntax, and information generally applied to CUs included in a single CTU is encoded as CTU syntax. In addition, information typically applied to all blocks in one section is encoded as section header syntax, and information applied to all blocks constituting one image is encoded in picture parameter set (PPS) or image header. In addition, information typically referenced by multiple images is encoded in a Sequence Parameter Set (SPS). In addition, information typically referenced by one or more SPSs is encoded in a Video Parameter Set (VPS). In the same way, information typically applied to a single fragment or group of fragments can be encoded as fragment header or fragment group header syntax.

The image division unit 110 determines the size of the CTU coding tree unit. CTU size information (CTU size) is encoded as SPS set syntax or PPS set syntax and transmitted to the video decoding apparatus.

The image division unit 110 divides each image constituting a video image into a plurality of CTUs having a predetermined size, and then uses a tree structure to recursively divide the CTUs. The leaf node in the tree structure becomes the CU coding unit, which is the basic coding unit.

The tree structure used can be a QuadTree (QT) in which the top node (or parent node) is divided into four equally sized lower nodes (or child nodes), a Binary Tree (BT) in which the top node divided into two lower nodes, by a ternary tree (TernaryTree (TT)), in which the top node is divided into three lower nodes with a size ratio of 1:2:1, or a combination of two or more of a QT structure, a BT structure, and a TT structure. For example, a structure including quadtree and binary tree (QTBT) or a structure including quadtree and binary tree ternary (QTBTTT) may be used. Here, the VTTT may be collectively referred to as a multiple type tree (MTT).

In fig. Figure 2 shows the tree structure of the QTBTTT partition. As shown in FIG. 2, the CTU unit can be first divided into a QT structure. Quadtree splitting can be repeated until the splitting block size reaches the minimum block size (MinQTSize) of a leaf node allowed in the QT structure. The first flag (QT_split_flag) indicating whether each node of the QT structure is divided into four lower-level nodes is encoded by the entropy encoding unit 155 and transmitted to the video decoding apparatus. If a leaf node of a QT structure does not exceed the maximum block size (MaxBTSize) of the root node allowed in a BT structure, it may be further divided into any one or more BT structures or TT structures. There can be multiple directions of separation in a VT structure and/or a CT structure. For example, there may be two directions in which the block of the corresponding node is divided horizontally and vertically. As shown in FIG. 2, when MTT splitting begins, a second flag (mtt_split_flag) indicating whether the nodes are split, and, if so, an additional flag indicating the split directions (vertical or horizontal), and/or a flag indicating the split or split type (binary or ternary ) is encoded by entropy encoding unit 155 and transmitted to the video decoding device.

Alternatively, before encoding the first flag (QT_split_flag) indicating whether each node is split into four lower-level nodes, a CU split flag (split_cu_flag) may be encoded indicating whether the node is split or not. When the value of a CU's split flag (split_cu_flag) indicates that a node has not been split, the block of that node becomes a leaf node in the split tree structure and becomes a coding unit (CU), which is the base coding unit. When the value of the CU split flag (split_cu_flag) indicates that a node has been split, the video encoding apparatus starts encoding from the first flag in the manner described above.

As another example of a tree structure that uses the QTBT structure, there are two types of partitioning, including a type that horizontally divides the block of the corresponding node into two blocks of the same size (i.e., symmetrical horizontal partitioning), and a type that it is also divided vertically (i.e. symmetrical vertical division). Encoded by the entropy encoding block 155 and transmitted to the video decoding device is a split flag (split_flag) indicating whether each node of the VT structure is divided into lower-level blocks, and split type information indicating its division type. Meanwhile, there may be an additional type in which the block of the corresponding node is split into two asymmetrically formed blocks. The asymmetrical shape may include a block shape of the corresponding node being divided into two rectangular blocks having an aspect ratio of 1:3, or a block shape of the corresponding node being divided in a diagonal direction.

The CU unit can have different sizes depending on the structure of the QTBT or QTBTTT division of the CTU unit. In the following, the block corresponding to the CU to be encoded or decoded (i.e., the QTBTTT end node) is called the “current block”. When using QTBTTT splitting, the current block can be shaped not only as a square, but also as a rectangle.

The prediction block 120 predicts the current block to generate a prediction block. The prediction block 120 includes an intra-prediction block 122 and an inter-prediction block 124.

In general, each of the current image blocks may be predictively encoded. Prediction of the current block can usually be done using an intra-prediction method or an inter-prediction method, with the intra-prediction method using data from the image containing the current block itself, and the inter-prediction method using data from the previous image encoded before the image containing current block. Inter-prediction includes both unidirectional and bidirectional prediction.

The intra-prediction unit 122 predicts the pixels in the current block using peripheral pixels (reference pixels) located around the current block in the current image. The different prediction directions represent several corresponding intra-prediction modes. For example, as shown in FIG. 3A, multiple intra-prediction modes may include 2 omnidirectional modes including planar mode and DC mode, as well as 65 directional modes. The corresponding prediction modes provide various appropriate definitions of neighboring pixels and calculation formulas to be used.

For efficient directional prediction of the current rectangular block, additional modes used may be the directional modes shown in FIG. 3B with dotted arrows of intra-prediction modes numbered 67-80 and numbers from -1 to -14. These may be called "wide-angle intra-prediction modes". Arrows in Fig. 3B indicate the corresponding reference samples used for prediction rather than the direction of prediction. The direction of the prediction is opposite to the direction indicated by the arrow. Wide-angle intra-prediction modes are modes applied when the current block is rectangular in shape to perform backward prediction of a specific directional mode without additional bit transmission. In this case, of the wide-angle intra-prediction modes, some of the wide-angle intra-prediction modes available for use in the current block may be determined by the width-to-height ratio of the rectangular current block. For example, wide-angle intra-prediction modes that have an angle of less than 45 degrees (intra-prediction modes numbered 67-80) are available for use in the current block when they have a rectangular shape in which the height is less than the width. Wide-angle intra-prediction modes that have an angle of -135 degrees or greater (intra-prediction modes numbered -1 through -14) are available for use in the current block if there is a rectangular shape with a height greater than the width.

The intra-prediction block 122 may determine the intra-prediction mode that will be used to encode the current block. In some examples, intra-prediction block 122 may encode the current block using multiple intra-prediction modes and select an appropriate intra-prediction mode from the tested modes to use. For example, intra-prediction block 122 may calculate speed distortion values by analyzing the speed-distortion of several tested intra-prediction modes and select the intra-prediction mode that has the best speed-distortion characteristics among the tested modes.

The intra prediction unit 122 selects one intra prediction mode from a plurality of intra prediction modes and predicts the current block using at least one neighboring pixel (reference pixel) determined according to the selected intra prediction mode and the calculation formula. Information about the selected intra-prediction mode is encoded by entropy encoding unit 155 and transmitted to the video decoding device.

