Image encoding / decoding method, storage medium, and data transmission method

Abstract

The invention relates to an image encoding / decoding method, a storage medium, and a data transmission method. An image decoding method according to the present document comprises a step of deriving a residual sample, in which the step of deriving the residual sample comprises the steps of: deriving a transform kernel for an inverse primary transform applied to a sub-partition block on the basis of a horizontal length or a vertical length of the sub-partition block when a current block is divided into the sub-partition block; and deriving residual samples from the transform coefficients based on the transform kernel.

Description

[0001] This application is a divisional application of the invention patent application with the original application number 202080072344.7 (International application number: PCT / KR2020 / 012707, application date: September 21, 2020, invention name: Transform-based image coding method and device). Technical Field

[0002] The present disclosure relates to an image coding technology, and more particularly, to a method and apparatus for encoding an image based on transformation in an image coding system. Background Art

[0003] Nowadays, the demand for high-resolution and high-quality images / videos such as 4K, 8K or higher ultra-high-definition (UHD) images / videos has been growing in various fields. As image / video data becomes higher resolution and higher quality, the amount of information or bit volume sent increases compared to traditional image data. Therefore, when using a medium such as a traditional wired / wireless broadband line to send image data or using an existing storage medium to store image / video data, its transmission cost and storage cost increase.

[0004] In addition, nowadays, interest and demand for immersive media such as virtual reality (VR) and artificial reality (AR) content or holograms are increasing, and broadcasting of images / videos having image characteristics different from real images such as game images is increasing.

[0005] Therefore, there is a need for an efficient image / video compression technology that effectively compresses and transmits or stores and reproduces information of high-resolution and high-quality images / videos having various characteristics as described above. Summary of the invention

[0006] Technical Purpose

[0007] A technical aspect of the present disclosure is to provide a method and an apparatus for increasing image encoding efficiency.

[0008] Another technical aspect of the present disclosure is to provide a method and apparatus for increasing the efficiency of transform index encoding.

[0009] Yet another technical aspect of the present disclosure is to provide an image encoding method and apparatus using MTS.

[0010] Still another technical aspect of the present disclosure is to provide a method and apparatus for encoding an image for applying an MTS to a sub-partitioned transform block.

[0011] Technical Solution

[0012] According to an embodiment of the present specification, an image decoding method performed by a decoding device is provided herein. The method may include deriving residual samples, wherein the step of deriving residual samples may include: when a current block is divided into sub-partition blocks, deriving a transform kernel for an inverse primary transform applied to the sub-partition blocks based on a horizontal length or a vertical length of the sub-partition blocks; and deriving residual samples from transform coefficients based on the transform kernel.

[0013] The transform kernel may be derived based on whether the horizontal length or the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, and the residual sample may be derived based on the size of the non-cleared block indicating an area within the sub-partition block to which the transform kernel is applied and in which valid transform coefficients exist.

[0014] If each of the horizontal length and the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, the transform kernel may be derived as DST-7, and the width and height of the non-zero block may be set to 16 each.

[0015] If each of the horizontal length and the vertical length of the sub-partition block does not satisfy the condition of being equal to or greater than 4 and less than or equal to 16, the transform kernel can be derived as DCT-2, and the width and height of the non-cleared block can each be set to 32 or less.

[0016] The MTS index indicating the transform core of the current block may not be signaled.

[0017] The step of deriving the residual sample may also include: deriving a modified transform coefficient by applying LFNST to the sub-partition block based on the LFNST index applied to the current block being not equal to 0, wherein the transform kernel may be derived as DCT-2, and wherein the width and height of the non-cleared block may each be set to 32 or less.

[0018] According to an embodiment of the present specification, an image encoding method performed by an encoding device is provided herein. The method may include deriving a transform coefficient by applying a primary transform to a residual sample, wherein when a current block is divided into sub-partition blocks, the step of deriving the transform coefficient may include: deriving a transform kernel of the primary transform applied to the sub-partition block based on a horizontal length or a vertical length of the sub-partition block; and deriving the transform coefficient from the residual sample based on the transform kernel.

[0019] According to still another embodiment of the present disclosure, a digital storage medium storing image data including a bit stream generated according to an image encoding method performed by an encoding device and encoded image information may be provided.

[0020] According to yet another embodiment of the present disclosure, a digital storage medium storing image data including encoded image information and a bit stream so that a decoding device performs an image decoding method may be provided.

[0021] Technical Effects

[0022] According to the present disclosure, the overall image / video compression efficiency can be increased.

[0023] According to the present disclosure, the efficiency of transform index encoding can be increased.

[0024] According to the present disclosure, the overall image / video compression efficiency can be increased.

[0025] According to the present disclosure, the efficiency of transform index encoding can be increased.

[0026] Technical aspects of the present disclosure may provide an image encoding method and apparatus using MTS.

[0027] Technical aspects of the present disclosure may provide a method and apparatus for encoding an image for applying an MTS to a sub-partitioned transform block.

[0028] The effects that can be obtained by the specific examples of the present disclosure are not limited to the effects listed above. For example, there may be various technical effects that can be understood by a person of ordinary skill in the relevant field or derived from the present disclosure. Therefore, the specific effects of the present disclosure are not limited to those explicitly described in the present disclosure, and may include various effects that can be understood or derived from the technical features of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 An example of a video / image encoding system to which the present disclosure can be applied is schematically illustrated.

[0030] Figure 2 is a diagram schematically illustrating a configuration of a video / image encoding device to which the present disclosure can be applied.

[0031] Figure 3 is a diagram schematically illustrating a configuration of a video / image decoding device to which the present disclosure can be applied.

[0032] Figure 4 This is a structural diagram of a content flow system to which this specification is applied.

[0033] Figure 5 The multi-conversion technology according to the embodiments of this specification is schematically illustrated.

[0034] Figure 6 An exemplary intra directional mode is shown for 65 prediction directions.

[0035] Figure 7is a diagram illustrating an RST according to an embodiment of the present specification.

[0036] Figure 8 is a diagram illustrating an order of arranging output data of a forward primary transform into a one-dimensional vector according to an example.

[0037] Fig. 9 is a diagram illustrating an order of arranging output data of a forward quadratic transform into two-dimensional blocks according to an example.

[0038] Fig.10 is a diagram illustrating a wide-angle intra prediction mode according to an embodiment of the present specification.

[0039] Fig.11 is a diagram illustrating a block shape to which LFNST is applied.

[0040] Fig.12 is a diagram illustrating an arrangement of output data of a forward LFNST according to an example.

[0041] Fig.13 A diagram showing that the number of output data of the forward LFNST is limited to a maximum value of 16 according to an example.

[0042] Fig.14 is a diagram illustrating clearing in a block to which 4×4 LFNST is applied according to an example.

[0043] Fig.15 is a diagram illustrating clearing in a block to which 8×8 LFNST is applied according to an example.

[0044] Fig.16 is a diagram illustrating clearing in a block to which 8×8 LFNST is applied according to another example.

[0045] Fig.17 An example of sub-blocks into which a coding block is divided is illustrated.

[0046] Fig.18 Another example of sub-blocks into which a coding block is divided is illustrated.

[0047] Fig.19 is a diagram illustrating symmetry between an M×2(M×1) block and a 2×M(1×M) block according to an example.

[0048] Fig. 20 is a diagram illustrating an example of transposing a 2×M block according to an example.

[0049] Fig.21 The scanning order of an 8×2 or 2×8 area according to an example is illustrated.

[0050] Fig. 22is a diagram showing an arrangement of 36 coefficients according to a 16×36 LFNST kernel according to an example.

[0051] Fig.23 is a flowchart illustrating the operation of a video decoding device according to an embodiment of the present specification.

[0052] Fig.24 is a flowchart illustrating the operation of a video encoding device according to an embodiment of the present specification. DETAILED DESCRIPTION

[0053] Although the present disclosure may be susceptible to various modifications and includes various embodiments, its specific embodiments have been shown by way of example in the accompanying drawings, and will now be described in detail. However, this is not intended to limit the present disclosure to the specific embodiments disclosed herein. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the technical ideas of the present disclosure. Unless the context clearly indicates otherwise, the singular form may include the plural form. Terms such as "including" and "having" are intended to indicate the presence of features, numbers, steps, operations, elements, components, or combinations thereof used in the following description, and therefore should not be understood as precluding the possibility of the presence or addition of one or more different features, numbers, steps, operations, elements, components, or combinations thereof.

[0054] In addition, for the convenience of describing the different characteristic functions, each component on the drawings described herein is illustrated independently, however, it is not intended that each component is implemented by separate hardware or software. For example, any two or more of these components can be combined to form a single component, and any single component can be divided into multiple components. Embodiments in which components are combined and / or divided will fall within the scope of the patent rights of the present disclosure as long as they do not depart from the essence of the present disclosure.

[0055] Hereinafter, preferred embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. In addition, in the accompanying drawings, the same reference numerals are used for the same components, and repeated description of the same components will be omitted.

[0056] This document relates to video / image coding. For example, the methods / examples disclosed in this document may relate to the VVC (Versatile Video Coding) standard (ITU-T Rec. H.266), the next generation video / image coding standard after VVC, or other video coding related standards (e.g., HEVC (High Efficiency Video Coding) standard (ITU-T Rec. H.265), EVC (Basic Video Coding) standard, AVS2 standard, etc.).

[0057] In this document, various embodiments related to video / image encoding may be provided, and unless otherwise specified, the embodiments may be combined with each other and performed.

[0058] In this document, video can refer to a collection of a series of images over a period of time. Generally, a picture refers to a unit of an image representing a specific time region, and a slice / tile is a unit that constitutes a part of a picture. A slice / tile may include one or more coding tree units (CTUs). A picture may consist of one or more slices / tiles. A picture may consist of one or more tile groups. A tile group may include one or more tiles.

[0059] A pixel or a picture element (pel) may refer to the smallest unit constituting a picture (or image). In addition, a "sample" may be used as a term corresponding to a pixel. A sample may generally represent a pixel or a pixel value, and may represent only a pixel / pixel value of a luminance component or only a pixel / pixel value of a chrominance component. Alternatively, a sample may mean a pixel value in a spatial domain, or when the pixel value is transformed into a frequency domain, it may mean a transform coefficient in a frequency domain.

[0060] A unit may represent a basic unit of image processing. A unit may include at least one of a specific region and information related to the region. A unit may include a luminance block and two chrominance (e.g., cb, cr) blocks. Depending on the situation, terms such as unit, block, region, etc. may be used interchangeably. In general, an M×N block may include a set (or array) of samples (or sample arrays) or transform coefficients consisting of M columns and N rows.

[0061] In this document, the terms " / " and "," should be interpreted as indicating "and / or". For example, the expression "A / B" may mean "A and / or B". In addition, "A, B" may mean "A and / or B". In addition, "A / B / C" may mean "at least one of A, B, and / or C". In addition, "A / B / C" may mean "at least one of A, B, and / or C".

[0062] In addition, in this document, the term "or" should be interpreted as indicating "and / or". For example, the expression "A or B" may include 1) only A, 2) only B, and / or 3) both A and B. In other words, the term "or" in this document should be interpreted as indicating "additionally or alternatively".

[0063] In the present disclosure, "at least one of A and B" may mean "only A", "only B", or "both A and B". In addition, in the present disclosure, the expression "at least one of A or B" or "at least one of A and / or B" may be interpreted as "at least one of A and B".

[0064] Furthermore, in the present disclosure, “at least one of A, B, and C” may mean “only A”, “only B”, “only C”, or “any combination of A, B, and C”. Furthermore, “at least one of A, B, or C” or “at least one of A, B and / or C” may mean “at least one of A, B, and C”.

[0065] In addition, the brackets used in the present disclosure may mean "for example". Specifically, when the indication is "prediction (intra-frame prediction)", it may mean that "intra-frame prediction" is proposed as an example of "prediction". In other words, the "prediction" of the present disclosure is not limited to "intra-frame prediction", and "intra-frame prediction" is proposed as an example of "prediction". In addition, when the indication is "prediction (ie, intra-frame prediction)", this may also mean that "intra-frame prediction" is proposed as an example of "prediction".

[0066] Technical features described separately in one drawing in the present disclosure may be implemented separately or may be implemented simultaneously.

[0067] Figure 1 An example of a video / image encoding system to which the present disclosure can be applied is schematically illustrated.

[0068] Reference Figure 1 The video / image coding system may include a first device (source device) and a second device (receiving device). The source device may transmit the coded video / image information or data to the receiving device in the form of a file or stream via a digital storage medium or a network.

