KR20240049203A - Method, apparatus and recording medium for encoding/decoding image - Google Patents

Abstract

A method, device, and recording medium for video encoding/decoding are disclosed. The method for video encoding/decoding includes the following steps: determining a template for a target block, selecting a template matching reference image, using a template to search for a template matching optimal block within the template matching reference image, and template matching optimal block. It includes the step of performing encoding/decoding using. Prediction for the target block may be performed using the template matching optimal block. The recording medium may store computer-executable code for an image decoding method and may store information generated by the image encoding method.

Description

Method, device and recording medium for image encoding/decoding {METHOD, APPARATUS AND RECORDING MEDIUM FOR ENCODING/DECODING IMAGE}

The present invention relates to a method, device, and recording medium for video encoding/decoding. Specifically, the present invention discloses a method, apparatus, and recording medium for video encoding/decoding using prediction.

Through the continued development of the information and communications industry, broadcasting services with HD (High Definition) resolution have spread globally. Through this proliferation, many users have become accustomed to high-resolution, high-definition images and/or videos.

In order to satisfy users' demand for high image quality, many organizations are accelerating the development of next-generation imaging devices. User interest in not only High Definition TV (HDTV) and Full HD (FHD) TV, but also Ultra High Definition (UHD) TV, which has a resolution more than four times that of FHD TV. has increased, and with this increase in interest, image encoding/decoding technology for images with higher resolution and image quality is required.

As video compression technology, there are various technologies such as inter prediction technology, intra prediction technology, transformation and quantization technology, and entropy coding technology.

Inter prediction technology is a technology that predicts the value of a pixel included in the current picture using pictures before and/or after the current picture. Intra prediction technology is a technology that predicts the value of a pixel included in the current picture using information about the pixel in the current picture. Transformation and quantization technology is a technology for compressing the energy of the residual image. Entropy coding technology is a technology that assigns short codes to values with a high frequency of occurrence and long codes to values with a low frequency of occurrence.

Using this video compression technology, video data can be effectively compressed, transmitted, and stored.

One embodiment may provide an apparatus, method, and recording medium for encoding/decoding an image using template matching.

One embodiment may provide an apparatus, method, and recording medium for encoding/decoding an image using prediction.

On one side, determining a template for a target block; selecting a template matching reference image; Searching for an optimal template matching block within the template matching reference image using the template; and performing decoding using the template matching optimal block.

Subsampling may be used in generating the template.

Among the plurality of template matching reference image construction methods, a plurality of template matching reference image construction methods may be selected to select the template matching reference image.

One search area among search areas having different shapes may be selected as a search area for deriving the optimal template matching block.

A template matching search starting point in the search area for deriving the optimal template matching block may be set.

Template matching-based intra residual signal prediction may be performed for the decoding.

For the decoding, blending prediction of inter and template matching can be used.

On the other hand, determining a template for a target block; selecting a template matching reference image; Searching for an optimal template matching block within the template matching reference image using the template; and performing encoding using the template matching optimal block.

Subsampling may be used in generating the template.

Among the plurality of template matching reference image construction methods, a plurality of template matching reference image construction methods may be selected to select the template matching reference image.

One search area among search areas having different shapes may be selected as a search area for deriving the optimal template matching block.

A template matching search starting point in the search area for deriving the optimal template matching block may be set.

Template matching-based intra residual signal prediction may be performed for the encoding.

For the above encoding, blending prediction of inter and template matching can be used.

A computer-readable recording medium that stores a bitstream generated by the video encoding method may be provided.

On another side, in a computer-readable recording medium storing a bitstream for image decoding, the bitstream includes coding information, a template for a target block is determined, and a template matching reference image is selected. A template matching optimal block is searched within the template matching reference image using the template, decoding is performed using the template matching optimal block, and the coding information is used to determine the template, select the reference image, A computer-readable recording medium used for at least one of deriving the template matching optimal block and decoding may be provided.

Subsampling may be used in generating the template.

Among the plurality of template matching reference image construction methods, a plurality of template matching reference image construction methods may be selected to select the template matching reference image.

One search area among search areas having different shapes may be selected as a search area for deriving the optimal template matching block.

A template matching search starting point in the search area for deriving the optimal template matching block may be set.

An apparatus, method, and recording medium for encoding/decoding an image using template matching are provided.

An apparatus, method, and recording medium for encoding/decoding an image using prediction are provided.

1 is a block diagram showing the configuration of an encoding device to which the present invention is applied according to an embodiment.
Figure 2 is a block diagram showing the configuration of a decoding device according to an embodiment to which the present invention is applied.
Figure 3 is a diagram schematically showing the division structure of an image when encoding and decoding an image.
Figure 4 is a diagram showing the form of a prediction unit that a coding unit can include.
Figure 5 is a diagram showing the form of a conversion unit that can be included in a coding unit.
Figure 6 shows division of a block according to an example.
Figure 7 is a diagram for explaining an embodiment of the intra prediction process.
Figure 8 is a diagram to explain a reference sample used in the intra prediction process.
Figure 9 is a diagram for explaining an embodiment of the inter prediction process.
Figure 10 shows spatial candidates according to an example.
Figure 11 shows the order of adding motion information of spatial candidates to a merge list according to an example.
Figure 12 explains the process of conversion and quantization according to an example.
13 shows diagonal scanning according to an example.
14 shows horizontal scanning according to an example.
15 shows vertical scanning according to one example.
Figure 16 is a structural diagram of an encoding device according to an embodiment.
Figure 17 is a structural diagram of a decoding device according to an embodiment.
Figure 18 is a flowchart of a method for predicting a target block and generating a bitstream according to an embodiment.
Figure 19 is a flowchart of a method for predicting a target block using a bitstream according to an embodiment.
20 is a flowchart of a prediction method according to an example.
Figure 21 shows areas of a target block according to an example.
Figure 22 shows template creation using some pixels according to an example.
Figure 23 shows template creation using multiple pixels according to an example.
Figure 24 illustrates template creation using multiple pixels according to an example.
Figure 25 shows template creation using multiple pixels according to one example.
Figure 26 shows template creation in which different starting position pixels and subsampling are used for lines according to an example.
Figure 27 shows target blocks, regions, and CTU boundaries according to an example.
Figure 28 shows areas of a target block according to an example.
FIGS. 29A to 29G show a plurality of templates of different target blocks as an example.
Figure 29a shows the configuration of a first template according to an example.
Figure 29b shows the configuration of a second template according to an example.
Figure 29c shows the configuration of a third template according to an example.
Figure 29D shows the configuration of a fourth template according to an example.
Figure 29e shows the configuration of a fifth template according to an example.
Figure 29f shows the configuration of a sixth template according to an example.
Figure 29g shows the configuration of a seventh template according to an example.
Figure 30 shows samples of areas of a target block according to an example.
31A shows samples for a first template according to an example.
31B shows samples for a second template according to one example.
Figure 31C shows samples for a third template according to an example.
Figure 32 is a table showing indices indicating templates according to an example.
Figure 33 shows the creation of a template using padding according to an example.
Figure 34a shows a template before flipping is applied according to an example.
Figure 34b shows a template after flipping is applied according to an example.
FIG. 35 may show a unique index of each candidate template matching reference image belonging to a template matching reference image list according to an example.
FIG. 36 may show a unique index of each candidate template matching reference image belonging to a template matching reference image list that does not include the target image according to an example.
FIG. 37A may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L0 according to an example.
FIG. 37B may show a unique index of each candidate template matching reference image belonging to the template matching reference image list L1 according to an example.
FIG. 38A may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L0 that does not include the target image according to an example.
FIG. 38B may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L1 that does not include the target image according to an example.
Figure 39 shows derivation of an optimal template matching block according to an example.
Figure 40 shows a first method of deriving a template matching motion vector according to an example.
Figure 41 shows a second method of deriving a template matching motion vector according to an example.
Figure 42 shows a first method of deriving a motion vector when a template matching reference image and a target image are different from each other, according to an example.
Figure 43 shows a second method of deriving a motion vector when a template matching reference image and a target image are different from each other, according to an example.
Figure 44 shows a case where the shape of the template matching search area in the target image is rectangular, according to an example.
Figure 45 shows a case where the shape of the template matching search area in a decoded past image or a decoded future image according to an example is rectangular.
Figure 46 shows that when a template matching reference image according to an example is the target image, some areas within the template matching search area within the template matching reference image are excluded from the search.
FIG. 47 illustrates that when the template matching reference image according to an example is a decoded past image or a decoded future image, a portion of an area within the template matching search area within the template matching reference image is excluded from the search.
Figure 48 shows a first method of defining the size of a template matching search area when the shape of the template matching search area is rectangular according to an example.
Figure 49 shows a second method of defining the size of the template matching search area when the shape of the template matching search area is rectangular, according to an example.
Figure 50 shows a case where the shape of the template matching search area in the target image according to an example is a diamond.
Figure 51 shows a case where the shape of the template matching search area in a decoded past image or a decoded future image according to an example is a diamond.
Figure 52 illustrates that when a template matching reference image according to an example is the target image, some areas within the template matching search area within the template matching reference image are excluded from the search.
FIG. 53 illustrates that when the template matching reference image according to an example is a decoded past image or a decoded future image, a portion of an area within the template matching search area within the template matching reference image is excluded from the search.
Figure 54 illustrates a first method of defining the size of a template matching search area when the shape of the template matching search area is a diamond according to an example.
Figure 55 shows a second method of defining the size of the template matching search area when the shape of the template matching search area is a diamond according to an example.
Figure 56 illustrates determination of a template matching search start pixel using corresponding coordinates of a template of a target block according to an example.
Figure 57 is a table showing indices indicating a method of initializing a template matching search start pixel according to an example.
FIG. 58A shows a target block, template matching optimal block, and intra prediction modes when the target image is a template matching reference image according to an example.
Figure 58b shows a target block, template matching optimal block, and intra prediction modes when the target image and the template matching reference image are different according to an example.
Figure 59 shows intra residual signal prediction based on limited template matching for a first specific region according to an example.
Figure 60 shows intra residual signal prediction based on limited template matching for a second specific region according to an example.
Figure 61 shows intra residual signal prediction based on limited template matching for a third specific region according to an example.
Figure 62 shows intra residual signal prediction based on limited template matching for a fourth specific region according to an example.
FIG. 63A shows a target block, a template of the target block, and pixels of the template according to an example.
FIG. 63B shows a template matching optimal block, a template of a target block, and pixels of the template in an example.
Figure 64A shows a target block, a template of the target block, and subsampled pixels of the template, according to an example.
Figure 64B shows a template matching optimal block, a template of a target block, and subsampled pixels of the template in an example.
Figure 65 shows representative values for a template of a target block according to an example.
Figure 66 discloses information for predicting intra and TM blending according to an example.
Figure 67 illustrates a template matching reference image, a target image, and an inter prediction reference image for blending prediction of inter and template matching according to an example.
Figure 68 shows syntax for template matching according to an example.
Figure 69 shows adjacent half-pel positions in eight directions according to one example.
Figure 70 shows a first binarization of a precision index according to an example.
Figure 71 shows a second binarization of a precision index according to an example.
Figure 72 shows the spatial portion of a filter according to one example.
Figure 73 shows search areas used to derive filter coefficients according to an example.
Figure 74 shows the procedure of IntraTMP synthesis according to one example.
Figure 75 shows a search area of IntraTMP according to an example.
Figure 76 shows an expanded search area of IntraTMP according to an example.
Figure 77 shows a limited search area of IntraTMP according to an example.
Figure 78a shows an IBC search area when the CTU size is 128 according to an example.
Figure 78b shows an IBC search area when the CTU size is 256 according to an example.
Figure 79 shows an IntraTMP search area when the block size is 64×64 according to an example.
Figure 80 shows code for limiting the search range according to an example.
Figure 81 shows IntraTMP search area and IBC search area with two different sampling rates according to an example.

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all changes, equivalents, and substitutes included in the spirit and technical scope of the present invention.

For a detailed description of the exemplary embodiments described below, refer to the accompanying drawings, which illustrate specific embodiments by way of example. These embodiments are described in sufficient detail to enable those skilled in the art to practice the embodiments. It should be understood that the various embodiments are different from one another but are not necessarily mutually exclusive. For example, specific shapes, structures and characteristics described herein with respect to one embodiment may be implemented in other embodiments without departing from the spirit and scope of the invention. Additionally, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the embodiment. Accordingly, the detailed description that follows is not to be taken in a limiting sense, and the scope of the exemplary embodiments, if properly described, is limited only by the appended claims, together with all equivalents to what those claims assert.

Similar reference numbers in the drawings refer to identical or similar functions across various aspects. The shapes and sizes of elements in the drawings may be exaggerated for clearer explanation.

In the present invention, terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, a first component may be named a second component without departing from the scope of the present invention, and similarly, the second component may also be named a first component. The term “and/or” may include any of a plurality of related stated items or a combination of a plurality of related stated items.

When a component is said to be "connected" or "connected" to another component, the two components may be directly connected or connected to each other, but It should be understood that other components may exist in the middle of the components. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that no other component exists in between the two components. something to do.

Components appearing in the embodiments are shown independently to represent different characteristic functions, and do not mean that each component consists of separate hardware or a single software component. In other words, each component is listed and included as separate components for convenience of explanation, and at least two of each component are combined to form one component, or one component is divided into multiple components to function. It can be performed, and integrated and separate embodiments of each of these components are included in the scope of the present invention as long as they do not deviate from the essence of the present invention.

The terms used in the examples are only used to describe specific examples and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In embodiments, terms such as “comprise” or “have” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are intended to indicate the presence of one or more other features, numbers, steps, operations, components, parts, or combinations thereof. It should be understood that this does not exclude in advance the presence or addition of features, numbers, steps, operations, components, parts, or combinations thereof. In other words, the description of “including” a specific configuration in the embodiments does not exclude configurations other than the configuration, and means that additional configurations may also be included in the practice of the present invention or the scope of the technical idea of the present invention. .

In embodiments, the term “at least one” may mean one of one or more numbers, such as 1, 2, 3, and 4. In embodiments, the term “a plurality of” may mean one of two or more numbers, such as 2, 3, and 4.

Some of the components of the embodiments may not be essential components that perform essential functions in the present invention, but may simply be optional components to improve performance. Embodiments may be implemented by including only components essential for implementing the essence of the embodiments, excluding components used only to improve performance. Structures that include only essential components excluding optional components used to improve performance are also included in the scope of the embodiments.

Hereinafter, embodiments will be described in detail with reference to the attached drawings so that those skilled in the art can easily implement the embodiments. In describing the embodiments, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present specification, the detailed description will be omitted. In addition, the same reference numerals are used for the same components in the drawings, and duplicate descriptions of the same components are omitted.

Hereinafter, an image may refer to a single picture constituting a video, and may also represent the video itself. For example, “encoding and/or decoding of an image” may mean “encoding and/or decoding of a video,” and may mean “encoding and/or decoding of one of the images that constitute a video.” It may be possible.

Hereinafter, the terms “video” and “motion picture(s)” may be used with the same meaning and may be used interchangeably.

Hereinafter, the target image may be an encoding target image that is the target of encoding and/or a decoding target image that is the target of decoding. Additionally, the target image may be an input image input to an encoding device or may be an input image input to a decoding device. Additionally, the target video may be a current video that is currently subject to encoding and/or decoding. For example, the terms “target image” and “current image” may be used interchangeably and may have the same meaning.

Hereinafter, the terms “image”, “picture”, “frame” and “screen” may be used with the same meaning and may be used interchangeably.

Hereinafter, the target block may be an encoding target block that is the target of encoding and/or a decoding target block that is the target of decoding. Additionally, the target block may be a current block that is currently the target of encoding and/or decoding. For example, the terms “target block” and “current block” may be used interchangeably and may be used interchangeably. The current block may mean an encoding target block that is the target of encoding during encoding and/or a decoding target block that is the target of decoding during decoding. Additionally, the current block may be at least one of a coding block, a prediction block, a residual block, and a transform block.

Hereinafter, the terms “block” and “unit” may be used with the same meaning and may be used interchangeably. Alternatively, “block” may refer to a specific unit.

Hereinafter, the terms “region” and “segment” may be used interchangeably.

In embodiments, each of the specified information, data, flag, index, element, attribute, etc. may have a value. The value "0" of information, data, flags, indexes, elements, and attributes may represent false, logical false, or a first predefined value. That is, the values “0”, false, logical false and the first predefined value can be used interchangeably. The value "1" in information, data, flags, indexes, elements, and attributes may represent true, logical true, or a second predefined value. That is, the value “1”, true, logical true and the second predefined value can be used interchangeably.

When a variable such as i or j is used to represent a row, column, or index, the value of i may be an integer greater than or equal to 0, or an integer greater than or equal to 1. That is, in embodiments rows, columns, indices, etc. may be counted from 0, and may be counted from 1.

In embodiments, the term “one or more” or the term “at least one” may mean the term “plural.” “One or more” or “at least one” can be used interchangeably with “plural.”

Below, terms used in the embodiments are explained.

Encoder: An encoder may refer to a device that performs encoding. In other words, an encoder may mean an encoding device.

Decoder: A decoder may refer to a device that performs decoding. In other words, a decoder may mean a decryption device.

Unit: A unit may represent a unit of video encoding and/or decoding. The terms “unit” and “block” may be used interchangeably and may be used interchangeably.

- A unit may be an MxN array of samples. M and N can each be positive integers. A unit can often refer to an array of samples in a two-dimensional form.

- In video encoding and decoding, a unit may be an area created by dividing one video. In other words, a unit may be a specified area within one image. One image may be divided into multiple units. Alternatively, a unit may refer to the divided parts when one image is divided into segmented parts and encoding or decoding is performed on the segmented parts.

- In encoding and decoding of images, predefined processing may be performed on the unit depending on the type of unit.

- Depending on the function, the types of units include Macro Unit, Coding Unit (CU), Prediction Unit (PU), Residual Unit, and Transform Unit (TU), etc. It can be classified as: Or, depending on the function, the unit may be a block, macroblock, Coding Tree Unit, Coding Tree Block, Coding Unit, Coding Block, or Prediction Unit. It may mean (Prediction Unit), Prediction Block, Residual Unit, Residual Block, Transform Unit, Transform Block, etc. For example, the target unit may be at least one of a CU, PU, residual unit, and TU that are the target of encoding and/or decoding.

- A unit may refer to information including a luma component block, a corresponding chroma component block, and a syntax element for each block in order to refer to it separately from a block.

- The size and shape of the unit may vary. Additionally, units may have various sizes and shapes. In particular, the shape of the unit may include geometric shapes that can be expressed in two dimensions, such as squares, rectangles, trapezoids, triangles, and pentagons.

- Additionally, the unit information may include at least one of the unit type, unit size, unit depth, unit encoding order, and unit decoding order. For example, the type of unit may indicate one of CU, PU, residual unit, and TU.

- One unit can be further divided into subunits with a smaller size than the unit.

Depth: Depth may refer to the degree to which a unit is divided. Additionally, the depth of a unit may indicate the level at which the unit(s) exist when the unit(s) are expressed as a tree structure.

- Unit division information may include depth regarding the depth of the unit. Depth may indicate the number and/or extent to which a unit is divided.

- In a tree structure, the root node has the shallowest depth, and the leaf node has the deepest depth. The root node may be the highest node. A leaf node may be the lowest node.

- One unit can be hierarchically divided into a plurality of sub-units while having depth information based on a tree structure. In other words, a unit and a sub-unit created by division of the unit may respectively correspond to a node and a child node of the node. Each divided sub-unit can have depth. Since depth indicates the number and/or extent to which a unit is divided, division information of a sub-unit may include information about the size of the sub-unit.

- In a tree structure, the highest node may correspond to the first undivided unit. The highest node may be referred to as the root node. Additionally, the highest node may have the minimum depth value. At this time, the highest node may have a depth of level 0.

- A node with a depth of level 1 may represent a unit created as the original unit is divided once. A node with a depth of level 2 may represent a unit created by dividing the original unit twice.

- A node with a depth of level n may represent a unit created as the original unit is divided n times.

- A leaf node may be the lowest node and may be a node that cannot be further divided. The depth of the leaf node may be at the maximum level. For example, the predefined value of the maximum level may be 3.

- QT depth can indicate the depth of quad division. BT depth may indicate the depth for binary division. TT depth can indicate depth to strikeout splits.

Sample: A sample may be the base unit that constitutes a block. Samples can be expressed as values from 0 to 2 ^Bd -1 depending on the bit depth (Bd).

- Samples can be pixels or pixel values.

- Hereinafter, the terms “pixel”, “pixel” and “sample” may be used with the same meaning and may be used interchangeably.

Coding Tree Unit (CTU): A CTU may be composed of one luma component (Y) coding tree block and two chroma component (Cb, Cr) coding tree blocks related to the luma component coding tree block. there is. Additionally, CTU may mean including the above blocks and syntax elements for each block of the above blocks.

- Each coding tree unit is composed of sub-units such as coding unit, prediction unit, and transformation unit, such as Quad Tree (QT), Binary Tree (BT), and Ternary Tree (TT). It can be partitioned using one or more partitioning methods. Quad tree may refer to a quarternary tree. Additionally, each coding tree unit may be partitioned using a MultiType Tree (MTT) using one or more partitioning methods.

- CTU can be used as a term to refer to a pixel block, which is a processing unit in the decoding and encoding process of an image, as in segmentation of an input image.

Coding Tree Block (CTB): Coding tree block can be used as a term to refer to any one of Y coding tree block, Cb coding tree block, and Cr coding tree block.

Neighbor block: A neighboring block may refer to a block adjacent to the target block. A neighboring block may also mean a reconstructed neighboring block.

- Hereinafter, the terms “neighboring block” and “adjacent block” may be used with the same meaning and may be used interchangeably.

- A neighbor block may mean a reconstructed neighbor block.

Spatial neighboring block: A spatial neighboring block may be a block spatially adjacent to the target block. Neighboring blocks may include spatial neighboring blocks.

- The target block and spatial neighboring blocks may be included in the target picture.

- A spatial neighboring block may refer to a block bordering the target block or a block located within a predetermined distance from the target block.

- A spatial neighboring block may refer to a block adjacent to the vertex of the target block. Here, the block adjacent to the vertex of the target block may be a block vertically adjacent to a neighboring block horizontally adjacent to the target block, or a block horizontally adjacent to a neighboring block vertically adjacent to the target block.

Temporal neighboring block: A temporal neighboring block may be a block temporally adjacent to the target block. A neighboring block may include temporal neighboring blocks.

- A temporal neighboring block may include a co-located block (col block).

- A call block may be a block in an already reconstructed co-located picture (col picture). The position of the call block within the call picture may correspond to the position of the target block within the target picture. Alternatively, the location of the collocated block in the collocated picture may be the same as the location of the target block in the target picture. A call picture may be a picture included in the reference picture list.

- A temporal neighboring block may be a block temporally adjacent to a spatial neighboring block of the target block.

Prediction mode: The prediction mode may be information indicating the mode used for intra prediction or the mode used for inter prediction.

Prediction unit: A prediction unit may refer to a base unit for prediction such as inter prediction, intra prediction, inter compensation, intra compensation, and motion compensation.

- One prediction unit may be divided into a plurality of partitions or sub-prediction units with smaller sizes. A plurality of partitions may also be a basic unit in performing prediction or compensation. A partition created by dividing a prediction unit may also be a prediction unit.

Prediction unit partition: Prediction unit partition may refer to the form in which the prediction unit is divided.

Reconstructed neighboring unit: The reconstructed neighboring unit may be a unit that has already been decrypted and rebuilt in the neighborhood of the target unit.

- The reconstructed neighboring unit may be a spatially adjacent unit or a temporally adjacent unit to the target unit.

- The reconstructed spatial neighboring unit may be a unit in the target picture and a unit that has already been reconstructed through encoding and/or decoding.

- The reconstructed temporal neighboring unit may be a unit in the reference image and a unit that has already been reconstructed through encoding and/or decoding. The location of the reconstructed temporal neighboring unit within the reference image may be the same as the location of the target unit within the target picture, or may correspond to the location of the target unit within the target picture. Alternatively, the reconstructed temporal neighboring unit may be a neighboring block of the corresponding block in the reference image. Here, the location of the corresponding block within the reference image may correspond to the location of the target block within the target image. Here, that the positions of the blocks correspond may mean that the positions of the blocks are the same, that one block is included in another block, and that one block occupies the specified position of the other block. It could mean doing it.

Sub-picture: A picture can be divided into one or more sub-pictures. A sub-picture may consist of one or more tile rows and one or more tile columns.

- A sub-picture may be an area with a square or rectangular (i.e. non-square) shape within the picture. Additionally, the sub-picture may include one or more CTUs. .

- A sub-picture may be a rectangular area of one or more slices within one picture.

- One sub-picture may include one or more tiles, one or more bricks, and/or one or more slices.

Tiles: A tile can be an area within a picture that has a square or rectangular shape (i.e., a non-square shape).

- A tile may contain one or more CTUs.

- A tile can be split into one or more bricks.

Brick: A brick may refer to one or more CTU rows within a tile.

- A tile can be split into one or more bricks. Each brick may contain one or more CTU rows.

- Tiles that are not divided into two or more can also refer to bricks.

Slice: A slice may contain one or more tiles within a picture. Alternatively, a slice may include one or more bricks within a tile.

- A sub-picture may contain one or more slices that collectively cover a rectangular area within the picture. Accordingly, each sub-picture boundary may always be a slice boundary. Additionally, each vertical sub-picture boundary may always be a vertical tile boundary.

Parameter set: The parameter set may correspond to header information among the structures in the bitstream.

- Parameter sets include Video Parameter Set (VPS), Sequence Parameter Set (SPS), Picture Parameter Set (PPS), Adaptation Parameter Set (APS), and decoding parameters. It may include at least one of a set (Decoding Parameter Set (DPS)), etc.

Information signaled through a parameter set can be applied to pictures referencing the parameter set. For example, information in the VPS can be applied to pictures referencing the VPS. Information in the SPS can be applied to pictures referencing the SPS. Information in the PPS can be applied to pictures referencing the PPS.

A parameter set can refer to a parent parameter set. For example, PPS may refer to SPS. SPS may refer to VPS.

- Additionally, the parameter set may include tile group, slice header information, and tile header information. A tile group may refer to a group including a plurality of tiles. Additionally, the meaning of a tile group may be the same as that of a slice.

Rate-distortion optimization: The coding device uses a combination of coding unit size, prediction mode, prediction unit size, motion information, and transformation unit size to provide high coding efficiency. Distortion optimization can be used.

- The rate-distortion optimization method can calculate the rate-distortion cost of each combination to select the optimal combination among the above combinations. The rate-distortion cost can be calculated using the formula “D+λ*R”. In general, the combination that minimizes the rate-distortion cost according to the formula "D+λ*R" can be selected as the optimal combination in the rate-distortion optimization method.

- D may indicate distortion. D may be the mean square error of the difference values between the original transform coefficients and the reconstructed transform coefficients within the transform unit.

- R can represent a rate. R can represent the bit rate using related context information.

- λ may represent a Lagrangian multiplier. R may include not only coding parameter information such as prediction mode, motion information, and coded block flag, but also bits generated by encoding of transform coefficients.

- The encoding device may perform processes such as inter prediction, intra prediction, transformation, quantization, entropy coding, inverse quantization, and/or inverse transformation to calculate accurate D and R. These processes can greatly increase the complexity of the encoding device.

Bitstream: A bitstream may refer to a string of bits containing encoded image information.

Parsing: Parsing may mean entropy decoding a bitstream to determine the value of a syntax element. Alternatively, parsing may mean entropy decoding itself.

Symbol: May refer to at least one of a syntax element, a coding parameter, and a transform coefficient of an encoding target unit and/or a decoding target unit. Additionally, a symbol may mean the object of entropy encoding or the result of entropy decoding.

Reference picture: A reference picture may refer to an image that a unit refers to for inter prediction or motion compensation. Alternatively, the reference picture may be an image that includes a reference unit referenced by the target unit for inter prediction or motion compensation.

Hereinafter, the terms “reference picture” and “reference image” may be used with the same meaning and may be used interchangeably.

Reference picture list: A reference picture list may be a list containing one or more reference pictures used for inter prediction or motion compensation.

- The types of reference picture lists are List Combined (LC), List 0 (L0), List 1 (L1), List 2 (L2), and List 3 (List 3; L3). ), etc.

- One or more reference picture lists can be used in inter prediction.

Inter prediction indicator: The inter prediction indicator may indicate the direction of inter prediction for the target unit. Inter prediction can be either one-way prediction or two-way prediction. Alternatively, the inter prediction indicator may indicate the number of reference pictures used when generating a prediction unit of the target unit. Alternatively, the inter prediction indicator may mean the number of prediction blocks used for inter prediction or motion compensation for the target unit.

Prediction list utilization flag: The prediction list utilization flag may indicate whether a prediction unit is generated using at least one reference picture in a specific reference picture list.

- An inter prediction indicator can be derived using the prediction list utilization flag. Conversely, the prediction list utilization flag can be derived using the inter prediction indicator. For example, when the prediction list utilization flag indicates the first value of 0, it may indicate that a prediction block is not generated using a reference picture in the reference picture list for the target unit. When the prediction list utilization flag indicates a second value of 1, it may indicate that a prediction unit is generated using a reference picture list for the target unit.

Reference picture index: The reference picture index may be an index that indicates a specific reference picture in the reference picture list.

Picture Order Count (POC): The POC of a picture may indicate the display order of the picture.

Motion Vector (MV): A motion vector may be a two-dimensional vector used in inter prediction or motion compensation. A motion vector may mean an offset between a target image and a reference image.

- For example, MV can be expressed in the form (mv _x , mv _y ). mv _x can represent the horizontal component, and mv _y can represent the vertical component.

Search range: The search range may be a two-dimensional area where a search for MV is performed during inter prediction. For example, the size of the search area may be MxN. M and N can each be positive integers.

Motion vector candidate: A motion vector candidate may refer to a block that is a prediction candidate or a motion vector of a block that is a prediction candidate when predicting a motion vector.

- The motion vector candidate may be included in the motion vector candidate list.

Motion vector candidate list: The motion vector candidate list may refer to a list constructed using one or more motion vector candidates.

Motion vector candidate index: The motion vector candidate index may refer to an indicator indicating a motion vector candidate in the motion vector candidate list. Alternatively, the motion vector candidate index may be the index of a motion vector predictor.

Motion information: Motion information includes motion vectors, reference picture indices, and inter prediction indicators, as well as reference picture list information, reference pictures, motion vector candidates, motion vector candidate indices, merge candidates, and merge indices. It may mean information containing at least one of the following.

Merge candidate list: The merge candidate list may refer to a list constructed using one or more merge candidates.

Merge candidates: Merge candidates include spatial merge candidates, temporal merge candidates, combined merge candidates, combined bi-prediction merge candidates, candidates based on history, candidates based on the average of two candidates, and zero. It may mean a merge candidate, etc. The merge candidate may include an inter prediction indicator and may include motion information such as a reference picture index for each list, a motion vector, a prediction list utilization flag, and an inter prediction indicator.

Merge index: The merge index may be an indicator pointing to a merge candidate in the merge candidate list.

- The merge index may indicate the reconstructed unit that derived the merge candidate among the reconstructed units that are spatially adjacent to the target unit and the reconstructed units that are temporally adjacent to the target unit.

- The merge index may indicate at least one of the motion information of the merge candidate.

Transform unit: A transform unit may be a basic unit in residual signal coding and/or residual signal decoding, such as transform, inverse transform, quantization, inverse quantization, transform coefficient coding, and transform coefficient decoding. One transformation unit may be divided into a plurality of sub-transformation units with smaller sizes. Here, the transformation may include one or more of a first-order transformation and a second-order transformation, and the inverse transformation may include one or more of a first-order inversion and a second-order inversion.

Scaling: Scaling may refer to the process of multiplying the transform coefficient level by a factor.

- As a result of scaling to the transform coefficient level, a transform coefficient can be generated. Scaling may also be referred to as dequantization.

Quantization Parameter (QP): A quantization parameter may refer to a value used when generating a transform coefficient level for a transform coefficient in quantization. Alternatively, the quantization parameter may refer to a value used when generating a transform coefficient by scaling the transform coefficient level in dequantization. Alternatively, the quantization parameter may be a value mapped to the quantization step size.

Delta quantization parameter: The delta quantization parameter may mean the difference between the predicted quantization parameter and the quantization parameter of the target unit.

Scan: Scan can refer to a method of sorting the order of coefficients within a unit, block, or matrix. For example, sorting a two-dimensional array into a one-dimensional array can be called scanning. Alternatively, arranging a one-dimensional array into a two-dimensional array can also be referred to as a scan or inverse scan.

Transform coefficient: The transform coefficient may be a coefficient value generated as the encoding device performs transformation. Alternatively, the transformation coefficient may be a coefficient value generated as the decoding device performs at least one of entropy decoding and inverse quantization.

- Quantized levels or quantized transform coefficient levels generated by applying quantization to the transform coefficient or residual signal may also be included in the meaning of the transform coefficient.

Quantized level: A quantized level may refer to a value generated by performing quantization on a transform coefficient or residual signal in an encoding device. Alternatively, the quantized level may mean a value that is the target of inverse quantization when performing inverse quantization in a decoding device.

- The quantized transform coefficient level, which is the result of transformation and quantization, can also be included in the meaning of the quantized level.

Non-zero transform coefficient: A non-zero transform coefficient may mean a transform coefficient with a non-zero value or a transform coefficient level with a non-zero value. Alternatively, a non-zero transform coefficient may mean a transform coefficient whose value size is not 0 or a transform coefficient level whose value size is not 0.

Quantization matrix: A quantization matrix may refer to a matrix used in a quantization process or dequantization process to improve the subjective or objective image quality of an image. The quantization matrix may also be referred to as a scaling list.

Quantization matrix coefficient: The quantization matrix coefficient may refer to each element in the quantization matrix. Quantization matrix coefficients may also be referred to as matrix coefficients.

Default matrix: The default matrix may be a quantization matrix predefined in the encoding device and the decoding device.

Non-default matrix: A non-default matrix may be a quantization matrix that is not predefined in the encoding device and the decoding device. A non-default matrix may refer to a quantization matrix signaled by a user from an encoding device to a decoding device.

Most Probable Mode (MPM): MPM may indicate an intra prediction mode that is likely to be used for intra prediction of the target block.

- The encoding device and the decoding device may determine one or more MPMs based on coding parameters related to the target block and properties of entities related to the target block.

- The encoding device and the decoding device may determine one or more MPMs based on the intra prediction mode of the reference block. There may be multiple reference blocks. The plurality of reference blocks may include a spatial neighboring block adjacent to the left of the target block and a spatial neighboring block adjacent to the top of the target block. In other words, one or more different MPMs may be determined depending on which intra prediction modes are used for the reference blocks.

- One or more MPMs can be determined in the same way in the encoding device and the decoding device. In other words, the encoding device and the decoding device may share an MPM list containing the same one or more MPMs.

MPM list: The MPM list may be a list containing one or more MPMs. The number of one or more MPMs in the MPM list may be predefined.

MPM indicator: The MPM indicator may indicate the MPM used for intra prediction of the target block among one or more MPMs in the MPM list. For example, the MPM indicator may be an index to the MPM list.

- Since the MPM list is determined in the same way in the encoding device and the decoding device, the MPM list itself may not need to be transmitted from the encoding device to the decoding device.

- The MPM indicator can be signaled from the encoding device to the decoding device. As the MPM indicator is signaled, the decoding device can determine the MPM to be used for intra prediction for the target block among the MPMs in the MPM list.

MPM usage indicator: The MPM usage indicator may indicate whether the MPM usage mode will be used for prediction of the target block. The MPM use mode may be a mode that uses the MPM list to determine the MPM to be used for intra prediction for the target block.

- The MPM use indicator can be signaled from the encoding device to the decoding device.

Signaling: Signaling may indicate that information is transmitted from an encoding device to a decoding device. Alternatively, signaling may mean that an encoding device includes information in a bitstream or recording medium. Information signaled by the encoding device can be used by the decoding device.

- The encoding device can generate encoded information by performing encoding on the signaled information. Encoded information can be transmitted from an encoding device to a decoding device. The decoding device can obtain information by performing decoding on the transmitted encoded information. Here, the encoding may be entropy encoding, and the decoding may be entropy decoding.

Selective signaling: Information can be signaled selectively. Selective signaling of information may mean that an encoding device selectively includes information (according to certain conditions) within a bitstream or recording medium. Selective signaling of information may mean that the decoding device selectively extracts information from the bitstream (according to specific conditions).

Omission of signaling: Signaling of information may be omitted. Omission of signaling about information may mean that the encoding device does not include the information (depending on certain conditions) in the bitstream or recording medium. Omission of signaling for information may mean that the decoding device does not extract information from the bitstream (according to certain conditions).

Statistical value: Variables, coding parameters, constants, etc. can have values that can be operated on. The statistical value may be a value generated by an operation on the values of these specified objects. For example, a statistical value may be the average value, weighted average value, weighted sum, minimum value, maximum value, mode, etc. for values of specified variables, specified coding parameters, and specified constants. Can be one or more of values, intermediate values, and interpolated values.

1 is a block diagram showing the configuration of an encoding device to which the present invention is applied according to an embodiment.

The encoding device 100 may be an encoder, a video encoding device, or an image encoding device. A video may contain one or more images. The encoding device 100 may sequentially encode one or more images of a video.

Referring to FIG. 1, the encoding device 100 includes an inter prediction unit 110, an intra prediction unit 120, a switch 115, a subtractor 125, a transform unit 130, a quantization unit 140, and an entropy encoding unit. It may include a unit 150, an inverse quantization unit 160, an inverse transform unit 170, an adder 175, a filter unit 180, and a reference picture buffer 190.

The encoding device 100 may perform encoding on the target image using intra mode and/or inter mode. In other words, the prediction mode for the target block may be one of intra mode and inter mode.

Hereinafter, the terms “intra mode”, “intra prediction mode”, “intra-picture mode” and “intra-picture prediction mode” may be used with the same meaning and may be used interchangeably.

Hereinafter, the terms “inter mode”, “inter prediction mode”, “inter-screen mode”, and “inter-screen prediction mode” may be used with the same meaning and may be used interchangeably.

Hereinafter, the term “image” may refer only to a portion of an image or may refer to a block. Additionally, processing of “image” may represent sequential processing of a plurality of blocks.

Additionally, the encoding device 100 can generate a bitstream including encoded information through encoding of a target image, and output and store the generated bitstream. The generated bitstream may be stored in a computer-readable recording medium and streamed through wired and/or wireless transmission media.

As the prediction mode, if intra mode is used, switch 115 can be switched to intra. As the prediction mode, if the inter mode is used, the switch 115 can be switched to inter.

The encoding device 100 may generate a prediction block for the target block. Additionally, after the prediction block is generated, the encoding device 100 may encode the residual block for the target block using the residual of the target block and the prediction block.

When the prediction mode is intra mode, the intra prediction unit 120 may use a pixel of a block that is already encoded and/or decoded, which is a neighbor of the target block, as a reference sample. The intra prediction unit 120 may perform spatial prediction for the target block using a reference sample and generate prediction samples for the target block through spatial prediction. A prediction sample may refer to a sample within a prediction block.

The inter prediction unit 110 may include a motion prediction unit and a motion compensation unit.

When the prediction mode is inter mode, the motion prediction unit can search for the area that best matches the target block from the reference image during the motion prediction process, and use the searched area to derive motion vectors for the target block and the searched area. can do. At this time, the motion prediction unit may use the search area as the area that is the target of the search.

The reference image may be stored in the reference picture buffer 190, and when encoding and/or decoding of the reference image is processed, the encoded and/or decoded reference image may be stored in the reference picture buffer 190.

As the decoded picture is stored, the reference picture buffer 190 may be a decoded picture buffer (DPB).

The motion compensation unit may generate a prediction block for the target block by performing motion compensation using a motion vector. Here, the motion vector may be a two-dimensional vector used for inter prediction. Additionally, the motion vector may indicate an offset between the target image and the reference image.

When the motion prediction unit and the motion compensation unit have a non-integer value, the motion prediction unit and the motion compensation unit may generate a prediction block by applying an interpolation filter to some areas in the reference image. In order to perform inter prediction or motion compensation, the methods of motion prediction and motion compensation of the PU included in the CU based on the CU include skip mode, merge mode, and advanced motion vector prediction. Prediction (AMVP) mode or the current picture reference mode can be determined, and inter prediction or motion compensation can be performed according to each mode.

The subtractor 125 may generate a residual block, which is the difference between the target block and the prediction block. A residual block may also be referred to as a residual signal.

The residual signal may refer to the difference between the original signal and the predicted signal. Alternatively, the residual signal may be a signal generated by transforming or quantizing, or transforming and quantizing, the difference between the original signal and the prediction signal. A residual block may be a residual signal on a block basis.