The inter-prediction block 124 generates a prediction block for the current block through a motion compensation process. The inter-prediction unit 124 searches for a block most similar to the current block in the encoded and decoded reference picture before the current picture, and generates a current block prediction block using the found block. Then, the inter-prediction block 124 generates a motion vector corresponding to the offset between the current block in the current image and the prediction block in the reference image. In general, motion estimation is performed on the luminance component, and a motion vector calculated based on the luminance component is used for both the luminance component and the chrominance component. Motion information including reference image information and motion vector information used for predicting the current block is encoded by the entropy encoding unit 155 and transmitted to the video decoding apparatus.

Subtraction unit 130 generates a residual block by subtracting from the current block a prediction block generated by intra-prediction unit 122 or inter-prediction unit 124.

The transform block 140 divides the residual block into one or more transform blocks, applies the transform to the one or more transform blocks, and thereby converts the residual values of the transform blocks from the pixel domain to the frequency domain. In the frequency domain, transformed blocks are called coefficient blocks containing one or more transform coefficient values. A two-dimensional transformation kernel can be used for transformation, and a one-dimensional transformation kernel can be used for each of the horizontal transformation and the transformation in the vertical direction. The transform kernels may be based on discrete cosine transform (DCT), discrete sine transform (DST), or the like.

The transform block 140 may transform the residual signals in the residual block using the entire size of the residual block as a transform block. In addition, the transform unit 140 may divide the residual block into two sub-blocks in the horizontal or vertical direction and perform the transformation on only one of the two sub-blocks, as will be described below with reference to Figures 5A to 5D. Accordingly, the size of the transform block may be different from the size of the residual block (and hence the size of the prediction block). The subblock that is not converted may have no or very few non-zero residual samples. The residual samples of the subblock in which the conversion is performed are not signaled, and all of them can be considered as "0" by the video decoding apparatus. Depending on the directions and ratios of separation, several types of separation can exist. The transform block 140 provides the entropy encoding block 155 with information about the encoding mode (or transform mode) of the residual block, such as information indicating whether the transformed residual block is a residual subblock, information indicating the type of division selected for dividing the residual block into subblocks, and information for identifying the subblock in which the transformation is performed, etc. The entropy encoding unit 155 may encode information about the encoding mode (or transform mode) of the residual block.

The quantization unit 145 quantizes the transform coefficients output from the transform unit 140 and outputs the quantized transform coefficients to the entropy encoding unit 155 . The quantization unit 145 may directly quantize the corresponding residual block for a particular block or frame without conversion.

Reordering unit 150 may reorder coefficient values on the quantized residual samples. Reorderer 150 may use coefficient scanning to transform a two-dimensional matrix of coefficients into a one-dimensional sequence of coefficients. For example, reordering unit 150 may scan DC coefficients to high-frequency domain coefficients through a zigzag scan or a diagonal scan to output a one-dimensional sequence of coefficients. Depending on the size of the transform block and the intra-prediction mode, the zigzag scan used can be replaced by a vertical scan to scan a two-dimensional array of coefficients in the column direction and a horizontal scan to scan the coefficients into a two-dimensional block shape in the row direction. In other words, the scanning method to be used can be selected from zigzag scanning, diagonal scanning, vertical scanning and horizontal scanning according to the transform block size and the intra-prediction mode.

Entropy encoding block 155 uses various encoding techniques, such as context-sensitive adaptive binary arithmetic coding (CABAC), exponential Golomb code, and the like, to encode a sequence of one-dimensional quantized transform coefficients output from reordering block 150 to generate a bit stream.

In addition, the entropy encoding unit 155 encodes block division information such as CTU size, CU division flag, QT division flag, MTT division type, and MTT division direction, allowing the video decoding apparatus to divide the block in the same manner as the encoding apparatus. video images. In addition, the entropy encoding unit 155 encodes prediction type information indicating whether the current block is encoded by intra-prediction or inter-prediction, and decodes, depending on the prediction type, intra-prediction information, that is, intra-prediction mode information or inter-prediction information, i.e. information about reference images and motion vectors.

The inverse quantization unit 160 inversely quantizes the quantized transform coefficients output from the quantization unit 145 to generate transform coefficients. The inverse transform block 165 converts the transform coefficients output from the inverse quantization block 160 from the frequency domain to the spatial domain to reconstruct the residual block.

Adder 170 adds the reconstructed residual block and the prediction block generated by prediction block 120 to recover the current block. The pixels in the reconstructed current block are used as reference pixels for intra-prediction of the next block.

The filtering unit 180 filters the reconstructed pixels to reduce blocking artifacts, ringing artifacts, blur artifacts, etc. generated due to block-based prediction and transform/quantization. Filtering block 180 may include a deblocking filter 182 and an adaptive sample offset (SAO) filter 184.

A deblocking filter 180 filters the boundary between reconstructed blocks to remove blocking artifact caused by block-by-block encoding/decoding, and an SAO filter 184 further filters the deblocked filtered image. The SAO 184 filter is a filter used to compensate for the difference between the reconstructed pixel and the original pixel caused by lossy encoding.

The reconstructed block is filtered through deblocking filter 182 and SAO filter 184 and stored in memory 190. When all blocks in one image are reconstructed, the reconstructed image can be used as a reference image for inter-prediction of blocks in the upcoming image to be encoded.

In FIG. 4 is a diagram illustrating a video decoding apparatus configured to implement the methods of the present invention. Hereinafter, the video decoding apparatus and components of the apparatus will be described with reference to FIG. 4.

The video decoding apparatus may be configured to include an entropy decoding unit 410, a reordering unit 415, an inverse quantization unit 420, an inverse transform unit 430, a prediction unit 440, an adder 450, a filtering unit 460, and a memory 470.

As in the video encoding apparatus of FIG. 1, corresponding components of the video decoding apparatus may be implemented as hardware or software, or a combination of hardware and software. In addition, the function of each component may be implemented in software, and the software function of each component may be implemented to be executed by a microprocessor.

The entropy decoding unit 410 decodes the bit stream generated by the video encoding apparatus and extracts block division information to determine the current block to be decoded, and extracts prediction information necessary for reconstructing the current block and residual signal information, etc.

Entropy decoding unit 410 extracts CTU size information from a Sequence Parameter Set (SPS) or Picture Parameter Set (PPS), determines the CTU size, and divides the image into CTUs of a given size. Then, the entropy decoding unit 410 determines the CTU as the highest level, that is, the root node of the tree structure, and extracts the division information on the CTU and thereby divides the CTU using the tree structure.

For example, when splitting a CTU using the QTBTTT structure, the first flag (QT_split_flag) associated with the QT split is first retrieved, and each node is split into four lower-level nodes. For a node corresponding to a leaf node QT, entropy decoding unit 410 extracts a second flag (MTT_split_flag) associated with the MTT split and information about the split direction (vertical/horizontal) and/or split type (binary/ternary) to split this leaf node into MTT structure. This allows the corresponding nodes below the end node QT to be recursively split into a BT or TT structure.