[0069] The source device may include a video source, an encoding device, and a transmitter. The receiving device may include a receiver, a decoding device, and a renderer. The encoding device may be referred to as a video / image encoding device, and the decoding device may be referred to as a video / image decoding device. The transmitter may be included in the encoding device. The receiver may be included in the decoding device. The renderer may include a display, and the display may be configured as a separate device or an external component.

[0070] The video source can obtain the video / image by capturing, synthesizing or generating the process of the video / image. The video source may include a video / image capturing device and / or a video / image generating device. The video / image capturing device may include, for example, one or more cameras, a video / image archive including previously captured videos / images, etc. The video / image generating device may include, for example, a computer, a tablet computer, and a smart phone, and may (electronically) generate the video / image. For example, a virtual video / image may be generated by a computer, etc. In this case, the video / image capturing process may be replaced by the process of generating the relevant data.

[0071] The encoding device can encode the input video / image. The encoding device can perform a series of processes such as prediction, transformation and quantization for compression and coding efficiency. The encoded data (encoded video / image information) can be output in the form of a bitstream.

[0072] The transmitter may transmit the encoded video / image information or data output in the form of a bit stream to a receiver of a receiving device in the form of a file or stream through a digital storage medium or a network. The digital storage medium may include various storage media such as USB, SD, CD, DVD, Blu-ray, HDD, SSD, etc. The transmitter may include an element for generating a media file in a predetermined file format, and may include an element for transmitting through a broadcast / communication network. The receiver may receive / extract a bit stream and transmit the received / extracted bit stream to a decoding device.

[0073] The decoding device may decode a video / image by performing a series of processes such as dequantization, inverse transformation, prediction, etc. corresponding to the operation of the encoding device.

[0074] The renderer can render the decoded video / image, and the rendered video / image can be displayed through a display.

[0075] Figure 2 is a diagram schematically illustrating a configuration of a video / image encoding device to which the present disclosure can be applied. Hereinafter, the so-called video encoding device may include an image encoding device.

[0076] Reference Figure 2 , the encoding device 200 may include an image divider 210, a predictor 220, a residual processor 230, an entropy encoder 240, an adder 250, a filter 260, and a memory 270. The predictor 220 may include an inter-frame predictor 221 and an intra-frame predictor 222. The residual processor 230 may include a transformer 232, a quantizer 233, a dequantizer 234, and an inverse transformer 235. The residual processor 230 may further include a subtractor 231. The adder 250 may be referred to as a reconstructor or a reconstructed block generator. According to an embodiment, the image divider 210, the predictor 220, the residual processor 230, the entropy encoder 240, the adder 250, and the filter 260 described above may be composed of one or more hardware components (e.g., an encoder chipset or processor). In addition, the memory 270 may include a decoded picture buffer (DPB) and may be composed of a digital storage medium. The hardware component may further include a memory 270 as an internal / external component.

[0077] The image divider 210 may divide the input image (or picture or frame) input to the encoding device 200 into one or more processing units. As an example, the processing unit may be referred to as a coding unit (CU). In this case, starting from a coding tree unit (CTU) or a maximum coding unit (LCU), the coding unit may be recursively divided according to a quadtree binary tree ternary tree (QTBTTT) structure. For example, based on a quadtree structure, a binary tree structure, and / or a ternary tree structure, a coding unit may be divided into a plurality of coding units with a deeper depth. In this case, for example, a quadtree structure may be applied first, and a binary tree structure and / or a ternary tree structure may be applied later. Alternatively, a binary tree structure may be applied first. The encoding process according to the present disclosure may be performed based on a final coding unit that is not further divided. In this case, based on the coding efficiency according to the image characteristics, the maximum coding unit may be directly used as the final coding unit. Alternatively, the coding unit may be recursively divided into coding units of a deeper depth as needed, whereby a coding unit of the best size may be used as the final coding unit. Here, the encoding process may include processes such as prediction, transformation, and reconstruction, which will be described later. As another example, the processing unit may further include a prediction unit (PU) or a transform unit (TU). In this case, the prediction unit and the transform unit may be separated or divided from the above-mentioned final coding unit. The prediction unit may be a unit for sample prediction, and the transform unit may be a unit for deriving a transform coefficient and / or a unit for deriving a residual signal from a transform coefficient.

[0078] Unit and terms such as block, region, etc. may be used to replace each other depending on the situation. In general, an M×N block may represent a set of samples or transform coefficients consisting of M columns and N rows. A sample may generally represent a pixel or a pixel value, and may represent only a pixel / pixel value of a luminance component, or only a pixel / pixel value of a chrominance component. A sample may be used as a term corresponding to a pixel or a pel of a picture (or image).

[0079] The subtractor 231 subtracts the prediction signal (prediction block, prediction sample array) output from the predictor 220 from the input image signal (original block, original sample array) to generate a residual signal (residual block, residual sample array), and the generated residual signal is sent to the transformer 232. The predictor 220 can perform prediction on the processing target block (hereinafter referred to as the "current block"), and can generate a prediction block including the prediction samples of the current block. The predictor 220 can determine whether to apply intra-frame prediction or inter-frame prediction based on the current block or CU. As discussed later in the description of each prediction mode, the predictor can generate various information related to the prediction, such as prediction mode information, and send the generated information to the entropy encoder 240. Information about the prediction can be encoded in the entropy encoder 240 and output in the form of a bitstream.

[0080] The intra-frame predictor 222 can predict the current block by referring to the samples in the current picture. Depending on the prediction mode, the reference sample can be located near the current block or separated from the current block. In intra-frame prediction, the prediction mode may include a variety of non-directional modes and a variety of directional modes. The non-directional mode may include, for example, a DC mode and a plane mode. Depending on the level of detail of the prediction direction, the directional mode may include, for example, 33 directional prediction modes or 65 directional prediction modes. However, this is only an example, and more or fewer directional prediction modes may be used depending on the setting. The intra-frame predictor 222 may determine the prediction mode applied to the current block by using the prediction mode applied to the neighboring blocks.

[0081] The inter-frame predictor 221 may derive a prediction block for the current block based on a reference block (reference sample array) specified by a motion vector on a reference picture. At this time, in order to reduce the amount of motion information sent in the inter-frame prediction mode, the motion information may be predicted based on a block, a sub-block, or a sample based on the correlation of the motion information between the neighboring block and the current block. The motion information may include a motion vector and a reference picture index. The motion information may also include inter-frame prediction direction (L0 prediction, L1 prediction, Bi prediction, etc.) information. In the case of inter-frame prediction, the neighboring blocks may include spatial neighboring blocks existing in the current picture and temporal neighboring blocks existing in the reference picture. The reference picture including the reference block and the reference picture including the temporal neighboring block may be the same as or different from each other. The temporal neighboring block may be referred to as a collocated reference block, a collocated CU (colCU), etc., and the reference picture including the temporal neighboring block may be referred to as a collocated picture (colPic). For example, the inter-frame predictor 221 may configure a motion information candidate list based on the neighboring blocks, and generate information indicating which candidate is used to derive the motion vector and / or reference picture index of the current block. Inter-frame prediction may be performed based on various prediction modes. For example, in the case of the skip mode and the merge mode, the inter-frame predictor 221 can use the motion information of the neighboring block as the motion information of the current block. In the skip mode, unlike the merge mode, the residual signal cannot be sent. In the case of the motion information prediction (motion vector prediction, MVP) mode, the motion vector of the neighboring block can be used as a motion vector predictor, and the motion vector of the current block can be indicated by signaling the motion vector difference.

[0082] The predictor 220 can generate a prediction signal based on various prediction methods. For example, the predictor can apply intra prediction or inter prediction to the prediction of a block, and can also apply intra prediction and inter prediction at the same time. This can be called combined inter and intra prediction (CIIP). In addition, the predictor can be based on an intra-block copy (IBC) prediction mode or a palette mode to perform prediction on the block. The IBC prediction mode or the palette mode can be used for content image / video encoding such as games such as screen content coding (SCC). Although IBC basically performs prediction in the current block, its execution method is similar to inter prediction in that it derives a reference block in the current block. That is, IBC can use at least one of the inter prediction techniques described in the present disclosure.

[0083] The prediction signal generated by the inter-frame predictor 221 and / or the intra-frame predictor 222 can be used to generate a reconstruction signal or to generate a residual signal. The transformer 232 can generate a transform coefficient by applying a transform technique to the residual signal. For example, the transform technique may include at least one of a discrete cosine transform (DCT), a discrete sine transform (DST), a Karhunen-Loève transform (KLT), a graph-based transform (GBT), or a conditional nonlinear transform (CNT). Here, GBT means a transform obtained from a curve graph when the relationship information between pixels is represented by a curve graph. CNT refers to a transform obtained based on a prediction signal generated using all previously reconstructed pixels. In addition, the transform process can be applied to square pixel blocks of the same size, or can be applied to blocks of variable size rather than square blocks.

[0084] The quantizer 233 may quantize the transform coefficients and send them to the entropy encoder 240, and the entropy encoder 240 may encode the quantized signal (information about the quantized transform coefficients) and output the encoded signal in the bitstream. The information about the quantized transform coefficients may be referred to as residual information. The quantizer 233 may rearrange the quantized transform coefficients of the block type into a one-dimensional vector form based on the coefficient scanning order, and generate information about the quantized transform coefficients based on the quantized transform coefficients in the one-dimensional vector form. The entropy encoder 240 may perform various encoding methods such as, for example, exponential Golomb, context adaptive variable length coding (CAVLC), context adaptive binary arithmetic coding (CABAC), etc. The entropy encoder 240 may encode the information required for video / image reconstruction other than the quantized transform coefficients (e.g., the value of the syntax element, etc.) together or separately. The encoded information (e.g., the encoded video / image information) may be transmitted or stored on a unit basis of a network abstraction layer (NAL) in the form of a bitstream. The video / image information may also include information about various parameter sets such as an adaptation parameter set (APS), a picture parameter set (PPS), a sequence parameter set (SPS), a video parameter set (VPS), etc. In addition, the video / image information may also include conventional constraint information. In the present disclosure, information and / or syntax elements sent from an encoding device to / signaled to a decoding device may be included in the video / image information. The video / image information may be encoded by the above-mentioned encoding process and included in a bitstream. The bitstream may be transmitted over a network or stored in a digital storage medium. Here, the network may include a broadcast network, a communication network, and / or the like, and the digital storage medium may include various storage media such as USB, SD, CD, DVD, Blu-ray, HDD, SSD, etc. A transmitter (not shown) that transmits a signal output from the entropy encoder 240 or a memory (not shown) that stores it may be configured as an internal / external element of the encoding device 200, or the transmitter may be included in the entropy encoder 240.

[0085] The quantized transform coefficients output from the quantizer 233 can be used to generate a prediction signal. For example, by applying dequantization and inverse transformation to the quantized transform coefficients using the dequantizer 234 and the inverse transformer 235, the residual signal (residual block or residual sample) can be reconstructed. The adder 155 adds the reconstructed residual signal to the prediction signal output from the inter-frame predictor 221 or the intra-frame predictor 222, so that a reconstructed signal (reconstructed picture, reconstructed block, reconstructed sample array) can be generated. When there is no residual for the processing target block as in the case of applying the jump mode, the prediction block can be used as a reconstructed block. The adder 250 can be referred to as a reconstructor or a reconstructed block generator. The generated reconstructed signal can be used for intra-frame prediction of the next processing target block in the target picture, and as described later, can be used for inter-frame prediction of the next picture performed by filtering.

[0086] Furthermore, in the picture encoding and / or reconstruction process, luminance mapping with chroma scaling (LMCS) may be applied.

[0087] The filter 260 can improve the subjective / objective video quality by applying filtering to the reconstructed signal. For example, the filter 260 can generate a modified reconstructed picture by applying various filtering methods to the reconstructed picture, and the modified reconstructed picture can be stored in the memory 270, especially in the DPB of the memory 270. Various filtering methods may include, for example, deblocking filtering, sample adaptive offset, adaptive ring filter, bilateral filter, etc. As discussed later in the description of each filtering method, the filter 260 can generate various information related to filtering and send the generated information to the entropy encoder 240. Information about filtering can be encoded in the entropy encoder 240 and output in the form of a bitstream.

[0088] The modified reconstructed picture transmitted to the memory 270 may be used as a reference picture in the inter-frame predictor 221. Accordingly, the encoding device can avoid prediction mismatch in the encoding device 100 and the decoding device when applying inter-frame prediction, and can also improve encoding efficiency.