The transform unit 130 may generate a transform coefficient by performing transformation on the residual block and output the generated transform coefficient. Here, the transformation coefficient may be a coefficient value generated by performing transformation on the residual block.

The conversion unit 130 may use one of a plurality of predefined conversion methods when performing conversion.

The plurality of predefined transformation methods may include Discrete Cosine Transform (DCT), Discrete Sine Transform (DST), and Karhunen-Loeve Transform (KLT) based transformation, etc. there is.

The transformation method used to transform the residual block may be determined according to at least one of coding parameters for the target block and/or the neighboring block. For example, the conversion method may be determined based on at least one of the inter prediction mode for the PU, the intra prediction mode for the PU, the size of the TU, and the shape of the TU. Alternatively, conversion information indicating a conversion method may be signaled from the encoding device 100 to the decoding device 200.

When the transform skip mode is applied, the transform unit 130 may omit transforming the residual block.

A quantized transform coefficient level or quantized level can be generated by applying quantization to the transform coefficient. Hereinafter, in embodiments, the quantized transform coefficient level and the quantized level may also be referred to as transform coefficients.

The quantization unit 140 may generate a quantized transform coefficient level (that is, a quantized level or a quantized coefficient) by quantizing the transform coefficient according to the quantization parameter. The quantization unit 140 may output the generated quantized transform coefficient level. At this time, the quantization unit 140 may quantize the transformation coefficient using a quantization matrix.

The entropy encoding unit 150 may generate a bitstream by performing entropy encoding according to a probability distribution based on the values calculated by the quantization unit 140 and/or coding parameter values calculated during the encoding process. . The entropy encoding unit 150 may output the generated bitstream.

The entropy encoding unit 150 may perform entropy encoding on information about pixels of an image and information for decoding the image. For example, information for decoding an image may include syntax elements, etc.

When entropy coding is applied, a small number of bits may be assigned to symbols with a high probability of occurrence, and a large number of bits may be assigned to symbols with a low probability of occurrence. As symbols are expressed through this allocation, the size of the bitstring for the symbols that are the target of encoding can be reduced. Therefore, the compression performance of video encoding can be improved through entropy coding.

In addition, the entropy encoding unit 150 uses exponential golomb, context-adaptive variable length coding (CAVLC), and context-adaptive binary arithmetic coding for entropy encoding. Coding methods such as Arithmetic Coding (CABAC) can be used. For example, the entropy encoding unit 150 may perform entropy encoding using a Variable Length Coding/Code (VLC) table. For example, the entropy encoding unit 150 may derive a binarization method for the target symbol. Additionally, the entropy encoder 150 can derive a probability model of the target symbol/bin. The entropy encoding unit 150 may perform arithmetic encoding using the derived binarization method, probability model, and context model.

The entropy encoder 150 can change the coefficients of the two-dimensional block form into the form of a one-dimensional vector through a transform coefficient scanning method to encode the quantized transform coefficient level.

Coding parameters may be information required for encoding and/or decoding. The coding parameter may include information encoded in the encoding device 100 and transmitted from the encoding device 100 to the decoding device, and may include information that can be derived during the encoding or decoding process. For example, information transmitted to the decoding device includes syntax elements.

Coding parameters are encoded in an encoding device, such as syntax elements, and may include information derived from the encoding process or decoding process, as well as information (or flags and indexes, etc.) signaled from the encoding device to the decoding device. there is. Additionally, coding parameters may include information required for encoding or decoding an image. For example, the size of the unit/block, the shape of the unit/block, the depth of the unit/block, the division information of the unit/block, the division structure of the unit/block, information indicating whether the unit/block is divided into a quad tree form, Information indicating whether the unit/block is split into a binary tree, the direction of the binary tree split (horizontal or vertical), the split type of the binary tree (symmetric split or asymmetric split), and whether the unit/block is split into a ternary tree. Information indicating whether the unit/block is divided into a ternary tree, the direction of division (horizontal or vertical), the division type of the ternary tree (symmetric division or asymmetric division, etc.), and whether the unit/block is a multi-type tree. Information indicating whether the division is divided in the form of a multi-type tree, combination and direction of the division in the multi-type tree form (horizontal or vertical direction, etc.), division type of the division in the multi-type tree form (symmetric division or asymmetric division), multi-type Splitting tree in the form of a tree (binary tree or ternary tree), type of prediction mode (intra prediction or inter prediction), intra prediction mode/direction, intra luma prediction mode/direction, intra chroma prediction mode/direction, intra splitting information, inter Segmentation information, coding block segmentation flag, prediction block segmentation flag, transformation block segmentation flag, reference sample filtering method, reference sample filter tab (tap), reference sample filter coefficient, prediction block filtering method, prediction block filter tab, prediction block filter coefficient , prediction block boundary filtering method, prediction block boundary filter tab, prediction block boundary filter coefficient, inter prediction mode, motion information, motion vector, motion vector difference, reference picture index, inter prediction direction, inter prediction indicator, prediction list utilization. ) Flag, reference picture list, reference image, POC, motion vector predictor, motion vector prediction index, motion vector prediction candidate, motion vector candidate list, information indicating whether merge mode is used, merge index, merge candidate, merge candidate list , information indicating whether skip mode is used, type of interpolation filter, filter tab of the interpolation filter, filter coefficient of the interpolation filter, motion vector size, motion vector expression accuracy, transformation type, transformation size, and first-order transformation. Information indicating whether to use, information indicating whether to use additional (secondary) transformation, primary transformation selection information (or primary transformation index), secondary transformation selection information (or secondary transformation index), residual Information indicating the presence or absence of a signal, coded block pattern, coded block flag, quantization parameter, residual quantization parameter, quantization matrix, information on intra-loop filter, intra-loop filter Information indicating whether to apply, coefficient of intra-loop filter, filter tab of intra-loop, shape/form of intra-loop filter, information indicating whether to apply deblocking filter, deblocking filter Coefficients, filter tab of deblocking filter, strength of deblocking filter, shape/form of deblocking filter, information indicating whether adaptive sample offset is applied, adaptive sample offset value, adaptive sample offset category, adaptive sample Offset type, information indicating whether to apply an adaptive in-loop filter, coefficients of the adaptive-loop filter, filter tab of the adaptive-loop filter, shape of the adaptive-loop filter/ Shape, binarization/debinarization method, context model, context model determination method, context model update method, information indicating whether regular mode is performed, information indicating whether bypass mode is performed, significant coefficient flag, and finally Significant coefficient flag, coefficient group unit coding flag, last significant coefficient position, flag indicating whether the coefficient value is greater than 1, flag indicating whether the coefficient value is greater than 2, flag indicating whether the coefficient value is greater than 3 , remaining coefficient value information, sign information, reconstructed luma sample, reconstructed chroma sample, context bin, bypass bin, residual luma sample, residual chroma sample, transformation coefficient, luma transformation coefficient, chroma transformation coefficient, Quantized level, luma quantized level, chroma quantized level, transform coefficient level, luma transform coefficient level, chroma transform coefficient level, transform coefficient level scanning method, size of motion vector search area on the side of the decoding device, size of the motion vector search area on the side of the decoding device Shape of motion vector search area on the side, number of motion vector searches on the side of the decoder, CTU size, minimum block size, maximum block size, maximum block depth, minimum block depth, video display/output order, slice identification information , slice type, slice division information, tile group identification information, tile group type, tile group division information, tile identification information, tile type, tile division information, picture type, bit depth, input sample bit depth, reconstructed sample bit depth. , at least one value of residual sample bit depth, transform coefficient bit depth, quantized level bit depth, information about luma signal, information about chroma signal, color space of target block, and color space of residual block, Combined forms or statistics may be included in coding parameters. Additionally, information related to the above-described coding parameters may also be included in the coding parameters. Information used to calculate and/or derive the coding parameters described above may also be included in the coding parameters. Information calculated or derived using the above-described coding parameters may also be included in the coding parameters.

Primary transformation selection information may indicate the primary transformation applied to the target block.

Secondary transformation selection information may indicate secondary transformation applied to the target block.

The residual signal may represent the difference between the original signal and the predicted signal. Alternatively, the residual signal may be a signal generated by transforming the difference between the original signal and the predicted signal. Alternatively, the residual signal may be a signal generated by converting and quantizing the difference between the original signal and the predicted signal. A residual block may be a residual signal for a block.

Here, signaling information may mean that the encoding device 100 includes entropy-encoded information generated by performing entropy encoding on a flag or index in a bitstream. , this may mean that the decoding device 200 obtains information by performing entropy decoding on entropy-encoded information extracted from the bitstream. Here, the information may include flags and indexes.

A signal may refer to signaled information. Hereinafter, information about images and blocks may be referred to as signals. Additionally, hereinafter, the terms “information” and “signal” may be used with the same meaning and may be used interchangeably. For example, a specific signal may be a signal representing a specific block. The original signal may be a signal representing the target block. A prediction signal may be a signal representing a prediction block. The residual signal may be a signal representing a residual block.

The bitstream may include information according to a specified syntax. The encoding device 100 may generate a bitstream including information according to a specified syntax. The encoding device 200 may obtain information from the bitstream according to the specified syntax.

Since encoding through inter prediction is performed by the encoding device 100, the encoded target image can be used as a reference image for other image(s) to be processed later. Accordingly, the encoding device 100 can reconstruct or decode the encoded target image, and store the reconstructed or decoded image as a reference image in the reference picture buffer 190. For decoding, inverse quantization and inverse transformation may be processed on the encoded target image.

The quantized level may be inversely quantized in the inverse quantization unit 160 and inversely transformed in the inverse transformation unit 170. The inverse quantization unit 160 may generate an inverse quantized coefficient by performing inverse quantization on the quantized level. The inverse transform unit 170 may generate inverse quantized and inverse transformed coefficients by performing inverse transformation on the inverse quantized coefficients.

The inverse-quantized and inverse-transformed coefficients can be combined with the prediction block through the adder 175. A reconstructed block can be generated by combining the inverse-quantized and inverse-transformed coefficients with the prediction block. Here, the dequantized and/or inverse-transformed coefficient may mean a coefficient on which at least one of dequantization and inverse-transformation has been performed, and may mean a reconstructed residual block. Here, the reconstructed block may mean a recovered block or a decoded block.

The reconstructed block may pass through the filter unit 180. The filter unit 180 includes at least a deblocking filter, a sample adaptive offset (SAO), an adaptive loop filter (ALF), and a non-local filter (NLF). One or more can be applied to a reconstructed sample, reconstructed block, or reconstructed picture. The filter unit 180 may also be referred to as an in-loop filter.

The deblocking filter can remove block distortion occurring at the boundaries between blocks in the reconstructed picture. To determine whether to apply a deblocking filter, it may be determined whether to apply a deblocking filter to the target block based on the pixel(s) included in a few columns or rows included in the block.

When applying a deblocking filter to a target block, the applied filter may vary depending on the strength of deblocking filtering required. In other words, among different filters, a filter determined according to the strength of deblocking filtering may be applied to the target block. When a deblocking filter is applied to the target block, a long-tap filter, strong filter, weak filter, and Gaussian filter are used depending on the strength of the deblocking filtering required. ) one or more filters may be applied to the target block.

Additionally, when vertical filtering and horizontal filtering are performed on the target block, horizontal filtering and vertical filtering may be processed in parallel.

SAO can add an appropriate offset to the pixel value of a pixel to compensate for coding errors. SAO can perform correction using an offset for the difference between the original image and the deblocked image in pixel units for the image to which deblocking has been applied. In order to perform offset correction on an image, a method is used to divide the pixels included in the image into a certain number of areas, determine the area where offset is to be performed among the divided areas, and apply the offset to the determined area. A method of applying an offset by considering edge information of each pixel of the image may be used.

ALF can perform filtering based on a comparison between the reconstructed image and the original image. After dividing the pixels included in the image into predetermined groups, a filter to be applied to each divided group can be determined, and filtering can be performed differentially for each group. Information related to whether to apply an adaptive loop filter may be signaled for each CU. This information can be signaled for the luma signal. The shape of the ALF and filter coefficients to be applied to each block may be different for each block. Alternatively, regardless of the characteristics of the block, a fixed form of ALF may be applied to the block.

A non-local filter can perform filtering based on reconstructed blocks similar to the target block. An area similar to the target block may be selected from the reconstructed image, and filtering of the target block may be performed using statistical properties of the selected similar area. Information related to whether to apply a non-local filter may be signaled to the CU. Additionally, the shapes and filter coefficients of non-local filters to be applied to blocks may be different depending on the block.

The reconstructed block or reconstructed image that has passed through the filter unit 180 may be stored in the reference picture buffer 190 as a reference picture. The reconstructed block that has passed through the filter unit 180 may be part of a reference picture. In other words, the reference picture may be a reconstructed picture composed of reconstructed blocks that have passed through the filter unit 180. The stored reference picture can then be used for inter prediction or motion compensation.

Figure 2 is a block diagram showing the configuration of a decoding device according to an embodiment to which the present invention is applied.

The decoding device 200 may be a decoder, a video decoding device, or an image decoding device.

Referring to FIG. 2, the decoding device 200 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, an inter prediction unit 250, and a switch 245. , may include an adder 255, a filter unit 260, and a reference picture buffer 270.

The decoding device 200 may receive the bitstream output from the encoding device 100. The decoding device 200 can receive a bitstream stored in a computer-readable recording medium and can receive a bitstream streaming through a wired/wireless transmission medium.

The decoding device 200 may perform intra-mode and/or inter-mode decoding on the bitstream. Additionally, the decoding device 200 can generate a reconstructed image or a decoded image through decoding, and output the generated reconstructed image or a decoded image.

For example, switching to intra mode or inter mode according to the prediction mode used for decoding may be performed by the switch 245. If the prediction mode used for decoding is intra mode, the switch 245 may be switched to intra mode. If the prediction mode used for decoding is the inter mode, the switch 245 may be switched to inter.

The decoding device 200 can obtain a reconstructed residual block by decoding the input bitstream and generate a prediction block. When the reconstructed residual block and the prediction block are obtained, the decoding device 200 can generate a reconstructed block that is the target of decoding by combining the reconstructed residual block and the prediction block.

The entropy decoding unit 210 may generate symbols by performing entropy decoding on the bitstream based on a probability distribution for the bitstream. The generated symbols may include symbols in the form of quantized transform coefficient levels (i.e., quantized levels or quantized coefficients). Here, the entropy decoding method may be similar to the entropy encoding method described above. For example, the entropy decoding method may be the reverse process of the entropy encoding method described above.

The entropy decoder 210 can change coefficients in the form of a one-dimensional vector into a two-dimensional block form through a transform coefficient scanning method in order to decode the quantized transform coefficient level.

For example, by scanning the coefficients of a block using a diagonal scan in the upper right corner, the coefficients can be changed into a two-dimensional block form. Alternatively, which scan to use among the upper right diagonal scan, vertical scan, and horizontal scan may be determined depending on the block size and/or intra prediction mode.

The quantized coefficient may be inverse quantized in the inverse quantization unit 220. The inverse quantization unit 220 may generate an inverse quantized coefficient by performing inverse quantization on the quantized coefficient. Additionally, the inverse quantized coefficient may be inversely transformed in the inverse transform unit 230. The inverse transform unit 230 may generate a reconstructed residual block by performing inverse transform on the inverse quantized coefficients. As a result of performing inverse quantization and inverse transformation on the quantized coefficients, a reconstructed residual block may be generated. At this time, the inverse quantization unit 220 may apply a quantization matrix to the quantized coefficients when generating a reconstructed residual block.

When the intra mode is used, the intra prediction unit 240 may generate a prediction block by performing spatial prediction on the target block using pixel values of already decoded blocks neighboring the target block.

The inter prediction unit 250 may include a motion compensation unit. Alternatively, the inter prediction unit 250 may be called a motion compensation unit.

When the inter mode is used, the motion compensation unit may generate a prediction block by performing motion compensation on the target block using a motion vector and a reference image stored in the reference picture buffer 270.

When the motion vector has a non-integer value, the motion compensation unit can apply an interpolation filter to some areas in the reference image and generate a prediction block using the reference image to which the interpolation filter has been applied. In order to perform motion compensation, the motion compensation unit may determine which of skip mode, merge mode, AMVP mode, and current picture reference mode is the motion compensation method used for the PU included in the CU based on the CU, and the determined mode. Motion compensation can be performed according to .

The reconstructed residual block and prediction block can be added through an adder 255. The adder 255 may generate a reconstructed block by adding the reconstructed residual block and the prediction block.

The reconstructed block may pass through the filter unit 260. The filter unit 260 may apply at least one of a deblocking filter, SAO, ALF, and non-local filter to the reconstructed block or the reconstructed image. The reconstructed image may be a picture containing reconstructed blocks.

The filter unit 260 may output a reconstructed image.

The reconstructed block and/or the reconstructed image that has passed through the filter unit 260 may be stored as a reference picture in the reference picture buffer 270. The reconstructed block that has passed through the filter unit 260 may be part of a reference picture. In other words, the reference picture may be a reconstructed image composed of reconstructed blocks that have passed through the filter unit 260. The stored reference picture can then be used for inter prediction and/or motion compensation.

Figure 3 is a diagram schematically showing the division structure of an image when encoding and decoding an image.

Figure 3 may schematically show an example in which one unit is divided into a plurality of sub-units.

In order to efficiently segment an image, a coding unit (CU) may be used in encoding and decoding. A unit may be a term that refers to a combination of 1) a block containing video samples and 2) a syntax element. For example, “division of a unit” may mean “division of a block corresponding to a unit.”

CU may be used as a base unit for video encoding and/or decoding. Additionally, a CU may be used as a unit to which a selected mode of intra mode and inter mode is applied in video encoding and/or decoding. In other words, in video encoding and/or decoding, it can be determined which mode among intra mode and inter mode will be applied to each CU.

Additionally, a CU may be a basic unit in prediction, transformation, quantization, inverse transformation, inverse quantization, and encoding and/or decoding of transformation coefficients.

Referring to FIG. 3, the image 300 may be sequentially divided into units of largest coding units (LCUs). For each LCU, a partition structure may be determined. Here, LCU may be used with the same meaning as Coding Tree Unit (CTU).

Division of a unit may mean division of a block corresponding to the unit. Block division information may include depth information regarding the depth of the unit. Depth information may indicate the number and/or extent to which a unit is divided. One unit may be hierarchically divided into a plurality of sub-units with depth information based on a tree structure.

Each divided sub-unit may have depth information. Depth information may be information indicating the size of the CU. Depth information may be stored for each CU.

Each CU may have depth information. When a CU is split, CUs created by splitting may have a depth that increases by 1 from the depth of the split CU.

The division structure may refer to the distribution of CUs within the LCU 310 for efficiently encoding images. This distribution may be determined depending on whether to divide one CU into multiple CUs. The number of divided CUs may be a positive integer greater than or equal to 2, including 2, 4, 8, and 16.

The horizontal and vertical sizes of the CU created by division may be smaller than the horizontal and vertical sizes of the CU before division, depending on the number of CUs created by division. For example, the horizontal and vertical sizes of the CU created by division may be half the horizontal size and half the vertical size of the CU before division.

A split CU can be recursively split into multiple CUs in the same manner. By recursive division, at least one of the horizontal and vertical sizes of the divided CU may be reduced compared to at least one of the horizontal and vertical sizes of the CU before division.

Division of the CU can be done recursively up to a predefined depth or predefined size.

For example, the depth of the CU may have a value of 0 to 3. The size of the CU can range from 64x64 to 8x8 depending on the depth of the CU.

For example, the depth of the LCU 310 may be 0, and the depth of the Smallest Coding Unit (SCU) may be a predefined maximum depth. Here, the LCU may be a CU with the maximum coding unit size as described above, and the SCU may be a CU with the minimum coding unit size.

Division may begin from the LCU 310, and the depth of the CU may increase by 1 whenever the horizontal and/or vertical size of the CU is reduced due to division.

For example, for each depth, an undivided CU may have a size of 2Nx2N. Additionally, in the case of a divided CU, a CU of 2Nx2N size may be divided into 4 CUs of NxN size. The size of N can be reduced by half each time the depth increases by 1.

Referring to FIG. 3, an LCU with a depth of 0 may be 64x64 pixels or a 64x64 block. 0 may be the minimum depth. A SCU with a depth of 3 may be 8x8 pixels or an 8x8 block. 3 may be the maximum depth. At this time, the CU of the 64x64 block, which is the LCU, can be expressed as depth 0. A CU in a 32x32 block can be expressed with a depth of 1. A CU in a 16x16 block can be expressed with a depth of 2. A CU of an 8x8 block, which is an SCU, can be expressed with a depth of 3.

Information about whether a CU is divided can be expressed through the division information of the CU. Segmentation information may be 1 bit of information. All CUs except SCU may include segmentation information. For example, the partition information value of a CU that is not divided may be a first value, and the partition information value of a divided CU may be a second value. When the division information indicates whether the CU is divided, the first value may be 0 and the second value may be 1.

For example, when one CU is divided into four CUs, the horizontal and vertical sizes of each of the four CUs created by division are half the horizontal size and half the vertical size of the CU before division, respectively. You can. When a CU of size 32x32 is divided into 4 CUs, the sizes of the 4 divided CUs may be 16x16. When one CU is divided into four CUs, it can be said that the CU is divided into a quad-tree form. In other words, it can be seen that quad-tree partitioning has been applied to the CU.

For example, when one CU is divided into two CUs, the horizontal or vertical size of each CU of the two CUs created by division is half the horizontal size or half the vertical size of the CU before division, respectively. You can. When a CU of size 32x32 is vertically divided into two CUs, the sizes of the two divided CUs may be 16x32. When a CU of size 32x32 is horizontally divided into two CUs, the sizes of the two divided CUs may be 32x16. When one CU is divided into two CUs, it can be said that the CU is divided in a binary-tree form. In other words, it can be seen that binary-tree partitioning has been applied to the CU.

For example, when one CU is divided into three CUs, three divided CUs can be created by dividing the horizontal or vertical size of the CU before division at a ratio of 1:2:1. For example, if a CU with a size of 16x32 is divided into three CUs in the horizontal direction, the three divided CUs may have sizes of 16x8, 16x16, and 16x8, respectively, from the top. For example, if a CU with a size of 32x32 is divided into three CUs in the vertical direction, the three divided CUs may have sizes of 8x32, 16x32, and 8x32, respectively, from the left. When one CU is divided into three CUs, it can be said that the CU is divided in a ternary-tree form. In other words, it can be seen that ternary-tree partitioning has been applied to the CU.

Both quad-tree type partitioning and binary-tree type partitioning were applied to the LCU 310 of FIG. 3.

In the encoding device 100, a Coding Tree Unit (CTU) of 64x64 size may be divided into a plurality of smaller CUs using a recursive Quad-Cree structure. One CU can be divided into four CUs with identical sizes. CUs can be divided recursively, and each CU can have a quad tree structure.

Through recursive partitioning of the CU, the optimal partitioning method that generates the minimum rate-distortion ratio can be selected.

The CTU 320 in FIG. 3 is an example of a CTU to which quad tree partitioning, binary tree partitioning, and ternary tree partitioning are all applied.

As described above, to partition a CTU, at least one of quad tree partitioning, binary tree partitioning, and ternary tree partitioning may be applied to the CTU. Partitions may be applied based on a specified priority.

For example, quad tree partitioning may be applied preferentially for CTU. A CU that can no longer be divided into a quad tree may correspond to a leaf node of the quad tree. The CU corresponding to the leaf node of the quad tree can be the root node of the binary tree and/or ternary tree. That is, the CU corresponding to the leaf node of the quad tree may be divided into a binary tree or a ternary tree, or may not be divided any further. At this time, quad tree division is not applied again to the CU created by applying binary tree division or ternary tree division to the CU corresponding to the leaf node of the quad tree, thereby preventing block division and/or signaling of block division information. It can be performed effectively.

The division of the CU corresponding to each node of the quad tree can be signaled using quad division information. Quad partition information with a first value (eg, “1”) may indicate that the CU is partitioned in a quad tree form. Quad partition information with a second value (eg, “0”) may indicate that the CU is not partitioned in a quad tree form. Quad split information may be a flag with a specified length (eg, 1 bit).

There may be no priority between binary tree partitioning and ternary tree partitioning. That is, the CU corresponding to the leaf node of the quad tree may be divided into a binary tree or a ternary tree. Additionally, the CU generated by binary tree partitioning or ternary tree partitioning may be partitioned again into a binary tree form or a ternary tree form, or may not be partitioned any further.

Partitioning when no priority exists between binary tree partitioning and ternary tree partitioning may be referred to as multi-type tree partitioning. In other words, the CU corresponding to the leaf node of the quad tree can become the root node of the multi-type tree. The division of the CU corresponding to each node of the multi-type tree may be signaled using at least one of information indicating whether the multi-type tree is divided, division direction information, and division tree information. To split the CU corresponding to each node of the multi-type tree, information indicating whether to split sequentially, split direction information, and split tree information may be signaled.

For example, information indicating whether a multi-type tree with a first value (eg, “1”) is split may indicate that the corresponding CU is split in the form of a multi-type tree. Information indicating whether a multi-type tree with a second value (eg, “0”) is divided may indicate that the corresponding CU is not divided into a multi-type tree.

When the CU corresponding to each node of the multi-type tree is split in the form of a multi-type tree, the corresponding CU may further include split direction information.

Splitting direction information may indicate the splitting direction of multi-type tree splitting. Division direction information with a first value (eg, “1”) may indicate that the corresponding CU is divided in the vertical direction. Division direction information with a second value (eg, “0”) may indicate that the corresponding CU is divided in the horizontal direction.

When the CU corresponding to each node of the multi-type tree is split into a multi-type tree, the corresponding CU may further include split tree information. Splitting tree information may indicate the tree used for multi-type tree splitting.

For example, split tree information with a first value (eg, “1”) may indicate that the corresponding CU is split in the form of a binary tree. Split tree information with a second value (eg, “0”) may indicate that the corresponding CU is split in a ternary tree form.

Here, each of the above-described information indicating whether to split, split tree information, and split direction information may be a flag with a specified length (eg, 1 bit).

At least one of the above-described quad split information, information indicating whether the multi-type tree is split, split direction information, and split tree information may be entropy encoded and/or entropy decoded. For entropy encoding/decoding of such information, information on a neighboring CU adjacent to the target CU can be used.

For example, the splitting form of the left CU and/or the upper CU (i.e., whether to split, splitting tree, and/or splitting direction) and the splitting form of the target CU may be considered highly likely to be similar to each other. Therefore, based on information on the neighboring CU, context information for entropy encoding and/or entropy decoding of information on the target CU may be derived. At this time, the information on the neighboring CU may include at least one of the neighboring CU's 1) quad split information, 2) information indicating whether the multi-type tree is split, 3) split direction information, and 4) split tree information.

As another embodiment, among binary tree partitioning and ternary tree partitioning, binary tree partitioning may be performed preferentially. That is, binary tree division is applied first, and the CU corresponding to the leaf node of the binary tree may be set as the root node of the ternary tree. In this case, quad tree division and binary tree division may not be performed on the CU corresponding to the node of the ternary tree.

A CU that is no longer split by quad tree splitting, binary tree splitting, and/or ternary tree splitting may become a unit of encoding, prediction, and/or transformation. That is, for prediction and/or transformation, the CU may no longer be split. Accordingly, a split structure and split information for splitting a CU into prediction units and/or transform units may not exist in the bitstream.

However, if the size of the CU that is the unit of division is larger than the size of the maximum conversion block, this CU may be recursively divided until the size of the CU becomes less than or equal to the size of the maximum conversion block. For example, if the size of the CU is 64x64 and the maximum conversion block size is 32x32, the CU may be divided into four 32x32 blocks for conversion. For example, if the size of the CU is 32x64 and the maximum conversion block size is 32x32, the CU may be divided into two 32x32 blocks for conversion.

In this case, information about whether the CU is divided for conversion may not be signaled separately. Without signaling, whether to split a CU may be determined by comparison between the horizontal size (and/or vertical size) of the CU and the horizontal size (and/or vertical size) of the maximum transform block. For example, if the horizontal size of the CU is larger than the horizontal size of the maximum transformation block, the CU may be divided vertically into two. Additionally, if the vertical size of the CU is larger than the vertical size of the maximum transformation block, the CU may be divided horizontally into two.

Information about the maximum size and/or minimum size of the CU and information about the maximum size and/or minimum size of the transform block may be signaled or determined at a higher level for the CU. For example, higher levels may be sequence level, picture level, tile level, tile group level, and slice level. For example, the minimum size of a CU may be determined to be 4x4. For example, the maximum size of a transform block may be determined to be 64x64. For example, the minimum size of the transform block may be determined to be 4x4.

Information about the minimum size of the CU corresponding to the leaf node of the quad tree (say, the quad tree minimum size) and/or the maximum depth of the path from the root node to the leaf node of the multi-type tree (say, the multi-type tree maximum Information about depth) may be signaled or determined at a higher level for the CU. For example, higher levels may be sequence level, picture level, slice level, tile group level, and tile level. Information about the quad tree minimum size and/or information about the multi-type tree maximum depth may be signaled or determined separately for each of the intra-slice and inter-slice.

Differential information about the size of the CTU and the maximum size of the transform block may be signaled or determined at a higher level for the CU. For example, higher levels may be sequence level, picture level, slice level, tile group level, and tile level. Information about the maximum size of the CU corresponding to each node of the binary tree (that is, the maximum size of the binary tree) can be determined based on the size and difference information of the CTU. The maximum size of the CU corresponding to each node of the ternary tree (that is, the maximum size of the ternary tree) may have different values depending on the type of slice. For example, within an intra slice, the maximum ternary tree size may be 32x32. Additionally, for example, within an inter slice, the maximum ternary tree size may be 128x128. For example, the minimum size of a CU corresponding to each node in a binary tree (say, the binary tree minimum size) and/or the minimum size of a CU corresponding to each node in a ternary tree (say, the ternary tree minimum size) is Can be set to the minimum size.

As another example, the binary tree maximum size and/or the ternary tree maximum size may be signaled or determined at the slice level. Additionally, the binary tree minimum size and/or ternary tree minimum size may be signaled or determined at the slice level.

Based on the various block sizes and various depths described above, quad split information, information indicating whether the multi-type tree is split, split tree information, and/or split direction information may or may not exist in the bitstream.

For example, if the size of the CU is not larger than the quad tree minimum size, the CU may not include quad partition information, and the quad partition information for the CU may be inferred as the second value.

For example, the size (horizontal and vertical size) of the CU corresponding to a node in a multi-type tree is larger than the binary tree maximum size (horizontal size and vertical size) and/or the ternary tree maximum size (horizontal size and vertical size). For larger cases, the CU may not be partitioned into binary and/or ternary tree form. According to this decision method, information indicating whether to split the multi-type tree may not be signaled and may be inferred as a second value.

Alternatively, the size (horizontal and vertical size) of the CU corresponding to the node of the multi-type tree is equal to the minimum size (horizontal and vertical size) of the binary tree, or the size of the CU (horizontal and vertical size) is equal to the minimum size (horizontal and vertical size) of the binary tree. If equal to twice the minimum size (horizontal and vertical sizes), the CU may not be partitioned into binary tree form and/or ternary tree form. According to this decision method, information indicating whether to split the multi-type tree may not be signaled and may be inferred as a second value. This is because, when dividing a CU into a binary tree form and/or a ternary tree form, a CU smaller than the minimum binary tree size and/or the minimum ternary tree size is generated.

Alternatively, binary tree partitioning or ternary tree partitioning may be limited based on the size of the virtual pipeline data unit (i.e., pipeline buffer size). For example, if a CU is split into sub-CUs that do not fit the pipeline buffer size by binary tree partitioning or ternary tree partitioning, binary tree partitioning or ternary tree partitioning may be limited. The pipeline buffer size may be equal to the size of the maximum conversion block (e.g., 64X64).

For example, when the pipeline buffer size is 64X64, the following partitions may be limited.

- ternary tree split for NxM (N and/or M is 128) CUs

- Horizontally directed binary tree split for 128xN (N <= 64) CUs

- Vertically oriented binary tree partitioning for Nx128 (N <= 64) CUs

Alternatively, if the depth within the multi-type tree of the CU corresponding to the node of the multi-type tree is equal to the maximum depth of the multi-type tree, the CU may not be divided into a binary tree form and/or a ternary tree form. According to this decision method, information indicating whether to split the multi-type tree may not be signaled and may be inferred as a second value.

Or, for a CU corresponding to a node of a multi-type tree, a multi-type tree only if at least one of vertical binary tree partitioning, horizontal binary tree partitioning, vertical ternary tree partitioning, and horizontal ternary tree partitioning is possible. Information indicating whether to divide may be signaled. Otherwise, the CU may not be partitioned into binary tree form and/or ternary tree form. According to this decision method, information indicating whether to split the multi-type tree may not be signaled and may be inferred as a second value.

Alternatively, split direction information only if both vertical binary tree splitting and horizontal binary tree splitting are possible for the CU corresponding to the node of the multi-type tree, or both vertical ternary tree splitting and horizontal ternary tree splitting are possible. can be signaled. Otherwise, the division direction information may not be signaled and may be inferred as a value indicating the direction in which the CU can be divided.

Alternatively, split tree information only if both vertical binary tree splitting and vertical ternary tree splitting are possible for the CU corresponding to the node of the multi-type tree, or both horizontal binary tree splitting and horizontal ternary tree splitting are possible. can be signaled. Otherwise, the split tree information may not be signaled and may be inferred as a value indicating a tree applicable to splitting the CU.

Figure 4 is a diagram showing the form of a prediction unit that a coding unit can include.

Among the CUs divided from the LCU, CUs that are no longer divided may be divided into one or more prediction units (PUs).

PU may be the basic unit for prediction. PU can be encoded and decoded in any one of skip mode, inter mode, and intra mode. PU can be divided into various forms depending on each mode. For example, the target block described above with reference to FIG. 1 and the target block described with reference to FIG. 2 may be a PU.

A CU may not be divided into PUs. If the CU is not divided into PUs, the size of the CU and the size of the PU may be the same.

In skip mode, there may be no partitions within the CU. In skip mode, 2Nx2N mode 410 in which the sizes of PU and CU are the same without division can be supported.

In inter mode, eight partition types can be supported within the CU. For example, in inter mode, 2Nx2N mode (410), 2NxN mode (415), Nx2N mode (420), NxN mode (425), 2NxnU mode (430), 2NxnD mode (435), nLx2N mode (440), and nRx2N Mode 445 may be supported.

In intra mode, 2Nx2N mode 410 and NxN mode 425 may be supported.

In 2Nx2N mode 410, a PU with a size of 2Nx2N can be encoded. A PU of size 2Nx2N may mean a PU of the same size as the size of the CU. For example, a PU of size 2Nx2N may have sizes of 64x64, 32x32, 16x16 or 8x8.

In NxN mode 425, PUs of NxN size can be encoded.

For example, in intra prediction, when the size of a PU is 8x8, four divided PUs can be encoded. The size of the divided PU may be 4x4.

When the PU is encoded by intra mode, the PU may be encoded using one intra prediction mode among a plurality of intra prediction modes. For example, High Efficiency Video Coding (HEVC) technology can provide 35 intra prediction modes, and a PU can be encoded with one intra prediction mode among the 35 intra prediction modes.

Which of the 2Nx2N mode 410 and NxN mode 425 will be used to encode the PU can be determined by the rate-distortion cost.

The encoding device 100 can perform an encoding operation on a PU of size 2Nx2N. Here, the encoding operation may be encoding the PU in each of a plurality of intra prediction modes that the encoding device 100 can use. Through encoding operations, the optimal intra prediction mode for a PU of size 2Nx2N can be derived. The optimal intra prediction mode may be an intra prediction mode that generates the minimum rate-distortion cost for encoding a PU of 2Nx2N size among a plurality of intra prediction modes that the encoding device 100 can use.

Additionally, the encoding device 100 may sequentially perform an encoding operation on each PU of the NxN divided PUs. Here, the encoding operation may be encoding the PU in each of a plurality of intra prediction modes that the encoding device 100 can use. The optimal intra prediction mode for a PU of NxN size can be derived through encoding operations. The optimal intra prediction mode may be an intra prediction mode that generates the minimum rate-distortion cost for encoding an NxN sized PU among a plurality of intra prediction modes that the encoding device 100 can use.

The encoding device 100 may determine which of the 2Nx2N sized PUs and the NxN sized PUs to encode based on comparison of the rate-distortion costs of the 2Nx2N sized PU and the rate-distortion costs of the NxN sized PUs.

One CU can be divided into one or more PUs, and a PU can also be divided into multiple PUs.

For example, when one PU is divided into four PUs, the horizontal and vertical sizes of each of the four PUs created by division are half the horizontal size and half the vertical size of the PU before division, respectively. You can. When a PU of size 32x32 is divided into 4 PUs, the sizes of the 4 divided PUs may be 16x16. When one PU is divided into four PUs, it can be said that the PU is divided into a quad-tree form.

For example, when one PU is divided into two PUs, the horizontal or vertical size of each PU of the two PUs created by division is half the horizontal size or half the vertical size of the PU before division, respectively. You can. When a PU of size 32x32 is vertically divided into two PUs, the sizes of the two divided PUs may be 16x32. When a PU of size 32x32 is horizontally divided into two PUs, the sizes of the two divided PUs may be 32x16. When one PU is divided into two PUs, it can be said that the PU is divided into a binary-tree form.

Figure 5 is a diagram showing the form of a conversion unit that can be included in a coding unit.

Transform Unit (TU) may be a basic unit used for the processes of transformation, quantization, inverse transformation, inverse quantization, entropy encoding, and entropy decoding within the CU.

TU may have a square or rectangular shape. The shape of the TU may be determined depending on the size and/or shape of the CU.

Among the CUs divided from the LCU, CUs that are no longer divided into CUs may be divided into one or more TUs. At this time, the division structure of the TU may be a quad-tree structure. For example, as shown in FIG. 5, one CU 510 may be divided one or more times according to a quad-tree structure. Through division, one CU 510 can be composed of TUs of various sizes.

If one CU is divided more than two times, the CU can be viewed as being divided recursively. Through partitioning, one CU can be composed of TUs with various sizes.

Alternatively, one CU may be divided into one or more TUs based on the number of vertical lines and/or horizontal lines dividing the CU.

A CU may be divided into symmetric TUs or may be divided into asymmetric TUs. For division into asymmetric TUs, information about the size and/or shape of the TU may be signaled from the encoding device 100 to the decoding device 200. Alternatively, the size and/or shape of the TU may be derived from information about the size and/or shape of the CU.

A CU may not be divided into TUs. If the CU is not divided into TUs, the size of the CU and the size of the TU may be the same.

One CU may be divided into one or more TUs, and a TU may also be divided into multiple TUs.

For example, when one TU is split into four TUs, the horizontal and vertical sizes of each of the four TUs created by the split are half the horizontal size and half the vertical size of the TU before splitting, respectively. You can. When a TU of size 32x32 is divided into 4 TUs, the sizes of the 4 divided TUs may be 16x16. When one TU is divided into four TUs, it can be said that the TU is divided into a quad-tree form.

For example, when one TU is split into two TUs, the horizontal or vertical size of each TU of the two TUs created by the split is half the horizontal size or half the vertical size of the TU before splitting, respectively. You can. When a TU of size 32x32 is vertically divided into two TUs, the sizes of the two divided TUs may be 16x32. When a TU of size 32x32 is horizontally divided into two TUs, the sizes of the two divided TUs may be 32x16. When one TU is divided into two TUs, it can be said that the TU is divided into a binary-tree form.

The CU may be divided in a manner other than that shown in FIG. 5.

For example, one CU can be divided into three CUs. The horizontal or vertical size of the three divided CUs may be 1/4, 1/2, and 1/4 of the horizontal or vertical size of the CU before division, respectively.

For example, when a CU of size 32x32 is vertically divided into 3 CUs, the sizes of the 3 divided CUs may be 8x32, 16x32, and 8x32, respectively. In this way, when one CU is divided into three CUs, the CU can be viewed as being divided in the form of a ternary tree.

One of the exemplified quad tree-type partitioning, binary tree-type partitioning, and ternary tree-type partitioning may be applied for partitioning the CU, and a plurality of partitioning methods may be combined together and used for partitioning the CU. . At this time, the case where a plurality of partitioning methods are used in combination can be referred to as partitioning in the form of a composite tree.

Figure 6 shows division of a block according to an example.

In the process of encoding and/or decoding an image, the target block may be divided as shown in FIG. 6. For example, the target block may be a CU.

To split the target block, an indicator indicating splitting information may be signaled from the encoding device 100 to the decoding device 200. Splitting information may be information indicating how the target block is divided.