As another example, when splitting a CTU using the QTBTTT structure, the entropy decoding unit 410 may first extract a CU split flag (split_cu_flag) indicating whether the CU is split, and after splitting the corresponding block, it may also extract the first flag. (QT_split_flag). In a partitioning process, each node may have zero or more recursive QT partitions followed by zero or more recursive MTT partitions. For example, a CTU may immediately enter an MTT partition or, conversely, have only a few QT partitions.

As another example, when splitting a CTU using a QTBT structure, the entropy decoding unit 410 extracts the first flag (QT_split_flag) associated with the QT split to split each node into four lower-level nodes. And for the node corresponding to the leaf node QT, the entropy decoding unit 410 extracts a split flag (split_flag) indicating whether this node is further divided into BT and split direction information.

Meanwhile, when the entropy decoding unit 410 determines the current block to be decoded by dividing the tree structure, it extracts prediction type information indicating whether the current block has been predicted by intra-prediction or inter-prediction. When the prediction type information indicates intra-prediction, the entropy decoding unit 410 extracts a syntax element for the intra-prediction information (intra-prediction mode) of the current block. When the prediction type information indicates inter-prediction, the entropy decoding unit 410 extracts a syntax element for the inter-prediction information, that is, information indicating the motion vector and the reference image referenced by the motion vector.

Meanwhile, the entropy decoding unit 410 extracts from the bitstream information about the encoding mode of the residual block, for example, information about whether the residual block has been encoded or only subblocks of the residual block have been encoded, information indicating the selected division type for dividing the residual block into subblocks, information , identifying the encoded residual subblocks, quantization parameters, etc. In addition, the entropy decoding unit 410 extracts information about the quantized transform coefficients of the current block as residual signal information.

The reordering unit 415 reorders, in reverse order of the coefficient scan sequence performed by the video encoding device, the one-dimensional quantized transform coefficients that are entropy decoded by the entropy decoding unit 410 into a two-dimensional coefficient array, i.e. block.

An inverse quantization unit 420 performs inverse quantization of the quantized transform coefficients. The inverse transform unit 430 performs inverse transform, based on the encoding mode information of the residual block, of the inversely quantized transform coefficients from the frequency domain to the spatial domain to recover the residual signals and thereby generate a reconstructed residual block of the current block.

The deconversion unit 430 responds when the residual block encoding mode information indicates that the residual block of the current block has been encoded in the video encoding apparatus by deconverting the inversely quantized transform coefficients using the size of the current block (thus, the size of the residual block should be restored) as a conversion unit to generate the restored residual block of the current block.

In addition, the inverse transform unit 430 responds when the residual block encoding mode information indicates that a single subblock of the residual block has been encoded in the video encoding apparatus by performing inverse transform for the inversely quantized transform coefficients using the size of the transformed subblock as the transform unit for restoring the residual signals of the transformed subblock while filling the residual signals of the untransformed subblock with "0" values, thereby generating a reconstructed residual block of the current block.

The prediction block 440 may include an intra-prediction block 442 and an inter-prediction block 444. The intra-prediction block 442 is activated when the prediction type of the current block is intra-prediction, and the inter-prediction block 444 is activated when the prediction type of the current block is inter-prediction.

The intra-prediction unit 442 determines, among a plurality of intra-prediction modes, the intra-prediction mode of the current block from the syntax element for the intra-prediction mode extracted by the entropy decoding unit 410, and according to the determined intra-prediction mode, it predicts the current block using adjacent reference pixels of the current block.

The inter-prediction block 444 uses the syntax element for the intra-prediction mode extracted by the entropy decoding block 410 to determine the motion vector of the current block and the reference image referenced by the motion vector, and it uses the motion vector and the reference image as determined to predict the current block. .

An adder 450 adds the residual block output from the inverse transform block and the prediction block output from the inter-prediction block or intra-prediction block to reconstruct the current block. The pixels in the reconstructed current block are used as reference pixels in intra-prediction of the incoming block to be decoded.

The filtering unit 460 may include a deblocking filter 462 and an SAO filter 464 . The deblocking filter 462 deblocks and filters the boundary between reconstructed blocks to remove block artifact caused by block decoding. The SAO filter 464 performs additional filtering on the reconstructed block after the deblocking filtering to compensate for the difference between the reconstructed pixel and the original pixel caused by lossy encoding. The reconstructed block is filtered through deblocking filter 182 and SAO filter 184 and stored in memory 190. When all blocks in one image are reconstructed, the reconstructed image can be used as a reference image for inter-prediction of blocks in the approximation image to be encoded.

The technologies of the embodiments illustrated herein generally relate to inter-prediction coding, that is, encoding and decoding of the current block. To efficiently encode the residual block obtained according to block-based motion compensation, the video encoding apparatus may divide the residual block into two residual subblocks in the horizontal or vertical direction, and encode only one residual subblock between the two residual subblocks. In other words, the transformation and quantization process can only be applied to one residual subblock. Residual values of another uncoded subblock are treated as "0" or ignored. Such an encoding tool can be effective in encoding a residual block when some area of it has no residual value or has a small residual value.

The video encoding apparatus can select an optimal encoding mode from the first and second encoding modes, wherein the first encoding mode encodes the residual block, thereby converting the residual signals in the residual block into a transform block that is equal in size to the prediction block, and the second encoding mode divides the residual block. block into two subblocks in the horizontal or vertical direction and encodes only one subblock, thus transforming the residual signals in the residual block into a transform block that is different in size from the prediction block.

The second mode associated with residual block encoding is described in detail next. The following description mainly relates to the decoding technology, in particular to the operation of the video decoding apparatus, while the encoding technology is described briefly since the latter is the inverse of the decoding technology described in detail. Additionally, the terms “residual block subblock(s)” and “residual subblock(s)” may be used interchangeably.

1. Split type and conversion position

In fig. 5A-5D show various types of partitioning or partitioning of the residual block and subblocks that are to be encoded (and thus transformed). The residual block can be divided into two subblocks. In the illustrated residual blocks 510-580, the subblocks that undergo transformation are shaded and labeled "A". A subblock located on the left or top side in a residual block can be expressed as located at "position 0" in the residual block, and a subblock located on the right or bottom side in a residual block can be expressed as located at "position 1" in residual block. The position of the sub-block in the residual block at which the transformation is performed may be referred to as the "transformation position". Therefore, a transformation position designated as "position 0" may represent a transformation performed on a subblock located on the left or top side in the residual block, a transformation position designated as "position 1" may represent a transformation performed on a subblock located on the right or below in the residual block.