[0089] The memory 270DPB may store the modified reconstructed picture so as to use it as a reference picture in the inter-frame predictor 221. The memory 270 may store the motion information of the block in the current picture from which the motion information has been derived (or encoded) and / or the motion information of the block in the reconstructed picture. The stored motion information may be sent to the inter-frame predictor 221 to be used as the motion information of the neighboring block or the motion information of the temporal neighboring block. The memory 270 may store the reconstructed samples of the reconstructed block in the current picture and send them to the intra-frame predictor 222.

[0090] Figure 3is a diagram schematically illustrating a configuration of a video / image decoding device to which the present disclosure can be applied.

[0091] Reference Figure 3 , the video decoding device 300 may include an entropy decoder 310, a residual processor 320, a predictor 330, an adder 340, a filter 350, and a memory 360. The predictor 330 may include an intra-frame predictor 331 and an inter-frame predictor 332. The residual processor 320 may include a dequantizer 321 and an inverse transformer 322. According to an embodiment, the entropy decoder 310, the residual processor 320, the predictor 330, the adder 340, and the filter 350 described above may be composed of one or more hardware components (e.g., a decoder chipset or a processor). In addition, the memory 360 may include a decoded picture buffer (DPB) and may be composed of a digital storage medium. The hardware component may also include a memory 360 as an internal / external component.

[0092] When a bit stream including video / image information is input, the decoding device 300 can Figure 2 The image is reconstructed correspondingly to the processing of the video / image information in the encoding device. For example, the decoding device 300 can derive the unit / block based on the information related to the block segmentation obtained from the bit stream. The decoding device 300 can perform decoding by using the processing unit applied in the encoding device. Therefore, the decoding processing unit can be, for example, a coding unit, which can be divided into a quadtree structure, a binary tree structure and / or a ternary tree structure using a coding tree unit or a maximum coding unit. One or more transformation units can be derived from the coding unit. And, the reconstructed image signal decoded and output by the decoding device 300 can be reproduced by a reproducer.

[0093] The decoding device 300 may receive the Figure 2The signal output by the encoding device of the encoding device, and the received signal can be decoded by the entropy decoder 310. For example, the entropy decoder 310 can parse the bitstream to derive the information (e.g., video / image information) required for image reconstruction (or picture reconstruction). The video / image information may also include information about various parameter sets such as adaptive parameter set (APS), picture parameter set (PPS), sequence parameter set (SPS), video parameter set (VPS), etc. In addition, the video / image information may also include conventional constraint information. The decoding device may further decode the picture based on the information about the parameter set and / or the conventional constraint information. In the present disclosure, the signaling / receiving information and / or syntax elements to be described later can be decoded and obtained from the bitstream through the decoding process. For example, the entropy decoder 310 can decode the information in the bitstream based on a coding method such as exponential Golomb coding, CAVLC, CABAC, etc., and can output the values ​​of the syntax elements required for image reconstruction and the quantized values ​​of the transform coefficients of the residual. More specifically, the CABAC entropy decoding method can receive the bin corresponding to each syntax element in the bitstream, use the decoding target syntax element information and the decoding information of the neighboring and decoding target blocks or the information of the symbol / bin decoded in the previous step to determine the context model, predict the bin generation probability according to the determined context model and perform arithmetic decoding on the bin to generate the symbol corresponding to each syntax element value. Here, the CABAC entropy decoding method can update the context model using the information of the symbol / bin decoded by the context model for the next symbol / bin after determining the context model. The information about prediction among the information decoded in the entropy decoder 310 can be provided to the predictor (inter-frame predictor 332 and intra-frame predictor 331), and the residual value (ie, quantized transform coefficient) and the associated parameter information that have been entropy decoded in the entropy decoder 310 can be input to the residual processor 320. The residual processor 320 can derive a residual signal (residual block, residual sample, residual sample array). In addition, the information about filtering among the information decoded in the entropy decoder 310 can be provided to the filter 350. In addition, a receiver (not shown) receiving a signal output from the encoding device may also constitute the decoding device 300 as an internal / external element, and the receiver may be a component of the entropy decoder 310. In addition, the decoding device according to the present disclosure may be referred to as a video / image / picture encoding device, and the decoding device may be divided into an information decoder (video / image / picture information decoder) and a sample decoder (video / image / picture sample decoder). The information decoder may include the entropy decoder 310, and the sample decoder may include at least one of the dequantizer 321, the inverse transformer 322, the adder 340, the filter 350, the memory 360, the inter-frame predictor 332, and the intra-frame predictor 331.

[0094] The dequantizer 321 may output the transform coefficient by dequantizing the quantized transform coefficient. The dequantizer 321 may rearrange the quantized transform coefficient into a two-dimensional block form. In this case, the rearrangement may be performed based on the order of coefficient scanning that has been performed in the encoding device. The dequantizer 321 may perform dequantization on the quantized transform coefficient using a quantization parameter (e.g., quantization step size information) and obtain the transform coefficient.

[0095] The inverse transformer 322 obtains a residual signal (residual block, residual sample array) by performing inverse transformation on the transformation coefficients.

[0096] The predictor may perform prediction on the current block and generate a prediction block including prediction samples for the current block. The predictor may determine whether to apply intra prediction or inter prediction to the current block based on information on prediction output from the entropy decoder 310, and may specifically determine an intra / inter prediction mode.

[0097] The predictor can generate a prediction signal based on various prediction methods. For example, the predictor can apply intra prediction or inter prediction to the prediction of a block, and can also apply intra prediction and inter prediction at the same time. This can be called combined inter and intra prediction (CIIP). In addition, the predictor can perform intra-block copying (IBC) for the prediction of the block. Intra-block copying can be used for content image / video coding such as games such as screen content coding (SCC). Although IBC basically performs prediction in the current block, its execution method is similar to inter prediction in that it derives a reference block in the current block. That is, IBC can use at least one of the inter prediction techniques described in the present disclosure.

[0098] The intra-frame predictor 331 can predict the current block by referring to samples in the current picture. Depending on the prediction mode, the reference sample can be located near the current block or separated from the current block. In intra-frame prediction, the prediction mode can include multiple non-directional modes and multiple directional modes. The intra-frame predictor 331 can determine the prediction mode applied to the current block by using the prediction mode applied to the neighboring block.

[0099] The inter-frame predictor 332 may derive a prediction block for the current block based on a reference block (reference sample array) specified by a motion vector on a reference picture. At this time, in order to reduce the amount of motion information sent in the inter-frame prediction mode, the motion information may be predicted based on a block, a sub-block, or a sample based on the correlation of the motion information between the neighboring block and the current block. The motion information may include a motion vector and a reference picture index. The motion information may also include inter-frame prediction direction (L0 prediction, L1 prediction, Bi prediction, etc.) information. In the case of inter-frame prediction, the neighboring blocks may include spatial neighboring blocks existing in the current picture and temporal neighboring blocks existing in the reference picture. For example, the inter-frame predictor 332 may configure a motion information candidate list based on neighboring blocks, and derive a motion vector and / or a reference picture index of the current block based on the received candidate selection information. Inter-frame prediction may be performed based on various prediction modes, and information about the prediction may include information indicating a mode of inter-frame prediction for the current block.

[0100] The adder 340 can generate a reconstructed signal (reconstructed picture, reconstructed block, reconstructed sample array) by adding the obtained residual signal to the prediction signal (prediction block, prediction sample array) output from the predictor 330. When there is no residual for the processing target block as in the case of applying the skip mode, the prediction block can be used as the reconstructed block.

[0101] The adder 340 may be referred to as a reconstructor or a reconstructed block generator. The generated reconstructed signal may be used for intra prediction of the next processing target block in the current block, and as described later, may be output through filtering or used for inter prediction of the next picture.

[0102] Furthermore, in the picture decoding process, luma mapping with chroma scaling (LMCS) may be applied.

[0103] The filter 350 can improve the subjective / objective video quality by applying filtering to the reconstructed signal. For example, the filter 350 can generate a modified reconstructed picture by applying various filtering methods to the reconstructed picture, and the modified reconstructed picture can be sent to the memory 360, especially to the DPB of the memory 360. The various filtering methods may include, for example, deblocking filtering, sample adaptive offset, adaptive ring filter, bilateral filter, etc.

[0104] The (modified) reconstructed picture stored in the DPB of the memory 360 may be used as a reference picture in the inter-frame predictor 332. The memory 360 may store the motion information of the block in the current picture from which the motion information has been derived (or decoded) and / or the motion information of the block in the reconstructed picture. The stored motion information may be sent to the inter-frame predictor 332 to be used as the motion information of the neighboring block or the motion information of the temporal neighboring block. The memory 360 may store the reconstructed samples of the reconstructed block in the current picture and send them to the intra-frame predictor 331.

[0105] In the present specification, the examples described in the predictor 330, the dequantizer 321, the inverse transformer 322, and the filter 350 of the decoding device 300 may be similarly or correspondingly applied to the predictor 220, the dequantizer 234, the inverse transformer 235, and the filter 260 of the encoding device 200, respectively.

[0106] As described above, prediction is performed in order to improve the compression efficiency when performing video encoding. Accordingly, a prediction block including prediction samples for a current block as a coding target block can be generated. Here, the prediction block includes prediction samples in a spatial domain (or a pixel domain). The prediction block can be derived identically in an encoding device and a decoding device, and the encoding device can improve the image coding efficiency by signaling to the decoding device information (residual information) about the residual between the original block and the prediction block rather than the original sample value of the original block itself. The decoding device can derive a residual block including residual samples based on the residual information, generate a reconstructed block including reconstructed samples by adding the residual block to the prediction block, and generate a reconstructed picture including the reconstructed block.

[0107] The residual information may be generated by a transform process and a quantization process. For example, the encoding device may derive a residual block between the original block and the prediction block, derive a transform coefficient by performing a transform process on the residual samples (residual sample array) included in the residual block, and derive a quantized transform coefficient by performing a quantization process on the transform coefficient, so that it can signal the associated residual information to the decoding device (through a bitstream). Here, the residual information may include value information, position information, transform technology, transform kernel, quantization parameter, etc. of the quantized transform coefficient. The decoding device may perform a quantization / dequantization process based on the residual information and derive a residual sample (or residual sample block). The decoding device may generate a reconstructed block based on the prediction block and the residual block. The encoding device may derive a residual block by dequantizing / inverse transforming the quantized transform coefficient so as to serve as a reference for inter-frame prediction of the next picture, and may generate a reconstructed picture based on this.

[0108] Figure 4 The structure of the content streaming system to which the present disclosure is applied is illustrated.

[0109] Furthermore, a content streaming system to which the present disclosure is applied may generally include an encoding server, a streaming server, a web server, a media storage device, a user device, and a multimedia input device.

[0110] The encoding server is used to compress the content input from a multimedia input device such as a smart phone, a camera, a camcorder, etc. into digital data to generate a bitstream, and send it to a streaming server. As another example, in the case where a multimedia input device such as a smart phone, a camera, a camcorder, etc. directly generates a bitstream, the encoding server can be omitted. The bitstream can be generated by applying the encoding method or the bitstream generation method disclosed in the present invention. And the streaming server can temporarily store the bitstream during the process of sending or receiving the bitstream.

[0111] The streaming server sends multimedia data to the user device through the web server based on the user's request, and the web server serves as a tool to inform the user of what services exist. When the user requests the service the user wants, the web server transmits the request to the streaming server, and the streaming server sends the multimedia data to the user. In this regard, the content streaming system may include a separate control server, and in this case, the control server is used to control the command / response between the corresponding devices in the content streaming system.

[0112] The streaming server can receive content from a media storage device and / or an encoding server. For example, in the case of receiving content from an encoding server, the content can be received in real time. In this case, in order to smoothly provide a streaming service, the streaming server can store the bit stream for a predetermined time.

[0113] For example, user devices may include mobile phones, smart phones, laptop computers, digital broadcast terminals, personal digital assistants (PDAs), portable multimedia players (PMPs), navigators, tablet PCs, tablet PCs, ultrabooks, wearable devices (e.g., watch-type terminals (smart watches), glasses-type terminals (smart glasses), head-mounted displays (HMDs)), digital TVs, desktop computers, digital signage, etc. Each server in the content streaming system may operate as a distributed server, and in this case, data received by each server may be processed in a distributed manner.

[0114] Figure 5 The multi-conversion technology according to the embodiment of the present disclosure is schematically illustrated.

[0115] Reference Figure 5 , the converter can correspond to the aforementioned Figure 2 The converter in the encoding device, and the inverse converter may correspond to the aforementioned Figure 2 an inverse transformer in an encoding device, or Figure 3 An inverse transformer in a decoding device.