Splitting information includes split flag (hereinafter referred to as “split_flag”), quad-binary flag (hereinafter referred to as “QB_flag”), quad tree flag (hereinafter referred to as “quadtree_flag”), and binary tree flag (hereinafter referred to as “binarytree_flag”). It may be one or more of a binary type flag (hereinafter denoted as "Btype_flag").

split_flag may be a flag indicating whether the block is split. For example, a value of 1 in split_flag may indicate that the block is split. A value of 0 for split_flag may indicate that the block is not split.

QB_flag may be a flag indicating whether the block is divided into a quad tree format or a binary tree format. For example, a value of 0 for QB_flag may indicate that the block is divided into a quad tree format. A value of 1 for QB_flag may indicate that the block is divided into a binary tree form. Alternatively, the value of QB_flag 0 may indicate that the block is divided into a binary tree form. A value of 1 in QB_flag may indicate that the block is divided into a quad tree format.

quadtree_flag may be a flag indicating whether the block is divided into a quad tree format. For example, a value of 1 in quadtree_flag may indicate that the block is divided into a quad tree format. A value of 0 for quadtree_flag may indicate that the block is not divided into a quad tree format.

binarytree_flag may be a flag indicating whether the block is divided in binary tree form. For example, a value of 1 in binarytree_flag may indicate that the block is split into a binary tree. A value of 0 for binarytree_flag may indicate that the block is not divided into a binary tree form.

Btype_flag may be a flag indicating whether the block is divided into vertical division or horizontal division when the block is divided into binary tree form. For example, a value of 0 in Btype_flag may indicate that the block is divided in the horizontal direction. A value of 1 in Btype_flag may indicate that the block is divided in the vertical direction. Alternatively, the value 0 of Btype_flag may indicate that the block is divided in the vertical direction. A value of 1 in Btype_flag may indicate that the block is divided in the horizontal direction.

For example, partition information for the block of FIG. 6 can be derived by signaling at least one of quadtree_flag, binarytree_flag, and Btype_flag as shown in Table 1 below.

[Table 1]

For example, split information for the block of FIG. 6 can be derived by signaling at least one of split_flag, QB_flag, and Btype_flag as shown in Table 2 below.

[Table 2]

The partitioning method may be limited to quad trees only, or only binary trees, depending on the size and/or shape of the block. When this restriction is applied, split_flag may be a flag indicating whether to split into a quad tree form or a flag indicating whether to split into a binary tree form. The size and shape of the block can be derived according to the depth information of the block, and the depth information can be signaled from the encoding device 100 to the decoding device 200.

If the size of the block falls within a specified range, only quad tree-type division may be possible. For example, the specified range may be defined by at least one of the maximum block size and minimum block size for which only quad tree-type division is possible.

Information indicating the maximum block size and/or minimum block size for which only quad tree-type division is possible may be signaled from the encoding device 100 to the decoding device 200 through a bitstream. Additionally, this information may be signaled for at least one unit among video, sequence, picture, parameter, tile group, and slice (or segment).

Alternatively, the maximum block size and/or minimum block size may be fixed sizes predefined in the encoding device 100 and the decoding device 200. For example, if the block size is 64x64 or larger and 256x256 or smaller, only quad tree type division may be possible. In this case, split_flag may be a flag indicating whether to split into quad tree form.

If the block size is larger than the maximum conversion block size, only quad tree-type division may be possible. At this time, the divided block may be at least one of CU and TU.

In this case, split_flag may be a flag indicating whether to split into quad tree form.

If the block size falls within a specified range, only binary tree or ternary tree division may be possible. Here, for example, the specified range may be defined by at least one of the maximum block size and minimum block size for which only division in the form of a binary tree or a ternary tree is possible.

Information indicating the maximum block size and/or minimum block size for which only binary tree-type splitting or ternary tree-type splitting is possible may be signaled from the encoding device 100 to the decoding device 200 through a bitstream. Additionally, this information may be signaled for at least one unit among sequence, picture, and slice (or segment).

Alternatively, the maximum block size and/or minimum block size may be fixed sizes predefined in the encoding device 100 and the decoding device 200. For example, if the block size is 8x8 or larger and 16x16 or smaller, only division in the form of a binary tree may be possible. In this case, split_flag may be a flag indicating whether to split in binary tree form or ternary tree form.

The above-described description of quad tree-type partitioning can be equally applied to binary tree-type and/or ternary-tree form partitioning.

Splitting of a block may be limited by previous splitting. For example, when a block is divided into a specified binary tree form and a plurality of divided blocks are created, each divided block can be further divided only into the specified tree form. Here, the specified tree form may be at least one of a binary tree form, a ternary tree form, and a quad tree form.

If the horizontal or vertical size of the divided block corresponds to a size that cannot be further divided, the above-described indicator may not be signaled.

Figure 7 is a diagram for explaining an embodiment of the intra prediction process.

Arrows from the center of the graph of FIG. 7 to the outside may indicate prediction directions of directional intra prediction modes. Additionally, numbers displayed close to the arrows may represent an example of a mode value assigned to the intra prediction mode or the prediction direction of the intra prediction mode.

In FIG. 7, the number 0 may represent Planar mode, which is a non-directional intra prediction mode. The number 1 may represent DC mode, which is a non-directional intra prediction mode.

Intra encoding and/or decoding may be performed using reference samples of neighboring blocks of the target block. The neighboring block may be a reconstructed neighboring block. A reference sample may refer to a neighboring sample.

For example, intra encoding and/or decoding may be performed using the value or coding parameter of a reference sample included in the reconstructed neighboring block.

The encoding device 100 and/or the decoding device 200 may generate a prediction block by performing intra prediction on the target block based on information on samples in the target image. When performing intra prediction, the encoding device 100 and/or the decoding device 200 may generate a prediction block for the target block by performing intra prediction based on information on samples in the target image. When performing intra prediction, the encoding device 100 and/or the decoding device 200 may perform directional prediction and/or non-directional prediction based on at least one reconstructed reference sample.

A prediction block may refer to a block generated as a result of performing intra prediction. A prediction block may correspond to at least one of CU, PU, and TU.

The unit of the prediction block may be the size of at least one of CU, PU, and TU. The prediction block may have a square shape with a size of 2Nx2N or NxN. NxN sizes can include 4x4, 8x8, 16x16, 32x32, and 64x64.

Alternatively, the prediction block may be a square-shaped block with a size of 2x2, 4x4, 8x8, 16x16, 32x32, or 64x64, or a rectangular block with a size of 2x8, 4x8, 2x16, 4x16, and 8x16. there is.

Intra prediction may be performed according to the intra prediction mode for the target block. The number of intra prediction modes that a target block can have may be a predefined fixed value or a value determined differently depending on the properties of the prediction block. For example, properties of the prediction block may include the size of the prediction block and the type of the prediction block. Additionally, properties of a prediction block may indicate coding parameters for the prediction block.

For example, the number of intra prediction modes may be fixed to N regardless of the size of the prediction block. Or, for example, the number of intra prediction modes may be 3, 5, 9, 17, 34, 35, 36, 65, 67, or 95.

The intra prediction mode may be a non-directional mode or a directional mode.

For example, an intra prediction mode may include 2 undirectional modes and 65 directional modes, corresponding to numbers 0 to 66 shown in FIG. 7 .

For example, when a specified intra prediction method is used, the intra prediction mode may include 2 undirectional modes and 93 directional modes, corresponding to numbers -14 to 80 shown in FIG. 7.

The two non-directional modes may include DC mode and Planar mode.

The directional mode may be a prediction mode with a specific direction or a specific angle. Directional mode may also be referred to as an argular mode.

The intra prediction mode may be expressed by at least one of a mode number, mode value, mode angle, and mode direction. That is to say, the terms “(mode) number of intra prediction mode”, “(mode) value of intra prediction mode”, “(mode) angle of intra prediction mode” and “(mode) direction of intra prediction mode” have the same meaning. can be used, and can be used interchangeably.

The number of intra prediction modes may be M. M may be 1 or more. In other words, the number of intra prediction modes may be M, including the number of non-directional modes and the number of directional modes.

The number of intra prediction modes may be fixed to M regardless of the size and/or color component of the block. For example, the number of intra prediction modes may be fixed to either 35 or 67, regardless of the block size.

Alternatively, the number of intra prediction modes may vary depending on the shape, size, and/or type of color component of the block.

For example, in FIG. 7, directional prediction modes shown in dotted lines can only be applied to prediction for non-square blocks.

For example, as the block size increases, the number of intra prediction modes may increase. Alternatively, as the block size increases, the number of intra prediction modes may decrease. If the block size is 4x4 or 8x8, the number of intra prediction modes may be 67. If the block size is 16x16, the number of intra prediction modes may be 35. If the block size is 32x32, the number of intra prediction modes may be 19. If the block size is 64x64, the number of intra prediction modes may be 7.

For example, the number of intra prediction modes may vary depending on whether the color component is a luma signal or a chroma signal. Alternatively, the number of intra prediction modes of the luma component block may be greater than the number of intra prediction modes of the chroma component block.

For example, in the case of vertical mode with a mode value of 50, prediction may be performed in the vertical direction based on the pixel value of the reference sample. For example, in the case of horizontal mode with a mode value of 18, prediction may be performed in the horizontal direction based on the pixel value of the reference sample.

Even in the case of a directional mode other than the above-described mode, the encoding device 100 and the decoding device 200 can perform intra prediction on the target unit using a reference sample according to the angle corresponding to the directional mode.

The intra prediction mode located to the right of the vertical mode may be named vertical-right mode. The intra prediction mode located below the horizontal mode may be named the horizontal-below mode. For example, in Figure 7, intra prediction modes with mode values one of 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, and 66 are vertical These may be the right modes. Intra prediction modes with mode values of one of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17 may be horizontal bottom modes.

Non-directional modes may include DC mode and planar mode. For example, the mode value of DC mode may be 1. The mode value of the planner mode may be 0.

Directional modes may include angular modes. Among the plurality of intra prediction modes, the remaining modes except DC mode and planner mode may be directional modes.

When the intra prediction mode is DC mode, a prediction block may be generated based on the average of pixel values of a plurality of reference samples. For example, the pixel value of the prediction block may be determined based on the average of pixel values of a plurality of reference samples.

The number of intra prediction modes described above and the mode value of each intra prediction mode may be merely exemplary. The number of intra prediction modes described above and the mode value of each intra prediction mode may be defined differently depending on embodiment, implementation, and/or need.

In order to perform intra prediction on the target block, a step may be performed to check whether samples included in the reconstructed neighboring block can be used as reference samples of the target block. If there is a sample among the samples of a neighboring block that cannot be used as a reference sample of the target block, a value generated by copying and/or interpolation using at least one sample value among the samples included in the reconstructed neighboring block. This can be replaced with the sample value of a sample that cannot be used as a reference sample. If the value generated by copying and/or interpolation is replaced with the sample value of the sample, the sample can be used as a reference sample of the target block.

When intra prediction is used, a filter may be applied to at least one of a reference sample or a prediction sample based on at least one of the intra prediction mode and the size of the target block.

The type of filter applied to at least one of the reference sample and the prediction sample may vary depending on at least one of the intra prediction mode of the target block, the size of the target block, and the shape of the target block. The type of filter can be classified according to one or more of the length of the filter tap, the value of the filter coefficient, and the filter strength. The length of the above filter tabs may mean the number of filter tabs. Additionally, the number of filter tabs may mean the length of the filter.

When the intra prediction mode is planar mode, when generating the prediction block of the target block, depending on the location of the prediction target sample in the prediction block, the upper reference sample of the target sample, the left reference sample of the target sample, and the upper right reference sample of the target block And the sample value of the prediction target sample may be generated using the weighted sum (weight-sum) of the lower left reference sample of the target block.

When the intra prediction mode is DC mode, when generating the prediction block of the target block, the average value of the top reference samples and the left reference samples of the target block can be used. Additionally, filtering using values of reference samples may be performed on specified rows or specified columns within the target block. The rows specified may be one or more top rows adjacent to the reference sample. The specified columns may be one or more left columns adjacent to the reference sample.

When the intra prediction mode is a directional mode, a prediction block may be generated using the top reference sample, left reference sample, top right reference sample, and/or bottom left reference sample of the target block.

Real-valued interpolation may be performed to generate the prediction samples described above.

The intra prediction mode of the target block may be predicted from the intra prediction mode of the target block's neighboring block, and information used for prediction may be entropy encoded/decoded.

For example, if the intra prediction modes of the target block and the neighboring block are the same, it may be signaled that the intra prediction modes of the target block and the neighboring block are the same using a predefined flag.

For example, an indicator indicating an intra prediction mode that is the same as the intra prediction mode of the target block among the intra prediction modes of a plurality of neighboring blocks may be signaled.

If the intra prediction modes of the target block and the neighboring block are different from each other, information on the intra prediction mode of the target block may be encoded and/or decoded using entropy coding and/or decoding.

Figure 8 is a diagram for explaining reference samples used in the intra prediction process.

The reconstructed reference samples used for intra prediction of the target block are below-left reference samples, left reference samples, above-left corner reference samples, and above reference samples. and above-right reference samples, etc.

For example, left reference samples may refer to reconstructed reference pixels adjacent to the left side of the target block. Top reference samples may refer to reconstructed reference pixels adjacent to the top of the target block. The upper left corner reference sample may refer to a reconstructed reference pixel located at the upper left corner of the target block. Additionally, the lower left reference samples may refer to a reference sample located at the bottom of the left sample line among samples located on the same line as the left sample line composed of left reference samples. The upper right reference samples may refer to reference samples located to the right of the upper pixel line among samples located on the same line as the upper sample line composed of upper reference samples.

When the size of the target block is NxN, the number of bottom left reference samples, left reference samples, top reference samples, and top right reference samples may each be N.

A prediction block may be generated through intra prediction for the target block. Generating a prediction block may include determining values of pixels of the prediction block. The sizes of the target block and prediction block may be the same.

The reference sample used for intra prediction of the target block may vary depending on the intra prediction mode of the target block. The direction of the intra prediction mode may indicate a dependency relationship between reference samples and pixels of the prediction block. For example, the value of a specified reference sample can be used as the value of one or more specified pixels of the prediction block. In this case, the specified reference sample and one or more specified pixels of the prediction block may be samples and pixels designated by a straight line in the direction of the intra prediction mode. In other words, the value of the specified reference sample can be copied to the value of the pixel located in the reverse direction of the intra prediction mode. Alternatively, the pixel value of the prediction block may be the value of a reference sample located in the direction of the intra prediction mode based on the position of the pixel.

For example, when the intra prediction mode of the target block is vertical mode, top reference samples can be used for intra prediction. When the intra prediction mode is a vertical mode, the pixel value of the prediction block may be the value of a reference sample located vertically above the position of the pixel. Therefore, top reference samples adjacent to the top of the target block can be used for intra prediction. Additionally, the values of pixels in one row of the prediction block may be the same as the values of the upper reference samples.

For example, when the intra prediction mode of the target block is horizontal mode, left reference samples can be used for intra prediction. When the intra prediction mode is a horizontal mode, the pixel value of the prediction block may be the value of a reference sample located horizontally to the left of the pixel. Therefore, left reference samples adjacent to the left of the target block can be used for intra prediction. Additionally, the values of pixels in one column of the prediction block may be the same as the values of the left reference samples.

For example, when the mode value of the intra prediction mode of the target block is 34, at least some of the left reference samples, the top left corner reference sample, and at least some of the top reference samples may be used for intra prediction. When the mode value of the intra prediction mode is 34, the pixel value of the prediction block may be the value of a reference sample located diagonally to the upper left with respect to the pixel.

Additionally, when an intra prediction mode with a mode value of one of 52 to 66 is used, at least some of the upper right reference samples may be used for intra prediction.

Additionally, when an intra prediction mode with a mode value of one of 2 to 17 is used, at least some of the lower left reference samples may be used for intra prediction.

Additionally, when an intra prediction mode with a mode value of one of 19 to 49 is used, the upper left corner reference sample can be used for intra prediction.

The reference sample used to determine the pixel value of one pixel of the prediction block may be one or two or more.

As described above, the pixel value of the pixel of the prediction block may be determined according to the location of the pixel and the location of the reference sample indicated by the direction of the intra prediction mode. If the position of the reference sample indicated by the pixel position and the direction of the intra prediction mode is an integer position, the value of one reference sample indicated by the integer position may be used to determine the pixel value of the pixel of the prediction block.

If the position of the reference sample indicated by the pixel position and the direction of the intra prediction mode is not an integer position, an interpolated reference sample can be generated based on the two reference samples closest to the position of the reference sample. there is. The value of the interpolated reference sample can be used to determine the pixel value of the pixel of the prediction block. In other words, when the position of the reference sample indicated by the position of the pixel of the prediction block and the direction of the intra prediction mode indicates the gap between two reference samples, an interpolated value is generated based on the values of the two samples. You can.

The prediction block generated by prediction may not be identical to the original target block. In other words, a prediction error, which is a difference between the target block and the prediction block, may exist, and a prediction error may also exist between the pixels of the target block and the pixels of the prediction block.

Hereinafter, the terms “difference”, “error” and “residual” may be used with the same meaning and may be used interchangeably.

For example, in the case of directional intra prediction, the larger the distance between the pixel of the prediction block and the reference sample, the larger the prediction error may occur. Discontinuity may occur between the prediction block and neighboring blocks generated due to such prediction errors.

Filtering on prediction blocks may be used to reduce prediction error. Filtering may be adaptively applying a filter to an area considered to have a large prediction error among prediction blocks. For example, an area considered to have a large prediction error may be the boundary of a prediction block. Additionally, depending on the intra-prediction mode, the area considered to have a large prediction error among prediction blocks may be different, and the characteristics of the filter may be different.

As shown in FIG. 8, at least one of reference lines 0 to 3 may be used for intra prediction of the target block.

Each reference line in FIG. 8 may represent a reference sample line including one or more reference samples. The smaller the reference line number, the closer the reference sample line may be to the target block.

Instead of being obtained from a reconstructed neighboring block, the samples of segment A and segment F may be obtained through padding using the closest samples of segment B and segment E, respectively.

Index information indicating a reference sample line to be used for intra prediction of the target block may be signaled. Index information may indicate a reference sample line used for intra prediction of a target block among a plurality of reference sample lines. For example, index information may have a value between 0 and 3.

If the upper boundary of the target block is the boundary of the CTU, only reference sample line 0 may be available. Therefore, in this case, index information may not be signaled. If a reference sample line other than reference sample line 0 is used, filtering on the prediction block, which will be described later, may not be performed.

In the case of inter-color intra prediction, a prediction block for the target block of the second color component may be generated based on the corresponding reconstructed block of the first color component.

For example, the first color component may be a luma component, and the second color component may be a chroma component.

For intra prediction between color components, parameters of a linear model between the first color component and the second color component may be derived based on the template.

The template may include a top reference sample and/or a left reference sample of the target block, and may include a top reference sample and/or a left reference sample of the reconstructed block of the first color component corresponding to these reference samples. there is.

For example, the parameters of a linear model are 1) the value of the sample of the first color component that has the maximum value among the samples in the template, 2) the value of the sample of the second color component corresponding to this sample of the first color component, 3) the value of the sample of the first color component having the minimum value among the samples in the template, and 4) the value of the sample of the second color component corresponding to the sample of the first color component.

Once the parameters of the linear model are derived, a prediction block for the target block can be generated by applying the corresponding reconstructed block to the linear model.

Depending on the video format, subsampling may be performed on neighboring samples of the reconstructed block of the first color component and the corresponding reconstructed block. For example, if 1 sample of the second color component corresponds to 4 samples of the first color component, 1 corresponding sample can be calculated by subsampling the 4 samples of the first color component. there is. When subsampling is performed, derivation of parameters of a linear model and intra prediction between color components can be performed based on the subsampled corresponding samples.

Whether to perform intra prediction between color components and/or the range of the template may be signaled as an intra prediction mode.

The target block may be divided into 2 or 4 sub-blocks in the horizontal and/or vertical directions.

The divided sub-blocks can be sequentially reconstructed. That is, as intra prediction is performed on the sub-block, a sub-prediction block for the sub-block may be generated. Additionally, as inverse quantization and/or inverse transformation is performed on the sub-block, a sub-residual block for the sub-block may be generated. A reconstructed sub-block can be generated by adding the sub-prediction block to the sub-residual block. The reconstructed subblock can be used as a reference sample for intra prediction of the lower priority subblock.

A subblock may be a block containing a specified number (eg, 16) or more samples. Therefore, for example, if the target block is an 8x4 block or a 4x8 block, the target block may be divided into two sub-blocks. Additionally, if the target block is a 4x4 block, the target block cannot be divided into sub-blocks. If the target block has a different size, the target block may be divided into 4 sub-blocks.

Information regarding whether intra prediction based on these subblocks is performed and/or the division direction (horizontal or vertical direction) may be signaled.

Such sub-block-based intra prediction may be limited to being performed only when using reference sample line 0. When sub-block-based intra prediction is performed, filtering on the prediction block, which will be described later, may not be performed.

A final prediction block can be generated by performing filtering on the prediction block generated by intra prediction.

Filtering may be performed by applying a specific weight to the filtering target sample, left reference sample, top reference sample, and/or top left reference sample that are the objects of filtering.

Weights and/or reference samples (or ranges of reference samples or positions of reference samples, etc.) used for filtering may be determined based on at least one of block size, intra prediction mode, and location within the prediction block of the sample to be filtered. there is.

For example, filtering may be performed only for specified intra prediction modes (eg, DC mode, planar mode, vertical mode, horizontal mode, diagonal mode, and/or adjacent diagonal mode).

The adjacent diagonal mode may be a mode with a number in which k is added to the number of the diagonal mode, or it may be a mode with a number in which k is subtracted from the number of the diagonal mode. That is, the number of adjacent diagonal modes may be the sum of the number of diagonal modes and k, or the difference between the number of diagonal modes and k. For example, k may be a positive integer of 8 or less.

The intra prediction mode of the target block may be derived using the intra prediction mode of a neighboring block existing around the target block, and this derived intra prediction mode may be entropy encoded and/or entropy decoded.

For example, if the intra prediction mode of the target block and the intra prediction mode of the neighboring block are the same, information that the intra prediction mode of the target block and the intra prediction mode of the neighboring block are the same may be signaled using the specified flag information. .

Additionally, for example, indicator information about a neighboring block having the same intra prediction mode as the intra prediction mode of the target block among the intra prediction modes of a plurality of neighboring blocks may be signaled.

For example, if the intra prediction mode of the target block and the intra prediction mode of the neighboring block are different from each other, entropy coding and/or entropy decoding based on the intra prediction mode of the neighboring block are performed to obtain information about the intra prediction mode of the target block. Entropy encoding and/or entropy decoding may be performed.

Figure 9 is a diagram for explaining an embodiment of the inter prediction process.

The square shown in FIG. 9 may represent an image (or picture). Additionally, in FIG. 9, the arrow may indicate the prediction direction. An arrow from the first picture to the second picture may indicate that the second picture refers to the first picture. That is, the image can be encoded and/or decoded according to the prediction direction.

Each image can be classified into I picture (Intra Picture), P picture (Uni-prediction Picture), and B picture (Bi-prediction Picture) depending on the encoding type. Each picture may be encoded and/or decoded according to the encoding type of each picture.

If the target image that is the target of encoding is an I picture, the target image can be encoded using data within the image itself without inter prediction referring to other images. For example, an I picture can be encoded only with intra prediction.

When the target image is a P picture, the target image can be encoded through inter prediction using only reference pictures that exist in one direction. Here, unidirectional can be forward or reverse.

When the target image is a B picture, the target image can be encoded through inter prediction using reference pictures existing in both directions or inter prediction using reference pictures existing in one of the forward and reverse directions. Here, the two directions can be forward and reverse.

P pictures and B pictures that are encoded and/or decoded using a reference picture may be considered images for which inter prediction is used.

Below, inter prediction in inter mode according to the embodiment is described in detail.

Inter prediction or motion compensation can be performed using reference images and motion information.

In inter mode, the encoding device 100 may perform inter prediction and/or motion compensation for the target block. The decoding device 200 may perform inter prediction and/or motion compensation corresponding to the inter prediction and/or motion compensation in the encoding device 100 on the target block.

Motion information about the target block may be derived by each of the encoding device 100 and the decoding device 200 during inter prediction. The motion information may be derived using motion information of a reconstructed neighboring block, motion information of a call block, and/or motion information of a block adjacent to the call block.

For example, the encoding device 100 or the decoding device 200 performs prediction and/or motion compensation by using motion information of a spatial candidate and/or temporal candidate as motion information of the target block. It can be done. The target block may refer to PU and/or PU partition.

The spatial candidate may be a reconstructed block that is spatially adjacent to the target block.

The temporal candidate may be a reconstructed block corresponding to a target block in an already reconstructed collocated picture (col picture).

In inter prediction, the encoding device 100 and the decoding device 200 can improve encoding efficiency and decoding efficiency by using motion information of spatial candidates and/or temporal candidates. The motion information of the spatial candidate may be referred to as spatial motion information. The motion information of the temporal candidate may be referred to as temporal motion information.

Hereinafter, the motion information of the spatial candidate may be the motion information of the PU including the spatial candidate. The motion information of the temporal candidate may be motion information of a PU including the temporal candidate. The motion information of the candidate block may be motion information of the PU including the candidate block.

Inter prediction can be performed using a reference picture.

A reference picture may be at least one of a picture before the target picture or a picture after the target picture. A reference picture may refer to an image used for prediction of a target block.

In inter prediction, an area within a reference picture can be specified by using a reference picture index (or refIdx) indicating the reference picture and a motion vector to be described later. Here, a specified area within the reference picture may represent a reference block.

Inter prediction can select a reference picture and select a reference block corresponding to the target block within the reference picture. Additionally, inter prediction can generate a prediction block for the target block using the selected reference block.

Motion information may be derived during inter prediction by each of the encoding device 100 and the decoding device 200.

The spatial candidate may be a block that 1) exists in the target picture, 2) has already been reconstructed through encoding and/or decoding, and 3) is adjacent to the target block or located at a corner of the target block. Here, the block located at the corner of the target block may be a block vertically adjacent to a neighboring block horizontally adjacent to the target block, or a block horizontally adjacent to a neighboring block vertically adjacent to the target block. “Block located at the corner of the target block” may have the same meaning as “block adjacent to the corner of the target block.” “Blocks located at the corners of the target block” may be included in “blocks adjacent to the target block.”

For example, spatial candidates include a reconstructed block located to the left of the target block, a reconstructed block located at the top of the target block, a reconstructed block located at the lower left corner of the target block, and a reconstructed block located at the upper right corner of the target block. It may be a reconstructed block or a reconstructed block located in the upper left corner of the target block.

Each of the encoding device 100 and the decoding device 200 can identify a block that exists at a location spatially corresponding to the target block within a coll picture. The location of the target block in the target picture and the location of the identified block in the call picture may correspond to each other.

Each of the encoding device 100 and the decoding device 200 may determine a col block existing at a predefined relative position with respect to the identified block as a temporal candidate. The predefined relative position may be a position inside and/or outside the identified block.

For example, a call block may include a first call block and a second call block. When the coordinates of the identified block are (xP, yP) and the size of the identified block is (nPSW, nPSH), the first call block may be a block located at the coordinates (xP + nPSW, yP + nPSH). The second call block may be a block located at the coordinates (xP + (nPSW >> 1), yP + (nPSH >> 1)). The second call block can be selectively used when the first call block is unavailable.

The motion vector of the target block may be determined based on the motion vector of the call block. Each of the encoding device 100 and the decoding device 200 can scale the motion vector of a call block. The scaled motion vector of the call block can be used as the motion vector of the target block. Additionally, the motion vector of the motion information of the temporal candidate stored in the list may be a scaled motion vector.

The ratio of the motion vector of the target block and the motion vector of the call block may be equal to the ratio of the first temporal distance and the second temporal distance. The first temporal distance may be the distance between the reference picture of the target block and the target picture. The second temporal distance may be the distance between the reference picture and the call picture of the call block.

The method of deriving motion information may vary depending on the inter prediction mode of the target block. For example, inter prediction modes applied for inter prediction include Advanced Motion Vector Predictor (AMVP) mode, merge mode and skip mode, merge mode with motion vector difference, There may be subblock merge mode, triangulation mode, inter-intra combined prediction mode, affine inter mode, and current picture reference mode. Merge mode may also be referred to as motion merge mode. Below, each of the modes is explained in detail.

1) AMVP mode

When AMVP mode is used, the encoding device 100 can search for similar blocks in the neighbors of the target block. The encoding device 100 may obtain a prediction block by performing prediction on the target block using motion information of the searched similar block. The encoding device 100 may encode a residual block that is the difference between the target block and the prediction block.

1-1) Creation of a predicted motion vector candidate list

When the AMVP mode is used as the prediction mode, each of the encoding device 100 and the decoding device 200 can generate a prediction motion vector candidate list using the motion vector of the spatial candidate, the motion vector of the temporal candidate, and the zero vector. there is. The predicted motion vector candidate list may include one or more predicted motion vector candidates. At least one of the motion vector of the spatial candidate, the motion vector of the temporal candidate, and the zero vector may be determined and used as the predicted motion vector candidate.

Hereinafter, the terms “predicted motion vector (candidate)” and “motion vector (candidate)” may be used with the same meaning and may be used interchangeably.

Hereinafter, the terms “predicted motion vector candidate” and “AMVP candidate” may be used with the same meaning and may be used interchangeably.

Hereinafter, the terms “predictive motion vector candidate list” and “AMVP candidate list” may be used with the same meaning and may be used interchangeably.

Spatial candidates may include reconstructed spatial neighboring blocks. In other words, the motion vector of the reconstructed neighboring block may be referred to as a spatial prediction motion vector candidate.

Temporal candidates may include call blocks and blocks adjacent to call blocks. That is, the motion vector of a call block or a motion vector of a block adjacent to a call block may be referred to as a temporal prediction motion vector candidate.

The zero vector may be a (0, 0) motion vector.

The predicted motion vector candidate may be a motion vector predictor for predicting a motion vector. Additionally, in the encoding device 100, a predicted motion vector candidate may be a motion vector initial search position.

1-2) Search for motion vector using predicted motion vector candidate list

The encoding device 100 may use the predicted motion vector candidate list to determine a motion vector to be used for encoding the target block within the search range. Additionally, the encoding device 100 may determine a prediction motion vector candidate to be used as the prediction motion vector of the target block among the prediction motion vector candidates in the prediction motion vector candidate list.

The motion vector to be used for encoding the target block may be a motion vector that can be encoded at minimal cost.

Additionally, the encoding device 100 may determine whether to use the AMVP mode when encoding the target block.

1-3) Transmission of inter prediction information

The encoding device 100 may generate a bitstream including inter prediction information required for inter prediction. The decoding device 200 may perform inter prediction on the target block using inter prediction information of the bitstream.

Inter prediction information includes 1) mode information indicating whether AMVP mode is used, 2) prediction motion vector index, 3) motion vector difference (MVD), 4) reference direction, and 5) reference picture index. can do.

Hereinafter, the terms “predicted motion vector index” and “AMVP index” may be used with the same meaning and may be used interchangeably.

Additionally, inter prediction information may include a residual signal.

When the mode information indicates that AMVP mode is used, the decoding device 200 may obtain a predicted motion vector index, motion vector difference, reference direction, and reference picture index from the bitstream through entropy decoding.

The prediction motion vector index may indicate a prediction motion vector candidate used for prediction of the target block among prediction motion vector candidates included in the prediction motion vector candidate list.

1-4) Inter prediction in AMVP mode using inter prediction information

The decoding apparatus 200 may derive a predicted motion vector candidate using the predicted motion vector candidate list and determine motion information of the target block based on the derived predicted motion vector candidate.

The decoding apparatus 200 may use the predicted motion vector index to determine a motion vector candidate for the target block among the predicted motion vector candidates included in the predicted motion vector candidate list. The decoding apparatus 200 may select the prediction motion vector candidate indicated by the prediction motion vector index from among the prediction motion vector candidates included in the prediction motion vector candidate list as the prediction motion vector of the target block.

The encoding device 100 may generate an entropy-encoded predicted motion vector index by applying entropy coding to the predicted motion vector index, and generate a bitstream including the entropy-encoded predicted motion vector index. The entropy-encoded predicted motion vector index may be signaled from the encoding device 100 to the decoding device 200 through a bitstream. The decoding device 200 can extract an entropy-encoded predicted motion vector index from a bitstream, and obtain the predicted motion vector index by applying entropy decoding to the entropy-encoded predicted motion vector index.

The motion vector actually used for inter prediction of the target block may not match the prediction motion vector. MVD may be used to represent the motion vector that will actually be used for inter prediction of the target block and the difference between the prediction motion vectors. The encoding device 100 may derive a prediction motion vector similar to the motion vector that will actually be used for inter prediction of the target block in order to use the MVD of the smallest size possible.

MVD may be the difference between the motion vector of the target block and the predicted motion vector. The encoding device 100 can calculate the MVD and generate an entropy-encoded MVD by applying entropy encoding to the MVD. The encoding device 100 may generate a bitstream including entropy-encoded MDV.

MVD may be transmitted from the encoding device 100 to the decoding device 200 through a bitstream. The decoding device 200 can extract the entropy-encoded MVD from the bitstream and obtain the MVD by applying entropy decoding to the entropy-encoded MVD.

The decoding device 200 can derive the motion vector of the target block by combining the MVD and the predicted motion vector. In other words, the motion vector of the target block derived from the decoding device 200 may be the sum of the MVD and the motion vector candidate.

Additionally, the encoding device 100 can generate entropy-encoded MVD resolution information by applying entropy encoding to the calculated MVD resolution information, and can generate a bitstream including the entropy-encoded MVD resolution information. The decoding device 200 can extract entropy-encoded MVD resolution information from the bitstream and obtain MVD resolution information by applying entropy decoding to the entropy-encoded MVD resolution information. The decoding device 200 can adjust the resolution of the MVD using the MVD resolution information.

Meanwhile, the encoding device 100 may calculate the MVD based on the affine model. The decoding device 200 may derive an affine control motion vector of the target block through the sum of the MVD and affine control motion vector candidates, and may derive a motion vector for the sub-block using the affine control motion vector. there is.

The reference direction may indicate a reference picture list used for prediction of the target block. For example, the reference direction may point to one of the reference picture list L0 and the reference picture list L1.

The reference direction only indicates a reference picture list used for prediction of the target block, and may not indicate that the directions of reference pictures are limited to the forward direction or backward direction. That is, each of the reference picture list L0 and the reference picture list L1 may include forward and/or reverse pictures.

A uni-direction reference direction may mean that one reference picture list is used. Bi-directional reference direction may mean that two reference picture lists are used. That is, the reference direction may indicate that only the reference picture list L0 is used, that only the reference picture list L1 is used, and one of the two reference picture lists.

The reference picture index may indicate a reference picture used for prediction of the target block among reference pictures in the reference picture list. The encoding device 100 can generate an entropy-coded reference picture index by applying entropy coding to the reference picture index and generate a bitstream including the entropy-coded reference picture index. The entropy-coded reference picture index may be signaled from the encoding device 100 to the decoding device 200 through a bitstream. The decoding device 200 can extract an entropy-coded reference picture index from a bitstream and obtain the reference picture index by applying entropy decoding to the entropy-coded reference picture index.

When two reference picture lists are used for prediction of the target block. One reference picture index and one motion vector can be used for each reference picture list. Additionally, when two reference picture lists are used for prediction of the target block, two prediction blocks may be specified for the target block. For example, a (final) prediction block of the target block may be generated through an average or weighted sum of two prediction blocks for the target block.

The motion vector of the target block can be derived by the predicted motion vector index, MVD, reference direction, and reference picture index.

The decoding device 200 may generate a prediction block for the target block based on the derived motion vector and reference picture index. For example, the prediction block may be a reference block pointed to by the derived motion vector in the reference picture indicated by the reference picture index.

By encoding the predicted motion vector index and MVD rather than encoding the motion vector of the target block itself, the amount of bits transmitted from the encoding device 100 to the decoding device 200 can be reduced and coding efficiency can be improved.

The motion information of the reconstructed neighboring block may be used for the target block. In a specific inter prediction mode, the encoding device 100 may not separately encode the motion information itself for the target block. The motion information of the target block is not encoded, and other information that can derive the motion information of the target block through the motion information of the reconstructed neighboring block may be encoded instead. As other information is encoded instead, the amount of bits transmitted to the decoding device 200 can be reduced and coding efficiency can be improved.

For example, as an inter prediction mode in which the motion information of the target block is not directly encoded, there may be a skip mode and/or a merge mode. At this time, the encoding device 100 and the decoding device 200 may use an identifier and/or index that indicates which unit's motion information among the reconstructed neighboring units is used as the motion information of the target unit.

2) Merge Mode

As a method of deriving the movement information of the target block, there is a merge. Merge may mean merging movements of multiple blocks. Merging may mean applying the movement information of one block to other blocks as well. In other words, the merge mode may mean a mode in which the motion information of the target block is derived from the motion information of the neighboring block.

When the merge mode is used, the encoding device 100 may perform prediction on the motion information of the target block using motion information of the spatial candidate and/or motion information of the temporal candidate. Spatial candidates may include reconstructed spatial neighboring blocks that are spatially adjacent to the target block. Spatial neighboring blocks may include left neighboring blocks and top neighboring blocks. Temporal candidates may include call blocks. The terms “spatial candidate” and “spatial merge candidate” may be used interchangeably and may be used interchangeably. The terms “temporal candidate” and “temporal merge candidate” may be used interchangeably and may be used interchangeably.

The encoding device 100 may obtain a prediction block through prediction. The encoding device 100 may encode a residual block that is the difference between the target block and the prediction block.

2-1) Creation of merge candidate list

When the merge mode is used, each of the encoding device 100 and the decoding device 200 may generate a merge candidate list using motion information of the spatial candidate and/or motion information of the temporal candidate. Motion information may include 1) a motion vector, 2) a reference picture index, and 3) a reference direction. The reference direction can be unidirectional or bidirectional. The reference direction may mean an inter prediction indicator.

The merge candidate list may include merge candidates. The merge candidate may be motion information. In other words, the merge candidate list may be a list in which motion information is stored.

Merge candidates may be motion information such as temporal candidates and/or spatial candidates. In other words, the merge candidate list may include motion information such as temporal candidates and/or spatial candidates.

Additionally, the merge candidate list may include a new merge candidate created by combining merge candidates that already exist in the merge candidate list. In other words, the merge candidate list may include new motion information generated by combining motion information that already exists in the merge candidate list.

Additionally, the merge candidate list may include history-based merge candidates. A history-based merge candidate may be motion information of a block that was encoded and/or decoded before the target block.

Additionally, the merge candidate list may include a merge candidate based on the average of two merge candidates.

Merge candidates may be specified modes that derive inter prediction information. A merge candidate may be information indicating a specified mode that derives inter prediction information. Inter prediction information of the target block can be derived according to the specified mode indicated by the merge candidate. At this time, the specified mode may include a process of deriving a series of inter prediction information. This specified mode may be an inter prediction information derivation mode or a motion information derivation mode.

Inter prediction information of the target block may be derived according to the mode indicated by the merge candidate selected by the merge index among the merge candidates in the merge candidate list.

For example, the motion information derivation modes in the merge candidate list may be at least one of 1) a motion information derivation mode on a sub-block basis and 2) an affine motion information derivation mode.

Additionally, the merge candidate list may include motion information of the zero vector. Zero vectors may also be referred to as zero merge candidates.

In other words, the motion information in the merge candidate list is: 1) motion information of the spatial candidate, 2) motion information of the temporal candidate, 3) motion information generated by a combination of motion information already existing in the merge candidate list, and 4) zero vector. It can be at least one of:

Motion information may include 1) a motion vector, 2) a reference picture index, and 3) a reference direction. The reference direction may also be referred to as an inter prediction indicator. The reference direction can be unidirectional or bidirectional. A unidirectional reference direction may represent L0 prediction or L1 prediction.

The merge candidate list can be created before prediction by merge mode is performed.

The number of merge candidates in the merge candidate list may be predefined. The encoding device 100 and the decoding device 200 may add merge candidates to the merge candidate list according to a predefined method and a predefined rank so that the merge candidate list has a predefined number of merge candidates. Through a predefined method and a predefined ranking, the merge candidate list of the encoding device 100 and the merge candidate list of the decoding device 200 may be the same.

Merge can be applied on a CU or PU basis. When merging is performed on a CU or PU basis, the encoding device 100 may transmit a bitstream containing predefined information to the decoding device 200. For example, predefined information includes 1) information indicating whether to perform a merge for each block partition, 2) which block among blocks that are spatial candidates and/or temporal candidates for the target block and which block to merge with. It may include information about whether

2-2) Search for motion vector using merge candidate list

The encoding device 100 may determine a merge candidate to be used for encoding the target block. For example, the encoding device 100 may perform predictions on a target block using merge candidates from a merge candidate list and generate residual blocks for the merge candidates. The encoding device 100 may use a merge candidate that requires the minimum cost in encoding the prediction and residual blocks to encode the target block.