Different types of separation can be considered according to the direction of separation and the separation ratio. Types of block divisions to consider include the first type in which the division occurs symmetrically in the horizontal direction, the second type in which the division occurs symmetrically in the vertical direction, the third type in which the division occurs asymmetrically in the horizontal direction (e.g. ratio 1:3 or 3:1), and a fourth type in which the separation occurs asymmetrically in the vertical direction (for example, in a ratio of 1:3 or 3:1).

Block 510 is vertically divided into a left subblock and a right subblock, and the transformation is not performed on the right subblock at position 1, but is performed on the left subblock at position 0. Block 520 is vertically divided into left and right subblocks, and the transformation is not performed on the left subblock, but is performed in the right subblock at position 1. Block 530 is horizontally divided into upper and lower subblocks and the transformation is performed in the upper subblock at position 0 rather than in the lower subblock. Block 540 is horizontally divided into two subblocks and the transformation is performed in the lower subblock at position 1 rather than in the upper subblock.

Block 550 is vertically divided into a left subblock and a right subblock with a ratio of 1:3, and the conversion is not performed on the right subblock at position 1, but is performed on the left subblock at position 0. Block 560 is vertically divided into left and right subblocks with a ratio of 3:1, and the transformation is performed not on the left subblock, but on the right subblock at position 1. The block 570 is horizontally divided into upper and lower subblocks having a ratio of 1:3, and the transformation is performed on the upper subblock at position 0, rather than on the lower subblock. Block 540 is divided into two subblocks in the horizontal direction, and the transformation is performed in the lower subblock at position 1 rather than in the upper subblock.

As with blocks 550-580 shown in FIGS. 5C and 5D, when a residual block is split asymmetrically (eg, 1:3 or 3:1), a default may be set to perform the transformation on the smaller subblock.

The available split ratios and split directions can be limited according to the size of the residual block to provide 4 or more pixels for the height or width of the subblock that is being transformed. For example, the first type of horizontally symmetrical division may be allowed for residual blocks having a width of 8 pixels or more, and the second type of vertically symmetrical division may be allowed for residual blocks having a height of 8 or more pixels. Likewise, the third type and fourth type, which provide horizontal or vertical asymmetrical division, may be allowed for residual blocks having a width or height of 16 pixels or more.

2. Selecting the transformation kernel.

Depending on the direction of division (horizontal or vertical) of the residual block and its transformation position, i.e. position of the transformed subblock, transformation kernels (horTransform and verTransform) in the horizontal and vertical directions can be selected and used to transform the subblock.

When vertical partitioning is applied to a residual block, the kernel to be used to transform the subblock can be implicitly obtained as follows. In the case of blocks 510 and 550, when the subblock being transformed is at position 0, then (horTransform, verTransform) = (DCT-8, DST-7). In the case of blocks 520 and 560, when the subblock being transformed is at position 1, then (horTransform, verTransform) = (DST-7, DST-7).

When vertical partitioning is applied to a residual block, the kernel to be used to transform the subblock can be implicitly obtained as follows. In the case of blocks 510 and 550, when the subblock being transformed is at position 0, then (horTransform, verTransform) = (DCT-8, DST-7). In the case of blocks 520 and 560, when the subblock being transformed is at position 1, then (horTransform, verTransform) = (DST-7, DST-7).

When one of the residual sub-blocks is larger than 32 pixels, the DCT-2 kernel can be applied to both horizontal and vertical directions. Moreover, for a chrominance component residual block, the DCT-2 kernel can always be applied to transform the residual subblock, regardless of the split direction (horizontal or vertical) of the residual block and its transform position, i.e. positions of the transformed subblock. In some cases, information identifying the transformation kernels used for transformation in the horizontal and vertical directions may be explicitly conveyed.

3. Residual coding mode signaling

When the current block is inter-prediction encoded and the corresponding residual block has at least one non-zero coefficient, a first flag may be explicitly signaled to indicate whether the relevant residual block of the current block is transformed or only one subblock of the residual block is encoded.

Accordingly, the video decoding apparatus, at the time of reconstructing a residual block of a current interprediction encoded block, can decode a first flag from the bitstream to determine whether the video encoding apparatus has divided the residual block into two subblocks and encoded only one subblock of the two subblocks.

In some embodiments, the first flag may be omitted depending on the width or height of the current block, or depending on the product of the width and height, the ratio of width to height, and the greater of the width and height. For example, when the width and height of the current block are less than 8 pixels, the video decoding apparatus may skip decoding the first flag and conclude that the entire residual block of the current block has been converted.

When only one subblock of a residual block is encoded, the division type and transform position of the residual block may be further signaled. Information about the division type and the transform position may be provided to the video encoding device in various ways. For example, one syntax element may be used to represent the split type of the residual block, or two flags may be used to represent the split ratio and split direction. In addition, one flag may be used to represent a transform position, which is the position of the subblock being converted. For example, you can set the flag indicating the transform position to "0" when the transform is applied to a subblock located on the left or top side in the remainder block, and set the flag indicating the transform position to "1" when the transform is applied to subblock located to the right or bottom in the residual block.

The remainder block, when partitioned 1:3 or 3:1, may add a constraint whereby the conversion is always performed in the smaller subblock, thereby eliminating signaling of the flag indicating the conversion position.

In addition, depending on the width and height of the residual block, or depending on the product of width and height, the ratio of width to height, and the greater of the width and height, it may be possible to limit the use of some type of partition or the direction of partition and/or partition ratio, thereby eliminating signaling syntax elements representing the same split type or split direction and/or split factor. For example, when the width and height of the residual block are less than 16 pixels, splitting the residual block "in a ratio of 1:3 or 3:1" cannot be allowed. Accordingly, the video decoding apparatus may respond to the residual block when the width and height are less than 16 pixels to output the division ratio as 1:1 without decoding the flag indicating the division ratio. In another example, when the width (or height) of the residual block is less than 8 pixels, the residual block cannot be allowed to be split in the vertical direction (or horizontal direction). Accordingly, the video decoding apparatus may respond to a residual block having a width of less than 8 pixels to infer that the residual block is divided in the horizontal direction without decoding flags indicating the direction of division. Similarly, the video decoding apparatus may respond to a residual block when it is less than 8 pixels in height to infer that division is performed in the vertical direction. According to yet another example, the device may react such that when the width is greater than the height, it is concluded that division is performed in the vertical direction, and when the height is greater than the width, it is concluded that division is performed in the horizontal direction.

The residual block associated with the luma component may have its single residual subblock exclusively coded, in which case the corresponding residual block associated with the chrominance component may also have its single residual subblock exclusively coded. Accordingly, the first flag transmitted at the CU unit level can be shared by the luma component residual block and the chrominance component residual blocks.