[0116] The transformer may derive (primary) transform coefficients by performing a primary transform based on the residual samples (residual sample array) in the residual block (S510). The primary transform may be referred to as a core transform. In this document, a primary transform may be based on a multi-transform selection (MTS), and when a multi-transform is used as a primary transform, it may be referred to as a multi-core transform.

[0117] The multi-core transform may refer to a method of performing transform using discrete cosine transform (DCT) type 2 and discrete sine transform (DST) type 7, DCT type 8, and / or DST type 1 in addition. That is, the multi-core transform may refer to a transform method of transforming a residual signal (or residual block) in the spatial domain into a transform coefficient (or primary transform coefficient) in the frequency domain based on a plurality of transform kernels selected from DCT type 2, DST type 7, DCT type 8, and DST type 1. In this article, the primary transform coefficient may be referred to as a temporary transform coefficient from the perspective of the transformer.

[0118] In other words, when a conventional transform method is applied, a transform coefficient can be generated by applying a transform from the spatial domain to the frequency domain to a residual signal (or residual block) based on DCT type 2. In contrast, when a multi-core transform is applied, a transform coefficient (or a primary transform coefficient) can be generated by applying a transform from the spatial domain to the frequency domain to a residual signal (or residual block) based on DCT type 2, DST type 7, DCT type 8, and / or DST type 1. In this document, DCT type 2, DST type 7, DCT type 8, and DST type 1 may be referred to as a transform type, a transform kernel, or a transform core. These DCT / DST transform types may be defined based on basis functions.

[0119] When performing a multi-core transform, a vertical transform core and a horizontal transform core for a target block may be selected from the transform cores, a vertical transform may be performed on the target block based on the vertical transform core, and a horizontal transform may be performed on the target block based on the horizontal transform core. Here, the horizontal transform may indicate a transform of a horizontal component of the target block, and the vertical transform may indicate a transform of a vertical component of the target block. The vertical transform core / horizontal transform core may be adaptively determined based on a prediction mode and / or a transform index of a target (CU or subblock) including a residual block.

[0120] In addition, according to the example, if a transformation is performed once by applying MTS, the mapping relationship of the transformation kernel can be set by setting a specific basis function to a predetermined value and combining the basis functions to be applied in the vertical transformation or the horizontal transformation. For example, when the horizontal transformation kernel is represented as trTypeHor and the vertical direction transformation kernel is represented as trTypeVer, trTypeHor or trTypeVer with a value of 0 can be set to DCT2, trTypeHor or trTypeVer with a value of 1 can be set to DST7, and trTypeHor or trTypeVer with a value of 2 can be set to DCT8.

[0121] In this case, the MTS index information may be encoded and signaled to the decoding device to indicate any one of the plurality of transform kernel sets. For example, an MTS index of 0 may indicate that both trTypeHor and trTypeVer values ​​are 0, an MTS index of 1 may indicate that both trTypeHor and trTypeVer values ​​are 1, an MTS index of 2 may indicate that both trTypeHor and trTypeVer values ​​are 2, an MTS index of 3 may indicate that both trTypeHor and trTypeVer values ​​are 1, and an MTS index of 4 may indicate that both trTypeHor and trTypeVer values ​​are 2.

[0122] In one example, the transformation core set according to the MTS index information is shown in the following table.

[0123] [Table 1]

[0124] tu_mts_idx[x0][y0] 0 1 2 3 4 trTypeHor 0 1 2 1 2 trTypeVer 0 1 1 2 2

[0125] The transformer may perform a secondary transform based on the (primary) transform coefficients to derive modified (secondary) transform coefficients (S520). A primary transform is a transform from the spatial domain to the frequency domain, while a secondary transform refers to a transform into a more compact representation using the correlation between the (primary) transform coefficients. The secondary transform may include an inseparable transform. In this case, the secondary transform may be referred to as a non-separable secondary transform (NSST) or a pattern-dependent non-separable secondary transform (MDNSST). NSST may represent a transform that performs a secondary transform on the (primary) transform coefficients derived by the primary transform based on an inseparable transform matrix to generate a modified transform coefficient (or secondary transform coefficient) for the residual signal. Here, based on the inseparable transform matrix, the transform may be applied once to the (primary) transform coefficients without separating the vertical transform and the horizontal transform (or applying the horizontal / vertical transforms independently). In other words, NSST is not applied separately to (primary) transform coefficients in the vertical and horizontal directions, and can represent, for example, a transform method in which a two-dimensional signal (transform coefficient) is rearranged into a one-dimensional signal through a specific predetermined direction (e.g., a row-first direction or a column-first direction) and then a modified transform coefficient (or secondary transform coefficient) is generated based on an inseparable transform matrix. For example, the row-first order is to arrange the M×N blocks in the order of the first row, the second row, ..., and the Nth row, and the column-first order is to arrange the M×N blocks in the order of the first column, the second column, ..., and the Mth column. NSST can be applied to the upper left area of ​​a block (hereinafter, referred to as a transform coefficient block) configured with (primary) transform coefficients. For example, when the width W and the height H of the transform coefficient block are both 8 or more, the 8×8 NSST can be applied to the upper left 8×8 area of ​​the transform coefficient block. In addition, when both the width (W) and the height (H) of the transform coefficient block are 4 or more, when the width (W) or the height (H) of the transform coefficient block is less than 8, the 4×4 NSST may be applied to the upper left min(8,W)×min(8,H) region of the transform coefficient block. However, the embodiment is not limited thereto, and for example, even if only the condition that the width W or the height H of the transform coefficient block is 4 or more is satisfied, the 4×4 NSST may be applied to the upper left min(8,W)×min(8,H) region of the transform coefficient block.

[0126] Specifically, for example, if a 4×4 input block is used, the non-separable secondary transform may be performed as follows.

[0127] A 4×4 input block X can be represented as follows.

[0128] [Formula 1]

[0129]

[0130] If X is represented as a vector, then the vector It can be expressed as follows.

[0131] [Formula 2]

[0132]

[0133] In formula 2, the vector is a one-dimensional vector obtained by rearranging the two-dimensional block X of Equation 1 according to the row-major order.

[0134] In this case, the non-separable quadratic transform can be calculated as follows.

[0135] [Formula 3]

[0136]

[0137] In this formula, denotes a transform coefficient vector and T denotes a 16×16 (non-separable) transform matrix.

[0138] By using the above formula 3, the 16×1 transform coefficient vector can be derived And the vector can be divided into two parts by scanning order (horizontally, vertically, diagonally, etc.) Reorganized into 4×4 blocks. However, the above calculation is an example, and Hypercube-Givens Transform (HyGT) or the like may also be used for the calculation of the inseparable secondary transform in order to reduce the calculation complexity of the inseparable secondary transform.

[0139] Furthermore, in the inseparable secondary transform, the transform kernel (or transform core, transform type) may be selected to be mode-dependent. In this case, the mode may include an intra prediction mode and / or an inter prediction mode.

[0140] As described above, an inseparable secondary transform may be performed based on an 8×8 transform or a 4×4 transform determined based on the width (W) and height (H) of the transform coefficient block. An 8×8 transform refers to a transform that can be applied to an 8×8 region included in a transform coefficient block when both W and H are equal to or greater than 8, and the 8×8 region may be the upper left 8×8 region in the transform coefficient block. Similarly, a 4×4 transform refers to a transform that can be applied to a 4×4 region included in a transform coefficient block when both W and H are equal to or greater than 4, and the 4×4 region may be the upper left 4×4 region in the transform coefficient block. For example, an 8×8 transform kernel matrix may be a 64×64 / 16×64 matrix, and a 4×4 transform kernel matrix may be a 16×16 / 8×16 matrix.

[0141] Here, in order to select the mode-dependent transform cores, two inseparable secondary transform cores may be configured for each transform set of inseparable secondary transforms for both the 8×8 transform and the 4×4 transform, and there may be four transform sets. That is, four transform sets may be configured for the 8×8 transform, and four transform sets may be configured for the 4×4 transform. In this case, each of the four transform sets for the 8×8 transform may include two 8×8 transform cores, and each of the four transform sets for the 4×4 transform may include two 4×4 transform cores.

[0142] However, as the size of the transform (ie, the size of the region to which the transform is applied) may be, for example, a size other than 8×8 or 4×4, the number of sets may be n, and the number of transform kernels in each set may be k.

[0143] The transform set may be referred to as an NSST set or a LFNST set. A specific set among the transform sets may be selected, for example, based on the intra prediction mode of the current block (CU or subblock). A low frequency non-separable transform (LFNST) may be an example of a reduced non-separable transform, which will be described later and represents a non-separable transform for a low frequency component.

[0144] For reference, for example, the intra prediction mode may include two non-directional (or non-angle) intra prediction modes and 65 directional (or angle) intra prediction modes. The non-directional intra prediction mode may include a plane intra prediction mode No. 0 and a DC intra prediction mode No. 1, and the directional intra prediction mode may include 65 intra prediction modes No. 2 to No. 66. However, this is an example, and this document may be applied even if the number of intra prediction modes is different. In addition, in some cases, intra prediction mode No. 67 may also be used, and intra prediction mode No. 67 may represent a linear model (LM) mode.

[0145] Figure 6 The intra directional mode for 65 prediction directions is schematically shown.

[0146] Reference Figure 6 , based on the intra prediction mode 34 having the upper left diagonal prediction direction, the intra prediction mode can be divided into an intra prediction mode having a horizontal directionality and an intra prediction mode having a vertical directionality. Figure 6In , H and V indicate horizontal directivity and vertical directivity, respectively, and numbers -32 to 32 indicate displacements of 1 / 32 units on sample grid positions. These numbers may represent offsets for mode index values. Intra-prediction modes 2 to 33 have horizontal directivity, and intra-prediction modes 34 to 66 have vertical directivity. Strictly speaking, intra-prediction mode 34 may be considered to be neither horizontal nor vertical, but may be classified as belonging to horizontal directivity when determining a transform set for a secondary transform. This is because the input data is transposed for a vertical orientation mode symmetrical based on intra-prediction mode 34, and an input data alignment method for a horizontal mode is used for intra-prediction mode 34. Transposing the input data means switching rows and columns of two-dimensional M×N block data to N×M data. Intra-prediction mode 18 and intra-prediction mode 50 may represent a horizontal intra-prediction mode and a vertical intra-prediction mode, respectively, and intra-prediction mode 2 may be referred to as an upper right diagonal intra-prediction mode because intra-prediction mode 2 has a left reference pixel and performs prediction in the upper right direction. Similarly, intra-prediction mode 34 may be referred to as a bottom-right diagonal intra-prediction mode, and intra-prediction mode 66 may be referred to as a bottom-left diagonal intra-prediction mode.

[0147] According to an example, four transform sets according to intra prediction modes may be mapped, for example, as shown in the following table.

[0148] [Table 2]

[0149] lfnstPredModeIntra lfnstTrSetIdx lfnstPredModeIntra<0 1 0<=lfinstPredModeIntra<=1 0 2<=lfnstPredModeIntra<=12 1 13<=lfnstPredModeIntra<=23 2 24<=lfnstPredModeIntra<=44 3 45<=lfnstPredModeIntra<=55 2 56<=lfnstPredModeIntra<=80 1 81<=lfnstPredModeIntra<=83 0

[0150] As shown in Table 2, depending on the intra prediction mode, any one of the four transform sets, i.e., lfnstTrSetId x , can be mapped to any of four indices (i.e., 0 to 3).

[0151] When it is determined that a specific set is used for an inseparable transform, one of the k transform cores in the specific set can be selected by an inseparable secondary transform index. The encoding device can derive an inseparable secondary transform index indicating a specific transform core based on a rate-distortion (RD) check, and can signal the inseparable secondary transform index to the decoding device. The decoding device can select one of the k transform cores in a specific set based on the inseparable secondary transform index. For example, lfnst index value 0 can refer to a first inseparable secondary transform core, lfnst index value 1 can refer to a second inseparable secondary transform core, and lfnst index value 2 can refer to a third inseparable secondary transform core. Alternatively, lfnst index value 0 can indicate that the first inseparable secondary transform is not applied to the target block, and lfnst index values ​​1 to 3 can indicate three transform cores.

[0152] The transformer can perform an inseparable secondary transform based on the selected transform kernel, and can obtain modified (secondary) transform coefficients. As described above, the modified transform coefficients can be derived as transform coefficients quantized by a quantizer, and can be encoded and signaled to a decoding device, and transmitted to a dequantizer / inverse transformer in an encoding device.

[0153] In addition, as described above, if the secondary transform is omitted, the (primary) transform coefficients that are the output of the primary (separable) transform can be derived as transform coefficients quantized by the quantizer as described above, and can be encoded and signaled to the decoding device and transmitted to the dequantizer / inverse transformer in the encoding device.