Additionally, the encoding device 100 may determine whether to use merge mode when encoding the target block.

2-3) Transmission of inter prediction information

The encoding device 100 may generate a bitstream including inter prediction information required for inter prediction. The encoding device 100 may generate entropy-encoded inter prediction information by performing entropy encoding on the inter prediction information, and may transmit a bitstream including the entropy-encoded inter prediction information to the decoding device 200. Entropy-encoded inter prediction information may be signaled from the encoding device 100 to the decoding device 200 through the bitstream. The decoding device 200 can extract entropy-encoded inter prediction information from a bitstream and obtain inter-prediction information by performing entropy decoding on the entropy-encoded inter prediction information.

The decoding device 200 may perform inter prediction on the target block using inter prediction information of the bitstream.

Inter prediction information may include 1) mode information indicating whether to use merge mode, 2) merge index, and 3) correction information.

Additionally, inter prediction information may include a residual signal.

The decoding device 200 can obtain a merge index from the bitstream only when the mode information indicates that the merge mode is used.

Mode information may be a merge flag. The unit of mode information may be a block. Information about the block may include mode information, and the mode information may indicate whether merge mode is applied to the block.

The merge index may indicate a merge candidate used to predict the target block among the merge candidates included in the merge candidate list. Alternatively, the merge index may indicate with which block among neighboring blocks spatially or temporally adjacent to the target block the merge is performed.

The encoding device 100 may select a merge candidate with the highest coding performance among the merge candidates included in the merge candidate list, and set the value of the merge index to indicate the selected merge candidate.

Correction information may be information used to correct a motion vector. The encoding device 100 can generate correction information. The decoding device 200 may correct the motion vector of the merge candidate selected by the merge index based on the correction information.

Correction information may include at least one of information indicating whether correction is made, correction direction information, and correction size information. The prediction mode that corrects the motion vector based on the signaled correction information may be called a merge mode with motion vector difference.

2-4) Inter prediction in merge mode using inter prediction information

The decoding device 200 may perform prediction on the target block using the merge candidate indicated by the merge index among the merge candidates included in the merge candidate list.

The motion vector of the target block can be specified by the motion vector of the merge candidate indicated by the merge index, the reference picture index, and the reference direction.

3) Skip mode

Skip mode may be a mode in which motion information of a spatial candidate or motion information of a temporal candidate is applied to the target block as is. Additionally, the skip mode may be a mode that does not use a residual signal. That is, when skip mode is used, the reconstructed block may be identical to the prediction block.

The difference between merge mode and skip mode may be whether or not residual signals are transmitted or used. That is to say, skip mode may be similar to merge mode except that residual signals are not transmitted or used.

When skip mode is used, the encoding device 100 sends information indicating which block's motion information among spatial candidate or temporal candidate blocks is used as motion information of the target block to the decoding device 200 through a bitstream. Can be transmitted. The encoding device 100 can generate entropy-coded information by performing entropy encoding on such information, and can signal the entropy-coded information to the decoding device 200 through a bitstream. The decoding device 200 can extract entropy-encoded information from a bitstream and obtain information by performing entropy decoding on the entropy-encoded information.

Additionally, when skip mode is used, the encoding device 100 may not transmit other syntax element information, such as MVD, to the decoding device 200. For example, when skip mode is used, the encoding device 100 may not signal syntax elements related to at least one of the MVD, the coded block flag, and the transform coefficient level to the decoding device 200.

3-1) Creation of merge candidate list

Skip mode can also use the merge candidate list. That is, the merge candidate list can be used in both merge mode and skip mode. In this respect, the merge candidate list may be named “skip candidate list” or “merge/skip candidate list.”

Alternatively, skip mode may use a separate candidate list than merge mode. In this case, in the description below, the merge candidate list and merge candidate may be replaced with the skip candidate list and skip candidate, respectively.

The merge candidate list can be created before prediction by skip mode is performed.

3-2) Search for motion vector using merge candidate list

The encoding device 100 may determine a merge candidate to be used for encoding the target block. For example, the encoding device 100 may perform predictions on the target block using merge candidates from the merge candidate list. The encoding device 100 may use a merge candidate that requires the minimum cost in prediction to encode the target block.

Additionally, the encoding device 100 may determine whether to use skip mode when encoding the target block.

3-3) Transmission of inter prediction information

The encoding device 100 may generate a bitstream including inter prediction information required for inter prediction. The decoding device 200 may perform inter prediction on the target block using inter prediction information of the bitstream.

Inter prediction information may include 1) mode information indicating whether skip mode is used, and 2) skip index.

The skip index may be the same as the merge index described above.

When skip mode is used, the target block can be encoded without a residual signal. Inter prediction information may not include residual signals. Alternatively, the bitstream may not include a residual signal.

The decoding device 200 can obtain a skip index from the bitstream only when the mode information indicates that skip mode is used. As described above, the merge index and skip index may be the same. The decoding device 200 can obtain a skip index from the bitstream only when the mode information indicates that merge mode or skip mode is used.

The skip index may indicate a merge candidate used to predict the target block among the merge candidates included in the merge candidate list.

3-4) Inter prediction in skip mode using inter prediction information

The decoding device 200 may perform prediction on the target block using the merge candidate indicated by the skip index among the merge candidates included in the merge candidate list.

The motion vector of the target block can be specified by the motion vector of the merge candidate indicated by the skip index, the reference picture index, and the reference direction.

4) Current picture reference mode

The current picture reference mode may refer to a prediction mode that uses a pre-reconstructed area within the target picture to which the target block belongs.

A motion vector may be used to specify a pre-reconstructed area. Whether the target block is encoded in the current picture reference mode can be determined using the reference picture index of the target block.

A flag or index indicating whether the target block is a block encoded in the current picture reference mode may be signaled from the encoding device 100 to the decoding device 200. Alternatively, whether the target block is a block encoded in the current picture reference mode may be inferred through the reference picture index of the target block.

When the target block is encoded in the current picture reference mode, the target picture may exist at a fixed position or a random position within the reference picture list for the target block.

For example, the fixed position may be a position where the reference picture index value is 0 or the very last position.

If the target picture exists at a random position in the reference picture list, a separate reference picture index indicating this random position may be signaled from the coding device 100 to the decoding device 200.

5) Subblock merge mode

Subblock merge mode may refer to a mode that derives motion information for a subblock of a CU.

When the sub-block merge mode is applied, the motion information of the call sub-block of the target sub-block in the reference image (i.e., sub-block based temporal merge candidate) and/or the affine control point motion vector A subblock merge candidate list may be created using a merge candidate (affine control point motion vector merge candidate).

6) Triangle partition mode

In the triangulation mode, divided target blocks can be created by dividing the target block diagonally. For each divided target block, motion information of each divided target block may be derived, and prediction samples for each divided target block may be derived using the derived motion information. The prediction sample of the target block may be derived through a weighted sum of the prediction samples of the divided target blocks.

7) Inter-intra combined prediction mode

The inter-intra combined prediction mode may be a mode in which a prediction sample of the target block is derived using a weighted sum of prediction samples generated by inter prediction and prediction samples generated by intra prediction.

In the above-described modes, the decoding device 200 may perform its own correction on the derived motion information. For example, the decoding device 200 may search a specified area based on the reference block indicated by the derived motion information and search for motion information with the minimum Sum of Absolute Differences (SAD). And, the searched motion information can be derived as corrected motion information.

In the above-described modes, the decoding device 200 may perform compensation for prediction samples derived through inter prediction using optical flow.

In the above-described AMVP mode, merge mode, and skip mode, etc., motion information to be used for prediction of the target block among motion information in the list can be specified through an index to the list.

To improve coding efficiency, the encoding device 100 may signal only the index of the element that causes the minimum cost in inter prediction of the target block among the elements of the list. The encoding device 100 can encode an index and signal the encoded index.

Accordingly, the above-described lists (i.e., the predicted motion vector candidate list and the merge candidate list) may have to be derived in the same way based on the same data in the encoding device 100 and the decoding device 200. Here, the same data may include a reconstructed picture and a reconstructed block. Additionally, in order to specify elements by index, the order of elements within the list may need to be constant.

Figure 10 shows spatial candidates according to an example.

In Figure 10, the locations of spatial candidates are shown.

The large block in the middle may represent the target block. Five small blocks may represent spatial candidates.

The coordinates of the target block may be (xP, yP), and the size of the target block may be (nPSW, nPSH).

Spatial candidate A ₀ may be a block adjacent to the lower left corner of the target block. A ₀ may be a block occupying a pixel with coordinates (xP - 1, yP + nPSH).

Spatial candidate A ₁ may be a block adjacent to the left of the target block. A ₁ may be the lowest block among blocks adjacent to the left of the target block. Alternatively, A ₁ may be a block adjacent to the top of A ₀ . A ₁ may be a block occupying a pixel with coordinates (xP - 1, yP + nPSH - 1).

Spatial candidate B ₀ may be a block adjacent to the upper right corner of the target block. B ₀ may be a block occupying a pixel with coordinates (xP + nPSW, yP - 1).

Spatial candidate B ₁ may be a block adjacent to the top of the target block. B ₁ may be the rightmost block among blocks adjacent to the top of the target block. Alternatively, B ₁ may be a block adjacent to the left of B ₀ . B ₁ may be a block occupying a pixel with coordinates (xP + nPSW - 1, yP - 1).

Spatial candidate B ₂ may be a block adjacent to the upper left corner of the target block. B ₂ may be a block occupying a pixel with coordinates (xP - 1, yP - 1).

Determination of availability of spatial and temporal candidates

In order to include the motion information of the spatial candidate or the motion information of the temporal candidate in the list, it must be determined whether the motion information of the spatial candidate or the motion information of the temporal candidate is available.

Hereinafter, candidate blocks may include spatial candidates and temporal candidates.

For example, the above determination can be made by sequentially applying steps 1) to 4) below.

Step 1) If the PU containing the candidate block is outside the boundary of the picture, the availability of the candidate block may be set to false. “Availability is set to false” may mean the same as “set to unavailable.”

Step 2) If the PU containing the candidate block is outside the boundary of the slice, the availability of the candidate block may be set to false. If the target block and the candidate block are located in different slices, the availability of the candidate block may be set to false.

Step 3) If the PU containing the candidate block is outside the boundary of the tile, the availability of the candidate block may be set to false. If the target block and the candidate block are located within different tiles, the availability of the candidate block may be set to false.

Step 4) If the prediction mode of the PU including the candidate block is intra prediction mode, the availability of the candidate block may be set to false. If the PU containing the candidate block does not use inter prediction, the availability of the candidate block may be set to false.

Figure 11 shows the order of adding motion information of spatial candidates to a merge list according to an example.

As shown in FIG. 11, when adding motion information of spatial candidates to the merge list, the order of A ₁ , B ₁ , B ₀ , A ₀ and B ₂ may be used. That is, motion information of available spatial candidates may be added to the merge list in the following order: A ₁ , B ₁ , B ₀ , A ₀ , and B ₂ .

Merge list derivation method in merge mode and skip mode

As described above, the maximum number of merge candidates in the merge list can be set. The set maximum number is indicated as N. The set number may be transmitted from the encoding device 100 to the decoding device 200. The slice header of a slice may include N. In other words, the maximum number of merge candidates in the merge list for the target block of the slice can be set by the slice header. For example, by default, the value of N may be 5.

Motion information (i.e., merge candidate) can be added to the merge list in the order of steps 1) to 4) below.

Step 1) Among the spatial candidates, available spatial candidates can be added to the merge list. Motion information of available spatial candidates can be added to the merge list in the order shown in FIG. 11. At this time, if the motion information of the available spatial candidate overlaps with other motion information that already exists in the merge list, the motion information may not be added to the merge list. Checking whether there is overlap with other motion information present in the list can be outlined as a “redundancy check.”

There may be a maximum of N pieces of added motion information.

Step 2) If the number of motion information items in the merge list is smaller than N and a temporal candidate is available, the motion information of the temporal candidate may be added to the merge list. At this time, if the motion information of the available temporal candidate overlaps with other motion information that already exists in the merge list, the motion information may not be added to the merge list.

Step 3) If the number of motion information in the merge list is smaller than N and the type of target slice is "B", the combined motion information generated by combined bi-prediction will be added to the merge list. You can.

The target slice may be a slice containing the target block.

The combined motion information may be a combination of L0 motion information and L1 motion information. L0 motion information may be motion information that refers only to the reference picture list L0. L1 motion information may be motion information that refers only to the reference picture list L1.

Within the merge list, there may be more than one L0 motion information. Additionally, within the merge list, there may be more than one L1 motion information.

There may be more than one piece of combined motion information. In generating the combined motion information, which L0 motion information and which L1 motion information to use among one or more L0 motion information and one or more L1 motion information may be predefined. One or more combined motion information may be generated in a predefined order by combined bidirectional prediction using pairs of different motion information in the merge list. One of the pairs of different motion information may be L0 motion information and the other may be L1 motion information.

For example, the combined motion information added with highest priority may be a combination of L0 motion information with a merge index of 0 and L1 motion information with a merge index of 1. If motion information with a merge index of 0 is not L0 motion information, or motion information with a merge index of 1 is not L1 motion information, the above combined motion information may not be generated and added. The motion information added next may be a combination of L0 motion information with a merge index of 1 and L1 motion information with a merge index of 0. The specific combination below may follow other combinations in the video encoding/decoding field.

At this time, if the combined motion information overlaps with other motion information that already exists in the merge list, the combined motion information may not be added to the merge list.

Step 4) If the number of motion information items in the merge list is smaller than N, zero vector motion information may be added to the merge list.

Zero vector motion information may be motion information in which the motion vector is a zero vector.

There may be one or more zero vector motion information. Reference picture indices of one or more pieces of zero vector motion information may be different from each other. For example, the value of the reference picture index of the first zero vector motion information may be 0. The value of the reference picture index of the second zero vector motion information may be 1.

The number of zero vector motion information may be equal to the number of reference pictures in the reference picture list.

The reference direction of zero vector motion information may be bidirectional. Both motion vectors may be zero vectors. The number of zero vector motion information may be the smaller of the number of reference pictures in the reference picture list L0 and the number of reference pictures in the reference picture list L1. Alternatively, if the number of reference pictures in the reference picture list L0 and the number of reference pictures in the reference picture list L1 are different from each other, a unidirectional reference direction may be used for a reference picture index that can be applied to only one reference picture list.

The encoding device 100 and/or the decoding device 200 may sequentially add zero vector motion information to the merge list while changing the reference picture index.

If the zero vector motion information overlaps with other motion information that already exists in the merge list, the zero vector motion information may not be added to the merge list.

The order of steps 1) to 4) described above is merely exemplary, and the order between steps may be changed. Additionally, some of the steps may be omitted depending on predefined conditions.

Method for deriving a predicted motion vector candidate list in AMVP mode

The maximum number of prediction motion vector candidates in the prediction motion vector candidate list may be predefined. The predefined maximum number is denoted by N. For example, the predefined maximum number may be 2.

Motion information (i.e., predicted motion vector candidate) may be added to the predicted motion vector candidate list in the order of steps 1) to 3) below.

Step 1) Available spatial candidates among spatial candidates may be added to the predicted motion vector candidate list. Spatial candidates may include a first spatial candidate and a second spatial candidate.

The first spatial candidate may be one of A ₀ , A ₁ , scaled A ₀ and scaled A ₁ . The second spatial candidate may be one of B ₀ , B ₁ , B ₂ , scaled B ₀ , scaled B ₁ and scaled B ₂ .

Motion information of available spatial candidates may be added to the predicted motion vector candidate list in the order of the first spatial candidate and the second spatial candidate. At this time, if the motion information of the available spatial candidate overlaps with other motion information that already exists in the prediction motion vector candidate list, the motion information may not be added to the prediction motion vector candidate list. In other words, when the value of N is 2, if the motion information of the second spatial candidate is the same as the motion information of the first spatial candidate, the motion information of the second spatial candidate may not be added to the prediction motion vector candidate list.

There may be a maximum of N pieces of added motion information.

Step 2) If the number of motion information items in the predicted motion vector candidate list is smaller than N and a temporal candidate is available, the motion information of the temporal candidate may be added to the predicted motion vector candidate list. At this time, if the motion information of the available temporal candidate overlaps with other motion information that already exists in the predicted motion vector candidate list, the motion information may not be added to the predicted motion vector candidate list.

Step 3) If the number of motion information pieces in the predicted motion vector candidate list is smaller than N, zero vector motion information may be added to the predicted motion vector candidate list.

There may be one or more zero vector motion information. Reference picture indices of one or more pieces of zero vector motion information may be different from each other.

The encoding device 100 and/or the decoding device 200 may sequentially add zero vector motion information to the prediction motion vector candidate list while changing the reference picture index.

If the zero vector motion information overlaps with other motion information that already exists in the prediction motion vector candidate list, the zero vector motion information may not be added to the prediction motion vector candidate list.

The description of the zero vector motion information described above for the merge list can also be applied to the zero vector motion information. Redundant descriptions are omitted.

The order of steps 1) to 3) described above is merely exemplary, and the order between steps may be changed. Additionally, some of the steps may be omitted depending on predefined conditions.

Figure 12 explains the process of conversion and quantization according to an example.

As shown in FIG. 12, a quantized level can be generated by performing a conversion and/or quantization process on the residual signal.

The residual signal can be generated as the difference between the original block and the prediction block. Here, the prediction block may be a block generated by intra prediction or inter prediction.

The residual signal can be converted to the frequency domain through a transformation process that is part of the quantization process.

Transformation kernels used for transformation may include various DCT kernels such as Discrete Cosine Transform (DCT) type 2 (DCT-II) and Discrete Sine Transform (DST) kernels. .

These transform kernels can perform a separable transform or a 2Dimensional (2D) non-separable transform on the residual signal. The separable transformation may be a transformation that performs one-dimensional (1D) transformation on the residual signal in each of the horizontal and vertical directions.

DCT types and DST types adaptively used for 1D conversion may include DCT-V, DCT-VIII, DST-I, and DST-VII in addition to DCT-II, as shown in Table 3 and Table 4 below, respectively. there is.

[Table 3]

[Table 4]

As shown in Tables 3 and 4, a transform set can be used to derive the DCT type or DST type to be used for transformation. Each transformation set may include multiple transformation candidates. Each transformation candidate may be a DCT type or a DST type.

Table 5 below shows an example of a transform set applied to the horizontal direction and a transform set applied to the vertical direction according to the intra prediction mode.

[Table 5]

In Table 5, the number of the vertical transformation set and the number of the horizontal transformation set applied to the horizontal direction of the residual signal according to the intra prediction mode of the target block are displayed.

As illustrated in Table 5, transformation sets applied to the horizontal and vertical directions may be predefined according to the intra prediction mode of the target block. The encoding device 100 may perform transformation and inverse transformation on the residual signal using the transformation included in the transformation set corresponding to the intra prediction mode of the target block. Additionally, the decoding apparatus 200 may perform inverse transformation on the residual signal using a transformation included in a transformation set corresponding to the intra prediction mode of the target block.

For these transformations and inverse transformations, the set of transformations applied to the residual signal may be determined as illustrated in Tables 3, 4, and 5, and may be unsignaled. Transformation instruction information may be signaled from the encoding device 100 to the decoding device 200. Transformation instruction information may be information indicating which transform candidate is used among a plurality of transform candidates included in a transform set applied to the residual signal.

For example, when the size of the target block is 64x64 or less, transform sets each having three transforms may be configured according to the intra prediction mode. The optimal transformation method can be selected among a total of 9 multiple transformation methods resulting from a combination of three transformations in the horizontal direction and three transformations in the vertical direction. Coding efficiency can be improved by encoding and/or decoding the residual signal using this optimal conversion method.

At this time, for at least one of the vertical transformation and the horizontal transformation, information about which transformation among the transformations belonging to the transformation set was used may be entropy encoded and/or decoded. Truncated unary binarization may be used to encode and/or decode this information.

The method using various transforms as described above can be applied to a residual signal generated by intra prediction or inter prediction.

Transformation may include at least one of primary transformation and secondary transformation. A transform coefficient can be generated by performing a first-order transform on the residual signal, and a second-order transform coefficient can be generated by performing a second-order transform on the transform coefficient.

A primary transformation may be named primary. Additionally, the first-order transform may be named Adaptive Multiple Transform (AMT). AMT may mean that different transformations are applied to each of the 1D directions (i.e., vertical and horizontal directions) as described above.

The secondary transformation may be a transformation to improve the energy concentration of the transformation coefficient generated by the primary transformation. Secondary transformations, like primary transformations, can be either separable transformations or non-separable transformations. The non-separable transform may be a Non-Separable Secondary Transform (NSST).

Primary transformation may be performed using at least one of a plurality of predefined transformation methods. For example, a plurality of predefined transformation methods include Discrete Cosine Transform (DCT), Discrete Sine Transform (DST), and Karhunen-Loeve Transform (KLT)-based transformation, etc. It can be included.

Additionally, the first-order transformation may be a transformation with various transformation types depending on the kernel function that defines DCT or DST.

For example, the transformation type is 1) prediction mode of the target block (e.g., one of intra prediction and inter prediction), 2) size of the target block, 3) shape of the target block, 4) intra prediction mode of the target block. , 5) a component of the target block (e.g., one of the luma component and a chroma component), and 6) the partition type applied to the target block (e.g., Quad Tree (QT), Binary Tree (BT) ) and one of a Ternary Tree (TT).

For example, the first-order transformation includes transformations such as DCT-2, DCT-5, DCT-7, DST-7, DST-1, DST-8, and DCT-8 according to the transformation kernels shown in Table 6 below. can do. Table 6 illustrates various transform types and transform kernel functions for multiple transform selection (MTS).

MTS may mean that a combination of one or more DCT and/or DST transformation kernels is selected to transform the residual signal in the horizontal and/or vertical directions.

[Table 6]

In Table 6, i and j may be integer values between 0 and N-1.

A secondary transform may be performed on the transformation coefficient generated by performing the primary transformation.

As with first-order transformations, a set of transformations can be defined for second-order transformations. Methods for deriving and/or determining a set of transformations such as those described above can be applied to secondary transformations as well as primary transformations.

Primary transformation and secondary transformation can be determined for a specified target.

For example, a first-order transform and a second-order transform may be applied to one or more signal components of a luma component and a chroma component. Whether to apply the first transform and/or the second transform may be determined according to at least one of coding parameters for the target block and/or the neighboring block. For example, whether to apply primary transformation and/or secondary transformation may be determined by the size and/or shape of the target block.

In the encoding device 100 and the decoding device 200, conversion information indicating the conversion method to be used for the target can be derived by using specified information.

For example, the transformation information may include an index of the transformation to be used for primary transformation and/or secondary transformation. Alternatively, the transformation information may indicate that the primary transformation and/or secondary transformation is not used.

For example, when the target of the primary transformation and secondary transformation is the target block, the transformation method(s) applied to the primary transformation and/or secondary transformation indicated by the transformation information is applied to the target block and/or neighboring blocks. It may be determined according to at least one of the coding parameters for.

Alternatively, conversion information indicating a conversion method for a specified target may be signaled from the encoding device 100 to the decoding device 200.

For example, for one CU, whether the primary transform is used, an index indicating the primary transform, whether the secondary transform is used, and an index indicating the secondary transform, etc. can be derived as transformation information in the decoding device 200. there is. Alternatively, for one CU, transformation information indicating whether to use the primary transformation, an index indicating the primary transformation, whether to use the secondary transformation, and an index indicating the secondary transformation may be signaled.

A quantized transform coefficient (i.e., a quantized level) may be generated by performing quantization on a result or a residual signal generated by performing a first-order transform and/or a second-order transform.

13 shows diagonal scanning according to an example.

14 shows horizontal scanning according to an example.

15 shows vertical scanning according to one example.

The quantized transform coefficients may be scanned according to at least one of (up-right) diagonal scanning, vertical scanning, and horizontal scanning, according to at least one of intra prediction mode, block size, and block type. A block may be a transformation unit.

Each scanning can start at a specified starting point and end at a specified ending point.

For example, by scanning the coefficients of a block using the diagonal scanning of FIG. 13, the quantized transform coefficients can be changed into a one-dimensional vector form. Alternatively, the horizontal scanning of FIG. 14 or the vertical scanning of FIG. 15 may be used instead of diagonal scanning, depending on the size of the block and/or the intra prediction mode.

Vertical scanning may be scanning two-dimensional block-shaped coefficients in a column direction. Horizontal scanning may be scanning two-dimensional block-shaped coefficients in the row direction.

That is, depending on the size of the block and/or the inter prediction mode, it may be determined which scanning among diagonal scanning, vertical scanning, and horizontal scanning will be used.

As shown in FIGS. 13, 14, and 15, the quantized transform coefficients can be scanned along the diagonal, horizontal, or vertical directions.

Quantized transform coefficients can be expressed in block form. A block may include multiple sub-blocks. Each subblock can be defined according to the minimum block size or minimum block type.

In scanning, the scanning order according to the type or direction of scanning can first be applied to sub-blocks. Additionally, a scanning order according to the direction of scanning may be applied to the quantized transform coefficients within the sub-block.

For example, as shown in FIGS. 13, 14, and 15, when the size of the target block is 8x8, the transform coefficients quantized by the first transform, second transform, and quantization of the residual signal of the target block are can be created. Thereafter, one of three scanning orders may be applied to the four 4x4 sub-blocks, and quantized transform coefficients may be scanned for each 4x4 sub-block according to the scanning order.

The encoding device 100 can generate entropy-encoded quantized transform coefficients by performing entropy encoding on the scanned quantized transform coefficients, and can generate a bitstream including the entropy-encoded quantized transform coefficients. .

The decoding device 200 can extract entropy-encoded quantized transform coefficients from a bitstream and generate quantized transform coefficients by performing entropy decoding on the entropy-encoded quantized transform coefficients. Quantized transformation coefficients can be arranged in a two-dimensional block form through inverse scanning. At this time, as a reverse scanning method, at least one of (upper right) diagonal scan, vertical scan, and horizontal scan may be performed.

In the decoding device 200, dequantization may be performed on the quantized transform coefficients. Depending on whether the secondary inverse transformation is performed, the secondary inverse transformation may be performed on the result generated by performing the inverse quantization. Additionally, depending on whether the first inversion is performed, the first inversion may be performed on the result generated by performing the second inversion. A reconstructed residual signal can be generated by performing a first-order inversion on the result generated by performing a second-order inversion.

For luma components reconstructed through intra prediction or inter prediction, inverse mapping of the dynamic range may be performed before in-loop filtering.

The dynamic range can be divided into 16 equal pieces, and a mapping function for each piece can be signaled. The mapping function can be signaled at the slice level or tile group level.

A reverse mapping function for performing reverse mapping may be derived based on the mapping function.

In-loop filtering, storage of reference pictures, and motion compensation can be performed in the demapped region.

A prediction block generated through inter prediction can be converted into a mapped area by mapping using a mapping function, and the converted prediction block can be used to generate a reconstructed block. However, since intra prediction is performed in a mapped region, the prediction block generated by intra prediction can be used to generate a reconstructed block without mapping and/or demapping.

For example, if the target block is a residual block of a chroma component, the residual block can be converted to a demapped region by performing scaling on the chroma component of the mapped area.

Whether scaling is available can be signaled at the slice level or tile group level.

For example, scaling can only be applied if mapping for the luma component is available and the splitting of the luma component and the splitting of the chroma component follow the same tree structure.

Scaling may be performed based on the average of the values of samples of the luma prediction block corresponding to the chroma prediction block. At this time, if the target block uses inter prediction, the luma prediction block may mean a mapped luma prediction block.

The value required for scaling can be derived by referring to the look-up table using the index of the piece to which the average value of the samples of the luma prediction block belongs.

By performing scaling on the residual block using the finally derived value, the residual block can be converted into a demapped area. Thereafter, for the chroma component block, reconstruction, intra prediction, inter prediction, in-loop filtering, and storage of the reference picture can be performed in the demapped region.

For example, information indicating whether mapping and/or de-mapping of such luma components and chroma components is available may be signaled through a sequence parameter set.

The prediction block of the target block may be generated based on the block vector. A block vector may indicate displacement between a target block and a reference block. The reference block may be a block in the target image.

In this way, the prediction mode that generates a prediction block with reference to the target image may be called an intra block copy (IBC) mode.

IBC mode can be applied to CUs of a specified size. For example, IBC mode can be applied to MxN CU. Here, M and N may be less than or equal to 64.

IBC mode may include skip mode, merge mode, and AMVP mode. In the case of skip mode or merge mode, a merge candidate list may be constructed, and a merge index may be signaled, thereby specifying one merge candidate among the merge candidates in the merge candidate list. The block vector of the specified merge candidate can be used as the block vector of the target block.

For AMVP mode, differential block vectors can be signaled. Additionally, the prediction block vector may be derived from the left neighboring block and the top neighboring block of the target block. Additionally, an index regarding which neighboring block will be used may be signaled.

The prediction block in IBC mode may be included in the target CTU or the left CTU, and may be limited to blocks within the previously reconstructed area. For example, the value of the block vector may be limited so that the prediction block of the target block is located within a specified area. The specified area may be an area of three 64x64 blocks that are encoded and/or decoded before the 64x64 block containing the target block. By limiting the value of the block vector in this way, memory consumption and device complexity according to the implementation of IBC mode can be reduced.

Figure 16 is a structural diagram of an encoding device according to an embodiment.

The encoding device 1600 may correspond to the encoding device 100 described above.

The encoding device 1600 includes a processing unit 1610, a memory 1630, a user interface (UI) input device 1650, a UI output device 1660, and storage that communicate with each other through a bus 1690. (1640) may be included. Additionally, the encoding device 1600 may further include a communication unit 1620 connected to the network 1699.

The processing unit 1610 may be a semiconductor device that executes processing instructions stored in a central processing unit (CPU), memory 1630, or storage 1640. The processing unit 1610 may be at least one hardware processor.

The processing unit 1610 may generate and process signals, data, or information that are input to the encoding device 1600, output from the encoding device 1600, or used inside the encoding device 1600. Inspection, comparison, and judgment related to data or information can be performed. That is, in an embodiment, generation and processing of data or information, and inspection, comparison, and judgment related to the data or information may be performed by the processing unit 1610.

The processing unit 1610 includes an inter prediction unit 110, an intra prediction unit 120, a switch 115, a subtractor 125, a transform unit 130, a quantization unit 140, an entropy encoding unit 150, and an inverse quantization unit. It may include a unit 160, an inverse transform unit 170, an adder 175, a filter unit 180, and a reference picture buffer 190.

Inter prediction unit 110, intra prediction unit 120, switch 115, subtractor 125, transform unit 130, quantization unit 140, entropy encoding unit 150, inverse quantization unit 160, At least some of the inverse transform unit 170, the adder 175, the filter unit 180, and the reference picture buffer 190 may be program modules and may communicate with an external device or system. Program modules may be included in the encoding device 1600 in the form of an operating system, application program module, and other program modules.

Program modules may be physically stored on various known storage devices. Additionally, at least some of these program modules may be stored in a remote memory device capable of communicating with the encoding device 1600.

Program modules are routines, subroutines, programs, objects, components, and data that perform a function or operation according to an embodiment or implement an abstract data type according to an embodiment. It may include data structures, etc., but is not limited thereto.

Program modules may be composed of instructions or codes that are executed by at least one processor of the encoding device 1600.

The processing unit 1610 includes an inter prediction unit 110, an intra prediction unit 120, a switch 115, a subtractor 125, a transform unit 130, a quantization unit 140, an entropy encoding unit 150, and an inverse quantization unit. Commands or codes of the unit 160, the inverse transform unit 170, the adder 175, the filter unit 180, and the reference picture buffer 190 can be executed.

The storage unit may represent memory 1630 and/or storage 1640. Memory 1630 and storage 1640 may be various types of volatile or non-volatile storage media. For example, the memory 1630 may include at least one of ROM 1631 and RAM 1632.

The storage unit may store data or information used for the operation of the encoding device 1600. In an embodiment, data or information held by the encoding device 1600 may be stored in the storage unit.

For example, the storage unit can store pictures, blocks, lists, motion information, inter prediction information, and bitstreams.

The encoding device 1600 may be implemented in a computer system that includes a recording medium that can be read by a computer.

The recording medium may store at least one module required for the encoding device 1600 to operate. The memory 1630 may store at least one module, and the at least one module may be configured to be executed by the processing unit 1610.

Functions related to communication of data or information of the encoding device 1600 may be performed through the communication unit 1620.

For example, the communication unit 1620 may transmit a bitstream to the decoding device 1700, which will be described later.

Figure 17 is a structural diagram of a decoding device according to an embodiment.

The decoding device 1700 may correspond to the decoding device 200 described above.

The decryption device 1700 includes a processing unit 1710, a memory 1730, a user interface (UI) input device 1750, a UI output device 1760, and storage that communicate with each other through a bus 1790. (1740). Additionally, the decryption device 1700 may further include a communication unit 1720 connected to the network 1799.

The processing unit 1710 may be a semiconductor device that executes processing instructions stored in a central processing unit (CPU), memory 1730, or storage 1740. The processing unit 1710 may be at least one hardware processor.

The processing unit 1710 may generate and process signals, data, or information that are input to the decoding device 1700, output from the decoding device 1700, or used inside the decoding device 1700. Inspection, comparison, and judgment related to data or information can be performed. That is, in an embodiment, generation and processing of data or information, and inspection, comparison, and judgment related to the data or information may be performed by the processing unit 1710.

The processing unit 1710 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, an inter prediction unit 250, a switch 245, an adder 255, and a filter. It may include a unit 260 and a reference picture buffer 270.

Entropy decoding unit 210, inverse quantization unit 220, inverse transform unit 230, intra prediction unit 240, inter prediction unit 250, switch 245, adder 255, filter unit 260, and At least some of the reference picture buffers 270 may be program modules and may communicate with an external device or system. Program modules may be included in the decryption device 1700 in the form of an operating system, application program module, and other program modules.

Program modules may be physically stored on various known storage devices. Additionally, at least some of these program modules may be stored in a remote memory device capable of communicating with the decoding device 1700.

Program modules are routines, subroutines, programs, objects, components, and data that perform a function or operation according to an embodiment or implement an abstract data type according to an embodiment. It may include data structures, etc., but is not limited thereto.

Program modules may be composed of instructions or codes that are executed by at least one processor of the decoding device 1700.

The processing unit 1710 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, an inter prediction unit 250, a switch 245, an adder 255, and a filter. Instructions or codes of the unit 260 and the reference picture buffer 270 may be executed.

The storage unit may represent memory 1730 and/or storage 1740. Memory 1730 and storage 1740 may be various types of volatile or non-volatile storage media. For example, the memory 1730 may include at least one of ROM 1731 and RAM 1732.

The storage unit may store data or information used for the operation of the decoding device 1700. In an embodiment, data or information held by the decoding device 1700 may be stored in the storage unit.

For example, the storage unit can store pictures, blocks, lists, motion information, inter prediction information, and bitstreams.

The decryption device 1700 may be implemented in a computer system that includes a recording medium that can be read by a computer.

The recording medium may store at least one module required for the decoding device 1700 to operate. The memory 1730 may store at least one module, and the at least one module may be configured to be executed by the processing unit 1710.

Functions related to communication of data or information of the decryption device 1700 may be performed through the communication unit 1720.

For example, the communication unit 1720 may receive a bitstream from the encoding device 1600.

Hereinafter, the processing unit may refer to the processing unit 1610 of the encoding device 1600 and/or the processing unit 1710 of the decoding device 1700. For example, for functions related to prediction, the processing unit may represent switch 115 and/or switch 245. In functions related to inter prediction, the processing unit may represent an inter prediction unit 110, a subtractor 125, and an adder 175, and may represent an inter prediction unit 250 and an adder 255. In the function related to intra prediction, the processing unit may represent an intra prediction unit 120, a subtractor 125, and an adder 175, and may represent an intra prediction unit 240 and an adder 255. In functions related to transformation, the processing unit may represent a transformation unit 130 and an inverse transformation unit 170, and may indicate an inverse transformation unit 230. In functions related to quantization, the processing unit may represent a quantization unit 140 and an inverse quantization unit 160, and may represent an inverse quantization unit 220. In functions related to entropy encoding and/or decoding, the processing unit may represent an entropy encoding unit 150 and/or an entropy decoding unit 210. In functions related to filtering, the processing unit may represent a filter unit 180 and/or a filter unit 260. For functions related to reference pictures, the processing unit may represent a reference picture buffer 190 and/or a reference picture buffer 270.

Figure 18 is a flowchart of an encoding method and a bitstream generation method according to an embodiment.

The encoding method and bitstream generation method of the embodiment may be performed by the encoding device 1600. An embodiment may be part of an encoding method.

In step 1810, the processor 1610 may perform encoding on the target block.

The processing unit 1610 may determine coding information to be used for encoding the target block.

Coding information may include information described in embodiments. For example, coding information may include prediction information. Prediction information may include information related to predictions described in embodiments. Prediction information may include inter prediction information and intra prediction information.

Coding information may include signaling/encoding/decoding information described in embodiments. In other words, the coding information may be information used to perform decoding in the decoding device 1700 corresponding to the encoding performed in the encoding device 1600. Here, encoding and decoding may include prediction.

Coding information may include syntax elements described in the embodiments.

Coding information may include coding parameters related to the target block.

The processing unit 1610 may use coding information when encoding the target block. Alternatively, coding information may be generated to correspond to information used in encoding the target block.

In step 1820, the processing unit 1610 may perform encoding on the coding information to generate encoded coding information.

In step 1830, processing unit 1610 may generate a bitstream.

Information included in the bitstream may be generated in step 1820, or may be at least partially generated in steps 1810 and 1820.

The processing unit 1610 may store the generated bitstream in the storage 1640. Alternatively, the communication unit 1620 may transmit the bitstream to the decoding device 1700.

The bitstream may include encoded information about the target block. The processing unit 1610 may generate encoded information about the target block by performing entropy encoding on the information about the target block.

Information about the target block may include transformed and quantized coefficients. Encoded information about the target block may include an encoded transform and quantized coefficients.

Figure 19 is a flowchart of a decoding method using a bitstream according to an embodiment.

The decoding method using the bitstream of the embodiment may be performed by the decoding device 1700. The embodiment may be part of a decryption method.

A computer-readable recording medium may include a bitstream. In step 1910, the communication unit 1720 may obtain a bitstream. The communication unit 1720 may receive a bitstream from the encoding device 1600. The processing unit 1710 may store the obtained bitstream in the storage 1740.

The processing unit 1710 can read a bitstream from the storage unit 1740.

The bitstream may include encoded information about the target block. The processing unit 1710 may generate information about the target block by performing entropy decoding on the encoded information about the target block.

Encoded information about the target block may include an encoded transform and quantized coefficients. Information about the target block may include transformed and quantized coefficients.

The bitstream may include encoded coding information or coding information.

In step 1920, the processor 1710 may obtain coding information from the bitstream.

The processing unit 1710 may generate coding information by performing decoding on the encoded coding information in the bitstream.

Coding information may include signaling/encoding/decoding information described in embodiments. In other words, coding information may be information used to perform prediction in the decoding device 1700 corresponding to prediction performed in the encoding device 1600.

Coding information may include syntax elements described in the embodiments.

Coding information may include coding parameters related to the target block.

In step 1930, the processing unit 1710 may perform decoding on the target block.

The processing unit 1710 may determine coding information to be used for decoding the target block.

Coding information may include information described in embodiments. For example, coding information may include prediction information. Prediction information may include information related to predictions described in embodiments. Prediction information may include inter prediction information and intra prediction information.

The processing unit 1710 may use coding information when decoding the target block. Alternatively, coding information in the decoding device 1700 may be generated to correspond to coding information used in the encoding device 1600 to encode the target block. Coding information in the decoding device 1700 may be the same as coding information in the encoding device 1600. In other words, the processing unit 1710 may generate coding information that is the same as the coding information used in step 1810 in order to perform decoding corresponding to the encoding performed in step 1810.

Processing unit 1710 may determine coding information using methods described in embodiments.

The processing unit 1710 may generate a reconstructed block by decoding the target block using information about the target block and coding information.

20 is a flowchart of a prediction method according to an example.

The embodiment may be performed in the encoding device 1600 and the decoding device 1700.

The forecast of embodiments may include stages 2010, 2020, 2030 and 2040. For example, step 1810 may include steps 2010, 2020, 2030, and 2040. Additionally, step 1930 may include steps 2010, 2020, 2030, and 2040.