The same residual encoding mode selected for the luma component residual block can be applied to a chrominance component residual block co-located with the luma component residual block. In other words, when a luma component residual block has only one coded subblock, a chrominance component residual block may likewise have only one coded subblock. In this case, information about the division type of the residual block and the position of the subblock to be encoded can be shared between the luma block and the chrominance blocks.

In at least one embodiment of a single tree structure, the same residual encoding mode may be applied to the luma component and the chrominance component, and the same information about the split direction determined for the luma component residual block and the position of the subblock to be encoded may be applied to the residual blocks of chromaticity components. According to another example, in the case of encoding only one subblock of a residual chrominance component block, when only one subblock of a combined residual luma component block is encoded in the same way, the same split direction information determined for the residual luma component block and the position of the subblock to be encoded is can be applied to the residual blocks of the chrominance component.

4. Restoring the residual block

The video decoding apparatus can determine, based on the division type and the transform position, the position of a sub-block in the residual block where to decode the corresponding transform coefficient information from the bit stream, as well as the size of the sub-block. The video decoding apparatus may decode transform coefficient information from the bitstream to recover transform coefficients of a subblock corresponding to the transform position. The video decoding apparatus applies an inverse quantization/inverse transform process to the reconstructed transform coefficients to generate residual signals of a sub-block corresponding to a transform position, and sets all residual signals of another sub-block corresponding to a position without transform to "0", thereby recovering a residual block of the current block. .

Similarly, the video encoding apparatus transforms and quantizes the residual subblock corresponding to the transform position, entropy encodes the quantized transform coefficients, and applies an inverse quantization/inverse transform process to the quantized transform coefficients to recover the residual subblock corresponding to the transform position and sets all residual signals of the other subblock. the corresponding positions without transformation are equal to "0", thereby restoring the residual block of the current block.

5. Calculation of quantization parameters.

The video encoding apparatus may determine a quantization parameter (QP) value for a current unit (CU), and determine, based on the QP value and the QP prediction value, a differential quantization parameter (Delta QP or DQP) value for the current block. The video encoder may be configured to signal the DQP value and quantize the current block using the specific QP value. The video encoding apparatus may adjust the QP value of the current block and thus adjust the degree of quantization applied to the coefficient blocks related to the current block.

DQP is defined as the difference between the current QP, i.e. the actual QP used in the current block, and the predicted value of the current QP or the predicted QP value. Based on the signaled DQP, the corresponding current QP value can be recovered by adding the DQP to the QP prediction value. Specifically, the video encoding apparatus outputs a DQP to be calculated by subtracting the QP prediction value from the actual QP of the current block, and the video decoding apparatus outputs the actual QP of the current block to be reconstructed by adding the received DQP to the QP prediction value.

In some embodiments, the average of the actual QP values for the top block and the left block with respect to the current block may be determined as the QP prediction value of the current block. In some other embodiments, representative values (eg, mean, median, etc.) of the actual QP values for the left block, the top left block, and the top block relative to the current block may be defined as the QP prediction value for the current block.

In some embodiments, only one subblock of the corresponding residual block of the current block is encoded, with information about the divisions of the residual block, the shape of the division or the position of the subblock to be encoded, etc. are used as a basis for selecting neighboring blocks to reference. The actual QP values for selecting neighboring blocks can then be used as the QP prediction value of the current block, or the actual QP values of the selected neighboring block(s) can be used as the basis for deriving the QP prediction value for the current block.

In fig. 6 is a diagram explaining the case of encoding one sub-block in a relevant residual block of a current block, illustrating the positions of neighboring blocks to be referred to for quantization parameter (QP) prediction.

When the residual block 610 is vertically divided and the residual signal is only present in the left sub-block, that is, only the left sub-block is encoded, at least one of the QP prediction values of the neighboring blocks "a, b, d, e" adjacent to the left sub-block can be used as the basis for calculating the QP prediction value of the current block. For example, the QP prediction value of the current block may be calculated based on the average of the QP values of block "a" and block "b".

When the residual block 620 is vertically divided and the residual signal is only in the right sub-block, that is, only the right sub-block is encoded, the QP value of block “c”, which is a neighboring block adjacent to the right sub-block, can be determined as the QP prediction value of the current block.

In fig. 7 shows a diagram of a chrominance block and at least one luma block, which is combined with the chrominance block.

The QP value for a chrominance block may be determined based on the determined QP value for the corresponding luminance block(s). The QP prediction value of a chrominance block can be determined based on the QP value of a luma block at the same location. As an example, the QP value of the luminance block "a" located in the upper left corner of the current block can be used as the QP prediction value for the residual chrominance component block. Alternatively, the residual chroma component block prediction value QP used may be the most frequent value, maximum value, minimum value, average value, or average value among the QP values of luminance blocks a to i.

When encoding one subblock in a chrominance component residual block, the chrominance block QP prediction value used may be a luminance block QP value selected based on a larger value (alternatively a smaller value) among the division direction of the residual block, the position of the subblock to be encoded, the size of the subblock, the horizontal to vertical ratio subblock, and the width or height of the subblock. In at least one embodiment, the used QP prediction value of a chrominance component block is the QP value of one of the "d" block and a block which are luminance blocks corresponding to the subblock A encoded in the chrominance component, a residual block, or the average of the QP values of the "d" block " and block

6. In-loop filtration

The video encoding apparatus and the video decoding apparatus may perform in-loop filtering, including deblocking filtering, on the reconstructed current block, and store the filtered current block in a buffer (eg, memory 190 in FIG. 1 and memory 470 in FIG. 4) for future use. as a reference image for inter-prediction of incoming blocks in the encoded image.

A deblocking filter is applied to pixels at a block boundary to reduce blocking artifacts caused by the boundary between blocks in the reconstructed image. Deblocking filtering is performed on the vertical edge and horizontal edge. Filtering performed first on the vertical boundary can serve as an input for subsequent filtering on the horizontal boundary.

In fig. 8 is a diagram of reconstructed current blocks (810, 820) illustrating their respective residual blocks undergoing vertical splits and transformations in their left subblocks (ie, transformation position is 0) to show their inter-subblock division boundaries and deblocking boundaries.

The video encoding apparatus and the video decoding apparatus may determine whether or not to apply deblocking filtering to the current block based on whether a 4x4 (or 8x8) deblocking boundary coincides with a boundary between subblocks of the residual block. For example, a video decoding apparatus may establish a grid of N samples at regular intervals in the horizontal and vertical directions and perform deblocking filtering only when the boundary between two subblocks in the current block coincides with the grid boundary. Accordingly, deblocking filtering cannot be performed on boundaries between subblocks that do not coincide with the grid boundary. Here N can be set to 2n (n is a natural number). Alternatively, the video encoding apparatus and the video decoding apparatus may determine whether to apply deblocking filtering to the current block based on pixels included in multiple columns or rows included in the current block. A grid in different units can be used depending on whether the current block is a luma block or a chrominance block. For example, the grid can be set in 4x4 units for a luma block, and the grid can be set in 8x8 units for a chrominance block.