[0154] The inverse transformer may perform a series of processes in the reverse order of the order already performed in the above-mentioned transformer. The inverse transformer may receive the (dequantized) transform coefficients, and derive the (primary) transform coefficients by performing a secondary (inverse) transform (S550), and may obtain the residual block (residual sample) by performing a (inverse) transform on the (primary) transform coefficients (S560). In this regard, from the perspective of the inverse transformer, the primary transform coefficients may be referred to as modified transform coefficients. As described above, the encoding device and the decoding device may generate a reconstructed block based on the residual block and the prediction block, and may generate a reconstructed picture based on the reconstructed block.

[0155] The decoding device may further include a secondary inverse transform application determiner (or an element for determining whether to apply a secondary inverse transform) and a secondary inverse transform determiner (or an element for determining a secondary inverse transform). The secondary inverse transform application determiner may determine whether to apply a secondary inverse transform. For example, the secondary inverse transform may be NSST, RST, or LFNST, and the secondary inverse transform application determiner may determine whether to apply a secondary inverse transform based on a secondary transform flag obtained by parsing the bitstream. In another example, the secondary inverse transform application determiner may determine whether to apply a secondary inverse transform based on a transform coefficient of a residual block.

[0156] The secondary inverse transform determiner may determine the secondary inverse transform. In this case, the secondary inverse transform determiner may determine the secondary inverse transform applied to the current block based on the LFNST (NSST or RST) transform set specified according to the intra prediction mode. In an embodiment, the secondary transform determination method may be determined depending on the primary transform determination method. Various combinations of the primary transform and the secondary transform may be determined according to the intra prediction mode. In addition, in an example, the secondary inverse transform determiner may determine the area to which the secondary inverse transform is applied based on the size of the current block.

[0157] In addition, as described above, if the secondary (inverse) transform is omitted, the (dequantized) transform coefficients can be received, a (separable) inverse transform can be performed, and a residual block (residual sample) can be obtained. As described above, the encoding device and the decoding device can generate a reconstructed block based on the residual block and the prediction block, and can generate a reconstructed picture based on the reconstructed block.

[0158] Furthermore, in the present disclosure, a reduced secondary transform (RST) in which the size of a transformation matrix (kernel) is reduced may be applied in the concept of NSST in order to reduce the amount of calculation and storage required for an inseparable secondary transform.

[0159] In addition, the transformation kernel, transformation matrix, and coefficients constituting the transformation kernel matrix described in the present disclosure, that is, the kernel coefficients or matrix coefficients, can be represented by 8 bits. This can be a condition for implementation in a decoding device and an encoding device, and compared with the existing 9 bits or 10 bits, the amount of storage required to store the transformation kernel can be reduced, and performance degradation can be reasonably adapted. In addition, representing the kernel matrix with 8 bits can allow the use of small multipliers and can be more suitable for single instruction multiple data (SIMD) instructions for optimal software implementation.

[0160] In the present specification, the term "RST" may refer to a transformation performed on residual samples of a target block based on a transformation matrix whose size is reduced according to a reduction factor. In the case of performing a reduced transformation, the amount of calculation required for the transformation can be reduced due to the reduction in the size of the transformation matrix. That is, RST can be used to solve the computational complexity problem that occurs when transforming a block of large size or an inseparable transform.

[0161] RST may be referred to as various terms such as reduced transform, reduced secondary transform, reduced transform, simplified transform, and simple transform, and the names that RST may be referred to as are not limited to the listed examples. Alternatively, since RST is mainly performed in a low-frequency region including non-zero coefficients in a transform block, it may be referred to as a low-frequency non-separable transform (LFNST). A transform index may be referred to as a LFNST index.

[0162] In addition, when performing a secondary inverse transform based on RST, the inverse transformer 235 of the encoding device 200 and the inverse transformer 322 of the decoding device 300 may include: an inverse reduction secondary transformer that derives a modified transform coefficient based on an inverse RST of the transform coefficient; and an inverse primary transformer that derives a residual sample of the target block based on an inverse primary transform of the modified transform coefficient. The inverse primary transform refers to an inverse transform of a primary transform applied to the residual. In the present disclosure, deriving a transform coefficient based on a transform may refer to deriving a transform coefficient by applying a transform.

[0163] Figure 7is a diagram illustrating an RST according to an embodiment of the present disclosure.

[0164] In the present disclosure, a “target block” may refer to a current block to be encoded, a residual block, or a transform block.

[0165] In the RST according to the example, an N-dimensional vector can be mapped to an R-dimensional vector located in another space, so that a reduction transformation matrix can be determined, where R is less than N. N may refer to the square of the length of the side of the block to which the transformation is applied, or the total number of transformation coefficients corresponding to the block to which the transformation is applied, and the reduction factor may refer to an R / N value. The reduction factor may be referred to as a reduction factor, a reduction factor, a simplification factor, a simple factor, or other various terms. In addition, R may be referred to as a reduction coefficient, but depending on the situation, the reduction factor may refer to R. In addition, depending on the situation, the reduction factor may refer to an N / R value.

[0166] In the example, the reduction factor or reduction coefficient may be signaled through the bitstream, but the example is not limited thereto. For example, a predetermined value for the reduction factor or reduction coefficient may be stored in each of the encoding device 200 and the decoding device 300, and in this case, the reduction factor or reduction coefficient may not be signaled separately.

[0167] The size of the reduced transform matrix according to an example may be R×N which is smaller than N×N (the size of a conventional transform matrix), and may be defined as in Equation 4 below.

[0168] [Formula 4]

[0169]

[0170] Figure 7 The matrix T in the reduced transform block shown in (a) may refer to the matrix T of Formula 4 R×N .like Figure 7 As shown in (a), when the reduction transformation matrix T R×N When multiplied by the residual samples of the target block, the transform coefficients of the current block can be derived.

[0171] In the example, if the size of the block to which the transform is applied is 8×8 and R=16 (ie, R / N=μ16 / 64=1 / 4), then according to Figure 7 The RST of (a) can be expressed as a matrix operation shown in the following Equation 5. In this case, the storage and multiplication calculations can be reduced to about 1 / 4 by a reduction factor.

[0172] In the present disclosure, a matrix operation may be understood as an operation of obtaining a column vector by multiplying a column vector by a matrix provided on the left side of the column vector.

[0173] [Formula 5]

[0174]

[0175] In formula 6, r1 to r 64 y may represent the residual sample of the target block, and specifically may be a transform coefficient generated by applying a transform. As a result of the calculation of Equation 5, the transform coefficient c of the target block may be derived. i , and derive c i The process can be shown as formula 6.

[0176] [Formula 6]

[0177]

[0178] As a result of the calculation of Equation 6, the transform coefficients c1 to c R That is, when R=16, the transform coefficients c1 to c 16 . If a conventional transform is applied instead of RST, and a transform matrix of a size of 64×64 (N×N) is multiplied by a residual sample of a size of 64×1 (N×1), only 16 (R) transform coefficients are derived for the target block because RST is applied, although 64 (N) transform coefficients are derived for the target block. Since the total number of transform coefficients for the target block is reduced from N to R, the amount of data transmitted by the encoding device 200 to the decoding device 300 is reduced, and thus the transmission efficiency between the encoding device 200 and the decoding device 300 can be improved.

[0179] When considered from the perspective of the size of the transformation matrix, the size of the conventional transformation matrix is ​​64×64 (N×N), but the size of the reduced transformation matrix is ​​reduced to 16×64 (R×N), so the storage usage in the case of performing RST can be reduced by the R / N ratio compared to the case of performing the conventional transformation. In addition, when compared with the number of multiplication calculations N×N in the case of using the conventional transformation matrix, the use of the reduced transformation matrix can reduce the number of multiplication calculations (R×N) by the R / N ratio.

[0180] In an example, the transformer 232 of the encoding device 200 may derive the transform coefficient of the target block by performing a primary transform and a secondary transform based on RST on the residual samples of the target block. These transform coefficients may be transmitted to the inverse transformer of the decoding device 300, and the inverse transformer 322 of the decoding device 300 may derive the modified transform coefficient based on the inverse reduced secondary transform (RST) for the transform coefficient, and may derive the residual sample of the target block based on the inverse primary transform for the modified transform coefficient.

[0181] According to the example, the inverse RST matrix T N×RThe size of is N×R which is smaller than the size of the conventional inverse transform matrix N×N, and is similar to the reduced transform matrix T shown in Equation 4. R×N Has a transposition relationship.

[0182] Figure 7 The matrix T in the reduced inverse transform block shown in (b) t It can refer to the inverse RST matrix T N×R T (The superscript T refers to transposition). Figure 7 As shown in (b), when the inverse RST matrix T N×R T When multiplied by the transform coefficients of the target block, the modified transform coefficients of the target block or the residual samples of the target block can be derived. R×N T It can be expressed as (T R×N ) T N×R .

[0183] More specifically, when the inverse RST is used as the secondary inverse transform, when the inverse RST matrix T N×R T When multiplied by the transform coefficients of the target block, the modified transform coefficients of the target block can be derived. In addition, the inverse RST can be used as an inverse primary transform, and in this case, when the inverse RST matrix T N×R T When multiplied by the transform coefficients of the target block, the residual samples of the target block can be derived.

[0184] In the example, if the size of the block to which the inverse transform is applied is 8×8 and R=16 (ie, R / N=16 / 64=1 / 4), then according to Figure 7 The RST of (b) can be expressed as the matrix operation shown in the following Equation 7.

[0185] [Formula 7]

[0186]

[0187] In formula 7, c1 to c 16 As a result of the calculation of Equation 7, r representing the modified transform coefficient of the target block or the residual sample of the target block can be derived. j , and derive r j The process can be shown as formula 8.

[0188] [Formula 8]

[0189]

[0190] As a result of the calculation of Equation 8, r1 to r2 representing the modified transform coefficients of the target block or the residual samples of the target block can be derived. N . From the perspective of the size of the inverse transformation matrix, the size of the conventional inverse transformation matrix is ​​64×64 (N×N), but the size of the inverse reduction transformation matrix is ​​reduced to 64×16 (R×N), so the storage usage rate in the case of performing the inverse RST can be reduced by the R / N ratio compared to the case of performing the conventional inverse transformation. In addition, when compared with the number of multiplication calculations N×N in the case of using the conventional inverse transformation matrix, the use of the inverse reduction transformation matrix can reduce the number of multiplication calculations (N×R) by the R / N ratio.

[0191] The transform set configuration shown in Table 2 can also be applied to 8×8 RST. That is, 8×8 RST can be applied according to the transform set in Table 2. Since a transform set includes two or three transforms (kernels) according to the intra prediction mode, it can be configured to select one of up to four transforms including the case where the secondary transform is not applied. In the transform where the secondary transform is not applied, the application of the identity matrix can be considered. Assuming that indexes 0, 1, 2, and 3 are assigned to the four transforms respectively (for example, index 0 can be assigned to the case where the identity matrix is ​​applied, that is, the case where the secondary transform is not applied), the transform index or lfnst index as a syntax element can be signaled for each transform coefficient block, thereby specifying the transform to be applied. That is, for the upper left 8×8 block, the 8×8 NSST in the RST configuration can be specified by the transform index, or the 8×8 lfnst can be specified when LFNST is applied. 8×8lfnst and 8×8RST refer to the transform applicable to an 8×8 region included in a transform coefficient block when W and H of a target block to be transformed are both equal to or greater than 8, and the 8×8 region may be the upper left 8×8 region in the transform coefficient block. Similarly, 4×4lfnst and 4×4RST refer to the transform applicable to a 4×4 region included in a transform coefficient block when W and H of a target block are both equal to or greater than 4, and the 4×4 region may be the upper left 4×4 region in the transform coefficient block.

[0192] According to an embodiment of the present disclosure, for the transformation in the encoding process, only 48 pieces of data can be selected, and a maximum 16×48 transform kernel matrix can be applied thereto, instead of applying a 16×64 transform kernel matrix to 64 pieces of data forming an 8×8 region. Here, "maximum" means that m has a maximum value of 16 in the m×48 transform kernel matrix for generating m coefficients. That is, when RST is performed by applying an m×48 transform kernel matrix (m≤16) to an 8×8 region, 48 pieces of data are input, and m coefficients are generated. When m is 16, 48 pieces of data are input and 16 coefficients are generated. That is, assuming that 48 pieces of data form a 48×1 vector, the 16×48 matrix and the 48×1 vector are multiplied in sequence, thereby generating a 16×1 vector. Here, the 48 pieces of data forming an 8×8 region can be appropriately arranged, thereby forming a 48×1 vector. For example, a 48×1 vector can be constructed based on 48 pieces of data constituting an area other than the lower right 4×4 area among the 8×8 area. Here, when the matrix operation is performed by applying the maximum 16×48 transform kernel matrix, 16 modified transform coefficients are generated and the 16 modified transform coefficients can be arranged in the upper left 4×4 region according to the scanning order, and the upper right 4×4 region and the lower left 4×4 region can be filled with zeros.