In an embodiment, the processing unit may represent the processing unit 1610 of the encoding device 1600 and/or the processing unit 1710 of the decoding device 1700.

In step 2010, the processing unit may generate a template for the target block.

The processing unit can create a template using coding information.

For example, the processing unit: 1) horizontal size of the target block, 2) vertical size of the target block, 3) vertical size of the upper template, 4) horizontal size of the left template, 5) number of pixels in the upper template area (= n _top ), 6) Number of pixels in the left template area (= n _left ), 7) Number of pixels in the top left template area (= n _{top_left} ), 8) Pixels whose positions in the top template area are multiples of α, 9) pixel whose position in the left template area is a multiple of β, 10) pixel whose position in the upper left template area is a multiple of γ, 11) pixel located on the diagonal line in the upper left template area, 12) upper template area 13) pixels on the ith adjacent horizontal line in the left template area (e.g., a starting position pixel and a pixel whose position is a multiple of α _i ), 13) pixels on the ith adjacent vertical line in the left template area (e.g., starting position pixel and pixels whose positions are multiples of β _i ), 14) pixels on the ith adjacent horizontal line in the upper left template area (e.g., starting position pixels and pixels whose positions are multiples of γ _i ), 15) Pixels on the ith adjacent vertical line in the upper left template area (e.g., a starting position pixel and a pixel whose position is a multiple of γ _i ), 16) template_idx, 17) template padding, 18) template_vertical_flipping, 19) template_horizontal_flipping, 20 A template can be created using at least one of ) template_flipping_flag and 21) template_flipping_vertical_flag.

In creating a template, upper pixels outside the boundary of a specific unit may not be used as a template. The specific unit may be a unit described in embodiments such as a Coding Tree Unit (CTU).

In step 2020, the processing unit may select a template matching (TM) reference image.

The processing unit may select a TM reference image using coding information.

For example, the processing unit includes 1) template matching reference image list (= template_matching_ref_list), 2) template matching reference image index (= template_matching_ref_idx), 3) template matching reference image list L0 (= template_matching_ref_l0), 4) template matching reference image list. At least one of L1 (= template_matching_ref_l1), 5) L0 template matching reference image index (= template_matching_ref_idx_l0), 6) L1 template matching reference image index (= template_matching_ref_idx_l1), 7) slice type (= slice_type), and 8) coding mode of the target block. You can use one to select a TM reference image.

For example, to select a template matching reference image, the processor may select one template matching reference image configuring method among a plurality of template matching reference image configuring methods. Here, the method of configuring a template matching reference image may include a method of configuring a template matching reference image list.

At step 2030, the processing unit may perform TM search.

The processing unit can derive the optimal template matching block by performing TM search within the template matching reference image using the template.

TM search may mean search for a TM optimal block.

The processing unit can search for the TM optimal block using coding information.

For example, the processing unit 1) TM motion vector (= MV _tm ), 2) TM search area (region), 3) rectangular TM search area, 4) diamond TM search area, 5) size of the rectangular TM search area (square Horizontal size of the TM search area (= rec_template_width) and/or vertical size of the rectangular TM search area (= rec_template_height)), 6) both directions from the template matching search start point (both horizontal directions from the template matching search start point (= del_width) and/or both vertical directions from the template matching search start point (= del_height)), 7) the size of the rectangular TM search area (horizontal size of the rectangular TM search area (= dia_template_width) and/or 10) the rectangular TM search area vertical size (= dia_template_height)), 8) shape of template matching search area (= template_matching_shape_idx), 9) corresponding position of target block, 10) corresponding position of template of target block (= (x _tm , y _tm )), 11) TM motion vector of a block encoded using TM among blocks adjacent to the target block (= (MV _{tm_x} , MV _{tm_y} ), 12) merge list, 13) merge index (= merge_idx), 14) TM search start point Index (= template_matching_init_idx), 15) TM search method, 16) TM search method index (= template_matching_search_idx), 17) Template matching correlation measure (e.g., Sum of Absolute Differences (SAD), Sum of Squared Differences (SSD) and Sum of Absolute Transformed Differences (SATD), etc.), 18) Template matching correlation scale index (= template_matching_criteria_idx), 19) Template matching search The TM optimal block can be searched using at least one of 20) whether it terminates early (= template_matching_early_termination_flag) and 20) the template matching search early termination threshold (= template_matching_early_termination_threshold).

For example, the processor may select one of search areas having different shapes as a search area for deriving the optimal template matching block.

For example, the processing unit can set the template matching search starting point of the search area to derive the optimal matching block.

In embodiments, the size of a specific unit may refer to the area of the specific unit. Alternatively, the size of a specific unit may mean one or more of the horizontal size and vertical size of the specific unit.

TM search methods may include full search, 2D logarithmic search, N-step search, diamond search, hexagonal search, and test zone search. .

In step 2040, the processing unit may perform encoding/decoding using the TM optimal block.

The processing unit can use coding information to perform encoding/decoding using the TM optimal block.

For example, the processing unit may perform encoding/decoding using at least one of a template, a TM reference image, a TM motion vector, and a TM optimal block.

Step 2040 may include one or more of steps 2041, 2041, 2043, 2044, and 2045.

In step 2041, the processor may perform TM-based intra residual signal prediction.

The processing unit may use the coding information to perform encoding of the target block using the TM matching optimal block discovered by TM matching.

For example, the processing unit processes 1) the intra prediction residual signal of the target block, 2) the intra prediction mode, 3) the intra prediction residual signal of the TM optimal block and the prediction error (= residual signal) of the target block, 4) the residual of the target block. The difference value between the signal and the residual signal of the TM optimal block, 5) the difference value between the intra prediction residual signal of the target block and the intra prediction residual signal of the TM optimal block, and 6) the TM explored by TM matching using at least one of tm_intra_residual_pred. Encoding of the target block using the matching optimal block can be performed.

In step 2042, the processor may perform limited TM-based intra residual signal prediction.

The processing unit may perform limited TM-based intra residual signal prediction using coding information.

For example, the processing unit 1) some area of the target block (e.g., lower right square, lower right triangle, lower partial block, or right partial block), 2) x-axis with the template (or left end of the target block) At least one _of the following: distance in direction (= _d Using this, limited TM-based intra residual signal prediction can be performed.

In step 2043, the processing unit may perform Local Illumination Compensation (LIC)-based TM prediction.

The processing unit may perform LIC-based TM prediction using coding information.

For example, the processing unit can: 1) LIC, 2) linear model of TM optimal block, 3) α and β of linear model, 4) some pixels of template, 5) representative values of subsampling for template pixels, 6) target Average of large representative values of block template (= T ₁ ^A ), 7) Average of small representative values of target block template (= T ₂ ^A ), 8) Average of large representative values of TM optimal block template (= T ₂ ^A ), 9) the average of small representative values of the TM optimal block template (= T ₂ ^B ), and 10) lic_template_matching_mode. LIC-based TM prediction can be performed.

At step 2044, the processor may use blending prediction of intra and TM.

The processing unit can perform intra and TM blending prediction using coding information.

For example, the processing unit performs blending prediction of intra and TM using at least one of 1) intra prediction signal, 2) weight of intra prediction, 3) weight of TM optimal block, 4) intra_template_matching_blending_mode, and 5) intra prediction mode. can do.

At step 2045, the processor may use blending prediction of Inter and TM.

The processing unit can perform inter and TM blending prediction using coding information.

For example, the processing unit 1) inter prediction weight, 2) weight of TM optimal block, 3) inter prediction signal, 4) first inter prediction motion vector, 5) second inter prediction motion vector, 6) first inter prediction motion vector. Prediction reference image, 7) second inter prediction reference image, 8) first reference image list, 9) second reference image list, 10) weight of first inter prediction reference image, 11) weight of second inter prediction reference image , 12) first TM reference image, 13) second TM reference image, 14) first TM motion vector, 15) second TM motion vector, 16) first TM optimal block, 17) second TM optimal block, 18 ) Blending prediction of inter and TM can be performed using one or more of the weight of the first TM reference image, 19) the weight of the second TM reference image, 20) inter_tm_sharing_ref_flag, 21) p_inter_template_matching_blending_mode, 22) b_inter_template_matching_blending_mode, and 23) b_inter_bi_template_matching_blending_mode. there is.

Step 2045 may include one or more of steps 2046, 2047, and 2048.

At step 2046, the processor may use blending prediction of P inter and TM.

At step 2047, the processor may use blending prediction of B inter and TM.

At step 2048, the processor may use blending prediction of B-inter and B-TM.

Whether to perform at least some of steps 2010, 2020, 2030, and 2040 may be determined based on coding information.

For example, whether to perform at least part of steps 2010, 2020, 2030, and 2040 depends on 1) information about the picture, 2) information about the slice, 3) information about the tile, 4) quantization parameters. Parameter; QP), 5) Coded Block Flag (CBF), 6) Block size, 7) Block depth, 8) Block shape, 9) Entropy encoding/decoding method, 10) Block shape (For example, whether the block is square or whether the block is rectangular) and 11) the temporal layer level.

In embodiments, the target block is a Coding Tree Block (CTB), a Macro Block (MB), a Coding Block (CB), a Prediction Block (PB), and a Transform Block. ; TB), a Virtual Pipeline Data Unit (VPDU), and a block with a predetermined size.

Create template

Below, template creation in step 2010 is described.

Figure 21 shows areas of a target block according to an example.

A template with a specific size and specific shape can be created using pixel values of decoded pixels surrounding the target block.

In embodiments, the surrounding pixels of the target block may be pixels located at a location that satisfies a specific condition related to the target block. For example, the surrounding pixels of the target block may be pixels in the upper left area, upper area, and/or left area adjacent to the target block. For example, pixels surrounding the target block may be pixels whose x-axis distance and/or y-axis distance from the target block are less than or equal to a specific value. The surrounding pixels of the target block may be limited to pixels adjacent to the target block.

In embodiments, “a second object surrounding a first object” means “a second object within an area that satisfies a specific condition related to the first object,” for example, the area that satisfies the specific condition is relative to the target block. It may mean a specific area. For example, a specific area may represent an area described in an embodiment. For example, a specific area may be an area with a specific horizontal length and a specific vertical length. For example, the specific area may be the area of one unit including the first object.

The upper-left peripheral pixels of the target block may refer to pixels in the upper-left area adjacent (diagonally) to the upper-left side of the target block. The upper peripheral pixels of the target block may refer to pixels in the upper area adjacent to the upper side of the target block. The left peripheral pixels of the target block may refer to pixels in the left area adjacent to the left side of the target block.

Additionally, a template can be created using surrounding coding information of the target block.

In an embodiment, the surrounding coding information of the target block may be coding parameters of surrounding pixels of the target block.

For example, in generating a template, at least one of coding information such as a motion vector and an intra prediction mode (or intra prediction direction) may be used instead of a pixel. In other words, in embodiments, using a pixel to create a template may mean that information about the pixel is used to create a template. Pixel information may include a pixel value of the pixel and a coding parameter for the pixel. Coding information may include a motion vector and an intra prediction mode (or intra prediction direction). The coding parameter for a pixel may mean the location of the pixel or a coding parameter set for a block including the pixel.

As shown in FIG. 21, a template may be created using decoded pixels in at least one of the upper area, left area, and upper left area adjacent to the target block.

The upper area adjacent to the target block may be a rectangular area adjacent to the top surface of the target block. The left area adjacent to the target block may be a rectangular area adjacent to the left end face of the target block. The upper left area adjacent to the target block may be a rectangular area (diagonally) adjacent to the upper left corner of the target block.

As shown in Figure 21, the horizontal size of the target block (= W), the vertical size of the target block (= H), the vertical size of the upper area adjacent to the target block (= HL), and the horizontal size of the left area adjacent to the target block. The size of the template may be determined by at least one of (= WL).

The area of the template may include an upper area, a left area, and an upper left area.

The upper area may correspond to the upper template. The upper template may be a template adjacent to the top surface of the target block. The upper template may be named the upper adjacent template. The upper template area or the upper adjacent template area may refer to the area of the upper template.

A template consisting of an upper area, a left area, and an upper left area may be referred to as an L-shape template.

The left area may correspond to the left template. The left template may be a template adjacent to the left end face of the target block. The left template may be named the left adjacent template. The left template area or the left adjacent template area may mean the area of the left template.

The top left area may correspond to the top left template. The top left template may be a template adjacent (diagonally) to the top left corner of the target block. The top left template may be named the top left adjacent template. The upper left template area or the upper left adjacent template area may refer to the area of the upper left template.

Here, that a specific area and a specific template correspond may indicate that pixel(s) within the specific area are used to generate a specific template.

Using a pixel to create a template may mean that the pixel is used as part of a template.

As shown in FIG. 21, a template can be created using decoded pixels in the upper area, left area, and upper left area adjacent to a W×H target block with a horizontal size of W and a vertical size of H.

In embodiments, “W×H target block” may represent a target block with a horizontal size of W and a vertical size of H.

The horizontal size of the upper area may be W, and the vertical size may be HL.

The horizontal size of the left area may be WL, and the vertical size may be H.

The horizontal size of the upper left area may be WL, and the vertical size may be HL.

Create a template using some pixels

Figure 22 shows template creation using some pixels according to an example.

A template may be generated using some of the pixels among the decoded pixels in at least one of the upper area, left area, and upper left area adjacent to the target block.

As shown in FIG. 22, n _top pixels among the W×HL pixels in the upper area 1) adjacent to a W×H target block with a horizontal size of W and a vertical size of H (here, n _{top < W ​} Here, a template can be created including at least one of n _{top_left} < WL×HL).

The number and positions of some pixels used as a template may be determined based on at least one of the horizontal size W of the target block, the vertical size H of the target block, the horizontal thickness WL of the template, and the vertical thickness HL of the template.

Creating a template using a fraction of pixels: multiples of pixels

Figure 23 shows template creation using multiple pixels according to an example.

“Template pixels” supported in the figures may represent pixels used to create a template.

A template may be created using specific pixels among pixels in a specific area adjacent to the W×H target block.

Specific pixels are 1) a pixel whose horizontal position is a multiple of α among the W ) Among the WL×HL pixels in the upper left area, it may include at least one pixel whose position is a multiple of γ. Here, the location may mean a horizontal location and/or a vertical location.

α, β and γ may be determined by at least one of W, H, WL and HL.

As shown in FIG. 23, among the 8×2 pixels in the upper area adjacent to the 8×8 target block, 4 pixels whose horizontal positions are multiples of 3, and among the 2×8 pixels in the left area, the vertical pixels A template may be created using at least one of four pixels whose directional positions are a multiple of 3 and two pixels whose positions are a multiple of 2 among the 2×2 pixels in the upper left area.

For example, α could be [W/3].

For example, β may be [H/3].

For example, γ may be [WL/2].

Figure 24 illustrates template creation using multiple pixels according to an example.

As shown in FIG. 24, among the 8×2 pixels in the upper area adjacent to the 8×8 target block, 4 pixels whose horizontal positions are multiples of 2 (that is, 4 pixels whose horizontal positions are even numbers) pixels), among the 2×8 pixels in the left area, 4 pixels whose vertical positions are multiples of 2 (that is, 4 pixels whose vertical positions are even numbers) and 2×2 in the upper left area A template may be created using at least one of the two pixels whose positions are a multiple of 2 (that is, two pixels whose positions are even numbers).

Figure 25 shows template creation using multiple pixels according to one example.

As shown in FIG. 25, among the 8×2 pixels in the upper area adjacent to the 8×8 target block, 4 pixels whose horizontal positions are multiples of 2 (that is, 4 pixels whose horizontal positions are even numbers) pixels), among the 2×8 pixels in the left area, 4 pixels whose vertical positions are multiples of 2 (that is, 4 pixels whose vertical positions are even numbers) and 2×2 in the upper left area A template may be created using at least one of the pixels located on a diagonal line (that is, two pixels whose horizontal and vertical positions are the same).

Create a template where different starting position pixels and subsampling are used for the lines

The starting position pixel and subsampling used to create the template may be determined depending on the position of the line.

In embodiments, the first closest (horizontal/vertical) line to the target block may represent the (horizontal/vertical) line adjacent to the target block. The n-th adjacent horizontal line to the target block may represent a line adjacent to the upper side of the n-1-th adjacent horizontal line. The n-th adjacent vertical line to the target block may represent a line adjacent to the left of the n-1-th adjacent vertical line. n may be an integer of 2 or more.

For example, among the _W pixels, 2) a starting position pixel in the horizontal line 2nd closest from the target block and pixels whose horizontal positions from the starting position pixel are multiples of α ₂ , and 3) a starting position pixel in the horizontal line i th adjacent to the target block A template may be generated including at least one of pixels whose horizontal position from the start position pixel is a multiple of α _i . i may be an integer greater than or equal to 1 and less than or equal to HL.

For example, among the _WL pixels, 2) a starting position pixel in the vertical line 2nd closest from the target block and pixels whose vertical positions from the starting position pixel are multiples of α ₂ , and 3) a starting position pixel in the vertical line i th adjacent to the target block A template may be generated including at least one of pixels whose vertical position from the start position pixel is a multiple of β _i . i may be an integer greater than or equal to 1 and less than or equal to WL.

For example, among the _WL pixels that are multiples, 2) pixels in the starting position pixel in the horizontal line 2nd adjacent to the target block and pixels whose horizontal position from the starting position pixel is a multiple of γ ₂ , and 3) in the surrounding horizontal line i th adjacent to the target block. A template may be generated including at least one of a start position pixel and pixels whose horizontal positions from the start position pixel are multiples of γ _i . i may be an integer greater than or equal to 1 and less than or equal to HL.

For example, among the _WL pixels, 2) pixels in the starting position pixel in the 2nd adjacent peripheral vertical line from the target block and pixels whose positions in the vertical direction from the starting position pixel are multiples of δ ₂ , and 3) in the vertical line i th adjacent to the target block. A template may be generated including at least one of a start position pixel and pixels whose vertical positions from the start position pixel are multiples of δ. i may be an integer greater than or equal to 1 and less than or equal to HL.

Here, α _i , β _i , γ _i and δ _i may be determined based on at least one of H, L, WL and HL.

In embodiments, the starting position pixel within the line may be the first pixel among the pixels within the line used for creation of the template. The starting position pixel within the line may be the first pixel among pixels used to create the template within the line. The starting position pixel within the line may be the pixel with the smallest index among the pixels used to create the template within the line. The index can be a horizontal or vertical position.

In embodiments, the pixel at the first position within a line may be the first pixel among the pixels within the line. The pixel at the first position within the horizontal line may be the leftmost pixel among the pixels within the horizontal line. The pixel at the nth position within the horizontal line may be a pixel adjacent to the right of the pixel at the n-1 nth position. The pixel at the first position within the vertical line may be the uppermost pixel among the pixels within the vertical line. The pixel at the nth position within the vertical line may be a pixel adjacent to the lower side of the pixel at the n-1th position. n may be an integer of 2 or more. That is, the position of a pixel within a line may indicate the index of the pixel within the line.

Figure 26 shows template creation in which different starting position pixels and subsampling are used for lines according to an example.

For example, as shown in FIG. 26, among the 8×2 pixels in the upper area adjacent to the 8×8 target block, 1) the pixel at the first position within the horizontal line 1st adjacent to the target block and the pixel at the first position pixels whose horizontal positions from the pixel are multiples of 2, 2) pixels at the second position within the surrounding horizontal line second closest to the target block, and pixels whose horizontal positions from the pixel at the second position are multiples of 2 A template may be created including at least one of the following.

For example, as shown in FIG. 26, among the 2×8 pixels in the left area adjacent to the 8×8 target block, 1) the pixel at the first position within the vertical line 1st adjacent to the target block and the pixel at the first position pixels whose vertical positions from the pixel are multiples of 2, 2) pixels at the second position within the vertical line second closest to the target block, and pixels whose vertical positions from the pixel at the second position are multiples of 2 A template may be created including at least one of:

For example, as shown in FIG. 26, among the 2×2 pixels in the upper left area adjacent to the 8×8 target block, 1) the pixel at the second position within the horizontal line 1st adjacent to the target block, and 2) the target A template may be generated including at least one of the pixels at the first position within the second adjacent horizontal line from the block.

Exclusion of specific pixels from template creation

Figure 27 shows target blocks, regions, and CTU boundaries according to an example.

If a specific area adjacent to the target block is located outside a specific boundary, pixels within the specific area may not be used to generate the template.

For example, the specific area may include an upper left area and an upper left area.

The specific boundary may be the boundary of a specific unit described in the embodiments. For example, a specific unit may be at least one of a picture, slice, tile, and CTU.

The other area may be an area other than a specific area among areas adjacent to the target block. Even if other areas adjacent to the target block are located outside the specific boundary, pixels within the other areas can be used to generate the template.

As shown in FIG. 27, when the target block is adjacent to at least one of a picture boundary, a slice boundary, a tile boundary, and a CTU boundary, pixels outside the boundary may not be used as a template.

In one embodiment, if the top surrounding pixel and top left surrounding pixel of the target block are outside the top border of the CTU, the top surrounding pixel and top left surrounding pixel may not be used to generate the template. At this time, only the left peripheral pixels of the target block can be used to create the template.

Even if the left peripheral pixel of the target block is outside the left boundary of the CTU, a template can be created using the left peripheral pixel.

For example, the specific area may include an upper left area, an upper area, and a left area.

When the upper neighboring pixels of the target block are outside the upper boundary of the CTU, the vertical size of the upper adjacent template area of the target block may be defined as HL_OUTCTU. In addition, when the upper neighboring pixels of the target block are not outside the upper border of the CTU, the vertical size of the upper adjacent template area of the target block may be defined as HL_INCTU. At this time, HL_OUTCTU may be smaller than HL_INCTU. Each of HL_OUTCTU and HL_INCTU may be an integer greater than or equal to 0.

Pixels included in HL_OUTCTU in the upper area of the template may be excluded when creating the template. Pixels included in HL_INCTU in the upper area of the template may be included when creating the template. Here, pixel may mean pixel information such as motion vector, intra prediction mode, and coding parameters.

When the left peripheral pixel of the target block is outside the left border of the CTU, the horizontal size of the left adjacent template area of the target block can be defined as WL_OUTCTU. In addition, when the left peripheral pixel of the target block is not outside the left boundary of the CTU, the horizontal size of the left adjacent template area of the target block can be defined as WL_INCTU. At this time, WL_OUTCTU may be smaller than WL_INCTU. Each of WL_OUTCTU and WL_INCTU may be an integer greater than or equal to 0.

Pixels included in WL_OUTCTU in the left area of the template may be excluded when creating the template. Pixels included in WL_INCTU in the left area of the template may be included when creating the template. Here, pixel may mean pixel information such as motion vector, intra prediction mode, and coding parameters.

Multiple template shapes

Figure 28 shows areas of a target block according to an example.

FIGS. 29A to 29G show a plurality of templates of different target blocks as an example.

Figure 29a shows the configuration of a first template according to an example.

Figure 29b shows the configuration of a second template according to an example.

Figure 29c shows the configuration of a third template according to an example.

Figure 29D shows the configuration of a fourth template according to an example.

Figure 29e shows the configuration of a fifth template according to an example.

Figure 29f shows the configuration of a sixth template according to an example.

Figure 29g shows the configuration of a seventh template according to an example.

A plurality of templates may be created using surrounding pixels of the target block. Template identification information indicating which template is used among a plurality of available templates may be signaled as coding information. The processing unit may use template identification information to identify a template used for encoding/decoding among a plurality of templates.

A plurality of templates may be created according to the shape of the template using surrounding pixels of the target block. Template shape identification information indicating which template shape among available shapes is used may be signaled as coding information. The processing unit may use the template shape identification information to identify a template used for encoding/decoding among a plurality of templates each having different shapes.

As shown in FIG. 28, a plurality of templates may each be created using at least one of the upper left area, upper area, and left area adjacent to the target block. Each of the upper left area, top area, and left area may have a specific size.

The plurality of available templates may include one or more of a first template, a second template, a third template, a fourth template, a fifth template, a sixth template, and a seventh template.

As shown in FIG. 29A, the first template can be created using the upper left area, upper area, and left area.

As shown in FIG. 29B, the second template can be created using the upper left area.

As shown in FIG. 29C, the third template can be created using the upper left area and the upper layer area.

As shown in FIG. 29D, the fourth template can be created using the upper left area and the left area.

As shown in FIG. 29E, the fifth template can be created using the upper area.

As shown in FIG. 29f, the sixth template can be created using the left area.

As shown in FIG. 29g, the seventh template can be created using the upper area and the left area.

Multiple subsamplings for a template

Figure 30 shows samples of areas of a target block according to an example.

31A shows samples for a first template according to an example.

31B shows samples for a second template according to one example.

Figure 31C shows samples for a third template according to an example.

And a plurality of templates may be generated by subsampling using surrounding pixels of the target block. Template subsampling identification information indicating which template to which subsampling is applied among available subsamplings is used may be signaled as coding information. The processing unit may use template subsampling identification information to identify a template used for encoding/decoding among a plurality of templates to which different subsamplings are applied.

The template may have a shape configured to include at least one of an upper left area, an upper area, and a left area adjacent to the target block.

In embodiments, the shape of the template may refer to the area of the template.

A first template can be created using all pixels belonging to the shape of the template.

A second template may be created using first partial pixels among pixels belonging to the shape of the template. The first partial pixels may be some of the total pixels.

A third template may be created using second partial pixels among pixels belonging to the shape of the template. The second partial pixels may be some of the total pixels.

The first partial pixels and second partial pixels may be different.

As shown in FIG. 30, the template may have a shape configured to include at least one of an upper left area, an upper area, and a left area adjacent to the target block.

As shown in FIG. 31A, a first template can be created using all pixels belonging to the shape of the template. The first template may include all pixels belonging to the shape of the template.

As shown in FIG. 31B, a second template may be created using first partial pixels among pixels belonging to the shape of the template. The second template may include first partial pixels.

The first partial pixels may be pixels selected from pixels-pairs.

In embodiments, a pixel-pair may be two adjacent pixels. The pixel-pair may be the 2n-1th pixel and the 2nth pixel in the horizontal line or vertical line in the upper left area. The pixel-pair may be the 2n-1th pixel and the 2nth pixel in the horizontal line in the upper region. The pixel-pair may be the 2n-1th pixel and the 2nth pixel in the vertical line in the left area. n may be an integer greater than or equal to 1.

For example, the first partial pixels are 1) the 2nth pixels in the horizontal line in the upper left area, 2) the 2n-1th pixels in the horizontal line in the upper left area, and 3) the 2n-1th pixels in the vertical line in the upper left area. May contain pixels. n may be an integer greater than or equal to 1.

As shown in FIG. 31C, a third template may be created using second partial pixels among pixels belonging to the shape of the template. The third template may include second partial pixels.

The second partial pixels may be pixels selected from pixels-pairs.

For example, the second partial pixels are 1) the 2n-th pixels in the horizontal line 2m-1th adjacent to the target block in the upper left area, 2) the 2n-1th pixels in the horizontal line 2m-th adjacent to the target block in the upper left area. pixels, 3) 2n-1th pixels in the horizontal line 2m-1th adjacent to the target block in the upper area, 4) 2nth pixels in the horizontal line 2m-th adjacent to the target block in the upper area, 5) in the left area It may include 2n-1th pixels in a vertical line 2m-1th adjacent to the target block, and 6) 2nth pixels in a vertical line 2m-th adjacent to the target block in the left area. m may be an integer of 1 or more. n may be an integer greater than or equal to 1.

The embodiment described with reference to FIG. 28 and the embodiment described with reference to FIG. 29 may be combined.

Using surrounding pixels of the target block, a plurality of templates may be created according to the shapes of the template and subsampling applied to the template. That is, the shapes and subsamplings of the plurality of available templates may be different.

The template identification information may indicate the shape and subsampling used for the template among the shapes of the embodiments and the subsamplings of the embodiments.

Signaling of an index indicating one of multiple templates

Figure 32 is a table showing indices indicating templates according to an example.

Template identification information, template shape identification information, and template subsampling identification information may be signaled as an index.

As shown in FIG. 32, N templates can be created using surrounding pixels of the target block, and template_idx, an index indicating which template among the N templates is used, can be signaled.

For example, if template_idx is 0, it may be signaled that encoding/decoding is performed using the first template.

For example, if template_idx is 1, it may be signaled that encoding/decoding is performed using the second template.

For example, if template_idx is i, it may be signaled that encoding/decoding is performed using the i+1th template. i may be an integer between 0 and N-1.

For example, if template_idx is N-1, it may be signaled that encoding/decoding is performed using the N-th template.

In one embodiment, num_of_templates, which is information indicating the total number of templates allowed for a specific high-level syntax element (i.e., N in the table of FIG. 32), may be signaled. Specific high-level syntax elements may include syntax elements, units, and coding information described in the embodiments. For example, a specific high-level syntax element may include at least one of a sequence parameter, a picture parameter, and a slice header.

Alternatively, num_of_templates may represent “the total number of templates allowed for a specific high-level syntax element - 1” (i.e., N-1 in the table of FIG. 32).

Creating a template with padding

Figure 33 shows the creation of a template using padding according to an example.

If a pixel belonging to the template of the target block is located outside a specific boundary or is encoded/decoded in a specific mode, the pixel belonging to the template may not be usable.

The specific boundary may be the boundary of a specific unit described in the embodiments. For example, a specific unit may be at least one of a picture, slice, tile, and CTU.

The specific mode may be at least one of intra mode and inter mode.

These unusable pixels may be referred to as invalid pixels.

If the pixels belonging to the template are not available, you can create a template using padding.

In embodiments, padding may mean that reference pixels used for padding are used to create a template, replacing invalid pixels. By padding, information on reference pixels can be used as information on invalid pixels.

If invalid pixels exist in the template, padding may be performed on the invalid pixels, as shown in FIG. 33.

(1) For an invalid pixel in the target line that is first adjacent to the target block, padding may be performed using the reference pixel closest to the invalid pixel in the reference line that is first adjacent to the target block. For example, the reference pixel closest to the invalid pixel may be R ₁ in FIG. 33 .

The target line may represent a line containing invalid pixels to which padding is applied. The target line may be a line in the upper left area, upper area, or left area.

For example, a reference line may represent a line containing reference pixels used for padding. The reference line may include a horizontal/vertical line in the upper left area, a horizontal line in the upper left area, and a vertical line in the left area. The i th proximate reference line may include the i th proximate horizontal/vertical line in the upper left area, the i th proximate horizontal line in the upper left area, and the i th proximate numerical line in the left area.

For example, the i-th adjacent target line may represent pixels whose distance from the target block is i.

For example, the i-th adjacent reference line may represent pixels whose distance from the target block is i.

In embodiments, the distance between a pixel and the target block may be the larger of 1) the horizontal distance between the pixel and the target block and 2) the vertical distance between the pixel and the target block. Alternatively, the distance between the pixel and the target block may be the larger of the horizontal length and vertical length of the shortest straight line between the pixel and the target block.

(1) For an invalid pixel in the i-th adjacent target line to the target block, padding may be performed using the reference pixel closest to the invalid pixel in the i-th adjacent reference line to the target block. For example, when i is 2, the reference pixel closest to the invalid pixel may be R ₂ in FIG. 33.

Flipping templates

Figure 34a shows a template before flipping is applied according to an example.

Figure 34b shows a template after flipping is applied according to an example.

Specific flipping may be applied to specific areas of the template. Specific flipping may include left-right symmetrical flipping and vertically symmetrical flipping.

For example, a template can be created by applying left-right symmetrical flipping to the upper area of the template of the target block.

For example, a template can be created by applying vertically symmetrical flipping to the left area of the template of the target block.

As shown in Figure 34a, when there is a target block, an upper area of the template, and an area to the left of the template, left-right symmetrical flipping of the samples in the upper area and vertically symmetrical flipping of the samples in the left area are performed. By applying this, a template to which flipping is applied as shown in FIG. 34B can be created.

The syntax element template_vertical_flipping may indicate whether left-right symmetrical flipping is applied to the upper area of the target block. template_vertical_flipping may be signaled to indicate whether a template is created using left-right symmetrical flipping for the upper area of the target block.

For example, when template_vertical_flipping is the first value, a template can be created by applying left-right symmetrical flipping to the upper area of the target block. If template_vertical_flipping is the second value, a template can be created using the upper area of the target block as is (that is, flipping may not be applied to the upper area).

The syntax element template_horizontal_flipping may indicate whether vertically symmetrical flipping is applied to the left area of the target block. template_horizontal_flipping may be signaled to indicate whether a template is created using vertically symmetrical flipping for the left area of the target block.

For example, when template_horizontal_flipping is the first value, a template can be created by applying vertically symmetrical flipping to the left area of the target block. If template_horizontal_flipping is the second value, a template can be created using the left area of the target block as is (i.e., flipping may not be applied to the left area).

The syntax element template_flipping_flag may indicate whether flipping is applied to surrounding samples of the target block. template_flipping_flag may be signaled to indicate whether a template is generated using flipping for surrounding samples of the target block.

For example, when template_flipping_flag is the first value, a template can be created by applying flipping to samples surrounding the target block. If template_flipping_flag is the second value, a template can be created using the surrounding samples of the target block as is (that is, flipping may not be applied to the surrounding samples).

When template_flipping_flag is the first value, a syntax element template_flipping_vertical_flag indicating one of horizontally symmetric flipping and vertically symmetrical flipping may be signaled.

For example, if template_flipping_flag is the first value, template_flipping_vertical_flag may be signaled. The fact that template_flipping_vertical_flag is the first value may indicate that left-right symmetrical flipping is applied to the upper area. The second value of template_flipping_vertical_flag may indicate that vertically symmetrical flipping is applied to the left area.

Template matching reference image selection

Below, the selection of the template matching reference image in step 2020 is described.

Template matching reference video list and index

The template matching reference image may be an image referenced in prediction using template matching.

The template matching reference image can be determined and signaled implicitly or explicitly.

A list containing template matching reference images may be constructed. The template matching reference video list (= template_matching_ref_list) may be a list containing template matching reference videos.

The template matching reference image list may include at least one of a target image, a decoded temporal past image, and a decoded temporal future image.

In embodiments, the target image may be an image including the target block. The decrypted temporally past image may be a decoded image of a time earlier than the time of the target image. The decoded temporal future image may be a decoded image of a later time than the time of the target image.

The template matching reference image list may include a plurality of template matching reference images. Each template matching reference image of the plurality of template matching reference images in the template matching reference image list may have a unique index. The template matching reference index (= template_matching_ref_idx) may be an index for a plurality of template matching reference images in the template matching reference image list.

template_matching_ref_idx may be signaled to indicate a specific template matching reference image among a plurality of template matching reference images in the template matching reference image list.

template_matching_ref_idx can specify the index of the template matching reference image used for the target block among the template matching reference images belonging to the template matching reference image list.

A plurality of template matching reference image lists may be configured. For example, the plurality of template matching reference image lists may include a template matching reference image list L0 (= template_matching_ref_l0) and a template matching reference image list L1 (= template_matching_ref_l1).

The template matching reference image list L0 may include at least one of a target image, a decoded temporal past image, and a decoded temporal future image.

The template matching reference image list L0 may include a plurality of template matching reference images. Each template matching reference image of the plurality of template matching reference images in the template matching reference image list L0 may have a unique index. The template matching reference index L0 (= template_matching_ref_idx_l0) may be an index for a plurality of template matching reference images in the template matching reference image list L0.

The template matching reference image list L1 may include at least one of a target image, a decoded temporal past image, and a decoded temporal future image.

The template matching reference image list L1 may include a plurality of template matching reference images. Each template matching reference image of the plurality of template matching reference images in the template matching reference image list L1 may have a unique index. The template matching reference index L1 (= template_matching_ref_idx_l1) may be an index for a plurality of template matching reference images in the template matching reference image list L1.

At least one of template_matching_ref_idx_l0 and template_matching_ref_idx_l1 may be signaled. template_matching_ref_idx_l0 may specify the index of the template matching reference image used for the target block among the template matching reference images belonging to the template matching reference image list L0. template_matching_ref_idx_l1 may specify the index of the template matching reference image used for the target block among the template matching reference images belonging to the template matching reference image list L1.

Selection of template matching reference image

If information indicating a template matching reference image is not signaled separately, an image determined by a specific method may be used as a template matching reference image. In other words, when the template matching reference image is the same as the image determined by a specific method, signaling of information indicating the template matching reference image may be omitted.

For example, if information indicating a template matching reference image is not signaled separately, the target image may be used as a template matching reference image.

For example, if information indicating the template matching reference image is not signaled separately, the template matching reference image may be the image that is temporally closest to the target image among the decoded images.

For example, if information indicating a template matching reference image is not signaled separately, the template matching reference image is the past image temporally closest to the target block among the decoded images and the past image temporally closest to the target block among the decoded images. This could be at least one of the near-future images.

In one embodiment, the template matching reference image may be determined by a predefined method in the encoding device 1600 and the decoding device 1700 based on coding information.

For example, a template matching reference image may be determined based on the coding mode of the target block.

For example, a template matching reference image may be determined based on the slice type of the slice containing the target block.

Signaling of the index to the template matching reference image list containing the target image

FIG. 35 may show a unique index of each candidate template matching reference image belonging to a template matching reference image list according to an example.

In embodiments, P(t) may represent the target image at time point t.

In embodiments, P(tn ₁ ), P(tn ₂ ), and P(tn ₃ ) may represent decoded past images for the target image at time t. Each of n ₁ , n ₂ and n ₃ may be an integer of 1 or more. n ₁ , n ₂ and n ₃ may be different from each other. The decoded past image for the target image at time t can be summarized as the decoded past image.

In embodiments, P(t+n ₄ ), P(t+n ₅ ), and P(t+n ₆ ) may represent decoded future images of the target image at time t. Each of n ₄ , n ₅ and n ₆ may be an integer of 1 or more. n ₄ , n ₅ and n ₆ may be different from each other. The decrypted future image for the target image at time t can be outlined as the decoded future image.

The template matching reference image list (= template_matching_ref_list) may be configured to include a target image at time t, one or more decoded past images, and one or more decoded future images.

Here, each of n ₁ , n ₂ , n ₃ , n ₄ , n ₅ and n ₆ may have a predefined value. For example, values can be determined as in Equation 1 below.

[Formula 1]

n ₁ = 1, n ₂ = 2, n ₃ = 3, n ₄ = 1, n ₅ = 2, n ₆ = 3

The template matching reference video list (= template_matching_ref_list) can be configured as shown in Equation 2 below.

[Formula 2]

template_matching_ref_list = {P(t), P(tn ₁ ), P(t+n ₄ ), P(tn ₂ ), P(t+n ₆ )}

As shown in FIG. 35, each candidate template matching reference image belonging to the template matching reference image list may have a unique index.

template_matching_ref_idx indicating the template matching reference image of the target block may be signaled.

Signaling of index for template matching reference image list that does not contain target image

FIG. 36 may show a unique index of each candidate template matching reference image belonging to a template matching reference image list that does not include the target image according to an example.

A template matching reference image list (= template_matching_ref_list) may be configured to include one or more decoded past images and one or more decoded future images. In other words, the target image at time t may not be included in the template matching reference image list.

Here, each of n ₁ , n ₂ , n ₃ , n ₄ , n ₅ and n ₆ may have a predefined value. For example, values can be determined as in Equation 3 below.

[Formula 3]

n ₁ = 1, n ₂ = 2, n ₃ = 3, n ₄ = 1, n ₅ = 2, n ₆ = 3

The template matching reference video list (= template_matching_ref_list) can be configured as shown in Equation 4 below.

[Formula 4]

template_matching_ref_list = {P(tn ₁ ), P(t+n ₄ ), P(tn ₂ ), P(t+n ₆ )}

As shown in FIG. 36, each candidate template matching reference image belonging to the template matching reference image list may have a unique index.

template_matching_ref_idx indicating the template matching reference image of the target block may be signaled.

Signaling of indices for a plurality of template matching reference image lists including the target image

FIG. 37A may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L0 according to an example.

FIG. 37B may show a unique index of each candidate template matching reference image belonging to the template matching reference image list L1 according to an example.

In embodiments, the plurality of template matching reference image lists including the target image may mean that at least one of the plurality of template matching reference image lists includes the target image.

Each of the template matching reference image list L0 (= template_matching_ref_list_l0) and the template matching reference image list L1 (= template_matching_ref_list_l1) may be configured to include a target image at time t, one or more decoded past images, and one or more decoded future images. there is.

Here, each of n ₁ , n ₂ , n ₃ , n ₄ , n ₅ and n ₆ may have a predefined value. For example, values can be determined as in Equation 5 below.

[Formula 5]

n ₁ = 1, n ₂ = 2, n ₃ = 3, n ₄ = 1, n ₅ = 2, n ₆ = 3

The template matching reference video list L0 (= template_matching_ref_list_l0) can be configured as shown in Equation 6 below.