In some embodiments, deblocking filtering may be performed at the reconstructed residual signal level. In fig. 9 is a diagram for explaining deblocking filtering performed at the boundary between residual subblocks at the residual signal level. In fig. 9 shows a 4x8 residual block that undergoes horizontal splitting and transformation in its bottom 4x4 subblock (ie, transformation position is 1). After the residual signals corresponding to the lower sub-block are recovered and the residual signals of the upper sub-block are filled with "O" to complete the reconstructed residual block, deblocking filtering can be performed on the residual block. The embodiment of FIG. 9 illustrates a one-dimensional average filter having a parameter [1 2 1] applied to the boundary between two subblocks.

When applying a deblocking filter to the current block, a strong filter or a weak filter can be selected according to the desired deblocking filtering strength. In addition, one can consider using various conventional filters in common image processing techniques, such as mean filter, median filter, and two-way filter.

Regarding filter selection, the video decoding apparatus may be informed by the video encoding apparatus of the filter to be used before selecting a filter to use in the residual deblocking filtering. For example, the filter index to be used may be reported among a plurality of predefined filters. Alternatively, one predefined filter can be used.

Once a filter is selected, embodiments calculate filtering parameters related to the selected filter, particularly parameters that determine filter intensity and cutoff values for pixel changes, and based on the calculated filtering parameters filter on the residual sample boundary.

For filter parameters, the location information of a neighboring encoding block, such as a previous frame, a neighboring block, a group of neighboring fragments, a neighboring fragment, etc., can be transmitted, and the filter parameters of a neighboring decoding block corresponding to the location for the current block, as it is present, or the filter parameters for the current block are reconstructed from the values of neighboring parameters and additional signaled difference values.

In fig. 10 is a flowchart of a method performed by a video encoding apparatus for inter-predictive encoding of a current video block according to at least one embodiment of the present invention.

In step S1010, the video encoding apparatus can generate a prediction block by inter-prediction of the current block of video data to be encoded, and it can generate a residual block by subtracting the prediction block from the current block of video data.

In step S1020, the video encoding apparatus may determine a residual encoding mode for a residual block of the current block. In particular, the video encoding apparatus may select whether to encode a residual block of the current block or encode only one subblock that is a partial area of the residual block. This choice may be based on a speed-distortion test.

In step S1030, the video encoding apparatus may encode information about the residual encoding mode selected for the corresponding residual block of the current block.

The residual coding mode information may include a first flag indicating whether the residual block or only one subblock of the residual block is coded.

When only one subblock of the residual block is encoded, the video encoding apparatus may further encode information indicating a division type for dividing the residual block into two subblocks. Different types of separation can be considered according to the direction of separation and the separation ratio. For example, an inter-prediction coding technique may use partition types including a first type that provides a horizontal symmetric partition into two sub-blocks, a second type that provides a vertical symmetric partition into two sub-blocks, a third type that provides a horizontal asymmetric partition into two sub-blocks , (for example, with a ratio of 1:3 or 3:1), and a fourth type, which provides an asymmetrical vertical division into two sub-blocks (for example, with a ratio of 1:3 or 3:1). One or more syntax elements may be used to signal the split type (or split direction and split factor) for the corresponding remainder block of the current block.

Depending on the width and height of the corresponding residual block of the current block, and thus depending on the width and height of the prediction block, the video encoding apparatus may limit the use of certain division directions, division coefficients, and/or division types, and thus it may omit signaling of syntax elements expressing split direction, split ratio, and/or split type depending on the width and height of the remainder block.

In at least one embodiment, when the width and height of the residual block are less than 16 pixels, the residual block cannot be allowed to be divided in a "1:3 or 3:1 ratio". Accordingly, when the width and height of the residual block are less than 16 pixels, the video decoding apparatus may infer that the division ratio may be 1:1, which may be taken into account by the video encoding apparatus to skip syntax element signaling (for example, a 1-bit flag) , expressing the direction of separation. In another embodiment, when the width of the current block is less than 8 pixels, vertical division of the current block may be prohibited. When the height of the current block is less than 8 pixels, horizontal division of the current block may be prohibited. Accordingly, when the width (or height) of the residual block is less than 8 pixels, the current block can be divided at a 1:1 ratio in the vertical (or horizontal) direction by the video decoding apparatus, which allows the video encoding apparatus to bypass syntax element signaling (e.g., 1-bit flag) expressing the direction of division.

In addition, the video encoding apparatus may explicitly signal a flag to identify a sub-block of two sub-blocks classified according to the division type to be encoded (and thus converted). In some embodiments, a video encoding device may respond when the division type of the current block is a third type or a fourth type, that is, when the current block is asymmetrically divided in the horizontal or vertical direction, to immediately decide that a smaller subblock of two blocks while passing flag signaling to identify the subblock to be encoded.

In steps S1040 to S1050, the video encoding apparatus encodes a residual block or one subblock of a residual block in accordance with the residual encoding mode selected for the corresponding residual block of the current block. When residual block encoding is selected, the video encoding apparatus may transform and quantize the residual block in a transform unit equal in size to the residual block, and entropy encode the quantized transform coefficients. When encoding of only one subblock of the residual block is selected, the video encoding apparatus may perform transform and quantization of the subblock corresponding to the transform position and entropy encode the quantized transform coefficients. The residual signals of the other subblocks corresponding to the untransformed positions are considered to be 0, eliminating the need to transform the coefficient information.

In step S1060, the video encoding apparatus may apply an inverse quantization/inverse transform process to the quantized transform coefficients to recover residual signals corresponding to transform positions, and set all residual signals of other subblocks corresponding to untransformed positions to 0, thereby recovering the residual block for current block. The video encoding apparatus may add the reconstructed residual block with the prediction block to generate a reconstructed current block.

At step S1070, the video encoding apparatus may perform deblocking filtering on at least the reconstructed current block and store the filtered current block in a buffer, for example, in memory 190 in FIG. 1 for use as a reference picture for inter-prediction of incoming blocks in the encoded picture. In particular, the video encoding apparatus may establish a grid of N samples at regular intervals in the horizontal and vertical directions and determine whether to apply deblocking filtering to the current block based on whether the boundary between subblocks of the residual block coincides with the boundary of the grid. The video encoding apparatus can perform deblocking filtering only when the boundary between two residual subblocks in the current block coincides with the grid boundary. In other words, deblocking filtering can only be performed when the 4x4 (or 8x8) deblocking boundary coincides with the boundary between residual subblocks.

In fig. 11 is a flowchart of a method performed by a video decoding apparatus for decoding a current inter-prediction encoded block from a coded video bitstream according to at least one embodiment of the present invention.