[0193] For the inverse transform in the decoding process, the transposed matrix of the aforementioned transform kernel matrix can be used. That is, when the inverse RST or LFNST is performed in the inverse transform process performed by the decoding device, the input coefficient data to which the inverse RST is applied is configured in a one-dimensional vector according to a predetermined arrangement order, and the modified coefficient vector obtained by multiplying the one-dimensional vector with the corresponding inverse RST matrix on the left side of the one-dimensional vector can be arranged in a two-dimensional block according to a predetermined arrangement order.

[0194] In summary, in the transformation process, when RST or LFNST is applied to an 8×8 region, 48 transform coefficients in the upper left region, the upper right region, and the lower left region except the lower right region of the 8×8 region are matrix-operated with a 16×48 transform kernel matrix. For the matrix operation, the 48 transform coefficients are input in a one-dimensional array. When the matrix operation is performed, 16 modified transform coefficients are derived, and the modified transform coefficients can be arranged in the upper left region of the 8×8 region.

[0195] On the contrary, in the inverse transform process, when the inverse RST or LFNST is applied to the 8×8 area, 16 transform coefficients corresponding to the upper left area of ​​the 8×8 area among the transform coefficients in the 8×8 area can be input in a one-dimensional array according to the scanning order, and can undergo a matrix operation with a 48×16 transform kernel matrix. That is, the matrix operation can be expressed as (48×16 matrix)*(16×1 transform coefficient vector)=(48×1 modified transform coefficient vector). Here, the n×1 vector can be interpreted as having the same meaning as the n×1 matrix, and can therefore be expressed as an n×1 column vector. In addition, * represents matrix multiplication. When the matrix operation is performed, 48 modified transform coefficients can be derived, and the 48 modified transform coefficients can be arranged in the upper left area, the upper right area, and the lower left area of ​​the 8×8 area except the lower right area.

[0196] When the inverse secondary transform is based on RST, the inverse transformer 235 of the encoding device 200 and the inverse transformer 322 of the decoding device 300 may include an inverse reduction secondary transformer for deriving modified transform coefficients based on an inverse RST of the transform coefficients and an inverse primary transformer for deriving residual samples of the target block based on an inverse primary transform of the modified transform coefficients. The inverse primary transform refers to an inverse transform of a primary transform applied to the residual. In the present disclosure, deriving transform coefficients based on a transform may refer to deriving transform coefficients by applying a transform.

[0197] The non-separable transform (LFNST) described above will be described in detail as follows: LFNST may include a forward transform performed by an encoding device and an inverse transform performed by a decoding device.

[0198] The encoding device receives as input a result (or a portion of a result) derived after applying a primary (core) transform, and applies a forward secondary transform (secondary transform).

[0199] [Formula 9]

[0200] y=G T x

[0201] In Equation 9, x and y are the input and output of the secondary transform, respectively, G is a matrix representing the secondary transform, and the transform basis vectors consist of column vectors. In the case of inverse LFNST, when the dimension of the transform matrix G is expressed as [number of rows × number of columns], in the case of forward LFNST, the transpose of the matrix G becomes G T Dimension.

[0202] For inverse LFNST, the dimensions of the matrix G are [48×16], [48×8], [16×16], [16×8], and the [48×8] matrix and the [16×8] matrix are partial matrices of 8 transformation basis vectors sampled from the left side of the [48×16] matrix and the [16×16] matrix, respectively.

[0203] On the other hand, for forward LFNST, the matrix G T The dimensions of are [16×48], [8×48], [16×16], [8×16], and the [8×48] matrix and the [8×16] matrix are partial matrices obtained by sampling 8 transformation basis vectors from the upper parts of the [16×48] matrix and the [16×16] matrix, respectively.

[0204] Therefore, in the case of forward LFNST, a [48×1] vector or a [16×1] vector can be used as input x, and a [16×1] vector or a [8×1] vector can be used as output y. In video encoding and decoding, the output of the forward primary transform is two-dimensional (2D) data, so in order to construct a [48×1] vector or a [16×1] vector as input x, it is necessary to construct a one-dimensional vector by appropriately arranging the 2D data as the output of the forward transform.

[0205] Figure 8 is a diagram illustrating an order of arranging output data of a forward primary transform into a one-dimensional vector according to an example. Figure 8 The left figures of (a) and (b) show the order for constructing a [48×1] vector, and Figure 8 The right figures of (a) and (b) show the order for constructing a [16×1] vector. In the case of LFNST, the 2D data can be constructed by Figure 8 The same order as in (a) and (b) is sequentially arranged to obtain a one-dimensional vector x.

[0206] The arrangement direction of the output data of the forward primary transform may be determined according to the intra-frame prediction mode of the current block. For example, when the intra-frame prediction mode of the current block is in the horizontal direction relative to the diagonal direction, the arrangement direction of the output data of the forward primary transform may be determined according to the intra-frame prediction mode of the current block. Figure 8 The output data of the forward primary transform are arranged in the order of (a), and when the intra prediction mode of the current block is in a vertical direction relative to the diagonal direction, the output data of the forward primary transform can be arranged in the order of (a). Figure 8 The output data of the forward primary transform are arranged in the order of (b).

[0207] According to the example, different from Figure 8 The arrangement order of (a) and (b) is shown in Figure 1, and in order to derive and apply Figure 8If the arrangement order of (a) and (b) is the same result (y vector), the column vectors of the matrix G can be rearranged according to the arrangement order. In other words, the column vectors of G can be rearranged so that each element constituting the x vector is always multiplied by the same transformation basis vector.

[0208] Since the output y derived by Formula 9 is a one-dimensional vector, when two-dimensional data is required as input data in a process using the result of a forward quadratic transform as input (for example, in a process of performing quantization or residual encoding), the output y vector of Formula 9 needs to be properly arranged as 2D data again.

[0209] Fig. 9 is a diagram illustrating an order of arranging output data of a forward quadratic transform into two-dimensional blocks according to an example.

[0210] In the case of LFNST, the output values ​​may be arranged in 2D blocks according to a predetermined scanning order. Fig. 9 (a) shows that when the output y is a [16×1] vector, the output values ​​are arranged at 16 positions of the 2D block according to the diagonal scanning order. Fig. 9 (b) shows that when the output y is a [8×1] vector, the output values ​​are arranged at 8 positions of the 2D block according to the diagonal scanning order, and the remaining 8 positions are filled with zeros. Fig. 9 The X in (b) indicates that it is filled with zeros.

[0211] According to another example, since the order of processing the output vector y when performing quantization or residual encoding can be preset, the output vector y may not be arranged in a sequence such as Fig. 9 However, in the case of residual coding, data encoding may be performed in 2D block (eg, 4×4) units (eg, CG (coefficient group)), and in this case, according to Fig. 9 The data is arranged in a specific order in the diagonal scan order of .

[0212] In addition, the decoding device may configure a one-dimensional input vector y by arranging two-dimensional data output through a dequantization process according to a preset scanning order for inverse transformation. The input vector y may be output as an output vector x through the following equation.

[0213] [Formula 10]

[0214] x=Gy

[0215] In the case of inverse LFNST, the output vector x can be derived by multiplying the input vector y, which is a [16×1] vector or a [8×1] vector, by the G matrix. For inverse LFNST, the output vector x can be a [48×1] vector or a [16×1] vector.

[0216] The output vector x is based on Figure 8 The order shown in is arranged in a two-dimensional block and is arranged as two-dimensional data, and the two-dimensional data becomes input data (or a part of the input data) of an inverse primary transform.

[0217] Thus, the inverse quadratic transform is overall the reverse of the forward quadratic transform process, and in the case of an inverse transform, unlike in the forward direction, the inverse quadratic transform is applied first and then the inverse primary transform.

[0218] In the inverse LFNST, one of 8 [48×16] matrices and 8 [16×16] matrices can be selected as the transformation matrix G. Whether to apply the [48×16] matrix or the [16×16] matrix depends on the size and shape of the block.

[0219] In addition, 8 matrices can be derived from the four transform sets shown in Table 2 above, and each transform set can be composed of two matrices. Which transform set to use among the 4 transform sets is determined according to the intra prediction mode, and more specifically, the transform set is determined based on the value of the intra prediction mode extended by considering wide-angle intra prediction (WAIP). Which matrix is ​​selected from the two matrices constituting the selected transform set is derived by index signaling. More specifically, 0, 1, and 2 can be used as the index values ​​sent, 0 can indicate that LFNST is not applied, and 1 and 2 can indicate any one of the two transform matrices constituting the transform set selected based on the intra prediction mode value.

[0220] Fig.10 is a diagram illustrating a wide-angle intra prediction mode according to an embodiment of this document.

[0221] The general intra prediction mode value may have a value from 0 to 66 and from 81 to 83, and the intra prediction mode value extended due to WAIP may have a value from -14 to 83 as shown. The value from 81 to 83 indicates the CCLM (Cross Component Linear Model) mode, and the value from -14 to -1 and the value from 67 to 80 indicate the intra prediction mode extended due to WAIP application.

[0222] When the width of the current prediction block is greater than the height, the upper reference pixel is generally closer to the position inside the block to be predicted. Therefore, prediction in the lower left direction can be more accurate than in the upper right direction. Conversely, when the height of the block is greater than the width, the left reference pixel is generally closer to the position inside the block to be predicted. Therefore, prediction in the upper right direction can be more accurate than in the lower left direction. Therefore, it can be advantageous to apply remapping (i.e., mode index modification) to the index of the wide-angle intra prediction mode.

[0223] When wide-angle intra prediction is applied, information about the existing intra prediction may be signaled, and after the information is parsed, the information may be remapped to the index of the wide-angle intra prediction mode. Therefore, the total number of intra prediction modes for a specific block (e.g., a non-square block of a specific size) may not be changed, that is, the total number of intra prediction modes is 67, and the intra prediction mode encoding for the specific block may not be changed.

[0224] Table 3 below shows a process of deriving a modified intra mode by remapping the intra prediction mode to the wide-angle intra prediction mode.

[0225] [Table 3]

[0226]

[0227] In Table 3, the extended intra prediction mode value is finally stored in the predModeIntra variable, and ISP_NO_SPLIT indicates that the CU block is not divided into sub-partitions by the intra sub-partitioning (ISP) technology currently adopted in the VVC standard, and the cIdx variable values ​​of 0, 1, and 2 indicate the cases of the luminance component, the Cb component, and the Cr component, respectively. The log2 function shown in Table 3 returns a log value with a base of 2, and the Abs function returns an absolute value.

[0228] The variable predModeIntra indicating the intra prediction mode and the height and width of the transform block are used as input values ​​of the wide-angle intra prediction mode mapping process, and the output value is the modified intra prediction mode predModeIntra. The height and width of the transform block or the coding block can be the height and width of the current block for remapping of the intra prediction mode. At this time, the variable whRatio reflecting the ratio of width to width can be set to Abs(Log2(nW / nH)).

[0229] For non-square blocks, the intra prediction mode can be divided into two cases and modified.

[0230] First, if all of conditions (1) to (3) are satisfied, (1) the width of the current block is greater than the height, (2) the intra prediction mode before modification is equal to or greater than 2, and (3) the intra prediction mode is less than the value derived as (8+2*whRatio) when the variable whRatio is greater than 1 and is less than 8 when the variable whRatio is less than or equal to 1 (predModeIntra is less than (whRatio>1)? (8+2*whRatio): 8), then the intra prediction mode is set to a value 65 greater than predModeIntra [predModeIntra is set equal to (predModeIntra+65)].

[0231] If different from the above, that is, if conditions (1) to (3) are satisfied, (1) the height of the current block is greater than the width, (2) the intra-frame prediction mode before modification is less than or equal to 66, and (3) the intra-frame prediction mode is greater than the value derived as (60-2*whRatio) when whRatio is greater than 1 and is greater than 60 when whRatio is less than or equal to 1 (predModeIntra is greater than (whRatio>1)? (60-2*whRatio):60), then the intra-frame prediction mode is set to a value that is 67 less than predModeIntra [predModeIntra is set equal to (predModeIntra-67)].