[Formula 6]

template_matching_ref_list_l0 = {P(t), P(tn ₁ ), P(tn ₂ ), P(tn ₃ )}

The template matching reference video list L1 (= template_matching_ref_list_l1) can be configured as shown in Equation 7 below.

[Formula 7]

template_matching_ref_list_l1 = {P(t+n ₄ ), P(t+n ₅ ), P(t+n ₆ ), P(tn ₁ )}

As shown in FIGS. 37A and 37B, each candidate template matching reference image belonging to the template matching reference image list L0 or the template matching reference image list L1 may have a unique index.

At least one of template_matching_ref_idx_l0 and template_matching_ref_idx_l1 may be signaled to indicate the template matching reference image of the target block.

Signaling of indices for multiple template matching reference image lists that do not include the target image

FIG. 38A may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L0 that does not include the target image according to an example.

FIG. 38B may represent a unique index of each candidate template matching reference image belonging to the template matching reference image list L1 that does not include the target image according to an example.

In embodiments, a plurality of template matching reference image lists that do not include the target image may mean that the target image is not included in any template matching reference image list among the plurality of template matching reference image lists.

Each of the template matching reference video list L0 (= template_matching_ref_list_l0) and the template matching reference video list L1 (= template_matching_ref_list_l1) may be configured to include one or more decoded past videos and one or more decoded future videos. In other words, the target image at time t may not be included in the template matching reference image list L0 and may not be included in the template matching reference image list L1.

Here, each of n ₁ , n ₂ , n ₃ , n ₄ , n ₅ and n ₆ may have a predefined value. For example, values can be determined as in Equation 8 below.

[Formula 8]

n ₁ = 1, n ₂ = 2, n ₃ = 3, n ₄ = 1, n ₅ = 2, n ₆ = 3

The template matching reference video list L0 (= template_matching_ref_list_l0) can be configured as shown in Equation 9 below.

[Formula 9]

template_matching_ref_list_l0 = {P(tn ₁ ), P(tn ₂ ), P(tn ₃ ), P(t+n ₄ )}

The template matching reference video list L1 (= template_matching_ref_list_l1) can be configured as in Equation 10 below.

[Formula 10]

template_matching_ref_list_l1 = {P(t+n ₄ ), P(t+n ₅ ), P(t+n ₆ ), P(tn ₁ )}

As shown in FIGS. 38A and 38B, each candidate template matching reference image belonging to the template matching reference image list L0 or the template matching reference image list L1 may have a unique index.

At least one of template_matching_ref_idx_l0 and template_matching_ref_idx_l1 may be signaled to indicate the template matching reference image of the target block.

Default determination of template matching reference image

In one embodiment, when the coding mode of the target block is a mode that refers to the decoded area of the target image, the target image may be inferred as a template matching reference image.

In one embodiment, an image inferred as a template matching reference image may be determined by coding information about the target image or target block.

The image inferred as the template matching reference image may be determined based on the slice type of the target image or target block.

For example, if the target image or target block corresponds to an I slice, the target image may be inferred as a template matching reference image.

For example, if the target image or target block corresponds to a P slice or B slice, the image closest to the target image among the decoded images can be inferred as the template matching reference image.

For example, if the target image or target block corresponds to a P slice or B slice, the past image closest to the target image and the future image closest to the target image among the decoded images can be inferred as template matching reference images. .

For example, if the target image or target block corresponds to a P slice or B slice, the template matching reference image may be the first reference image in the first reference image list of inter prediction.

For example, if the target image or target block corresponds to a P slice or B slice, the template matching reference image may be the first reference image in the second reference image list of inter prediction.

For example, if the target image or target block corresponds to a P slice or B slice, the template matching reference images are the first reference image in the first reference image list and the first reference image in the second reference image list of inter prediction. You can.

Template matching exploration

Below, the selection of a template matching search in step 2030 is described.

Template matching search, correlation, definition of template matching optimal blocks and template matching motion vectors.

Figure 39 shows derivation of an optimal template matching block according to an example.

In embodiments, the target template may refer to the template of the target block. In other words, the target template may mean a template adjacent to the target block configured for the target block.

Given a target template and a template matching reference image, a search may be performed to find the optimal block (or optimal pixel) having a template most similar to the target template of the target block within the template matching reference image.

Given a target template and a template matching reference image, a search may be performed to find the optimal block (or optimal pixel) having a template most similar to the target template of the target block within the search area within the template matching reference image.

In embodiments, the search area may be a decoded area within a template matching reference image.

Here, the template of the optimal block may have the highest correlation with the target template. In other words, the correlation of the template of the optimal block with the target template may be the largest among the correlations with the target template of the searched templates.

In embodiments, for searched blocks within a search area, a correlation between the searched template and the target template of the target block may be calculated. Among the searched templates, the template with the highest correlation with the target template may be selected as the optimal template. A block corresponding to the optimal template may be selected as the optimal block. The relationship between the location of the optimal template and the location of the optimal block may correspond to the relationship between the location of the target template and the location of the target block.

This selected optimal block may be named a template matching optimal block.

In embodiments, the optimal template may be the template with the highest correlation with the target template. Additionally, the optimal template may be a template of the template matching optimal block.

In other words, through matching the target block with the target template, the optimal template with the highest correlation can be searched, and the block adjacent to the optimal template can be set as the template matching optimal block for encoding/decoding of the target block.

Derivation of template matching motion vector

The template matching motion vector (= MV _tm ) may represent the difference between the position of the target template and the position between the optimal templates.

Alternatively, the template matching motion vector may indicate the difference between the positions of the target blocks and the positions of the optimal template matching blocks.

Alternatively, the template matching motion vector may indicate the difference between the position of the leftmost and uppermost pixel in the target block and the position of the optimal pixel. The optimal pixel may be the top left pixel of the template matching optimal block.

In embodiments, the optimal pixel may be the pixel with the highest correlation with the target template.

Figure 40 shows a first method of deriving a template matching motion vector according to an example.

The position of the top left pixel within the target block may be (x _t , y _t ).

The position of the top left pixel in the template matching optimal block may be (x _tm , y _tm ).

In this case, the template matching motion vector (= MV _tm ) may represent the difference between the position of the top left pixel in the target block and the position of the top left pixel in the template matching optimal block.

For example, the template matching motion vector (= MV _tm ) may be at least one of the values of Equation 11 and Equation 12 below.

[Formula 11]

(x _t - x _tm , y _t - y _tm )

[Formula 12]

(x _tm - x _t , y _tm - y _t )

Figure 41 shows a second method of deriving a template matching motion vector according to an example.

The position of the top left pixel within the target template of the target block may be (x _t , y _t ).

The position of the top left pixel in the optimal template may be (x _tm , y _tm ).

In this case, the template matching motion vector (= MV _tm ) may represent the difference between the position of the top-left pixel in the target template and the position of the top-left pixel in the optimal template.

For example, the template matching motion vector (= MV _tm ) may be at least one of the values of Equation 13 and Equation 14 below.

[Formula 13]

(x _t - x _tm , y _t - y _tm )

[Formula 14]

(x _tm - x _t , y _tm - y _t )

Figure 42 shows a first method of deriving a motion vector when a template matching reference image and a target image are different from each other, according to an example.

Template matching reference image and target image may be different from each other. If the template matching reference image and the target image are different from each other, a position corresponding to the target block in the template matching reference image may be set. For example, the area occupied by the target block in the target image and the area specified by the corresponding position in the template matching reference image may be the same.

When the position of the top left pixel of the target block in the target image is (x _t , y _t ), the corresponding position in the template matching reference image may also be (x _t , y _t ).

The position of the top left pixel in the template matching optimal block may be (x _tm , y _tm ).

In this case, the template matching motion vector (= MV _tm ) may represent the difference between the position of the top left pixel in the template matching optimal block and the corresponding position within the template matching optimal block.

For example, the template matching motion vector (= MV _tm ) may be at least one of the values of Equation 15 and Equation 16 below.

[Formula 15]

(x _t - x _tm , y _t - y _tm )

[Formula 16]

(x _tm - x _t , y _tm - y _t )

Figure 43 shows a first method of deriving a motion vector when a template matching reference image and a target image are different from each other, according to an example.

Template matching reference image and target image may be different from each other. If the template matching reference image and the target image are different from each other, a position corresponding to the target block in the template matching reference image may be set. For example, the area occupied by the target template in the target image and the area specified by the corresponding position in the template matching reference image may be the same.

When the position of the top left pixel of the target template in the target image is (x _t , y _t ), the corresponding position in the template matching reference image may also be (x _t , y _t ).

The position of the top left pixel of the optimal template in the template matching reference image may be (x _tm , y _tm ).

In this case, the template matching motion vector (= MV _tm ) may represent the difference between the position of the top-left pixel in the target template and the position of the top-left pixel in the optimal template.

For example, the template matching motion vector (= MV _tm ) may be at least one of the values of Equation 17 and Equation 18 below.

[Formula 17]

(x _t - x _tm , y _t - y _tm )

[Formula 18]

(x _tm - x _t , y _tm - y _t )

Template matching navigation area specific

When a template and a template matching reference image are given, a template matching search area for searching for an optimal template matching block within the template matching reference image may be specified.

The template matching search area may be 1) the entire area of the template matching reference image or 2) a specific partial area of the template matching reference image.

The template matching search area may be an area composed of a specific number of blocks in the template matching reference image. In other words, an area other than the area of the above-mentioned specific number of blocks may not be a template matching search area. Accordingly, encoding/decoding using template matching may not be allowed for the remaining area where the area of a specific number of blocks surrounding the target block is excluded.

The block may be one of the blocks described in the embodiments. For example, blocks include Coding Tree Block (CTB), Macro Block (MB), Coding Block (CB), Prediction Block (PB), Transform Block; TB), a Virtual Pipeline Data Unit (VPDU), and a block with a predetermined size.

The undecrypted area, the target block, and the area of the target block's template may be excluded from the template matching search area.

The shape of the template matching search area may be one of the shapes described in the embodiments. For example, the shape of the template matching search area may be at least one of a square shape, a rectangular shape, and a diamond shape.

For example, the size of the rectangular template matching search area may be defined by the horizontal size and vertical size.

For example, the size of a diamond-shaped template matching search area can be defined by two diagonal sizes.

For example, the size of the template matching search area can be defined by the horizontal size and vertical size from the template matching search starting point.

At least one of the shape and size of the template matching search area may be determined based on coding information. At least one of the shape and size of the template matching search area may be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. Information indicating the shape and/or size of the template matching search area may be signaled within the parent syntax element.

Each of the information indicating the shape and/or size of the template matching search area and the coding information may be signaled for a specific unit. For example, specific units include Coding Tree Block (CTB), Macro Block (MB), Coding Block (CB), Prediction Block (PB), and Transform block (Transform). It may be one of Block (TB), Virtual Pipeline Data Unit (VPDU), and a block with a predetermined size.

Specific of template matching navigation area using specific units

In one embodiment, a VPDU may be used to specify a template matching search area.

The template matching search area may be an area of one VPDU surrounding the target block in the template matching reference image. In other words, an area other than the area of one VPDU above may not be a template matching search area. Accordingly, encoding/decoding using template matching may not be allowed in the remaining area excluding the area of one VPDU surrounding the target block.

In one embodiment, CTB may be used to specify a template matching search area.

The template matching search area may be an area of one CTB around the target block in the template matching reference image. In other words, areas other than the area of one CTB above may not be a template matching search area. Therefore, encoding/decoding using template matching may not be allowed in the remaining area excluding the area of one CTB surrounding the target block.

The template matching search area may be an area of one CTB to which the target block in the template matching reference image belongs. In other words, areas other than the area of one CTB above may not be a template matching search area. Therefore, encoding/decoding using template matching may not be allowed in the remaining area excluding the area of one CTB surrounding the target block.

Rectangular template matching navigation area

Figure 44 shows a case where the shape of the template matching search area in the target image is rectangular, according to an example.

In one embodiment, when the template matching reference image is the target image, the shape of the template matching search area may be rectangular.

The template matching search area may be the entire area or a partial area of the decoded area in the template matching reference image.

Figure 45 shows a case where the shape of the template matching search area in a decoded past image or a decoded future image according to an example is rectangular.

In one embodiment, when the template matching reference image is a decoded past image or a decoded future image, the shape of the template matching search area may be rectangular.

The template matching search area may be the entire area or a partial area of the decoded area in the template matching reference image.

Figure 46 shows that when a template matching reference image according to an example is the target image, some areas within the template matching search area within the template matching reference image are excluded from the search.

In one embodiment, when the template matching reference image is the target image, some areas within the template matching search area may be excluded from the search.

Some areas may include a target block, an undecoded area, and a template area of the target block.

FIG. 47 illustrates that when the template matching reference image according to an example is a decoded past image or a decoded future image, a portion of an area within the template matching search area within the template matching reference image is excluded from the search.

In one embodiment, when the template matching reference image is a decoded past image or a decoded future image, some areas within the template matching search area may be excluded from the search.

Some areas may include areas outside the boundaries of the image. The border may include one or more of an upper border and a left border.

Signaling of the size of the template matching search area when the shape of the template matching search area is rectangular.

Figure 48 shows a first method of defining the size of a template matching search area when the shape of the template matching search area is rectangular according to an example.

When the shape of the template matching search area is rectangular, the size of the template matching search area can be defined by the horizontal size (=rec_template_width) of the template matching search area and the vertical size (=rec_template_height) of the template matching search area.

At least one of rec_template_width and rec_template_height may be determined based on coding information. At least one of rec_template_width and rec_template_height may be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. rec_template_width and rec_template_height can be signaled within the parent syntax element.

Each of rec_template_width, rec_template_height, and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

Figure 49 shows a second method of defining the size of the template matching search area when the shape of the template matching search area is rectangular, according to an example.

When the shape of the template matching search area is rectangular, the size of the template matching search area can be defined by the horizontal size (= del_width) and the vertical size (= del_height). The horizontal size (= del_width) may represent the distance from the template matching search start point to the left border of the template matching search area and the distance from the template matching search start point to the right border of the template matching search area. The vertical size (= del_height) may represent the distance from the template matching search start point to the upper border of the template matching search area and the distance from the template matching search start point to the lower border of the template matching search area.

The template matching search start point may be located at the center of the template matching search area. Alternatively, the template matching search area can be defined by the horizontal size (= del_width) and vertical size (= del_height) so that the template matching search start point is the center.

At this time, the size of the template matching search area may be equal to Equation 19 below.

[Formula 19]

(1 + 2 * del_width) * (1 + 2 * del_height)

At least one of del_width and del_height may be determined based on coding information. At least one of del_width and del_height may be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. del_width and del_height can be signaled within the parent syntax element.

Each of del_width, del_height, and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

Rhombus template matching navigation area

Figure 50 shows a case where the shape of the template matching search area in the target image according to an example is a diamond.

In one embodiment, when the template matching reference image is the target image, the shape of the template matching search area may be a diamond.

The template matching search area may be the entire area or a partial area of the decoded area in the template matching reference image.

Figure 51 shows a case where the shape of the template matching search area in a decoded past image or a decoded future image according to an example is a diamond.

In one embodiment, when the template matching reference image is a decoded past image or a decoded future image, the shape of the template matching search area may be a diamond.

The template matching search area may be the entire area or a partial area of the decoded area in the template matching reference image.

Figure 52 illustrates that when a template matching reference image according to an example is the target image, some areas within the template matching search area within the template matching reference image are excluded from the search.

In one embodiment, when the template matching reference image is the target image, some areas within the template matching search area may be excluded from the search.

Some areas may include a target block, an undecoded area, and a template area of the target block.

FIG. 53 illustrates that when the template matching reference image according to an example is a decoded past image or a decoded future image, a portion of an area within the template matching search area within the template matching reference image is excluded from the search.

In one embodiment, when the template matching reference image is a decoded past image or a decoded future image, some areas within the template matching search area may be excluded from the search.

Some areas may include areas outside the boundaries of the image. The border may include one or more of an upper border and a left border.

Signaling of the size of the template matching search area when the shape of the template matching search area is a diamond.

Figure 54 illustrates a first method of defining the size of a template matching search area when the shape of the template matching search area is a diamond according to an example.

When the shape of the template matching search area is a diamond, the size of the template matching search area can be defined by the horizontal size (=dia_template_width) of the template matching search area and the vertical size (=dia_template_height) of the template matching search area. Here, the horizontal size may be the distance between the left and right vertices of the rhombus. The vertical size may be the distance between the top and bottom vertices of the rhombus.

At least one of dia_template_width and dia_template_height may be determined based on coding information. At least one of dia_template_width and dia_template_height may be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. dia_template_width and dia_template_height can be signaled within the parent syntax element.

Each of dia_template_width, dia_template_height, and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

Figure 55 shows a second method of defining the size of a template matching search area when the shape of the template matching search area is a diamond according to an example.

When the shape of the template matching search area is a diamond, the size of the template matching search area can be defined by the horizontal size (= del_width) and the vertical size (= del_height). The horizontal size (= del_width) may represent the distance from the template matching search start point to the left vertex of the template matching search area and the distance from the template matching search start point to the right vertex of the template matching search area. The vertical size (= del_height) may represent the distance from the template matching search start point to the upper vertex of the template matching search area and the distance from the template matching search start point to the lower vertex of the template matching search area.

The template matching search start point may be located at the center of the template matching search area. Alternatively, the template matching search area can be defined by the horizontal size (= del_width) and vertical size (= del_height) so that the template matching search start point is the center.

At this time, the size of the template matching search area may be equal to Equation 20 below.

[Formula 20]

1/2 * (1 + 2 * del_width) * (1 + 2 * del_height)

At least one of del_width and del_height may be determined based on coding information. At least one of del_width and del_height may be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. del_width and del_height can be signaled within the parent syntax element.

Each of del_width, del_height, and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

Signaling of the index that indicates the shape of the template matching search area

The shape of the template matching search area may be one of a plurality of shapes. The plurality of shapes may include rectangular and diamond shapes.

An index (= template_matching_shape_idx) that specifies the shape of the template matching search area may be signaled.

For example, template_matching_shape_idx being the first value may indicate that the shape of the template matching search area is a square. The fact that template_matching_shape_idx is the second value may indicate that the shape of the template matching search area is a diamond.

template_matching_shape_idx can be determined based on coding information. template_matching_shape_idx can be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. template_matching_shape_idx can be signaled within the parent syntax element.

Each of template_matching_shape_idx and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

Determination of template matching search start pixel using corresponding coordinates of template of target block

Figure 56 illustrates determination of a template matching search start pixel using corresponding coordinates of a template of a target block according to an example.

Given a template and a template matching reference image, a template matching search start pixel within the template matching reference image can be specified. The template matching search start pixel can be used to encode/decode the target block.

Here, the optimal pixel may represent an optimal block, a template matching optimal pixel, and a template matching optimal block.

The template matching search starting point may be determined using information on the coordinates of the template of the target block.

In embodiments, the top-left coordinates of a template may refer to the coordinates of the top-left pixel within the template.

When the top left coordinates in the template of the target block are (x _tm , y _tm ), the corresponding position in the template matching reference image may also be (x _tm , y _tm ).

In embodiments, the corresponding location may be a location in the template matching reference image that corresponds to the template of the target block. Alternatively, the corresponding position may be a position in the template matching reference image corresponding to the top left pixel of the template of the target block.

As the corresponding position is used as the template matching search start point, the template matching search start point of the target block can also be set to (x _tm , y _tm ).

Determination of template matching search start pixel using template matching coding block

In one embodiment, a template matching coding block among neighboring blocks of the target block may be used to determine the template matching search start pixel. A template matching coding block may refer to a block encoded/decoded using template matching.

If a template matching coding block exists among blocks adjacent to the target block, a template matching search starting point can be obtained from the template matching motion vector of the template matching coding block.

For example, if there is a template matching coding block among blocks adjacent to the target block, and the template matching motion vector of the template matching coding block is (MV _{tm_x} , MV _{tm_y} ), the template matching search starting point for the target block is ( It can be set as MV _{tm_x} , MV _{tm_y} ). That is, the template matching search start point of the target block may be set to be the same as the template matching motion vector of the template matching coding block.

For example, when the template of the target block is located at (x _tm , y _tm ), the corresponding position in the template matching reference image corresponding to the template of the target block may be (x _tm , y _tm ). For example, if there is a template matching coding block among blocks adjacent to the target block, and the template matching motion vector of the template matching coding block is (MV _{tm_x} , MV _{tm_y} ), the template matching search starting point for the target block is ( It can be set as x _tm + MV _{tm_x} , y _tm + MV _{tm_y} ).

Determination of template matching search start pixel using merge

The template matching search starting point can be determined using motion information belonging to the merge list of the target block.

The motion vector of a specific element in the merge list of the target block can be used as a template matching search starting point. A specific element may be the first element.

In embodiments, an element of the merge list may refer to a candidate for the merge list.

When the template of the target block is located at (x _tm , y _tm ), the corresponding position in the template matching reference image corresponding to the template of the target block may be (x _tm , y _tm ). If the motion vector of the first element in the merge list of the target block is (x _m1 , y _m1 ), the template matching search start point of the target block can be set to (x _tm + x _m1 , y _tm + y _m1 ). .

The template matching search starting point can be selected using the merge list of the target block. The method of determining motion information using a merge list described in the embodiments may be used to determine a template matching search starting point. For example, motion information determined using a merge list can be used as a template matching search starting point.

A merge index (= merge_index) may be signaled to specify the template matching starting point. merge_index can specify one of the merge index candidates. A template matching search start pixel may be set using the motion vector of the candidate specified by the signaled merge_index.

When the template of the target block is located at (x _tm , y _tm ), the corresponding position in the template matching reference image corresponding to the template of the target block may be (x _tm , y _tm ). If the motion vector specified by the signaled merge_index is (x _mi , y _mi ), the template matching search start point of the target block may be set to (x _tm + x _mi , y _tm + y _mi ).

Signaling of information to select template matching search starting point

Figure 57 is a table showing indices indicating a method of initializing a template matching search start pixel according to an example.

Template matching search start information may be information used to select a template matching search start point. In other words, the template matching search start information may indicate the template matching search starting point.

Template matching search start information may be determined based on coding information. Template matching search start information can be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. Template matching search start information may be signaled within the parent syntax element.

Each of the template matching search start information and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

To obtain the template matching search starting point of the target block, usage information indicating whether to use at least one of the coordinates of the target block, the merge list, and the coding information of the surrounding block (or adjacent block) may be determined.

Usage information may be determined based on coding information. Template matching search start information can be specified by signaling about coding information. For example, coding information may include high-level syntax elements such as sequence parameters, picture parameters, and slice headers. Usage information may be signaled within the parent syntax element.

Each of the usage information and coding information may be signaled for a specific unit. For example, a specific unit may be one of CTB, MB, CB, PB, TB, VPDU, and a block with a predetermined size.

The template matching initialization index (= template_matching_init_idx) may indicate one of a plurality of methods for initializing the template matching search start point. The template matching search start point can be initialized by the method indicated by the template matching initialization index.

template_matching_init_idx may indicate a method of initializing the template matching search start point as described in the table of FIG. 57.

For example, when template_matching_init_idx is the first value, the corresponding position in the template matching reference image corresponding to the template of the target block can be used as the template matching search starting point. If template_matching_init_idx is the second value, the template matching search start point can be obtained using the motion vector of the merge list. When template_matching_init_idx is the third value, the template matching search starting point can be obtained using coding information of a template matching coded block among adjacent blocks of the target block.

Signaling of template matching search method

Given a template, a template matching reference image, a template matching search area, and a template matching search start point, a template matching search method based on these templates, a template matching reference image, a template matching search area, and a template matching search start point may be determined.

For example, the template matching search method may be information indicating which pixels in the template matching search area and how to search.

Template matching search methods may include full search, 2D logarithmic search, N-step search, diamond search, hexagonal search, and test zone search. there is.

In an embodiment, global search may mean searching all pixels within the template matching search area.

A syntax element (= template_matching_search_idx) that specifies the template matching search method may be signaled.

For example, if template_matching_search_idx is the first value, global search may be performed. If template_matching_search_idx is the second value, 2D algebraic search may be performed. If template_matching_search_idx is the third value, N-step search may be performed. If template_matching_search_idx is the fourth value, a diamond search may be performed. If template_matching_search_idx is the fifth value, hexagon search may be performed. If template_matching_search_idx is the sixth value, test zone search may be performed.

Template matching correlation criteria

The correlation described in the embodiments may be calculated based on a cost function.

For example, the higher the correlation, the smaller the resulting cost function may be.

For example, cost functions include SAD, SATD, Mean Removed Sum of Absolute Differences (MR-SAD), Mean Squared Error (MSE), and Sum of Squared Errors ( Sum of Squared Errors; SSE). The type of cost function is not limited to the items listed above.

In embodiments, the “cost function” may be named “correlation measure.”

In embodiments, “highest correlation” may mean that the value associated with the correlation is the largest. The highest correlation may mean that the resulting cost function (or correlation measure) has the smallest value.

In embodiments, it has been described that objects such as templates and pixels can be correlated.

In embodiments, the object with the highest correlation may mean the object with the value associated with the greatest correlation. However, the object with the best correlation may not be limited to one object.

In embodiments, objects may be sorted in descending order of values related to their correlation. Among the sorted objects, a plurality of objects located at the front may be considered as the objects with the highest correlation. Here, there may be N plural objects. N may represent the number of objects to be selected.

Alternatively, the objects may be sorted in ascending order of the result values of the cost function (or correlation measure) of the objects, and a plurality of objects located in front among the sorted objects may be considered as the object with the highest correlation. Here, there may be N plural objects. N may represent the number of objects to be selected.

For example, N may be 2. Alternatively, N may be 3.

In other words, the highest correlation may mean that the resulting cost function (or correlation measure) is the smallest.

When a template with the highest correlation is searched for through template matching, the correlation measure may be at least one of SAD, SSD, and SATD).

The template with the lowest correlation measure value may be determined to have the highest correlation with the template of the target block.

In an embodiment, when calculating at least one of correlation, cost function, and correlation measure, coding parameters such as motion vector and intra prediction mode (or intra prediction direction) may be used instead of pixels. That is, in an embodiment, using a pixel may mean that not only the pixel value of the pixel is used, but also that information set about the location of the pixel or the block containing the pixel is used. Here, the information may include coding information.

For example, at least one of a correlation and a cost function (or correlation measure) may be calculated using at least one of coding parameters such as a motion vector and an intra-screen prediction mode (direction).

Additionally, at least one of the coding parameter with the highest correlation, the coding parameter with the lowest cost function, and the coding parameter with the lowest correlation measure determined by template matching may be included in the candidate list.

For example, the candidate list may be a list for encoding/decoding in intra prediction mode. The candidate list may be a Most Probable Mode (MPM) list.

For example, the candidate list may be a list for encoding/decoding motion information. The candidate list may be at least one of an advanced motion vector prediction (AMVP) and a merge list.

A syntax element (= template_matching_criteria_idx) that specifies the correlation measure (or cost function) used in template matching search may be signaled.

template_matching_criteria_idx may indicate a correlation measure used in a template matching search among a plurality of correlation measures. The plurality of correlation measures may include SAD, SSD, SATD, MR_SAD, and SSE.

For example, if template_matching_criteria_idx is the first value, the correlation measure (or cost function) may be SAD. If template_matching_criteria_idx is the second value, the correlation measure (or cost function) may be SSD. If template_matching_criteria_idx is the fifth value, the correlation measure (or cost function) may be SATD. If template_matching_criteria_idx is the fourth value, the correlation measure (or cost function) may be MR_SAD. If template_matching_criteria_idx is the fifth value, the correlation measure (or cost function) may be SSE.

Early termination of template matching search

Even if not all of the template matching search area has been searched, the template matching search may be terminated early if a specific condition is met. This early termination can be named “early termination of template matching search.”

In one embodiment, certain conditions are 1) the value of the correlation measure is less than a certain threshold and 2) the difference between the value of the correlation measure derived by the search and the value of the previous correlation measure is more than a certain threshold. It may be at least one of the small cases.

A syntax element (= template_matching_early_termination_flag) indicating whether early termination of the template matching search is allowed may be signaled.

For example, if template_matching_early_termination_flag is the first value, early termination of template matching search may not be allowed. If template_matching_early_termination_flag is the second value, early termination of template matching search may be allowed.

Threshold information (= template_matching_early_termination_threshold) used in conditions for determining early termination may be signaled. template_matching_early_termination_threshold1, which uses the value of the correlation measure as the threshold, and template_matching_early_termination_threshold2, which uses the difference between the values of the correlation measures as the threshold, can be used.

For example, when template_matching_early_termination_threshold1 is Th1 _tm , if the value of the correlation scale of a specific pixel is smaller than Th1 _tm , early termination of the template matching search may proceed, and the specific pixel is determined to be the optimal pixel with the highest correlation. It can be. At this time, the adjacent block corresponding to the template containing the specific pixel may be the optimal template matching block for encoding/decoding the target block.

For example, when template_matching_early_termination_threshold2 is Th2 _tm , if the difference between the value of the correlation measure of a specific pixel and the value of the correlation measure in the previous search is smaller than Th2 _tm , early termination of the template matching search may proceed, and A specific pixel in can be determined to be the optimal pixel with the highest correlation. At this time, the adjacent block corresponding to the template containing the specific pixel may be the optimal template matching block for encoding/decoding the target block.

Encoding/decoding of the target block may be performed using at least one of a template, a template matching reference image, a template matching optimal block, and a template matching motion vector.

Encoding/decoding using template matching optimal blocks

Below, encoding/decoding using the template matching optimal block in step 2040 is described.

Intra residual signal prediction based on template matching

Below, template matching based intra residual signal prediction in step 2041 is described.

FIG. 58A shows a target block, template matching optimal block, and intra prediction modes when the target image is a template matching reference image according to an example.

Figure 58b shows a target block, template matching optimal block, and intra prediction modes when the target image and the template matching reference image are different according to an example.

Prediction of the prediction error (or residual signal) of the target block may be performed using the template matching optimal block. Through this prediction, a predicted residual signal can be generated.

By predicting the residual signal of the target block using the template matching optimal block, the difference value between the residual signal and the predicted residual signal can be derived. The derived difference value may be encoded/decoded/signaled. A reconstructed residual signal can be obtained by summing the difference value and the predicted residual signal.

When a residual signal according to intra prediction or intra prediction mode is obtained for the target block, the residual signal of the template matching optimal block can be obtained by applying the same intra prediction mode to the template matching optimal block derived by template matching. . In other words, intra predictions using the same intra prediction mode can be performed on the target block and the optimal block, respectively. A residual signal for the target block and a residual signal for the template matching optimal block may be generated through intra predictions.

The residual signal of the template matching optimal block can be used to predict the residual signal of the target block. In other words, the residual signal of the template matching optimal block can be regarded as the predicted residual signal for the target block. The difference value between this predicted residual signal and the residual signal of the target block may be encoded/decoded/signaled. Prediction of this residual signal and/or signaling of the difference value may be referred to as TM-based intra residual signal prediction.

In embodiments, an intra residual signal may refer to a residual signal generated by intra prediction. A signal may refer to a block, and a signal at a specific location may refer to a pixel or pixel value at the specific location.

For a specific location (x, y), o _t (x, y) may represent the signal of the target block. m_i may indicate the i-th intra prediction mode. P _{t,m_i} (x, y) may be a signal generated by performing prediction on the target block using the i-th intra prediction mode. At this time, the intra prediction residual signal R _{t,m_i} (x, y) of the target block can be determined as in Equation 21 below.

[Formula 21]

R _{t,m_i} (x, y) = o _t (x, y) - P _{t,m_i} (x, y)

For a specific location (x, y), Rec _s (x, y) may represent the signal of the template matching optimal block. P _{s,m_i} (x, y) may be a signal generated by performing prediction on the template matching optimal block using the ith intra prediction mode. At this time, the intra prediction residual signal R _{s,m_i} (x, y) of the template matching optimal block generated by predicting the template matching optimal block using the i th intra prediction mode can be determined as in Equation 22 below. .

[Formula 22]

R _{s,m_i} (x, y) = Rec _s (x, y) - P _{s,m_i} (x, y)

The template matching motion vector MV _tm may be expressed as Equation 23 below.

[Formula 23]

MV _tm = (x _tm , y _tm )

For a specific location (x, y), the residual signal R _{s,m_i} (x + x _tm , y + y _tm ) of the template matching optimal block is a prediction of the residual signal R _{t,m_i} (x, y) of the target block. It can be used for. In other words, the residual signal R _{s,m_i} (x + x _tm , y + y _tm ) of the template matching optimal block can be regarded as the residual signal prediction value of the target block.

D _TMRP (x, y) is the residual signal of the target block R _{t,m_i} (x, y) and the predicted value of the residual signal of the target block (i.e., the residual signal of the template matching optimal block) R _{s,m_i} (x + x _tm , y + y _tm ). D _TMRP (x, y) can be defined as Equation 24 below.

[Formula 24]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm )

Given D _TMRP (x, y) for a specific position (x, y), 1) template matching reference image, 2) template matching motion vector MV _tm (= (x _tm , y _tm )), 3) template matching optimal block, 4) signal P _{t,m_i} (x, y) generated by performing prediction on the target block using the i-th intra prediction mode, 5) signal Rec _s (x, y) of template matching optimal block, 6) Signal P _{s,m_i} (x, y) generated by performing prediction for the template matching optimal block using the i-th intra prediction mode and 7) predicting the template matching optimal block using the i-th intra prediction mode. The reconstructed signal of the target block can be obtained using at least one of the intra prediction residual signals R _{s,m_i} (x, y) of the template matching optimal block generated by performing the process.

Rec _t (x, y) may be the reconstructed signal of the target block. Rec _t (x, y) can be obtained according to Equation 25 below.

[Formula 25]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm )

As shown in FIG. 58A, the template matching reference image may be a target image including a target block.

As shown in FIG. 58B, the template matching reference image may not be the target image including the target block. For example, the template matching reference image may be a reference image in the reference image list L0 or the reference image list L1.

Forecast when intra prediction mode is Planar mode

For a specific location (x, y), o _t (x, y) may represent the signal of the target block. The (optimal) intra prediction mode for the target block may be planner mode. At this time, P _t,Planar (x, y) may be a signal generated by performing prediction on the target block using planar mode. At this time, the intra prediction residual signal R _t,Planar (x, y) of the target block generated by performing prediction using the planar mode can be determined as in Equation 26 below.

[Formula 26]

R _t,Planar (x, y) = o _t (x, y) - P _t,Planar (x, y)

For a specific location (x, y), Rec _s (x, y) may represent the signal of the template matching optimal block. P _s,Planar (x, y) may be a signal generated by performing prediction on the template matching optimal block using planar mode. At this time, the intra prediction residual signal R _s,Planar (x, y) of the template matching optimal block generated by performing prediction on the template matching optimal block using the planner mode can be determined as in Equation 27 below.

[Formula 27]

R _s,Planar (x, y) = Rec _s (x, y) - P _s,Planar (x, y)

The template matching motion vector MV _tm may be expressed as Equation 28 below.

[Formula 28]

MV _tm = (x _tm , y _tm )

For a specific location (x, y), the residual signal R _s,Planar (x + x _tm , y + y _tm ) of the template matching optimal block is a prediction of the residual signal R _t,Planar (x, y) of the target block. It can be used for. In other words, the residual signal R _s,Planar (x + x _tm , y + y _tm ) of the template matching optimal block can be regarded as the residual signal prediction value of the target block.

D _TMRP (x, y) is the residual signal of the target block R _t,Planar (x, y) and the predicted value of the residual signal of the target block (i.e., the residual signal of the template matching optimal block) R _s,Planar (x + x _tm , y + y _tm ). D _TMRP (x, y) can be defined as Equation 29 below.

[Formula 29]

D _TMRP (x, y) = R _t,Planar (x, y) - R _s,Planar (x + x _tm , y + y _tm )

Given D _TMRP (x, y) for a specific position (x, y), 1) template matching reference image, 2) template matching motion vector MV _tm (= (x _tm , y _tm )), 3) template matching Optimal block, 4) Signal P _t,Planar (x, y), generated by performing prediction on target block using Planar mode, 5) Signal of template matching optimal block Rec _s (x, y), 6) Planar The signal P _s,Planar (x,y) generated by performing prediction on the template matching optimal block using mode 7) and 7) the signal of the template matching optimal block generated by performing prediction on the template matching optimal block using planar mode. The reconstructed signal of the target block can be obtained using at least one of the intra prediction residual signals R _{s, Planar} (x, y).

Rec _t (x, y) may be the reconstructed signal of the target block. Rec _t (x, y) can be obtained according to Equation 30 below.

[Formula 30]

Rec _t (x, y) = P _t,Planar (x, y) + D _TMRP (x, y) + R _s,Planar (x + x _tm , y + y _tm ) = P _t,Planar (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _s,Planar (x + x _tm , y + y _tm )

Signaling of syntactic elements for intra residual signal prediction based on template matching.

A syntax element (=tm_intra_residual_pred) indicating whether the target block is encoded/decoded using TM-based intra residual signal prediction may be signaled.

For example, tm_intra_residual_pred being the first value may indicate that the target block has been encoded/decoded using a TM-based intra residual signal prediction method. The second value of tm_intra_residual_pred may indicate that the target block has not been encoded/decoded using the TM-based intra residual signal prediction method.

For example, tm_intra_residual_pred being the first value may indicate that the difference value D _TMRP (x, y) is signaled. tm_intra_residual_pred being the second value may indicate that D _TMRP (x, y) is not signaled. D _TMRP (x, y) is the intra prediction residual signal R _{t,m_i} (x, y) of the target block and the predicted value of the intra prediction residual signal of the target block (i.e. the residual signal of the template matching optimal block) R _{s,m_i} ( It may be the difference value between x + x _tm , y + y _tm ).

D _TMRP (x, y) may be encoded/decoded/signaled in the same manner as the method of encoding/decoding/signaling other intra prediction residual signals in the embodiment. For example, D _TMRP (x, y) can be encoded/decoded/signaled by transformation, quantization, entropy encoding, entropy encoding, inverse quantization, and inverse transformation of the embodiment.

Constrained template-based intra residual signal prediction

Constrained TM-based intra residual signal prediction in step 2042 is described below.

When a residual signal according to intra prediction or intra prediction mode is obtained for the target block, the residual signal of the template matching optimal block can be obtained by applying the same intra prediction mode to the template matching optimal block derived by template matching. .

A part of the residual signal of the template matching optimal block may be used to predict a part of the residual signal of the target block. In other words, a part of the residual signal of the template matching optimal block can be regarded as the predicted residual signal for the target block. The difference value between this predicted residual signal and a portion of the residual signal of the target block may be encoded/decoded/signaled. This prediction of the residual signal and/or signaling of the difference value may be referred to as limited TM-based intra residual signal prediction. In other words, limited TM-based intra residual signal prediction may be TM-based intra residual signal prediction applied to a portion of the target block.

For a pixel at a specific location (x, y) belonging to part of the area of the target block, D _TMRP (x, y) can be derived according to Equation 31 below.

[Formula 31]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ a region

Encoding/decoding/signaling using Equation 32 below may be performed on pixels at specific positions (x, y) that belong to part of the area of the target block.

[Formula 32]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ a region

Even in limited TM-based intra residual signal prediction, as shown in FIG. 58A, the template matching reference image may be a target image including the target block.

Even in limited TM-based intra residual signal prediction, as shown in FIG. 58B, the template matching reference image may not be the target image including the target block. For example, the template matching reference image may be a reference image in the reference image list L0 or the reference image list L1.

Figure 59 shows intra residual signal prediction based on limited template matching for a first specific region according to an example.

Encoding/decoding/signaling using TM-based intra residual signal prediction may be applied only to pixels belonging to a specific region among the pixels of the target block.

In Figure 59, R1 represents a specific region.

The specific area may be a block adjacent to the lower right corner of the target block.

A specific area is a pixel whose 1) horizontal distance from the left side (or template) of the target block is more than d _x,th , and 2) the vertical distance from the top of the target block (or template) is more than d _y,th It could be their territory.

For example, d _x,th may be w/2. w may be the horizontal size of the target block.

For example, d _y,th may be h/2. h may be the vertical size of the target block.

In other words, TM-based intra only for pixels that 1) have a horizontal distance from the target block (or template) of d _x,th or more, and 2) have a vertical distance from the target block (or template) of d _y,th or more. Encoding/decoding/signaling using residual signal prediction may be applied.

For a pixel at a position (x, y) belonging to a specific area R1, D _TMRP (x, y) can be derived according to Equation 33 below.

[Formula 33]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R1

Encoding/decoding according to Equation 34 below may be performed on pixels at positions (x, y) belonging to a specific area R1.

[Formula 34]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R1

The (optimal) intra prediction mode for the target block may be planner mode.