In step S1110, a video decoding apparatus such as the entropy decoding unit in FIG. 4 can decode the residual encoding mode information of the corresponding residual block of the current block.

For example, a video decoding apparatus may decode from the bit stream a first flag indicating whether the relevant residual block of the current block or only one subblock of the residual block has been encoded. The first flag having a first value (eg, "0") may indicate that the residual block has been converted, and the first flag having a second value (eg, "1") may indicate that only one subblock of the residual block has been converted.

When the first flag indicates that only one subblock of the residual block has been converted, the video decoding apparatus may further decode from the bitstream one or more syntax elements representing the division type (division direction and/or division coefficient) of the residual block and may determine the division type of the residual block , at least in part, based on the decoded one or more syntax elements. Different types of separation can be considered according to the direction of separation and the separation ratio. For example, the division types may include a first type that provides a horizontal symmetric division into two sub-blocks, a second type that provides a vertical symmetric division into two sub-blocks, a third type that provides a horizontal asymmetric division into two sub-blocks, (for example, with a 1:3 ratio or 3:1), and a fourth type that provides an asymmetrical vertical division into two sub-units (for example, with a ratio of 1:3 or 3:1).

When determining the split type, the width and height of the current block may additionally be taken into account along with information indicating the split type. The use of some split types (or split direction and/or split ratio) may be restricted depending on the width and height of the current block, which may preclude signaling of syntax elements expressing the same split types (or split direction and/or split ratio) .

In at least one embodiment, the device may respond when the width and height of the current block is less than 16 pixels by disallowing the current block to be split in a "1:3 or 3:1 ratio." Accordingly, when the width and height of the residual block are less than 16 pixels, the video decoding apparatus can simply infer that the division ratio is 1:1 without having to decode a syntax element indicating the division direction, such as a 1-bit flag.

According to another embodiment, the device may respond when the width (or height) of the current block is less than 8 pixels by prohibiting the current block from being split in the vertical direction (or horizontal direction). Therefore, when the width of the residual block is less than 8 pixels, the video decoding apparatus can simply infer the dividing direction as the horizontal direction without having to decode the syntax element indicating the dividing direction. In the same way, when the height of the residual block is less than 8 pixels, the video decoding apparatus can simply infer the dividing direction as the vertical direction without having to decode the syntax element indicating the direction.

To identify the subblock from which transform coefficient information is to be decoded, the video decoding apparatus may further decode from the bitstream a syntax element (eg, a 1-bit flag) indicating the position of the encoded subblock, and thus the position of the subblock at which conversion completed. In some embodiments, a video decoding apparatus may respond when the division type of the current block is a third or fourth type by immediately inferring that the smaller subblock of the two blocks has undergone the transformation, skipping decoding the syntax element to indicate the position of the subblock where the transformation was performed. transformation.

In steps S1120 to S1130, the video decoding apparatus can recover a relevant residual block of the current block based on the decoded residual encoding mode information.

When the first flag indicates that the corresponding residual block of the current block has been encoded, the video decoding apparatus may decode transform coefficient information for the residual block from the bit stream to recover the transform coefficients. Thereafter, the video decoding apparatus inversely quantizes the transform coefficients and inversely transforms the quantized transform coefficients in a transform unit equal in size to the residual block, thereby reconstructing the residual block of the current block.

When the first flag indicates that only one subblock of the residual block has been encoded, the video decoding apparatus may decode transform coefficient information for the subblock corresponding to the transform position from the bit stream to recover the transform coefficients. Thereafter, the video decoding apparatus performs inverse quantization and inverse transform transformation coefficients to generate residual signals for a sub-block corresponding to the transform position, and sets all residual signals of the other sub-block corresponding to the untransformed position to 0, thereby restoring the residual block of the current block.

In step S1140, the video decoding apparatus can generate a prediction block by inter-predicting the current block, and can reconstruct the current block by summing the prediction block and the residual block.

In step S1150, the video decoding apparatus may perform deblocking filtering on at least the reconstructed current block and store the filtered current block in a buffer, etc., for example, in memory 470 of FIG. 4, for use as a reference picture for inter-prediction of incoming blocks in a coded picture. The video decoding apparatus may establish a grid of N samples at regular intervals in the horizontal and vertical directions and determine, based on the boundary of the set grid, whether to apply deblocking filtering to the current block. For example, deblocking filtering may be applied to the boundary between the current block and other blocks that have already been decoded when that boundary coincides with a grid boundary. Additionally, deblocking filtering can be applied to the boundary between subblocks of a residual block when that boundary coincides with the grid boundary. A grid in different units can be used depending on whether the current block is a luma block or a chrominance block. For example, the grid can be set in 4x4 units for a luma block, and the grid can be set in 8x8 units for a chrominance block.

It should be understood that the above description presents illustrative embodiments that may be implemented in various other ways. The functions described in some embodiments may be implemented by hardware, software, firmware, and/or a combination thereof. It should also be understood that the functional components described herein are labeled "...block" to emphasize their ability to be implemented independently.

Meanwhile, various methods or functions described herein may be implemented as instructions stored in a non-volatile storage medium that can be read and executed by one or more processors. Non-volatile recording media includes, for example, all types of recording devices in which data is stored in a form readable by a computer system. For example, the nonvolatile storage medium may include, but is not limited to, storage media such as an erasable programmable read-only memory (EPROM), a flash drive, an optical drive, a magnetic hard disk, and a solid-state drive (SSD).

Although exemplary embodiments of the present invention have been described for illustrative purposes, those skilled in the art will appreciate that various modifications, additions, and substitutions are possible without departing from the spirit and scope of the claimed invention. Therefore, exemplary embodiments of the present invention have been described for brevity and clarity. The scope of the technical idea of the present embodiments is not limited to the illustrations. Accordingly, it is clear to an ordinary specialist that the scope of the claimed invention is limited not by the embodiments that were directly described above, but by the claims and their equivalents.

Link to related application

This application claims priority to Korean Patent Application No. 10-2019-0028364, filed March 12, 2019, the disclosure of which is incorporated herein by reference in its entirety.