[0232] Table 2 above shows how to select a transform set based on the intra prediction mode value extended by WAIP in LFNST. Fig.10 As shown, modes 14 to 33 and modes 35 to 80 are symmetric about the prediction direction around mode 34. For example, mode 14 and mode 54 are symmetric about the direction corresponding to mode 34. Therefore, the same set of transformations is applied to modes located in mutually symmetric directions, and this symmetry is also reflected in Table 2.

[0233] In addition, it is assumed that the forward LFNST input data of mode 54 is symmetric with the forward LFNST input data of mode 14. For example, for mode 14 and mode 54, according to Figure 8 (a) and Figure 8 The arrangement order shown in (b) rearranges the two-dimensional data into one-dimensional data. In addition, it can be seen that Figure 8 (a) and Figure 8 The pattern of the sequence shown in (b) is symmetrical about the direction indicated by the pattern 34 (diagonal direction).

[0234] Furthermore, as described above, which transform matrix of the [48×16] matrix and the [16×16] matrix is ​​applied to the LFNST is determined by the size and shape of the transform target block.

[0235] Fig.11 is a diagram illustrating a block shape to which LFNST is applied. Fig.11 (a) shows a 4×4 block, Fig.11 (b) shows a 4×8 block and an 8×4 block, Fig.11 (c) shows a 4×N block or an N×4 block, where N is 16 or greater, Fig.11 (d) shows an 8×8 block, Fig.11 (e) shows an M×N block, where M≥8, N≥8 and N>8 or M>8.

[0236] exist Fig.11In , the blocks with thick borders indicate the areas where LFNST is applied. Fig.11 For the blocks (a) and (b), LFNST is applied to the top left 4×4 region, and for Fig.11 In the block (c), LFNST is applied separately to the two upper left 4×4 regions that are arranged consecutively. Fig.11 In (a), (b), and (c), since LFNST is applied in units of 4×4 regions, this LFNST will be referred to as “4×4LFNST” hereinafter. Depending on the matrix dimension of G, a [16×16] or [16×8] matrix may be applied.

[0237] More specifically, the [16×8] matrix is ​​applied to Fig.11 (a) 4×4 block (4×4TU or 4×4CU), and the [16×16] matrix is ​​applied to Fig.11 This is to adjust the worst case computational complexity to 8 multiplications per sample.

[0238] about Fig.11 In (d) and (e), LFNST is applied to the upper left 8×8 region, and this LFNST is hereinafter referred to as "8×8LFNST". As the corresponding transformation matrix, a [48×16] matrix or a [48×8] matrix may be applied. In the case of the forward LFNST, since a [48×1] vector (the X vector in Equation 9) is input as input data, not all sample values ​​of the upper left 8×8 region are used as input values ​​of the forward LFNST. That is, as can be seen from Figure 8 The left order of (a) or Figure 8 As can be seen from the left side sequence of (b), a [48×1] vector can be constructed based on samples belonging to the remaining 3 4×4 blocks while leaving the lower right 4×4 block as is.

[0239] The [48×8] matrix can be applied to Fig.11 The 8×8 block (8×8TU or 8×8CU) in (d) and the [48×16] matrix can be applied to Fig.11 The 8×8 block in (e) is also to adjust the worst-case computational complexity to 8 multiplications per sample.

[0240] Depending on the block shape, when the corresponding forward LFNST (4×4 or 8×8 LFNST) is applied, 8 or 16 output data (Y vector in Eq. 9, [8×1] or [16×1] vector) are generated. In the forward LFNST, since the matrix G T The characteristic is that the number of output data is equal to or less than the number of input data.

[0241] Fig.12 is a diagram illustrating arrangement of output data of a forward LFNST according to an example, and shows blocks in which the output data of the forward LFNST is arranged according to a block shape.

[0242] exist Fig.12 The shaded area on the upper left of the block shown corresponds to the area where the output data of the forward LFNST is located, the positions marked with 0 indicate samples filled with the value 0, and the remaining area represents the area not changed by the forward LFNST. In the area not changed by the LFNST, the output data of the forward primary transform remains unchanged.

[0243] As described above, since the size of the applied transformation matrix varies according to the shape of the block, the amount of output data also varies. Fig.12 , the output data of the forward LFNST may not completely fill the upper left 4×4 block. Fig.12 In the cases of (a) and (d) of FIG. 1 , the [16×8] matrix and the [48×8] matrix are applied to the block or the partial area inside the block indicated by the bold line, respectively, and an [8×1] vector is generated as the output of the forward LFNST. That is, according to Fig. 9 The scanning order shown in (b) can only fill 8 output data, such as Fig.12 As shown in (a) and (d), the remaining 8 positions can be filled with 0. Fig.11 (d) The case of LFNST application blocks, such as Fig.12 As shown in (d), the two 4×4 blocks on the upper right and lower left adjacent to the upper left 4×4 block are also filled with values ​​of 0.

[0244] As described above, basically, by signaling the LFNST index, it is specified whether to apply LFNST and the transformation matrix to be applied. Fig.12 As shown, when LFNST is applied, since the number of output data of the forward LFNST may be equal to or less than the number of input data, an area filled with zero values ​​appears as follows.

[0245] 1) If Fig.12 As shown in (a), samples from the eighth position and subsequent positions in the scanning order in the upper left 4×4 block, that is, samples from the ninth to sixteenth positions.

[0246] 2) If Fig.12 As shown in (d) and (e), when the [48×16] matrix or the [48×8] matrix is ​​applied, two 4×4 blocks adjacent to the upper left 4×4 block or the second and third 4×4 blocks in the scanning order.

[0247] Therefore, if non-zero data exists by checking areas 1) and 2), it is determined that LFNST is not applied, so that signaling of the corresponding LFNST index can be omitted.

[0248] According to an example, for example, in the case of LFNST adopted in the VVC standard, since the signaling of the LFNST index is performed after the residual coding, the encoding device can know whether there is non-zero data (significant coefficient) at all positions within the TU or CU block through the residual coding. Therefore, the encoding device can determine whether to perform signaling about the LFNST index based on the presence of non-zero data, and the decoding device can determine whether to parse the LFNST index. When non-zero data does not exist in the area specified in 1) and 2) above, the signaling of the LFNST index is performed.

[0249] Since the truncated unary code is applied as the binarization method of the LFNST index, the LFNST index consists of up to two bins, and 0, 10, and 11 are assigned as binary codes for possible LFNST index values ​​0, 1, and 2, respectively. In the case of LFNST currently used for VVC, context-based CABAC coding is applied to the first bin (conventional coding), and bypass coding is applied to the second bin. The total number of contexts for the first bin is 2, and when (DCT-2, DCT-2) is applied as a transform pair for horizontal and vertical directions and the luminance component and the chrominance component are encoded in a dual tree type, one context is allocated and the other context is applied to the remaining cases. The encoding of the LFNST index is shown in the following table.

[0250] [Table 4]

[0251]

[0252] In addition, for the adopted LFNST, the following simplification method can be applied.

[0253] (i) According to an example, the number of output data of the forward LFNST may be limited to a maximum value of 16.

[0254] exist Fig.11 In the case of (c), 4×4 LFNST can be applied to two 4×4 regions adjacent to the upper left, respectively, and in this case, a maximum of 32 LFNST output data can be generated. When the number of output data of the forward LFNST is limited to a maximum of 16, in the case of a 4×N / N×4 (N≥16) block (TU or CU), 4×4 LFNST is applied only to one 4×4 region in the upper left, and LFNST can be applied only to Fig.11 By doing this, the implementation of image coding can be simplified.

[0255] Fig.13 It is shown that the number of output data of the forward LFNST according to the example is limited to a maximum value of 16. Fig.13 , when LFNST is applied to the upper leftmost 4×4 region in a 4×N or N×4 block (where N is 16 or greater), the output data of the forward LFNST becomes 16.

[0256] (ii) According to an example, clearing may be additionally applied to an area to which LFNST is not applied. In this document, clearing may mean filling all positions belonging to a specific area with a value of 0. That is, clearing may be applied to an area that is unchanged due to LFNST, and the result of the forward primary transformation is maintained. As described above, since LFNST is divided into 4×4 LFNST and 8×8 LFNST, clearing may be divided into two types ((ii)-(A) and (ii)-(B)) as follows.

[0257] (ii)-(A) When 4×4 LFNST is applied, an area to which 4×4 LFNST is not applied may be cleared. Fig.14 is a diagram illustrating clearing in a block to which 4×4 LFNST is applied according to an example.

[0258] like Fig.14 As shown, for the block to which 4×4 LFNST is applied, that is, for Fig.12 For all blocks in (a), (b), and (c), the entire area where LFNST is not applied can be filled with zeros.

[0259] on the other hand, Fig.14 (d) shows that when the maximum value of the number of output data of the forward LFNST is limited to 16 (as Fig.13 ), the remaining blocks to which 4×4 LFNST is not applied are cleared.

[0260] (ii)-(B) When 8×8 LFNST is applied, an area to which 8×8 LFNST is not applied may be cleared. Fig.15 is a diagram illustrating clearing in a block to which 8×8 LFNST is applied according to an example.

[0261] like Fig.15 As shown, for the block with 8×8 LFNST applied, that is, Fig.12 For all blocks in (d) and (e), the entire area where LFNST is not applied can be filled with zeros.

[0262] (iii) Due to the zeroing presented in (ii) above, the area filled with zeros may not be the same as when LFNST is applied. Fig.12In the case of LFNST, the wider region performs the zeroing proposed in (ii) to check whether there is non-zero data.

[0263] For example, when (ii)-(B) is applied, in checking Fig.12 After the zero-filled areas in (d) and (e) have non-zero data, additionally check Fig.15 Whether there is non-zero data in the area filled with 0, the signaling for the LFNST index can be performed only when there is no non-zero data.

[0264] Of course, even if the clearing proposed in (ii) is applied, the presence of non-zero data can be checked in the same way as the existing LFNST index signaling. Fig.12 After checking whether there is non-zero data in the block filled with zeros in , LFNST index signaling can be applied. In this case, the encoding device only performs zero clearing and the decoding device does not assume zero clearing, that is, it only checks whether non-zero data exists only in Fig.12 In the regions explicitly marked as 0 in , LFNST index parsing can be performed.

[0265] Alternatively, according to another example, the following may be performed: Fig.16 Cleared as shown. Fig.16 is a diagram illustrating clearing in a block to which 8×8 LFNST is applied according to another example.

[0266] like Fig.14 and Fig.15 As shown, zeroing can be applied to all regions except the region where LFNST is applied, or zeroing can be applied only to local regions, as shown in Fig.16 Clearing to zero applies only to Fig.16 For areas outside the upper left 8×8 area, clearing may not be applied to the lower right 4×4 block within the upper left ×8 area.

[0267] Various embodiments of the combination of the simplified methods ((i), (ii)-(A), (ii)-(B), (iii)) applying LFNST can be derived. Of course, the combination of the simplified methods described above is not limited to the following embodiments, and any combination can be applied to LFNST.

[0268] Implementation

[0269] -Limit the number of output data of the forward LFNST to a maximum of 16 → (i)

[0270] -When 4×4 LFNST is applied, all regions where 4×4 LFNST is not applied are cleared → (II)-(A)

[0271] - When 8×8 LFNST is applied, all regions where 8×8 LFNST is not applied are cleared → (II)-(B)

[0272] - After checking whether non-zero data also exists in the existing areas filled with zero values ​​and in the areas filled with zeros due to the additional clearing ((ii)-(A), (ii)-(B)), the LFNST index is signaled only if non-zero data does not exist → (iii).

[0273] In the case of an embodiment, when LFNST is applied, the area where non-zero output data can exist is limited to the interior of the upper left 4×4 area. Fig.14 (a) and Fig.15 In the case of (a), the eighth position in the scan order is the last position where non-zero data can exist. Fig.14 (b) and (c) and Fig.15 In the case of (b), the sixteenth position in the scanning order (ie, the position of the lower right edge of the upper left 4×4 block) is the last position in which data other than 0 may exist.

[0274] Therefore, after applying LFNST, after checking whether non-zero data exists at a position that is not allowed by the residual encoding process (at a position beyond the last position), it can be determined whether to signal the LFNST index.

[0275] In the case of the zeroing method proposed in (ii), the amount of calculation required to perform the entire transformation process can be reduced due to the amount of data ultimately generated when both the primary transformation and LFNST are applied. That is, when LFNST is applied, since zeroing is applied to the forward primary transformation output data exists in the area where LFNST is not applied, it is not necessary to generate data for the area that becomes zeroed during the execution of the forward primary transformation. Therefore, the amount of calculation required to generate the corresponding data can be reduced. The additional effects of the zeroing method proposed in (ii) are summarized as follows.