For a pixel (x, y) of the target block, the horizontal distance between the pixel and the left side (or template) of the target block is greater than or equal to w/2, and the vertical distance between the pixel and the top side (or template) of the target block is h If it is greater than /2, D _TMRP (x, y) can be derived using TM-based intra residual signal prediction. D _TMRP (x, y) can be derived according to Equation 35 below.

[Formula 35]

D _TMRP (x, y) = R _t,Planar (x, y) - R _s,Planar (x + x _tm , y + y _tm ) where x ≥ x _{left_top} + w/2 and y ≥ y _{left_top} + h/ 2

Here, (x _{left_top} , y _{left_top} ) may be the upper left coordinates of the target block.

For a pixel (x, y) of the target block, the horizontal distance between the pixel and the left side (or template) of the target block is greater than or equal to w/2, and the vertical distance between the pixel and the top side (or template) of the target block is h If it is more than /2, the pixel can be encoded/decoded according to Equation 36 below.

[Formula 36]

Rec _t (x, y) = P _t,Planar (x, y) + D _TMRP (x, y) + R _s,Planar (x + x _tm , y + y _tm ) = P _t,Planar (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _s,Planar (x + x _tm , y + y _tm ) where x ≥ x _{left_top} + w/2 and y ≥ y _{left_top} + h/2

Figure 60 shows intra residual signal prediction based on limited template matching for a second specific region according to an example.

In Figure 60, R2 may represent a specific area.

The specific area may be a triangular area adjacent to the lower right corner of the target block.

A specific area may be determined based on a diagonal line within the target block.

The specific area may be an area of pixels in which the sum of 1) the horizontal distance from the left side (or template) of the target block and 2) the vertical distance from the top (or template) of the target block is greater than or equal to the reference value.

For example, the reference value may be the width or height of the target block.

For a pixel at a position (x, y) belonging to a specific area R2, D _TMRP (x, y) can be derived according to Equation 37 below.

[Formula 37]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R2

For a pixel at a position (x, y) belonging to a specific area R2, D _TMRP (x, y) can be derived according to Equation 38 below.

[Formula 38]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R2

Figure 61 shows intra residual signal prediction based on limited template matching for a third specific region according to an example.

Encoding/decoding/signaling using TM-based intra residual signal prediction may be applied only to pixels belonging to a specific region among the pixels of the target block.

In Figure 61, R3 may represent a specific area.

The specific area may be a block adjacent to the bottom of the target block.

The specific area may be an area of pixels whose vertical distance from the top (or template) of the target block is d _y,th or more.

For example, d _y,th may be h/2. h may be the vertical size of the target block.

In other words, encoding/decoding/signaling using TM-based intra residual signal prediction can be applied only to pixels whose vertical distance from the target block (or template) is d _y,th or more.

For a pixel at a position (x, y) belonging to a specific area R3, D _TMRP (x, y) can be derived according to Equation 39 below.

[Formula 39]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R3

Encoding/decoding according to Equation 40 below may be performed on pixels at positions (x, y) belonging to a specific area R3.

[Formula 40]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R3

The (optimal) intra prediction mode for the target block may be planner mode.

For a pixel (x, y) of the target block, if the vertical distance between the pixel and the upper side (or template) of the target block is greater than or equal to h/2, D _TMRP (x, y) is obtained using TM-based intra residual signal prediction. can be derived. D _TMRP (x, y) can be derived according to Equation 41 below.

[Formula 41]

D _TMRP (x, y) = R _t,Planar (x, y) - R _s,Planar (x + x _tm , y + y _tm ) where y ≥ y _{left_top} + h/2

Here, y _{left_top} may be the top y coordinate of the target block.

For the pixel (x, y) of the target block, if the vertical distance between the pixel and the upper side (or template) of the target block is more than h/2, the pixel can be encoded/decoded according to Equation 42 below.

[Formula 42]

Rec _t (x, y) = P _t,Planar (x, y) + D _TMRP (x, y) + R _s,Planar (x + x _tm , y + y _tm ) = P _t,Planar (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _s,Planar (x + x _tm , y + y _tm ) where y ≥ y _{left_top} + h/2

Figure 62 shows intra residual signal prediction based on limited template matching for a fourth specific region according to an example.

Encoding/decoding/signaling using TM-based intra residual signal prediction may be applied only to pixels belonging to a specific region among the pixels of the target block.

In Figure 62, R4 represents a specific region.

The specific area may be a block adjacent to the bottom of the target block.

The specific area may be an area of pixels whose horizontal distance from the left side (or template) of the target block is d _x,th or more.

For example, d _x,th may be w/2. w may be the horizontal size of the target block.

In other words, encoding/decoding/signaling using TM-based intra residual signal prediction can be applied only to pixels whose horizontal distance from the target block (or template) is d _x,th or more.

For a pixel at a position (x, y) belonging to a specific area R4, D _TMRP (x, y) can be derived according to Equation 43 below.

[Formula 43]

D _TMRP (x, y) = R _{t,m_i} (x, y) - R _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R4

Encoding/decoding according to Equation 44 below may be performed on pixels at positions (x, y) belonging to a specific area R4.

[Formula 44]

Rec _t (x, y) = P _{t,m_i} (x, y) + D _TMRP (x, y) + R _{s,m_i} (x + x _tm , y + y _tm ) = P _{t,m_i} (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _{s,m_i} (x + x _tm , y + y _tm ) where (x, y) ∈ R4

The (optimal) intra prediction mode for the target block may be planner mode.

For a pixel (x, y) of the target block, if the horizontal distance between the pixel and the left side (or template) of the target block is greater than or equal to w/2, D _TMRP (x, y) is obtained using TM-based intra residual signal prediction. can be derived. D _TMRP (x, y) can be derived according to Equation 45 below.

[Formula 45]

D _TMRP (x, y) = R _t,Planar (x, y) - R _s,Planar (x + x _tm , y + y _tm ) where x ≥ x _{left_top} + w/2

Here, x _{left_top} may be the left x coordinate of the target block.

For the pixel (x, y) of the target block, if the horizontal distance between the pixel and the left side (or template) of the target block is greater than or equal to w/2, the pixel can be encoded/decoded according to Equation 46 below.

[Formula 46]

Rec _t (x, y) = P _t,Planar (x, y) + D _TMRP (x, y) + R _s,Planar (x + x _tm , y + y _tm ) = P _t,Planar (x, y ) + D _TMRP (x, y) + Rec _s (x + x _tm , y + y _tm ) - P _s,Planar (x + x _tm , y + y _tm ) where x ≥ x _{left_top} + w/2

Signaling of syntactic elements for restricted TM-based intra residual signal prediction.

In one embodiment, a syntax element (= restrict_tm_intra_residual_pred) indicating whether the target block is encoded/decoded using restricted TM-based intra residual signal prediction may be signaled.

For example, restrict_tm_intra_residual_pred being the first value may indicate that the target block has been encoded/decoded using a restricted TM-based intra residual signal prediction method. restrict_tm_intra_residual_pred being the second value may indicate that the target block is not encoded/decoded using the restricted TM-based intra residual signal prediction method.

For example, restrict_tm_intra_residual_pred being the first value may indicate that the difference value D _TMRP (x, y) is signaled for a specific area of the target block. The fact that tm_intra_residual_pred is the second value may indicate that D _TMRP (x, y) is not signaled for a specific area of the target block. D _TMRP (x, y) is the intra prediction residual signal R _{t,m_i} (x, y) of the target block and the predicted value of the intra prediction residual signal of the target block (i.e. the residual signal of the template matching optimal block) R _{s,m_i} ( It may be the difference value between x + x _tm , y + y _tm ).

D _TMRP (x, y) may be encoded/decoded/signaled in the same manner as the method of encoding/decoding/signaling other intra prediction residual signals in the embodiment. For example, D _TMRP (x, y) can be encoded/decoded/signaled by transformation, quantization, entropy encoding, entropy encoding, inverse quantization, and inverse transformation of the embodiment.

In one embodiment, when the target block is encoded/decoded using TM-based intra residual signal prediction, a syntax element (= restrict_tm_intra_residual_flag) indicating the type of TM-based intra residual signal prediction may be signaled.

For example, restrict_tm_intra_residual_flag being the first value may indicate that the TM-based intra residual signal prediction method is used only for a specific area of the target block. The fact that restrict_tm_intra_residual_flag is the second value may indicate that the TM-based intra residual signal prediction method is used for the entire area of the target block. The fact that restrict_tm_intra_residual_flag is the third value may indicate that the TM-based intra residual signal prediction method is not used for the target block.

For example, restrict_tm_intra_residual_flag being the first value may indicate that the difference value D _TMRP (x, y) is signaled only for a specific area of the target block. The second value of restrict_tm_intra_residual_flag may indicate that the difference value D _TMRP (x, y) is signaled for the entire area of the target block. The third value of restrict_tm_intra_residual_flag may indicate that D _TMRP (x, y) is not signaled for the target block. D _TMRP (x, y) is the intra prediction residual signal R _{t,m_i} (x, y) of the target block and the predicted value of the intra prediction residual signal of the target block (i.e. the residual signal of the template matching optimal block) R _{s,m_i} ( It may be the difference value between x + x _tm , y + y _tm ).

In one embodiment, restrict_tm_intra_residual_flag may be signaled only when a TM-based intra residual signal prediction method is used for the target block.

For example, restrict_tm_intra_residual_flag can be signaled only when tm_intra_residual_pred is the first value.

LIC-based template matching prediction

Below, LIC-based template matching prediction in step 2043 is described.

FIG. 63A shows a target block, a template of the target block, and pixels of the template according to an example.

FIG. 63B shows a template matching optimal block, a template of a target block, and pixels of the template in an example.

LIC-based TM prediction may correct the prediction signal by applying LIC when using the template matching optimal block as the prediction signal of the target block. That is, when LIC-based TM prediction is used, correction using LIC can be applied to the prediction signal.

As shown in FIG. 63A, O _t may represent the signal of the target block. O _t may be the prediction signal of the target block. Template A may be a template of a target block adjacent to the target block.

As shown in FIG. 63B, template B may be the template with the highest correlation with template A. P _TM may represent the signal of the template matching optimal block adjacent to template B. P _TM may be a prediction signal of the template matching optimal block.

In embodiments, a template matching optimal block adjacent to a specific template may mean a template matching optimal block corresponding to a specific template.

The template matching motion vector MV _tm may be expressed as Equation 47 below.

[Formula 47]

MV _tm = (x _tm , y _tm )

At this time, as shown in Equation 48 below, the signal O(x, y) on the position (x, y) of the target block is predicted by a linear model using the signal P _TM (x, y) of the template matching optimal block. It can be.

[Formula 48]

Res _{LIC_TM} (x, y) = O _t (x, y) - (α·P _TM (x + x _tm , y + y _tm ) +β)

Here, Res _{LIC_TM} (x, y) may be the prediction error (or residual signal) of location (x, y).

α and β may be coefficients of a linear model.

Prediction error Res _{LIC_TM} (x, y), at least one of α and β may be encoded/decoded/signaled.

Among the prediction error Res _{LIC_TM} (x, y), α, β, template matching reference image, signal P _TM (x, y) of template matching optimal block, and template matching motion vector MV _tm (= (x _tm , y _tm )) The reconstructed signal Rec _t (x, y) of the target block can be obtained using at least one. Rec _t (x, y) can be obtained according to Equation 49 below.

[Formula 49]

Res _t (x, y) = (α·P _TM (x + x _tm , y + y _tm ) +β) + Res _{LIC_TM} (x, y)

Calculation of linear model coefficients for LIC-based TM predictions

Referring to FIGS. 63A and 63B, Y _A (i) may be the pixel value of pixel i belonging to template A. Y _B (i) may be the pixel value of i pixel belonging to template B.

In embodiments, the pixel value may be a luma value or a chroma value. The chroma value may be a Cb value or a Cr value.

Here, each of template A and template B may include a total of N pixels. In other words, i may be one of values between 1 and N.

At this time, the linear model coefficients α and β of LIC-based TM prediction can be obtained by Equation 50 and Equation 51 below, respectively.

[Formula 50]

[Formula 51]

Subsampling of linear model coefficient calculations for LIC-based TM predictions

Figure 64A shows a target block, a template of the target block, and subsampled pixels of the template, according to an example.

Figure 64B shows an example template matching optimal block, a template of a target block, and subsampled pixels of the template.

Template A may be the template of the target block. Y _A (i) may be the pixel value of pixel i belonging to template A.

Template B may be the template with the highest correlation with template A. Y _B (i) may be the pixel value of i pixel belonging to template B.

In embodiments, the pixel value may be a luma value or a chroma value. The chroma value may be a Cb value or a Cr value.

Here, each of template A and template B may include a total of N pixels. In other words, i may be one of values between 1 and N.

At this time, using some pixels belonging to template A and some pixels belonging to template B corresponding to some pixels of template A, the linear model coefficient α of LIC-based TM prediction is calculated as Equation 52 and Equation 53 below: and β can be obtained respectively.

[Formula 52]

[Formula 53]

Here, M may be a value equal to or less than N.

Some pixels belonging to template A may be configured by skipping pixels of template A one by one, as shown in FIG. 64A. At this time, M may be N/2.

Some pixels belonging to template B corresponding to template A may also be configured by skipping pixels of template B one by one, as shown in FIG. 64B.

Calculation of linear model coefficients for LIC-based template matching predictions

Figure 65 shows representative values for a template of a target block according to an example.

In calculating the linear model coefficients of LIC-based template matching prediction, subsampling can be used, and α and β can be derived as in a Cross-Component Linear Model (CCLM).

Template A may be the template of the target block. Y _A (i) may be the pixel value of pixel i belonging to template A.

Template B may be the template with the highest correlation with template A. Y _B (i) may be the pixel value of i pixel belonging to template B.

Here, each of template A and template B may include a total of N pixels. In other words, i may be one of values between 1 and N.

By performing downsampling on template A using at least one of the upper left coordinates of the target block (x ₀ , y ₀ ), the horizontal size W of the target block, the vertical size H of the target block, and the template A of the target block, a plurality of representative Values can be obtained. There may be N _sub representative values. N _sub may be an integer less than or equal to N.

For example, as shown in FIG. 65, the plurality of representative values may include t ₁ , t ₂ , t ₃ and t ₄ .

A plurality of representative values are t ₁ , t ₂ , t ₃ and t ₄ , each of which is the upper left coordinates of the target block (x ₀ , y ₀ ), the horizontal size W of the target block, the vertical size H of the target block, and the Each can be obtained using at least one of template A.

t ₁ may be the average value of samples at positions indicated by Equation 54 below.

[Formula 54]

(x ₀ + w/4 - 1, y ₀ - 1), (x ₀ + W/4 - 1, y ₀ - 2), (x ₀ + W/4, y ₀ - 1), (x ₀ + W/4, y ₀ - 2), (x ₀ + W/4 + 1, y ₀ - 1), (x ₀ + W/4 + 1, y ₀ - 2)

t ₂ may be the average value of samples at positions indicated by Equation 55 below.

[Formula 55]

(x ₀ + 3w/4 - 1, y ₀ - 1), (x ₀ + 3W/4 - 1, y ₀ - 2), (x ₀ + 3W/4, y ₀ - 1), (x ₀ + 3W/4, y ₀ - 2), (x ₀ + 3W/4 + 1, y ₀ - 1), (x ₀ + 3W/4 + 1, y ₀ - 2)

t ₃ may be the average value of samples at positions indicated by Equation 56 below.

[Formula 56]

(x ₀ - 1, y ₀ + H/4 - 1), (x ₀ - 2, y ₀ + H/4 - 1), (x ₀ - 1, y ₀ + H/4), (x ₀ - 2, y ₀ + H/4), (x ₀ - 1, y ₀ + H/4 + 1), (x ₀ - 2, y ₀ + H/4 + 1)

t ₄ may be the average value of samples at positions indicated by Equation 57 below.

[Formula 57]

(x ₀ - 1, y ₀ + 3H/4 - 1), (x ₀ - 2, y ₀ + 3H/4 - 1), (x ₀ - 1, y ₀ + 3H/4), (x ₀ - 2, y ₀ + 3H/4), (x ₀ - 1, y ₀ + 3H/4 + 1), (x ₀ - 2, y ₀ + 3H/4 + 1)

N _sub representative values can be classified according to the sizes of the representative values. The N _sub representative values can be classified into N _sub /2 representative values with larger values and N _sub /2 representative values with smaller values, depending on the sizes of the representative values.

For example, if the sizes of t ₁ , t ₂ , t ₃ and t ₄ are as in Equation 58 below, the two larger representative values (= t ₁ and t ₃ ) and the two smaller representative values (= t ₂ and t ₄ ).

[Formula 58]

t ₁ ≥ t ₂ ≥ t ₃ ≥ t ₄

The average T ₁ ^A of N _sub /2 representative values with larger values and the average T ₂ ^A of N _sub /2 representative values with smaller values can be obtained.

For example, T ₁ ^A and T ₂ ^A can be obtained according to Equation 59 and Equation 60 below, respectively.

[Formula 59]

T ₁ ^A = (t ₁ + t ₃ + 1) >> 1

[Formula 60]

T ₂ ^A = (t ₂ + t ₄ + 1) >> 1

By the same method as described above, the average T ₁ ^B of N _sub /2 representative values with larger values and N _sub /2 with smaller values from the template matching optimal block and the template B adjacent to the template matching optimal block are obtained. The average T ₂ ^B of the representative values can be obtained.

Using at least one of T ₁ ^A , T ₂ ^A T ₁ ^B and T ₂ ^B , the linear model coefficients α and β of LIC-based TM prediction can be obtained, respectively, according to Equations 61 and 62 below.

[Formula 61]

α = (T ₁ ^A - T ₂ ^A ) / (T ₁ ^B - T ₂ ^B )

[Formula 62]

β = T ₂ ^A - α·T ₂ ^B

Signaling in LIC-based TM prediction

A syntax element (= lic_template_matching_mode) indicating whether the target block has been encoded/decoded using LIC-based TM prediction may be signaled.

The fact that lic_template_matching_mode is the first value may indicate that the target block has been encoded/decoded using LIC-based TM prediction. The fact that lic_template_matching_mode is the second value may indicate that the target block has not been encoded/decoded using LIC-based TM prediction.

Blending prediction for intra and template matching

Below, the blending prediction of intra and template matching in step 2044 is described.

Figure 66 discloses information for predicting intra and TM blending according to an example.

Encoding/decoding of the target block can be performed by using the intra-mode prediction signal and the template matching optimal block together.

Intra and TM blending prediction may mean prediction using the intra mode prediction signal and template matching optimal block together. At this time, the prediction block of the target block may be calculated by a weighted combination or weighted summation of the intra-mode prediction signal and the template matching optimal block.

O _t may represent the target block. Template A may be a template of a target block adjacent to the target block. Template B may be the template with the highest correlation with template A. P _TM may represent a template matching optimal block adjacent to template B.

The template matching motion vector MV _tm may be expressed as Equation 63 below.

[Formula 63]

MV _tm = (x _tm , y _tm )

Given the above information, the signal O _t (x, y) of the target block at position (x, y) is P _TM (x + x _tm , y + y _tm ) and P _{t,m_i} (x, y) can be predicted using

P _TM (x + x _tm , y + y _tm ) may be the reconstructed signal of the template matching optimal block.

P _{t,m_i} (x, y) may be a signal generated by performing prediction on the target block using the i-th intra prediction mode.

According to Equation 64 below, the signal O _t (x, y) of the target block can be predicted.

[Formula 64]

Res _{TM_Intra} (x, y) = O _t (x, y) - (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _{t,m_i} (x, y))

Here, w ₁ may be the weight of the template matching optimal block. w ₂ may be the weight of the intra prediction signal.

The prediction signal of the target block may be as in Equation 65 below.

[Formula 65]

w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _{t,m_i} (x, y)

Res _{TM_Intra} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

The residual signal Res _{TM_Intra} (x, y) of the target block may be encoded/decoded/signaled.

Given the residual signal Res _{TM_Intra} (x, y) of the target block, the template matching reference image, the template matching motion vector MV _tm , and the reconstructed signal P _TM (x + x _tm , y + y _tm ) of the template matching optimal block. , the i th intra prediction mode, the signal P _{t,m_i} (x, y) generated by performing prediction on the target block using the i th intra prediction mode, the weight w ₁ of the template matching optimal block, and the weight of the intra prediction signal. The signal Rec _t (x, y) of the target block reconstructed using at least one of w ₂ may be obtained.

Rec _t (x, y) can be obtained according to Equation 66 below.

[Formula 66]

Rec _t (x, y) = (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _{t,m_i} (x, y)) + Res _{TM_Intra} (x, y)

Example of blending prediction of intra and template matching

In one embodiment, the template matching motion vector MV _tm may be (5, 5). P _t,Planar (x, y) may be a signal generated by performing prediction on the target block using planar mode. Weight w ₁ may be 1/2. Weight w ₂ may be 1/2.

In this case, the signal O _t (x, y) of the target block can be predicted according to Equation 67 below.

[Formula 67]

Res _{TM_Intra} (x, y) = O _t (x, y) - (P _TM (x + 5, y + 5) + P _{t,m_i} (x, y) + offset) >> 1

The prediction signal of the target block may be as in Equation 68 below.

[Formula 68]

(P _TM (x + 5, y + 5) + P _{t,m_i} (x, y) + offset) >> 1

Res _{TM_Intra} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

offset may be a value for rounding. For example, when encoding/decoding is performed on a 10-bit image, the offset may be 512. For example, when encoding/decoding is performed on an 8-bit video, the offset may be 128. If rounding is not applied, offset can be 0.

The residual signal Res _{TM_Intra} (x, y) of the target block may be encoded/decoded/signaled.

Given Res _{TM_Intra} (x, y), Rec _t (x, y) of the reconstructed target block can be obtained according to Equation 69 below.

[Formula 69]

Rec _t (x, y) = (P _TM (x + 5, y + 5) + P _{t,m_i} (x, y)) >> 1 + Res _{TM_Intra} (x, y)

Signaling in blending prediction of intra and template matching

A syntax element (= intra_template_matching_blending_mode) indicating whether the target block has been encoded/decoded using blending prediction of intra and template matching may be signaled.

For example, intra_template_matching_blending_mode being the first value may indicate that the target block was encoded/decoded using blending prediction of intra and template matching. The fact that intra_template_matching_blending_mode is the second value may indicate that the target block has not been encoded/decoded using blending prediction of intra and template matching.

Blending prediction of inter and template matching

Below, the blending prediction of inter and template matching in step 2045 is described.

Encoding/decoding of the target block can be performed by using the inter-mode prediction signal and the template matching optimal block together.

Blended prediction of inter and TM may mean prediction using the inter mode prediction signal and template matching optimal block together. At this time, the prediction block of the target block may be calculated by a weighted combination or weighted summation of the inter-mode prediction signal and the template matching optimal block.

Figure 67 illustrates a template matching reference image, a target image, and an inter prediction reference image for blending prediction of inter and template matching according to an example.

Blending prediction of inter and template matching can be divided as follows:

At step 2046, predict the blending of P inter and TM.

Blending prediction of B inter and TM at step 2047

Blending prediction of B-inter and B-TM at step 2048

Below, with reference to FIG. 67, 1) blending prediction of P-inter and TM, 2) blending prediction of B-inter and TM, and 3) blending prediction of B-inter and B-TM are described, respectively.

In the embodiments described below, “signal” may mean “prediction signal” and “inter prediction signal.”

In embodiments, blocks with weights may each be included in different images. In this respect, the weight of a specific image can be understood as the weight included in the specific image.

Blending prediction of P inter and template matching

Below, the blending prediction of P inter and template matching in step 2046 is described.

Encoding/decoding of the target block can be performed by using the prediction signal of uni-predictive inter mode and the template matching optimal block together.

Blending prediction of P inter and TM may mean prediction using the prediction signal of the inter mode of this single prediction and the template matching optimal block together.

As shown in FIG. 67, O _t may represent a target block. Template A may be a template of a target block adjacent to the target block. Template B may be the template with the highest correlation with template A. P _TM may represent a template matching optimal block adjacent to template B. MV _tm may be a template matching motion vector. P _Inter may be an inter prediction block generated by inter prediction for the target block. MV _Inter may be an inter prediction motion vector, which is a motion vector used for inter prediction.

The template matching motion vector MV _tm may be expressed as Equation 70 below.

[Formula 70]

MV _tm = (x _tm , y _tm )

The inter prediction motion vector MV _Inter may be expressed as Equation 71 below.

[Formula 71]

MV _Inter = (x _inter , y _inter )

Given the above information, the signal O _t (x, y) of the target block at position (x, y) is P _TM (x + x _tm , y + y _tm ) and P _Inter (x + x _inter , y + y _inter ) can be predicted using.

P _TM (x + x _tm , y + y _tm ) may be the reconstructed signal of the template matching optimal block.

P _Inter (x + x _inter , y + y _inter ) may be a signal generated by performing prediction on the target block using the inter prediction mode.

According to Equation 72 below, the signal O _t (x, y) of the target block can be predicted.

[Formula 72]

Res _{TM_Inter} (x, y) = O _t (x, y) - (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _Inter (x + x _inter , y + y _inter ) )

Here, w ₁ may be the weight of the template matching optimal block. w ₂ may be the weight of the inter prediction signal.

The prediction signal of the target block may be as in Equation 73 below.

[Formula 73]

w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _Inter (x + x _inter , y + y _inter )

Res _{TM_Inter} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

The residual signal Res _{TM_Inter} (x, y) of the target block may be encoded/decoded/signaled.

Given the residual signal Res _{TM_Inter} (x, y) of the target block, the template matching reference image, the template matching motion vector MV _tm , and the reconstructed signal P _TM (x + x _tm , y + y _tm ) of the template matching optimal block. , inter prediction reference image, inter prediction block P _inter , inter prediction motion vector MV _Inter , template matching signal of target block reconstructed using at least one of the optimal block weight w ₁ and the weight w ₂ of inter prediction signal Rec _t (x, y) can be obtained.

Rec _t (x, y) can be obtained according to Equation 74 below.

[Formula 74]

Rec _t (x, y) = (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _Inter (x + x _inter , y + y _inter )) + Res _{TM_Inter} (x, y )

Example of blending prediction of P inter and template matching

In one embodiment, the template matching motion vector MV _tm may be (5, 5). The inter prediction motion vector MV _Inter may be (1, 2). Weight w ₁ may be 1/2. Weight w ₂ may be 1/2.

In this case, the signal O _t (x, y) of the target block can be predicted according to Equation 75 below.

[Formula 75]

Res _{TM_Inter} (x, y) = O _t (x, y) - (P _TM (x + 5, y + 5) + P _Inter (x + 1, y + 2) + offset) >> 1

The prediction signal of the target block may be as in Equation 76 below.

[Formula 76]

(P _TM (x + 5, y + 5) + P _Inter (x + 1, y + 2) + offset) >> 1

Res _{TM_Inter} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

offset may be a value for rounding. For example, when encoding/decoding is performed on a 10-bit image, the offset may be 512. For example, when encoding/decoding is performed on an 8-bit video, the offset may be 128. If rounding is not applied, offset can be 0.

The residual signal Res _{TM_Inter} (x, y) of the target block may be encoded/decoded/signaled.

Given Res _{TM_Inter} (x, y), Rec _t (x, y) of the reconstructed target block can be obtained according to Equation 77 below.

[Formula 77]

Rec _t (x, y) = (P _TM (x + 5, y + 5) + P _Inter (x + 1, y + 2)) >> 1 + Res _{TM_Inter} (x, y)

Inference from template matching reference images

Once the inter prediction reference image is constructed, a template matching reference image can be inferred from the inter prediction reference image.

In one embodiment, the template matching reference image and the inter prediction reference image may be the same. In other words, the reference image used for inter prediction of the target block can be considered a template matching reference image.

For example, when the target block is encoded/decoded using the first reference image in the first reference image list, the first reference image in the first reference image list may be inferred as the template matching reference image.

For example, when the target block is encoded/decoded using the second reference image in the first reference image list, the second reference image in the first reference image list may be inferred as the template matching reference image.

In one embodiment, the template matching reference image and the inter prediction reference image may be different from each other.

For example, a specific reference image in a reference image list that does not include an inter-prediction reference image may be inferred as a template matching reference image. A specific reference image may be the first reference image.

For example, if the inter prediction reference image is a reference image in the first reference image list, the first reference image in the second reference image list may be inferred as the template matching reference image.

For example, if the inter-prediction reference image is a reference image in the second reference image list, the first reference image in the first reference image list may be inferred as the template matching reference image.

In one embodiment, the first reference image in the second reference image list of inter prediction may be set as the template matching reference image.

Signaling in blending prediction of P-inter and template matching

A syntax element (= p_inter_template_matching_blending_mode) indicating whether the target block has been encoded/decoded using blending prediction of P inter and template matching may be signaled.

For example, p_inter_template_matching_blending_mode being the first value may indicate that the target block was encoded/decoded using blending prediction of P inter and template matching. The fact that p_inter_template_matching_blending_mode is the second value may indicate that the target block has not been encoded/decoded using blending prediction of P inter and template matching.

A syntax element (=inter_tm_sharing_ref_flag) indicating whether the inter prediction reference image and the template matching reference image of the target block are the same may be signaled.

For example, inter_tm_sharing_ref_flag being the first value may indicate that the inter prediction reference image and template matching reference image of the target block are the same. The fact that inter_tm_sharing_ref_flag is the second value may indicate that the inter prediction reference image and template matching reference image of the target block are not the same. Since inter_tm_sharing_ref_flag is the third value, the template matching reference image may be the first reference image in the second reference image list of inter prediction.

If the inter-prediction reference image and the template matching reference image of the target block are the same, signaling of the syntax element indicating the template matching reference image may be omitted.

For example, when inter_tm_sharing_ref_flag is the first value, a syntax element indicating the template matching reference image of the target block may not be signaled.

Blending prediction of inter and template matching

Below, the blending prediction of B inter and template matching in step 2047 is described.

Encoding/decoding of the target block can be performed by using the prediction signal of the bi-predictive inter mode and the template matching optimal block together.

Blended prediction of B inter and TM may refer to prediction using the prediction signal of the inter mode of this bi-prediction and the template matching optimal block together.

As shown in FIG. 67, O _t may represent a target block. Template A may be a template of a target block adjacent to the target block. Template B may be the template with the highest correlation with template A. P _TM may represent a template matching optimal block adjacent to template B. MV _tm may be a template matching motion vector. P _{Inter_l0} may be the first inter prediction block of the first inter prediction reference image generated by inter prediction for the target block. P _{Inter_l1} may be a second inter prediction block of the second inter prediction reference image generated by inter prediction for the target block. MV _{Inter_l0} may be the first inter prediction motion vector of the first inter prediction reference image. MV _{Inter_l1} may be a second inter prediction motion vector of the second inter prediction reference image.

The template matching motion vector MV _tm may be expressed as Equation 78 below.

[Formula 78]

MV _tm = (x _tm , y _tm )

The first inter prediction motion vector MV _{Inter_l0} may be expressed as Equation 79 below.

[Formula 79]

MV _{Inter_l0} = (x _{inter_l0} , y _{inter_l0} )

The second inter prediction motion vector MV _{Inter_l1} may be expressed as Equation 80 below.

[Formula 80]

MV _{Inter_l1} = (x _{inter_l1} , y _{inter_l1} )

Given the above information, the signal O _t (x, y) of the target block at position (x, y) is P _TM (x + x _tm , y + y _tm ), P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) and P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} ).

P _TM (x + x _tm , y + y _tm ) may be the reconstructed signal of the template matching optimal block.

P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) may be an L0 signal generated by performing inter prediction in the L0 direction for the target block.

P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} ) may be an L1 signal generated by performing inter prediction in the L1 direction for the target block.

According to Equation 81 below, the signal O _t (x, y) of the target block can be predicted.

[Formula 81]

Res _{TM_Inter_B} (x, y) = O _t (x, y) - (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) + w ₃ ·P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} ))

Here, w ₁ may be the weight of the template matching optimal block. w ₂ may be the weight of the first inter prediction reference image (or L0 inter prediction signal). w ₃ may be the weight of the second inter prediction reference image (or L1 inter prediction signal).

The prediction signal of the target block may be as shown in Equation 82 below.

[Formula 82]

w ₁ ·P _TM (x + x _tm , y + y _tm ) ₊ w ₂ ·P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) + w ₃ ·P _{Inter_l1} ( _x +

Res _{TM_Inter_B} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

The residual signal Res _{TM_Inter_B} (x, y) of the target block may be encoded/decoded/signaled.

Given the residual signal Res _{TM_Inter_B} (x, y) of the target block, the template matching reference image, the template matching optimal block P _TM , the template matching motion vector MV _tm , the inter prediction block P _{inter_l0} of the first inter prediction reference image, the second Inter prediction block P _{inter_l1} of the inter prediction reference image, inter prediction motion vector MV _{Inter_l0} of the first inter prediction reference image, inter prediction motion vector MV _{Inter_l1} of the second inter prediction reference image, template matching reference image (or template matching optimal block) ) of the weight w ₁ , the weight w ₂ of the inter prediction signal of the first inter prediction reference image, and the weight w ₃ of the inter prediction signal of the second inter prediction reference image Rec _t of the target block reconstructed using at least one of (x, y) can be obtained.

Rec _t (x, y) can be obtained according to Equation 83 below.

[Formula 83]

Rec _t (x, y) = (w ₁ ·P _TM (x + x _tm , y + y _tm ) + w ₂ ·P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) + w ₃ ·P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} )) + Res _{TM_Inter_B} (x, y)

Example of blending prediction of B-inter and template matching

In one embodiment, the template matching motion vector MV _tm may be (5, 5). The inter prediction motion vector MV _{Inter_l0} of the first reference image may be (1, 2). The inter prediction motion vector MV _{Inter_l1} of the second reference image may be (2, 3). Weight w ₁ may be 1/3. Weight w ₂ may be 1/3. Weight w ₃ may be 1/3.

In this case, the signal O _t (x, y) of the target block can be predicted according to Equation 84 below.

[Formula 84]

Res _{TM_Inter_B} (x, y) = O _t (x, y) - ((1/3)·P _TM (x + 5, y + 5) + (1/3)·P _{Inter_l0} (x + 1, y + 2) + (1/3)·P _{Inter_l1} (x + 2, y + 3))

The prediction signal of the target block may be as shown in Equation 85 below.

[Formula 85]

(1/3)·P _TM (x + 5, y + 5) + (1/3)·P _{Inter_l0} (x + 1, y + 2) + (1/3)·P _{Inter_l1} (x + 2, y + 3)

Res _{TM_Inter_B} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

The residual signal Res _{TM_Inter_B} (x, y) of the target block may be encoded/decoded/signaled.

Given Res _{TM_Inter_B} (x, y), Rec _t (x, y) of the reconstructed target block can be obtained according to Equation 86 below.

[Formula 86]

Rec _t (x, y) = ((1/3)·P _TM (x + 5, y + 5) + (1/3)·P _{Inter_l0} (x + 1, y + 2) + (1/3) ·P _{Inter_l1} (x + 2, y + 3)) + Res _{TM_Inter_B} (x, y)

B Inference of template matching reference images in blending of inter and template matching.

Once the inter prediction reference image is constructed, a template matching reference image can be inferred from the inter prediction reference image.

In one embodiment, a reference image in the first reference image list of the target block may be inferred as a template matching reference image. In other words, when a specific reference image in the specific reference image list is used for encoding/decoding the target block, the specific reference image can be inferred as a template matching reference image.

For example, when the target block is encoded/decoded using the first reference image in the first reference image list, the first reference image in the first reference image list may be inferred as the template matching reference image.

For example, when the target block is encoded/decoded using the first reference image in the second reference image list, the first reference image in the second reference image list may be inferred as the template matching reference image.

Signaling in blending prediction of B-inter and template matching

A syntax element (=b_inter_template_matching_blending_mode) indicating whether the target block has been encoded/decoded using blending prediction of B inter and template matching may be signaled.

For example, b_inter_template_matching_blending_mode being the first value may indicate that the target block was encoded/decoded using blending prediction of B inter and template matching. The fact that b_inter_template_matching_blending_mode is the second value may indicate that the target block has not been encoded/decoded using blending prediction of B inter and template matching.

Blending prediction of B-Inter and B-TM

Below, the blending prediction of B-Inter and B-TM in step 2048 is described.

Encoding/decoding on the target block can be performed by using the prediction signal of the bi-predictive inter mode and the two template matching optimal blocks together.

Blending prediction of B-inter and bidirectional-TM (B-TM) may mean prediction using the prediction signal and template of the bi-prediction inter mode together with a plurality of matching optimal blocks.

O _t may represent the target block. Template A may be a template of a target block adjacent to the target block. As the template matching reference image, a first template matching reference image and a second template matching reference image may be used. Template B may be the template with the highest correlation with template A in the first template matching reference image. P _{TM_l0} may indicate the first template matching optimal block adjacent to template B in the first template matching reference image. Template C may be the template with the highest correlation with template A in the second template matching reference image. P _{TM_l1} may indicate a second template matching optimal block adjacent to template C in the second template matching reference image. MV _{tm_l0} may be the first template matching motion vector between P _{TM_l1} and the target block. MV _{tm_l0} may be a second template matching motion vector between P _{TM_l2} and the target block. P _{Inter_l0} may be the first inter prediction block of the first inter prediction reference image generated by inter prediction for the target block. P _{Inter_l1} may be a second inter prediction block of the second inter prediction reference image generated by inter prediction for the target block. MV _{Inter_l0} may be the first inter prediction motion vector of the first inter prediction reference image. MV _{Inter_l1} may be a second inter prediction motion vector of the second inter prediction reference image.

The first template matching motion vector MV _{tm_l0} may be expressed as Equation 87 below.

[Formula 87]

MV _{tm_l0} = (x _{tm_l0} , y _{tm_l0} )

The second template matching motion vector MV _{tm_l1} may be expressed as Equation 88 below.

[Formula 88]

MV _{tm_l1} = (x _{tm_l1} , y _{tm_l1} )

The first inter prediction motion vector MV _Inter may be expressed as Equation 89 below.

[Formula 89]

MV _{Inter_l0} = (x _{inter_l0} , y _{inter_l0} )

The second inter prediction motion vector MV _Inter may be expressed as Equation 90 below.

[Formula 90]

MV _{Inter_l1} = (x _{inter_l1} , y _{inter_l1} )

Given the above information, the signal O _t (x, y) of the target block at position (x, y) is the reconstructed signal of the template matching optimal block (i.e., P _{TM_l0} (x + x _{tm_l0} , y + y _{tm_l0} ) and P _{TM_l1} (x + x _{tm_l1} , y + y _{tm_l1} )) and signals generated by performing prediction on the target block using _the inter prediction mode (i.e., P _{Inter_l0} (x + y + y _{inter_l0} ) and P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} )).

P _{TM_l0} (x + x _{tm_l0} , y + y _{tm_l0} ) may be a reconstructed signal of the first template matching optimal block.

P _{TM_l1} (x + x _{tm_l1} , y + y _{tm_l1} ) may be a reconstructed signal of the second template matching optimal block.

P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) may be an L0 signal generated by performing inter prediction in the L0 direction for the target block.

P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} ) may be an L1 signal generated by performing inter prediction in the L1 direction for the target block.

According to Equation 91 below, the signal O _t (x, y) of the target block can be predicted.

[Formula 91]

Res _{S-TM_Inter_B} (x, y) = O _t (x, y) - (w ₁ ·P _{TM_l0} (x + x _{tm_l0} , y + y _{tm_l0} ) + w ₂ ·P _{TM_l1} (x + x _{tm_l1} , y + y _{tm_l1} ) + w ₃ ·P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) + w ₄ ·P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} ))

Here, w ₁ may be the weight of the first template matching optimal block. w ₂ may be the weight of the second template matching optimal block. w ₃ may be the weight of the first inter prediction reference image (or L0 inter prediction signal). w ₄ may be the weight of the second inter prediction reference image (or L1 inter prediction signal).

The prediction signal of the target block may be as in Equation 92 below.

[Formula 92]

w ₁ ·P _{TM_l0} (x + x _{tm_l0} , y + y _{tm_l0} ) + w ₂ ·P _{TM_l1} (x + x _{tm_l1 ,} y + y _{tm_l1} ) + w ₃ ·P _{Inter_l0} (x ₊ + w ₄ ·P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} )

Res _{S-TM_Inter_B} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

The residual signal Res _{S-TM_Inter_B} (x, y) of the target block may be encoded/decoded/signaled.