Claims (77)

1. A video decoding device containing: a decoding unit configured to decoding from the bitstream a flag indicating whether residual signals corresponding to only a partial area of the current block have been encoded, decoding from the bit stream motion information for the current block, and decoding from the bitstream information about transform coefficients for one subblock of two subblocks extracted from the current block to recover transform coefficients when the flag indicates that residual signals corresponding to only a partial area of the current block have been encoded; a prediction block configured to inter-predict a current block based on the motion information to generate a prediction block; an inverse quantization and inverse transform unit configured to perform inverse quantization and inverse transform transformation of transform coefficients in one sub-block for which the transform coefficient information has been decoded to generate a residual block for the current block; an adder configured to add the prediction block and the residual block to restore the current block; And a filtering block configured to set a grid of N samples at equal intervals in the horizontal and vertical directions and perform deblocking filtering at the boundary between two subblocks in the current block, which coincides with the grid boundary, wherein, when the current block is a luminance block, the horizontal transform kernel and the vertical transform kernel for the one sub-block are determined based on the relative position of the one sub-block in the current block. 2. The apparatus of claim 1, wherein the inverse quantization and deconversion block is further configured to generate a residual block for the current block by performing inverse quantization and deconversion of transform coefficients in one subblock for which transform coefficient information has been decoded to generate residuals. signals for one sub-block of two sub-blocks, and by setting the residual signals for the other sub-block of two sub-blocks to zero. 3. The device according to claim 1, in which the decoding unit is additionally configured to decoding from the bit stream one or more syntax elements indicating the division type of the current block, determining a division type for the current block to be divided into two subblocks, depending on one or more syntax elements indicating the division type of the current block, and the width and height of the current block, and decoding transform coefficient information corresponding to only one subblock of the two subblocks, which is determined according to the division type. 4. The device according to claim 3, in which the type of division for the current block includes: the first type, in which the current block is symmetrically divided in the horizontal direction; the second type, in which the current block is symmetrically divided in the vertical direction; the third type, in which the current block is asymmetrically divided in the horizontal direction; And the fourth type, in which the current block is asymmetrically divided in the vertical direction. 5. The apparatus of claim 4, wherein the decoding unit is further configured to decode transform coefficient information only for a relatively small subblock of two subblocks when the division type is a third type or a fourth type. 6. The device according to claim 1, additionally containing: a residual filtering unit configured to filter the boundary between two subblocks in a residual block generated by the inverse quantization and inverse transform unit. 7. The device according to claim 1, wherein the decoding block is further configured to decode the differential quantization parameter for the current block, and wherein the inverse quantization and inverse transformation block is additionally configured to obtaining a quantization parameter prediction value from a quantization parameter of at least one neighboring block adjacent to one subblock for which the transform coefficient information has been decoded, and performing inverse quantization using the quantization parameter prediction value and the differential quantization parameter for the transform coefficients in one subblock for which the transform coefficient information has been decoded. 8. A non-volatile computer-readable medium storing instructions readable by a processor to perform a decoding method, the method comprising: decoding from the bitstream motion information for the current block, decoding from the bitstream a flag indicating whether residual signals corresponding to only a partial area of the current block have been encoded, recovering transform coefficients by decoding from the bitstream information about transform coefficients for one sub-block of two sub-blocks extracted from the current block, when the flag indicates that residual signals corresponding to only a partial area of the current block have been encoded; inter-predicting the current block based on the motion information to generate a prediction block; generating a residual block for the current block by performing inverse quantization and inverse transformation of the transform coefficients in one sub-block for which the transform coefficient information has been decoded; restoring the current block by adding the prediction block and the residual block; And setting a grid of N samples at equal intervals in the horizontal and vertical directions and performing deblocking filtering on the boundary between two subblocks in the current block that coincides with the grid boundary, wherein, when the current block is a luminance block, the horizontal transform kernel and the vertical transform kernel for the one sub-block are determined based on the relative position of the one sub-block in the current block. 9. Non-volatile media, made readable by a computer, according to claim 8, in which the restoration of conversion coefficients includes: generating a residual block for the current block by performing inverse quantization and inverse transformation of transform coefficients in one subblock for which transform coefficient information has been decoded to generate residual signals for one subblock of two subblocks, and by setting the residual signals of another subblock of two subblocks equal zero. 10. Non-volatile media, computer-readable, according to claim 8, in which the restoration of conversion coefficients includes: decoding from the bitstream one or more syntax elements indicating the division type of the current block, determining a division type for the current block to be divided into two subblocks, depending on one or more syntax elements indicating the division type of the current block and the width and height of the current block, and decoding transform coefficient information corresponding to only one subblock of the two subblocks, which is determined according to the division type. 11. The computer-readable non-volatile medium of claim 10, wherein the partition type for the current block includes: the first type, in which the current block is symmetrically divided in the horizontal direction; the second type, in which the current block is symmetrically divided in the vertical direction; the third type, in which the current block is asymmetrically divided in the horizontal direction; And the fourth type, in which the current block is asymmetrically divided in the vertical direction. 12. A non-volatile medium, computer-readable, according to claim 11, in which decoding information about conversion coefficients includes: decoding transform coefficient information only for a relatively small subblock of two subblocks when the division type is the third type or the fourth type. 13. A method for encoding video images, including: encoding into a bitstream motion information for the current block; inter-predicting the current block based on the motion information to generate a prediction block; generating a residual block for the current block based on the prediction block; encoding into the bitstream a flag indicating that residual signals corresponding to only a partial area of the current block are encoded; a residual block encoding for the current block, wherein said residual block encoding includes: performing transformation of the residual signals in one sub-block of the two sub-blocks extracted from the current block to generate transform coefficients for one sub-block; quantizing the transform coefficients for one sub-block to generate quantized transform coefficients for one sub-block; And encoding quantized conversion coefficients into a bit stream, recovering a residual block for the current block by performing inverse quantization and inverse transformation of the quantized transform coefficients for one subblock; And restoring the current block by adding the prediction block and the reconstructed residual block; And setting up a grid of N samples at equal intervals in the horizontal and vertical directions and performing deblocking filtering on the boundary between two subblocks in the current block that coincides with the grid boundary, wherein, when the current block is a luminance block, the horizontal transform kernel and the vertical transform kernel for the one sub-block are determined based on the relative position of the one sub-block in the current block. 14. A method for storing a bitstream of encoded video data, the method comprising: encoding the video data into a bitstream by performing an encoding process; and storing the bit stream in a non-volatile computer readable medium, the encoding process includes: encoding into a bitstream motion information for the current block; inter-predicting the current block based on the motion information to generate a prediction block; generating a residual block for the current block based on the prediction block; encoding into the bitstream a flag indicating that residual signals corresponding to only a partial area of the current block are encoded; a residual block encoding for the current block, wherein said residual block encoding includes: performing transformation of the residual signals in one sub-block of the two sub-blocks extracted from the current block to generate transform coefficients for one sub-block; quantizing the transform coefficients for one sub-block to generate quantized transform coefficients for one sub-block; And encoding quantized conversion coefficients into a bit stream, recovering a residual block for the current block by performing inverse quantization and inverse transformation of the quantized transform coefficients for one subblock; And restoring the current block by adding the prediction block and the reconstructed residual block; And setting up a grid of N samples at equal intervals in the horizontal and vertical directions and performing deblocking filtering on the boundary between two subblocks in the current block that coincides with the grid boundary, wherein, when the current block is a luminance block, the horizontal transform kernel and the vertical transform kernel for the one sub-block are determined based on the relative position of the one sub-block in the current block.