Given the residual signal Res _{B-TM_Inter_B} (x, y) of the target block, the first template matching reference image, the first template matching optimal block P _{TM_l0} , the first template matching motion vector MV _{tm_l0} , the second template matching reference image, Second template matching optimal block P _{TM_l1} , second template matching motion vector MV _{tm_l1} , inter prediction block P _{inter_l0} of the first inter prediction reference image, inter prediction block P _{inter_l1} of the second inter prediction reference image, first inter prediction reference image inter prediction motion vector MV _{Inter_l0} , inter prediction motion vector MV _{Inter_l1} of the second inter prediction reference image, weight w ₁ of the first template matching reference image (or, first template matching optimal block), second template matching reference image ( Or, re-processing using at least one of the weight w ₂ of the second template matching optimal block, the weight w ₃ of the inter prediction signal of the first inter prediction reference image, and the weight w ₄ of the inter prediction signal of the second inter prediction reference image. The signal Rec _t (x, y) of the constructed target block can be obtained.

Rec _t (x, y) can be obtained according to Equation 93 below.

[Formula 93]

Rec _t (x, y) = (w ₁ ·P _{TM_l0} (x + x _{tm_l0} , y + y _{tm_l0} ) + w ₂ ·P _{TM_l1} (x + x _{tm_l1} , y + y _{tm_l1} ) + w ₃ ·P _{Inter_l0} (x + x _{inter_l0} , y + y _{inter_l0} ) + w ₄ ·P _{Inter_l1} (x + x _{inter_l1} , y + y _{inter_l1} )) + Res _{B-TM_Inter_B} (x, y)

Example of blending prediction of B-Inter and B-TM

In one embodiment, the template matching motion vector MV _{tm_l0} of the first template matching reference image may be (5, 5). The template matching motion vector MV _{tm_l1} of the second template matching reference image may be (4, 3). The inter prediction motion vector MV _{Inter_l0} of the first reference image may be (1, 2). The inter prediction motion vector MV _{Inter_l1} of the second reference image may be (2, 3). Weight w ₁ may be 1/4. Weight w ₂ may be 1/4. Weight w ₃ may be 1/4. Weight w ₄ may be 1/4.

In this case, the signal O _t (x, y) of the target block can be predicted according to Equation 94 below.

[Formula 94]

Res _{TM_Inter_B} (x, y) = O _t (x, y) - (P _{TM_l0} (x + 5, y + 5) + P _{TM_l1} (x + 4, y + 3) + P _{Inter_l0} (x + 1, y + 2) + P _{Inter_l1} (x + 2, y + 3) + offset) >> 2

The prediction signal of the target block may be as shown in Equation 95 below.

[Formula 95]

(P _{TM_l0} (x + 5, y + 5) + P _{TM_l1} (x + 4, y + 3) + P _{Inter_l0} (x + 1, y + 2) + P _{Inter_l1} (x + 2, y + 3) + offset ) >> 2

Res _{B-TM_Inter_B} (x, y) may be the difference (i.e., residual signal) between the prediction signal of the target block and the target block.

offset may be a value for rounding. For example, when encoding/decoding is performed on a 10-bit image, the offset may be 512. For example, when encoding/decoding is performed on an 8-bit video, the offset may be 128. If rounding is not applied, offset can be 0.

The residual signal Res _{B-TM_Inter_B} (x, y) of the target block may be encoded/decoded/signaled.

Given Res _{B-TM_Inter_B} (x, y), Rec _t (x, y) of the reconstructed target block can be obtained according to Equation 96 below.

[Formula 96]

Rec _t (x, y) = (P _{TM_l0} (x + 5, y + 5) + P _{TM_l1} (x + 4, y + 3) + P _{Inter_l0} (x + 1, y + 2) + P _{Inter_l1} (x + 2, y + 3) + offset) >> 2 + Res _{B-TM_Inter_B} (x, y)

Inference of template matching reference images in blending of B-inter and B-TM

When the first inter prediction reference image is constructed, the first template matching reference image can be inferred from the first inter prediction reference image.

When the second inter prediction reference image is constructed, the second template matching reference image can be inferred from the second inter prediction reference image.

In one embodiment, the inter prediction reference image and the template matching reference image used for inter prediction of the target block may be the same.

For example, when the target block is encoded/decoded using the first reference image in the first reference image list and the second reference image in the second reference image list, the first reference image in the first reference image list is the first reference image. may be inferred as a template matching reference image, and a second reference image in the second reference image list may be inferred as a second template matching reference image,

Signaling in blending prediction of B-inter and B-TM

A syntax element (=b_inter_bi_template_matching_blending_mode) indicating whether the target block has been encoded/decoded using blending prediction of B-inter and B-TM may be signaled.

For example, b_inter_bi_template_matching_blending_mode being the first value may indicate that the target block has been encoded/decoded using blending prediction of B-inter and B-TM. The fact that b_inter_bi_template_matching_blending_mode is the second value may indicate that the target block has not been encoded/decoded using blending prediction of B-inter and B-TM.

Suggestions for template matching-based encoding/decoding methods

Hereinafter, the steps of the template matching-based encoding/decoding method include one or more of the steps (2010, 2020, 2030, 2040, 2041, 2042, 2043, 2044, 2045, 2046, 2047, and 2048). The steps of the decryption method are outlined.

Whether performance of the steps of the template matching-based encoding/decoding method is permitted may be determined based on coding information. For example, whether performance of the steps of the template matching-based encoding/decoding method is permitted may be determined depending on the size of the target block.

In other words, performance of the steps of the template matching-based encoding/decoding method may not be permitted depending on the size of the target block.

For example, a template matching-based encoding/encoding method may or may not be allowed depending on the size of the target block.

The target block may be at least one of CTB, MB, CB, PB, TB, VPDU, and a block of a predetermined size.

At this time, if the size of the block having the predetermined size is MxN, the steps of the template matching-based encoding/decoding method can be performed only when the target block size is MxN. Here, each of M and N may be a positive integer.

Additionally, the size of the target block may be at least one of the maximum transform size and the maximum transform block size allowed by the encoding device 1600 and the decoding device 1700.

If the steps of the template matching-based encoding/decoding method are allowed only in blocks with a predetermined size, the prediction block may be the same as the block with the predetermined size, but the transform block may be a block divided to have a smaller size than the block with the predetermined size. You can. At this time, transform block partition information indicating what type of partition was applied for the transform block may be entropy encoded/decoded. Additionally, transform block partition information may be indicated by partition information of a coding block.

If the steps of the template matching-based encoding/decoding method are allowed only in blocks with a predetermined size, the prediction block may be a block divided to have a smaller size than the block with the predetermined size, but the transform block is the same as the block with the predetermined size. can do. At this time, prediction block division information indicating what type of division was applied for the prediction block may be entropy encoded/decoded. Additionally, prediction block partition information may be indicated by partition information of a coding block.

Performance of the steps of the template matching-based encoding/decoding method may be permitted when the size of the target block is larger than a predetermined size.

In other words, performance of the steps of the template matching-based encoding/decoding method may not be permitted if the size of the target block is smaller than a predetermined size.

A template matching-based encoding/decoding method can be accepted when the size of the target block is larger than a predetermined size. Template matching-based encoding/decoding methods may not be allowed if the size of the target block is smaller than a predetermined size.

The target block may be at least one of CTB, MB, CB, PB, TB, VPDU, and a block of a predetermined size.

For example, the predetermined size may be the maximum CTU size.

For example, the predetermined size may be "maximum CTU size >> 1" (=maximum CTU size / 2".

For example, the predetermined size may be 1 VPDU.

For example, the predetermined size may be “VPDU size >> 2” (=VPDU size / 4).

For example, the predetermined size may be MxN. Each of M and N may be a positive integer.

Performance of the steps of the template matching-based encoding/decoding method may be permitted when the size of the target block is smaller than a predetermined size.

In other words, performance of the steps of the template matching-based encoding/decoding method may not be permitted if the size of the target block is larger than a predetermined size.

The template matching-based encoding/decoding method can be accepted when the size of the target block is smaller than a predetermined size. Template matching-based encoding/decoding methods may not be allowed if the size of the target block is larger than a predetermined size.

The target block may be at least one of CTB, MB, CB, PB, TB, VPDU, and a block of a predetermined size.

For example, the predetermined size may be the maximum CTU size.

For example, the predetermined size may be "maximum CTU size >> 1" (=maximum CTU size / 2".

For example, the predetermined size may be 1 VPDU.

For example, the predetermined size may be “VPDU size >> 2” (=VPDU size / 4).

For example, the predetermined size may be MxN. Each of M and N may be a positive integer.

Whether performance of the steps of the template matching-based encoding/decoding method is permitted may be determined depending on the coding mode of the block surrounding the target block. Coding mode may include prediction mode. Coding mode may include coding information.

In other words, performance of the steps of the template matching-based encoding/decoding method may be permitted depending on the coding mode of the surrounding block. Performance of the steps of the template matching-based encoding/decoding method may not be permitted depending on the coding mode of the neighboring block.

A template matching-based encoding/decoding method may be allowed depending on the coding mode of the block surrounding the target block. Template matching-based encoding/decoding methods may not be allowed depending on the encoding mode of the surrounding block.

The surrounding block may be at least one of CTB, MB, CB, PB, TB, VPDU, and a block of a predetermined size.

In one embodiment, performance of the steps of the template matching-based encoding/decoding method may be permitted only when a block included in the template of the target block has been encoded/decoded using intra prediction.

Performance of the steps of the template matching-based encoding/decoding method may not be permitted if the block included in the template of the target block is not encoded/decoded using intra prediction.

A template matching-based encoding/decoding method can be accepted when a block included in the template of the target block is encoded/decoded using intra prediction.

Template matching-based encoding/decoding methods may not be allowed if blocks included in the template of the target block are not encoded/decoded using intra prediction.

For example, in the step of the template matching-based encoding/decoding method, if a block included in the template of the target block is not encoded/decoded using intra prediction, encoding is performed using intra prediction when constructing the template of the target block. /Undecoded pixels may be excluded.

In one embodiment, performance of the steps of the template matching-based encoding/decoding method may be permitted only when a block included in the template of the target block has been encoded/decoded using inter prediction.

Performance of the steps of the template matching-based encoding/decoding method may not be permitted if the block included in the template of the target block is not encoded/decoded using inter prediction.

The template matching-based encoding/decoding method can be allowed when a block included in the template of the target block is encoded/decoded using inter prediction.

Template matching-based encoding/decoding methods may not be allowed if blocks included in the template of the target block are not encoded/decoded using inter prediction.

For example, in the step of the template matching-based encoding/decoding method, if a block included in the template of the target block is not encoded/decoded using inter prediction, it is encoded using inter prediction when constructing the template of the target block. /Undecoded pixels may be excluded.

In one embodiment, performance of the steps of the template matching-based encoding/decoding method may be permitted only when the block included in the template of the target block has been encoded/decoded using the intra block copy (IBC) mode. .

Performance of the steps of the template matching-based encoding/decoding method may not be permitted if the block included in the template of the target block is not encoded/decoded using the IBC mode.

The template matching-based encoding/decoding method can be allowed if the block included in the template of the target block is encoded/decoded using IBC mode.

Template matching-based encoding/decoding methods may not be allowed if blocks included in the template of the target block are not encoded/decoded using IBC mode.

For example, in the step of the template matching-based encoding/decoding method, if the block included in the template of the target block is not encoded/decoded using the IBC mode, the IBC mode is used to construct the template of the target block. /Undecoded pixels may be excluded.

Two or more of the above-described embodiments may be combined. Here, the combination may include a combination of two or more of inter prediction, intra prediction, and IBC modes.

In one embodiment, performance of the steps of the template matching-based encoding/decoding method may be permitted only when the block included in the template of the target block has been encoded/decoded using one of intra prediction and IBC modes.

Performance of the steps of the template matching-based encoding/decoding method may not be permitted if the block included in the template of the target block has not been encoded/decoded using one of the intra prediction and IBC modes.

A template matching-based encoding/decoding method may be allowed if a block included in the template of the target block is encoded/decoded using one of intra prediction and IBC modes.

Template matching-based encoding/decoding methods may not be allowed if blocks included in the template of the target block are not encoded/decoded using one of intra prediction and IBC modes.

For example, in the step of the template matching-based encoding/decoding method, if a block included in the template of the target block is not encoded/decoded using one of the intra prediction and IBC modes, the intra prediction and IBC modes are not used to configure the template of the target block. Pixels that are not encoded/decoded using either prediction or IBC modes may be excluded.

Subsampling of Examples

In embodiments, generating a template using subsampling may mean that the template is constructed using at least one pixel among surrounding pixels of the target block.

In embodiments, when a template is created by subsampling, a filter may be applied to surrounding pixels of the target block. A template can be constructed using surrounding pixels with a filter applied.

For example, when 1/2 times subsampling is applied, one filtered value can be obtained by applying a bilinear filter to a pair of surrounding pixels. A template can be constructed using the filtered values.

For example, one value can be obtained by applying a bilinear filter to the two values described in the embodiments. The obtained value may be included in the template.

Multiple template types

There may be multiple templates in the embodiments. The plurality of templates may include an L-shaped template, a left template, and an upper template. That is, template matching can be performed using only the left template. Template matching can be performed using only the upper template.

Multiple modes may be used in template matching in embodiments. The plurality of modes are 1) a template matching mode that uses only the left template to derive candidates, 2) a template matching mode that uses only the top template to derive candidates, and 3) a template that synthesizes multiple candidates using linear combination. May include matching mode.

For each template of the plurality of templates, a plurality of candidates may be derived using the template matching of the embodiment. For example, four optimal candidates selected according to SAD cost can be derived.

That is, multiple candidates can be derived for multiple template types.

To select a candidate, the template type and the index of the candidate may need to be signaled. The template type may indicate a template used for template matching among a plurality of templates.

To signal the selected candidate, a flag indicating the type of template may be signaled. Next, the candidate's index may be signaled.

A template can be specified through a flag indicating the type of the template and an index of the candidate.

In one embodiment, multiple candidates may be combined. Weights for multiple candidates for combination may be: 1) derived by the template matching costs of the multiple candidates, and 2) MSE minimization in a Cross-Component Convolutional Model (CCCM) It can be derived using the derivation method of

Candidates to be fused may be candidates of an optimal plurality of L-forms. The plurality of L-form candidates can be 2 or 5. A flag indicating the number of candidates to be fused may be signaled. The flag can point to either 2 or 5.

Signaling for Candidates

A candidate list can be used to specify the template used for template matching. Candidate BVs within the candidate list may be sorted in ascending order of template matching costs of the candidate BVs.

A block vector (BV) may refer to a motion vector used in intra template matching prediction (IntraTMP).

IntraTMP can generate prediction values of the target block by copying specific values of the matching block (or matched block) in the target image. For example, certain values may be reconstructed values. Here, the matching of the matching block may be determined by template matching in the embodiments. Template matching can be performed based on the L-shaped template of the target block.

An index may be signaled within the bitstream to indicate which candidate BV among the candidate BVs in the candidate list is actually used for the target block. This method can use template matching to select a shortlist of promising candidates from a number of available BVs. This method allows the encoding device 1600 to access the original target block and select the final BM as a candidate.

Sparse search and refinement search can be performed to build a candidate list of size N. First, in sparse search, the subsampling factor can be set to a specific value. The specific value can be 2 or 3. In sparse search, the top 2N BVs with the smallest template matching costs can be maintained. Next, in refinement search, for each BV of 2N BVs, 3×3 blocks surrounding the BV can be checked. According to this check, the top N BVs can be selected to form the candidate list. For example, N may be 15 or 19.

Figure 68 shows syntax for template matching according to an example.

intra_tmp_flag indicates whether the intra prediction type for the target block is IntraTMP. IntraTMP may be a method of deriving an optimal template matching block within the target picture described in the embodiments.

A value of intra_tmp_flag of 1 may indicate that the target block uses IntraTMP.

intra_tmp_fusion_flag can indicate whether fusion is used for the target block.

intra_tmp_fusion_idx can specify a set of candidates used for IntraTMP synthesis. intra_tmp_fusion_idx can be an integer between 0 and 2. intra_tmp_fusion_idx can be used to indicate one of three sets. The three sets can be {BV0 to BV4}, {BV5 to BV9}, and {BV10 to BV14}.

intra_tmp_fusion_weight_type may indicate whether a SAD-based weight derivation method or a Wiener-filter-based derivation method is used.

intra_tmp_idx can specify the BV that is (actually) used for the target block among the BVs in the candidate list. intra_tmp_idx can be an integer between 0 and 18. Candidates derived from the L-shaped template, upper template, and left template may be included in the same candidate list.

In other words, one candidate among candidates derived by a plurality of templates can be selected through intra_tmp_idx.

intra_tmp_sub_pel_precision_idx can specify the precision index for the target block. intra_tmp_sub_pel_precision_idx may be an integer between 0 and 3. A value of 0 may indicate integer-pel precision. A value of 1 may indicate 1/2-pel precision. A value of 2 may indicate 1/4-pel precision. A value of 3 may indicate 3/4-pel precision.

intra_tmp_sub_pel_direction_idx can specify the sub-pel direction index for the target block. intra_tmp_sub_pel_direction_idx can be an integer between 0 and 7.

Intra-template matching prediction synthesis

Intra template matching prediction synthesis including the following features can be used.

1) Generation of multiple matched blocks during IntraTMP search

While the subsampled IntraTMP search process is performed, a candidate list may be initially created. The candidate list may represent the 30 matched blocks with the smallest template SAD.

For each of the 30 matched blocks, a full pixel refinement search can be performed within a small range. The sampling factor of the subsampled search processor may be 3. The improvement area may be a 3×3 block around each of the 30 matched blocks.

Next, the optimal three candidate matched blocks as measured by the template SAD around all improvement areas can be selected.

2) Selection of candidate matched blocks for synthesis

For each of the three optimal candidate matched blocks, a threshold can be used to determine whether the candidate matched block is used for synthesis, as shown in Equation 97 below.

[Formula 97]

Threshold = SAD ₁ << 1

Here, SAD ₁ may be the minimum template SAD of 3 candidate matched blocks. If the SAD of the candidate matched block is less than or equal to the Threshold, the candidate matched block can be used for synthesis. Accordingly, the number of candidate matched blocks can be determined.

3) Calculation of each composite weight of selected matched blocks

Once the blocks are determined to be composited, the blocks can be composited with composite weights. Composite weights can be calculated by the SAD of the blocks. Composite weights can be calculated as Equation 98 and Equation 99 below.

[Formula 98]

[Formula 99]

To reduce implementation cost, division operators can be replaced with an integer look-up table (LUT).

4) The final synthesized predictor can be determined according to Equation 100 below. A predictor may be a prediction block.

[Formula 100]

Here, p _i may be the i-th matched block. n may be the number of blocks selected by synthesis.

In step 2) described above, if only one matched block remains, the final predictor can be calculated according to Equation 101 below.

[Formula 101]

P _tmp may be one matched block. P _intra may be a predictor derived by planner mode. w ₁ and w ₂ may be weights. For example, w ₁ may be 7/8. w ₂ may be 1/8.

A flag for the target block may be signaled to indicate whether the target block is predicted using the above-described synthesis method or the existing method. The target block may be a CU.

Use of sub-pel precision in IntraTMP

Figure 69 shows adjacent half-pel positions in eight directions according to one example.

Figure 70 shows a first binarization of a precision index according to an example.

Figure 71 shows a second binarization of a precision index according to an example.

As shown in Figure 69, multiple sub-pel precisions may be used in IntraTMP. The plurality of sub-pel precisions may be three. The plurality of sub-pel precisions may include half-pel, quarter-pel, and 3-quarter-pel. Eight directions around the integer-pel position can be supported. That is, a quarter-pel position, a half-pel position, a 3-quarter-pel position, and an integer-pel position in each of the eight directions with the integer-pel positions as the center can be used.

The eight directions can be sorted by template cost with half-pel precision for each target block. The target block may be a CU. Among the eight directions, four directions with the lowest template costs can be used for each sub-pel precision.

If IntraTMP mode is selected for the target block, a precision index may be signaled to indicate which of the integer-pel and three sub-pel precisions is used. Figure 70 may illustrate signaling of precision index.

If one of the three sub-pel precisions is used, a direction index may be signaled to indicate which of the four directions is used. Figure 71 may illustrate signaling of a precision index.

A 4-tap Discrete Cosine Transform-Interpolation Filter (DCT-IF) can be used for sub-pel interpolation within IntraTMP.

Template matching using linear filters

Figure 72 shows the spatial portion of a filter according to one example.

Figure 73 shows search areas used to derive filter coefficients according to an example.

A linear filter model can be applied to the prediction of IntraTMP.

In one embodiment, the filter may be 6 taps. The filter may contain a 5-tab component. The 5-tap component may be a spatial component that has the form of a 5-tap plus sign. The filter may include a bias term.

The inputs to the spatial 5-tap component of the filter are: the center sample C within the reference block, the neighbor sample N above/north of the center sample, the neighbor sample S below/south of the center sample, and the neighbor sample S to the left/of the center sample. It may include a neighbor sample W to the west and a neighbor sample E to the right/east of the central sample.

The reference block may be located at positions corresponding to samples within the target block that is the target of prediction. That is, the reference block may be a template matching optimal block described in the embodiments. The reference template may be a template of the template matching optimal block.

The bias term B may represent a scalar offset between the input and output. The bias term B can be set to a middle luma value. For example, the bias term B for a 10-bit image may be 512.

The output of the filter can be calculated according to Equation 102 below.

[Formula 102]

predLumaVal = c0C + c1N + c2S + c3E + c4W + c5B

Filter coefficients ci can be calculated by minimizing the MSE between the reference template and target template.

The size and shape of the template may be the same as those described in the embodiments.

For example, the size of the template used for training may be determined to include four lines adjacent to the top of the target block and four lines adjacent to the left end of the target block. At this time, whether the line is included in the template may be determined based on the availability of the line.

Extensions to the area shown in Figure 73 are required to support side samples of the plus sign shaped spatial filter. The side samples may be one or more of neighboring sample N, neighboring sample S, neighboring sample W, and neighboring sample E. In other words, a side sample may be a sample that does not exist within the original region of the reference template.

If the side sample is within an unavailable area, padding for the side sample may be performed. The value determined by padding can be used as the value of the side sample.

Filter parameters may be derived using regression based MSE minimization (e.g., a Gaussian solver). Filter parameters can be used for other methods such as CCCM.

Information indicating whether the above-described linear filter model is applied to prediction of IntraTMP may be signaled. For example, the above information may be a CU level flag.

The mode using the linear filter model may be a sub-mode of IntraTMP. In other words, information indicating whether the above-described linear filter model is applied to the prediction of IntraTMP can be signaled only when the value of the IntraTMP flag is true.

Composition of reference blocks

Figure 74 shows the procedure of IntraTMP synthesis according to one example.

The optimal matching template can be obtained within a predefined search area with a step size of 2.

A refined search with a step size of 1 can be performed around the optimal matching template, and the final optimal matching template can be obtained by the refined search.

The block vector may be an offset between the final optimal matching template and the target template.

For the IntraTMP synthesis method, BVs corresponding to N matching templates can be searched with a step size of 2.

Next, N candidate reference blocks corresponding to N candidate matching templates can be obtained through BVs determined by search.

Next, to predict the target block, candidate reference blocks may be synthesized according to weights.

The quantity N of final candidates can be determined by the number of valid candidates actually found. N may be 1 or more and 3 or less.

The synthesis method may be as shown in Equation 103 below.

[Formula 103]

Here, predSamples may represent the final prediction block.

refBlock _n may be one of N candidate reference blocks.

midValue may be an offset. When the target image is a 10-bit image, midValue may be 512.

Weights W _n can be derived from the candidate matching templates and the target template using a Gaussian solution. Gaussian solution can be used in CCCM.

IntraTMP usage information (= intra_tmp_flag) indicating whether IntraTMP is used may be signaled.

IntraTMP synthesis usage information indicating whether IntraTMP synthesis of the embodiment is used may be signaled. Each of the IntraTMP usage information and IntraTMP composite usage information may be a flag and may be signaled at the CU level. The IntraTMP synthesis method of the embodiment may be a sub-mode of IntraTMP.

IntraTMP composite usage information may be signaled along with IntraTMP usage information. If the signaled IntraTMP usage information indicates that IntraTMP is used, IntraTMP composite usage information may be signaled next.

Adjustment of navigation area

The search area described in the embodiments may be adjusted.

For example, the size of the search area may be (5W, 5H). W may be the horizontal size of the target block. H may be the vertical size of the target block. This search area may include five areas. The search area may include the decrypted area of the target CTU including the target block.

The size of this search area may be limited. The size of the search area may be (min(aW, n), min(bHW, m)). Each of a, b, n, and m may be an integer of 1 or more. For example, a may be 5. b may be 5. c may be 64. d may be 64. In other words, the size of the search area can be adjusted to increase in proportion to the size of the target block and not exceed a specific rectangular area.

Figure 75 shows a search area of IntraTMP according to an example.

IntraTMP can use the template of a predefined target block to search for a candidate template with the same shape that optimally matches the template of the target block within a predetermined search area.

A cost function such as SAD can be defined to search for the optimal template. Both the encoding device 1600 and the decoding device 1700 may use the same search strategy to obtain an optimally matched block. Once the optimally matching block is found, the corresponding reference block can be selected as a prediction block for coding of the target block.

As shown in FIG. 75, the search area of IntraTMP may include four areas R1, R2, R3, and R4. The four regions may contain reconstructed samples from the top and/or left CTUs. The four regions may include reconstructed samples in the target CTU located above and to the left of the target block.

The search order of the four regions may be R4, R1, R2, and R3.

In Figure 75, a target block with a horizontal length of W and a vertical length of H is shown. The gray portion may represent areas of the upper left corners of all reference blocks within the reconstructed area.

searchRangeWidth can be the maximum horizontal size. searchRangeHeight can be the maximum vertical size.

searchRangeWidth can be set to max(5W, 64). searchRangeHeight can be set to max(5W, 64).

Figure 76 shows an expanded search area of IntraTMP according to an example.

The lower left area and upper right area relative to the target block may also be available and may be very close to the target block.

The search area of IntraTMP may include six areas R1, R2, R3, R4, R5, and R6. The search area of IntraTMP may be expanded as the search area further includes areas R5 and R6.

The search order of the six regions may be R4, R5, R6, R1, R2, and R3.

Figure 77 shows a limited search area of IntraTMP according to an example.

In one embodiment, a search may be performed across all available locations for a particular object. For example, a specific target may be a target CTU. This navigation can go beyond a certain navigation range or width/height.

Any search area can be limited to a specific search range. Alternatively, all search areas may be limited to a certain width and/or a certain height.

As shown in FIG. 77, the search range may be limited by searchRangeHeight and searchRangeHeight.

searchRangeHeight may be the vertical distance from the top left of the target block to the top of the limited search range. searchRangeHeight may be the vertical distance from the bottom left of the target block to the bottom of the limited search range. searchRangeWidth may be the horizontal distance from the upper left corner of the target block to the left end of the limited search range. searchRangeWidth may be the horizontal distance from the upper right corner of the target block to the right end of the limited search range.

That is, the size of the search range can be limited to (2·searchRangeWidth + H, 2·searchRangeHeight + W).

When the coordinates of the top left pixel of the target block are (x _top-left , y _top-left ), pixels with x coordinates greater than x _top-left and y coordinates greater than y _top-left will be excluded from the search range. You can. That is, the search range may consist of pixels whose x-coordinate is smaller than x _top-left or whose y-coordinate is smaller than y _top-left .

Adjustment of block vector resolution

Block vectors can be used for coding in IntraTMP mode. After this use, the block vector of IntraTMP can be stored for coding for IBC blocks.

In one embodiment, the resolution of the block vector of the stored IntraTMP may be full-pel resolution. That is, only block vectors with full-pel resolution can be selectively stored.

In one embodiment, the resolution of the block vector of the stored IntraTMP may include 1/2-pel resolution and 1/4-pel resolution. That is, block vectors with full-pel resolution, block vectors with 1/2-pel resolution, and block vectors with 1/4-pel resolution can be stored. Accordingly, search and storage of block vectors can be performed using either 1/2-pel resolution or 1/4-pel resolution.

The stored block vectors can be used for history-based motion vector prediction (HMVP).

Restrictions on navigation areas using IBC reference areas

Figure 78a shows an IBC search area when the CTU size is 128 according to an example.

Figure 78b shows an IBC search area when the CTU size is 256 according to an example.

Regardless of whether the CTU size is 128 or 256, the IBC reference region may contain 245 samples above the target CTU. Additionally, the IBC reference area may include all left neighboring CTUs within the row containing the target CTU. Additionally, when the CTU size is 128, the IBC reference area may include two CTU rows above the target CTU. If the CTU size is 256, the IBC reference area may include one CTU row above the target CTU.

In these cases, the top row of CTUs may be constrained to have one or two CTUs further to the left than the target CTU.

The upper left corner of the IBC reference area may be located 256 samples to the left of the left end of the target CTU. The upper left corner of the IBC reference area may be located 256 samples above the top of the target CTU.

Figure 79 shows an IntraTMP search area when the block size is 64×64 according to an example.

In Figure 79, the CTU size may be 128.

The IntraTMP search area may include multiple sub-areas (=R1, R2, and R3).

The IntraTMP search area can be set to be proportional to the block size (CbWidth, CbHeight). In this setting, a factor of 5 can be used.

In other words, the IntraTMP search area can be set as shown in Equation 104 below.

[Formula 104]

(SearchRange_w=5*CbWidth, SearchRange_h=5*CbHeight)

As shown in Figure 79, the block size may be 64×64. The top sub-area R1 may exceed the left and top boundaries of the IBC search area. These excesses may require larger memory compared to the IBC rebuilt buffer.

Figure 80 shows code for limiting the search range according to an example.

IntraTMP search area and IBC search area can be harmonized. The same memory size for reconstructed samples that is used for both the IntraTMP search area and the IBC search area can be used.

Figure 81 shows a TMP search area and an IBC search area with two different sampling rates according to an example.

In one embodiment, the size of IntraTMP sub-regions may be limited. Through constraints, it can be confirmed whether IntraTMP sub-regions are included within the IBC boundaries of the search area. As a result, for a block size of 64×64, the top and left boundaries of sub-area R1 can be constrained within the boundaries of the IBC search area.

In one embodiment, the boundaries between overlapping R1 and R3 may be changed. This overlap can increase the number of template matching calculations for all block sizes. This modification can increase the upper boundary (iVerMin) of R3 by one sample.

Figure 80 discloses code to increment one sample from the upper boundary of R3.

Figure 81 shows IntraTMP search area and IBC search area with two different sampling rates according to an example.

If the size of the target block is large, the IntraTMP search area may exceed the IBC search area. In this case, the IntraTMP search area may be limited by the IBC search area.

In one embodiment, IntraTMP sub-regions can be tailored to the specific shape of the IBC search area. Through these adjustments, the IntraTMP search area can be further expanded. Figure 81 shows the placement of new IntraTMP sub-regions.

Different subsampling may be applied to sub-regions. The sampling rates of subsamplings may be different.

For example, a sampling rate of SR ₁ may be used for a specific area of the IntraTMP search area. For example, SR ₁ may be 3. The specific area may be an area whose size is 5 times the size of the target block. A sampling rate of SR ₂ may be used for areas other than the specific area among the IntraTMP search areas. For example, SR ₂ may be the power of 8.

In one embodiment, the IntraTMP search area may be extended to the IBC search area. At this time, when performing IntraTMP search, a sampling rate of 3 can be used within the IntraTMP search area (before expansion). A sampling rate of 5 can be used within the IBC search area (added as the IntraTMP search area by extension).

The above embodiments may be performed in the encoding device 1600 and the decoding device 1700 using the same method and/or corresponding method. Additionally, a combination of one or more of the above embodiments may be used in encoding and/or decoding an image.

The order in which the above embodiments are applied may be different in the encoding device 1600 and the decoding device 1700. Alternatively, the order in which the above embodiments are applied may be (at least partially) the same in the encoding device 1600 and the decoding device 1700.

The above embodiments can be performed for each of the luma signal and the chroma signal. The above embodiments can be performed in the same way for luma signals and chroma signals.

The shape of the block to which the above embodiments are applied may have a square shape or a non-square shape.

Whether to apply and/or perform at least one of the above embodiments may be determined based on conditions regarding the size of the block. That is, at least one of the above embodiments can be applied and/or performed when the conditions for the size of the block are met. Conditions may include minimum block size and maximum block size. The block may be one of the blocks described above in the embodiments and the units described above in the embodiments. The block to which the minimum block size is applied and the block to which the maximum block size is applied may be different.

For example, when the size of a block is greater than or equal to the minimum size and/or when the size of the block is less than or equal to the maximum size, the above-described embodiments may be applied and/or performed. If the size of the block is larger than the minimum size and/or if the size of the block is less than or equal to the maximum size, the above-described embodiments may be applied and/or performed.

For example, the above-described embodiment can be applied only when the block size is a predefined block size. Predefined block sizes can be 2x2, 4x4, 8x8, 16x16, 32x32, 64x64 or 128x128. The predefined block size may be ( ₂ * _{SIZE SIZE} SIZE _Y may be one of integers greater than or equal to 1.

For example, the above-described embodiment can be applied only when the block size is greater than or equal to the minimum block size. The above-described embodiment can be applied only when the size of the block is larger than the minimum block size. Block minimum sizes can be 2x2, 4x4, 8x8, 16x16, 32x32, 64x64 or 128x128. Alternatively, the minimum block size may be (2*SIZE _{MIN_X} )x(2*SIZE _{MIN_Y} ). SIZE _{MIN_X} can be one of integers greater than 1. SIZE _{MIN_Y} can be one of integers greater than 1.

For example, the above-described embodiment can be applied only when the block size is less than or equal to the maximum block size. The above-described embodiment can be applied only when the block size is smaller than the maximum block size. The maximum block size can be 2x2, 4x4, 8x8, 16x16, 32x32, 64x64 or 128x128. Alternatively, the maximum block size may be (2*SIZE _{MAX_X} )x(2*SIZE _{MAX_Y} ). SIZE _{MAX_X} can be one of integers greater than 1. SIZE _{MAX_Y} can be one of integers greater than 1.

For example, the above-described embodiment can be applied only when the block size is greater than or equal to the minimum block size and less than or equal to the maximum block size. The above-described embodiment can be applied only when the block size is greater than the minimum block size and less than or equal to the maximum block size. The above-described embodiment can be applied only when the block size is greater than or equal to the minimum block size and smaller than the maximum block size. The above-described embodiment can be applied only when the block size is larger than the minimum block size and smaller than the maximum block size.

In the above-described embodiments, the size of the block may mean the horizontal size of the block or the vertical size of the block. The size of the block may refer to both the horizontal size of the block and the vertical size of the block. Additionally, the size of the block may mean the area of the block. Each of the area, minimum block size, and maximum block size can be one of integers greater than or equal to 1. Additionally, the size of the block may mean the result (or value) of a known formula using the horizontal and vertical sizes of the block or the result (or value) of a formula in an embodiment.

Additionally, in the above embodiments, the first embodiment may be applied to the first size, and the second embodiment may be applied to the second size.

The above embodiments can be applied according to the temporal layer. A separate identifier may be signaled to identify the temporal layer to which the above embodiments can be applied, and the above embodiments may be applied to the temporal layer specified by the identifier. The identifier here may be defined as the lowest layer and/or highest layer to which the above embodiment is applicable, or may be defined to indicate a specific layer to which the above embodiment is applicable. Additionally, a fixed temporal hierarchy to which the above embodiments are applied may be defined.

For example, the above embodiments can be applied only when the temporal layer of the target image is the lowest layer. For example, the above embodiments can be applied only when the temporal layer identifier of the target image is 1 or more. For example, the above embodiments can be applied only when the temporal layer of the target image is the highest layer.

A slice type or tile group type to which the above embodiments are applied may be defined, and the above embodiments may be applied depending on the corresponding slice type or tile group type.

In the above-described embodiments, in applying a specified process to a specified object, specified conditions may be required, and when it is described that the specified processing is processed under a specified decision, the specified coding parameters If it has been described that it is determined whether a specified condition is satisfied or that a specified decision is made based on a specified coding parameter, the specified coding parameter may be interpreted as being replaceable with another coding parameter. That is, coding parameters affecting specified conditions or specified decisions may be considered merely exemplary, and combinations of one or more other coding parameters in addition to the specified coding parameters will be understood to play the role of the specified coding parameters. You can.

In the above-described embodiments, the methods are described based on flowcharts as a series of steps or units, but the present invention is not limited to the order of steps, and some steps may occur in a different order or simultaneously with other steps as described above. You can. Additionally, a person of ordinary skill in the art will recognize that the steps shown in the flowchart are not exclusive and that other steps may be included or one or more steps in the flowchart may be deleted without affecting the scope of the present invention. You will understand.

The above-described embodiments include examples of various aspects. Although not all possible combinations for representing various aspects can be described, those skilled in the art will recognize that other combinations are possible in addition to those explicitly described. Accordingly, the present invention is intended to include all other substitutions, modifications and changes falling within the scope of the following claims.

Embodiments according to the present invention described above can be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention, or may be known and usable by those skilled in the computer software field.

A computer-readable recording medium may contain information used in embodiments according to the present invention. For example, a computer-readable recording medium may include a bitstream, and the bitstream may include information described in embodiments according to the present invention.

A bitstream may contain computer-executable code and/or programs. Computer-executable code and/or program may include information described in the embodiments and may include syntax elements described in the embodiments. That is, the information and syntax elements described in the actual example may be considered computer-executable code within the bitstream, and may be considered at least part of the computer-executable code and/or program represented by the bitstream.

Computer-readable recording media may include non-transitory computer-readable medium.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks. media) and specially configured hardware devices such as ROM, RAM, flash memory, etc. to store and execute program instructions. Examples of program instructions include not only machine language code such as that created by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The above hardware devices may be configured to operate as one or more software modules to perform processing according to the invention and vice versa.

In the above, the present invention has been described with specific details such as specific components and limited embodiments and drawings, but this is only provided to facilitate a more general understanding of the present invention, and the present invention is not limited to the above embodiments. No, those skilled in the art can make various modifications and changes based on this description.

Accordingly, the spirit of the present invention should not be limited to the above-described embodiments, and the scope of the patent claims described later as well as all modifications equivalent to or equivalent to the scope of the claims shall fall within the scope of the spirit of the present invention. It will be said that it belongs.

Claims (20)

determining a template for a target block;
selecting a template matching reference image;
Searching for an optimal template matching block within the template matching reference image using the template; and
Performing decoding using the template matching optimal block
A video decoding method including. According to paragraph 1,
A video decoding method in which subsampling is used in generating the template. According to paragraph 1,
An image decoding method in which a plurality of template matching reference image construction methods among a plurality of template matching reference image construction methods are selected to select the template matching reference image. According to paragraph 1,
A video decoding method in which one search area among search areas having different shapes is selected as a search area for deriving the optimal template matching block. According to paragraph 1,
A video decoding method in which a template matching search starting point in a search area for deriving the template matching optimal block is set. According to paragraph 1,
A video decoding method in which template matching-based intra residual signal prediction is performed for the decoding. According to paragraph 1,
A video decoding method in which blending prediction of inter and template matching is used for the decoding. determining a template for a target block;
selecting a template matching reference image;
Searching for an optimal template matching block within the template matching reference image using the template; and
Performing encoding using the template matching optimal block
A video encoding method including. According to clause 8,
A video encoding method in which subsampling is used in generating the template. According to clause 8,
An image encoding method in which a plurality of template matching reference image construction methods among a plurality of template matching reference image construction methods are selected to select the template matching reference image. According to clause 8,
A video encoding method in which one search area among search areas having different shapes is selected as a search area for deriving the template matching optimal block. According to clause 8,
A video encoding method in which a template matching search starting point in a search area for deriving the template matching optimal block is set. According to clause 8,
A video encoding method in which template matching-based intra residual signal prediction is performed for the decoding. According to clause 8,
A video decoding method in which blending prediction of inter and template matching is used for the encoding. A computer-readable recording medium storing a bitstream generated by the video encoding method of claim 8. In the computer-readable recording medium storing a bitstream for video decoding, the bitstream includes:
coding information
Including,
The template for the target block is determined,
A template matching reference image is selected,
Using the template, an optimal template matching block is searched within the template matching reference image,
Decryption using the template matching optimal block is performed,
The coding information is used for at least one of determining the template, selecting the reference image, deriving the optimal block matching the template, and decoding. According to clause 16,
A computer-readable recording medium in which subsampling is used in generating the template. According to clause 16,
A computer-readable recording medium in which a plurality of template matching reference image construction methods are selected among a plurality of template matching reference image construction methods to select the template matching reference image. According to clause 16,
A computer-readable recording medium in which one search area among search areas having different shapes is selected as a search area for deriving the optimal template matching block. According to clause 16,
A computer-readable recording medium in which a template matching search starting point of a search area for deriving the template matching optimal block is set.

Images