[0276] First, as mentioned above, the amount of computation required to perform the entire transformation process is reduced.

[0277] In particular, when (ii)-(B) is applied, the amount of calculation in the worst case is reduced, so that the transformation process can be made lighter. In other words, in general, a large amount of calculation is required to perform a large-size primary transformation. By applying (ii)-(B), the amount of data derived as a result of performing forward LFNST can be reduced to 16 or less. In addition, as the size of the entire block (TU or CU) increases, the effect of reducing the amount of transformation operations further increases.

[0278] Second, the amount of computation required for the entire transformation process can be reduced, thereby reducing the power consumption required to perform the transformation.

[0279] Third, the delay involved in the transformation process is reduced.

[0280] Secondary transforms such as LFNST add computational effort to the existing primary transform, thus increasing the overall latency involved in performing the transform. In particular, in the case of intra prediction, since reconstructed data of neighboring blocks is used in the prediction process, during encoding, the increase in latency due to the secondary transform results in an increase in latency until reconstruction. This can result in an increase in the overall latency of intra prediction encoding.

[0281] However, if the clearing proposed in (ii) is applied, the delay time for performing one transform can be greatly reduced when LFNST is applied, maintaining or reducing the delay time of the entire transform, so that the encoding device can be implemented more simply.

[0282] In conventional intra prediction, the current block to be encoded is regarded as one coding unit, and encoding is performed without segmentation. However, intra sub-partitioning (ISP) encoding means performing intra prediction encoding by dividing the current block to be encoded in the horizontal direction or the vertical direction. In this case, a reconstructed block can be generated by performing encoding / decoding in units of divided blocks, and the reconstructed block can be used as a reference block for the next divided block. According to an embodiment, in ISP encoding, one coding block may be divided into two or four sub-blocks and encoded, and in ISP, in one sub-block, intra prediction is performed with reference to the reconstructed pixel values ​​of the sub-block located on the adjacent left side or the adjacent upper side. Hereinafter, "encoding" may be used as a concept including both encoding performed by an encoding device and decoding performed by a decoding device.

[0283] Table 5 represents the number of sub-blocks divided according to the block size when ISP is applied, and the sub-partitions divided according to ISP may be referred to as transform blocks (TUs).

[0284] [Table 5]

[0285] Block size (CU) Number of divisions 4×4 Unavailable 4×8、8×4 2 All other cases 4

[0286] ISP is to divide the block predicted as a luma frame into two or four sub-partitions in the vertical or horizontal direction according to the size of the block. For example, the minimum block size to which ISP can be applied is 4×8 or 8×4. When the block size is larger than 4×8 or 8×4, the block is divided into 4 sub-partitions.

[0287] Fig.17 and Fig.18 An example of sub-blocks into which a coding block is divided is illustrated, and more specifically, Fig.17An example of division in which a coding block (width (W)×height (H)) is a 4×8 block or an 8×4 block is illustrated, and Fig.18 An example of division in a case where a coding block is not a 4×8 block, an 8×4 block, or a 4×4 block is illustrated.

[0288] When ISP is applied, subblocks are encoded sequentially from left to right or from top to bottom according to the partition type (e.g., horizontally or vertically), and after performing reconstruction processing via inverse transform and intra prediction for one subblock, encoding of the next subblock can be performed. For the leftmost or topmost subblock, the reconstructed pixels of the already encoded coding block are referenced, as in the conventional intra prediction method. In addition, when each side of the subsequent internal subblock is not adjacent to the previous subblock, in order to derive the reference pixels adjacent to the corresponding side, the reconstructed pixels of the already encoded adjacent coding blocks are referenced, as in the conventional intra prediction method.

[0289] In ISP coding mode, all sub-blocks can be encoded with the same intra prediction mode, and a flag indicating whether ISP coding is used and a flag indicating in which direction to split (horizontally or vertically) can be signaled. Fig.17 and Fig.18 As shown, the number of sub-blocks can be adjusted to 2 or 4 according to the shape of the block. When the size (width × height) of a sub-block is less than 16, it can be restricted so that division into corresponding sub-blocks is not allowed or ISP encoding itself is not applied.

[0290] In case of the ISP prediction mode, one coding unit is divided into two or four partition blocks (ie, subblocks) and predicted, and the same intra prediction mode is applied to the divided two or four partition blocks.

[0291] As described above, in the division direction, both the horizontal direction (when an M×N coding unit having a horizontal length and a vertical length of M and N, respectively, is divided in the horizontal direction, if the M×N coding unit is divided into two, the M×N coding unit is divided into M×(N / 2) blocks, and if the M×N coding unit is divided into four blocks, the M×N coding unit is divided into M×(N / 4) blocks) and the vertical direction (when the M×N coding unit is divided in the vertical direction, if the M×N coding unit is divided into two, the M×N coding unit is divided into (M / 2)×N blocks, and if the M×N coding unit is divided into four, the M×N coding unit is divided into (M / 4)×N blocks) are possible. When the M×N coding unit is divided in the horizontal direction, the partition blocks are encoded in a top-to-bottom order, and when the M×N coding unit is divided in the vertical direction, the partition blocks are encoded in a left-to-right order. In the case of horizontal (vertical) partitioning, the currently encoded partition block can be predicted by referring to the reconstructed pixel values ​​of the upper (left) partition block.

[0292] A transformation may be applied to the residual signal generated in units of partition blocks by the ISP prediction method. A multi-transform selection (MTS) technique based on a DST-7 / DCT-8 combination and the existing DCT-2 may be applied to a forward-based primary transform (core transform), and a forward low-frequency non-separable transform (LFNST) may be applied to the transform coefficients generated from the primary transform to generate the final modified transform coefficients.

[0293] That is, LFNST can be applied to the partition blocks divided by applying the ISP prediction mode, and the same intra prediction mode is applied to the divided partition blocks, as described above. Therefore, when the LFNST set derived based on the intra prediction mode is selected, the derived LFNST set can be applied to all partition blocks. That is, because the same intra prediction mode is applied to all partition blocks, the same LFNST set can be applied to all partition blocks.

[0294] According to an embodiment, LFNST may be applied only to a transform block having both a horizontal length and a vertical length of 4 or more. Therefore, when the horizontal length or the vertical length of a partition block divided according to the ISP prediction method is less than 4, LFNST is not applied and the LFNST index is not signaled. In addition, when LFNST is applied to each partition block, the corresponding partition block may be regarded as one transform block. When the ISP prediction method is not applied, LFNST may be applied to the coding block.

[0295] A method of applying LFNST to each partition block will be described in detail.

[0296] According to an embodiment, after applying forward LFNST to each partition block, only a maximum of 16 (8 or 16) coefficients are left in the upper left 4×4 area in the transform coefficient scanning order, and then zeroing can be applied, where the remaining positions and areas are all filled with 0.

[0297] Alternatively, according to an embodiment, when the length of one side of the partition block is 4, LFNST is applied only to the upper left 4×4 region, and when the lengths of all sides of the partition block (i.e., width and height) are 8 or greater, LFNST can be applied to the remaining 48 coefficients within the upper left 8×8 region except for the lower right 4×4 region.

[0298] Alternatively, according to an embodiment, in order to adjust the worst-case computational complexity to 8 multiplications per sample, when each partition block is 4×4 or 8×8, only 8 transform coefficients may be output after applying the forward LFNST. That is, when the partition block is 4×4, an 8×16 matrix may be applied as the transform matrix, and when the partition block is 8×8, an 8×48 matrix may be applied as the transform matrix.

[0299] In the current VVC standard, LFNST index signaling is performed in units of coding units. Therefore, in ISP prediction mode and when LFNST is applied to all partition blocks, the same LFNST index value can be applied to the corresponding partition blocks. That is, when the LFNST index value is sent once at the coding unit level, the corresponding LFNST index can be applied to all partition blocks in the coding unit. As described above, the LFNST index value can have values ​​of 0, 1, and 2, where 0 represents the case where LFNST is not applied, and 1 and 2 represent two transform matrices present in one LFNST set when LFNST is applied.

[0300] As described above, the LFNST set is determined by the intra prediction mode, and in the case of the ISP prediction mode, since all partition blocks in a coding unit are predicted in the same intra prediction mode, the partition blocks can refer to the same LFNST set.

[0301] As another example, LFNST index signaling is still performed in units of coding units, but in the case of the ISP prediction mode, it is not determined whether LFNST is uniformly applied to all partition blocks, and for each partition block, whether to apply the LFNST index value signaled at the coding unit level and whether to apply LFNST can be determined by a separate condition. Here, a separate condition can be signaled in the form of a flag for each partition block through the bitstream, and when the flag value is 1, the LFNST index value signaled at the coding unit level is applied, and when the flag value is 0, LFNST may not be applied.

[0302] In a coding unit to which the ISP mode is applied, an example of applying LFNST when the length of one side of a partition block is less than 4 is described as follows.

[0303] First, when the size of the partition block is N×2 (2×N), LFNST can be applied to the upper left M×2 (2×M) region (where M≤N). For example, when M=8, the upper left region becomes 8×2 (2×8), so the region where there are 16 residual signals can be the input of the forward LFNST, and a forward transformation matrix of R×16 (R≤16) can be applied.

[0304] Here, the forward LFNST matrix can be a separate additional matrix in addition to the matrix included in the current VVC standard. In addition, for the worst case complexity control, an 8×16 matrix in which only the upper 8 row vectors of the 16×16 matrix are sampled can be used for the transformation. The complexity control method will be described in detail later.

[0305] Secondly, when the size of the partition block is N×1 (1×N), LFNST can be applied to the upper left M×1 (1×M) region (where M≤N). For example, when M=16, the upper left region becomes 16×1 (1×16), so the region where there are 16 residual signals can be the input of the forward LFNST, and a forward transformation matrix of R×16 (R≤16) can be applied.

[0306] Here, the corresponding forward LFNST matrix can be a separate additional matrix in addition to the matrix included in the current VVC standard. In addition, in order to control the complexity of the worst case, an 8×16 matrix in which only the upper 8 row vectors of the 16×16 matrix are sampled can be used for the transformation. The complexity control method will be described in detail later.

[0307] The first embodiment and the second embodiment may be applied simultaneously, or either of the two embodiments may be applied. In particular, in the case of the second embodiment, because one transformation is considered in LFNST, it is experimentally observed that the compression performance improvement that can be obtained in the existing LFNST is relatively small compared to the LFNST index signaling cost. However, in the case of the first embodiment, a compression performance improvement similar to that that can be obtained from the conventional LFNST is observed. That is, in the case of ISP, the contribution of the application of 2×N and N×2 LFNST to the actual compression performance can be checked through experiments.

[0308] In the LFNST of the current VVC, symmetry between intra prediction modes is applied. The same set of LFNST is applied to two directional modes set around mode 34 (prediction in the 45-degree diagonal direction at the lower right corner), for example, the same set of LFNST is applied to mode 18 (horizontal prediction mode) and mode 50 (vertical prediction mode). However, in modes 35 to 66, when forward LFNST is applied, the input data is transposed and then LFNST is applied.

[0309] VVC supports wide angle intra prediction (WAIP) mode, and the LFNST set is derived based on the modified intra prediction mode taking into account the WAIP mode. For the modes extended by WAIP, the LFNST set is determined by using symmetry, just like in the general intra prediction direction mode. For example, because mode-1 is symmetric with mode 67, the same LFNST set is applied, and because mode-14 is symmetric with mode 80, the same LFNST set is applied. Modes 67 to 80 apply the LFNST transform after transposing the input data before applying the forward LFNST.

[0310] In the case of applying LFNST to the top left M×2 (M×1) block, since the block to which LFNST is applied is non-square, the symmetry to LFNST cannot be applied. Therefore, instead of applying the symmetry based on the intra prediction mode, as in the LFNST of Table 2, the symmetry between the M×2 (M×1) block and the 2×M (1×M) block can be applied.

[0311] Fig.19 is a diagram illustrating symmetry between an M×2 (M×1) block and a 2×M (1×M) block according to an embodiment.

[0312] like Fig.19 As shown, since mode 2 in the M×2(M×1) block can be considered symmetric to mode 66 in the 2×M(1×M) block, the same LFNST set can be applied to the 2×M(1×M) block and the M×2(M×1) block.

[0313] In this case, in order to apply the LFNST set applied to the M×2 (M×1) block to the 2×M (1×M) block, the LFNST set is selected based on mode 2 instead of mode 66. That is, before applying the forward LFNST, the LFNST may be applied after transposing the input data of the 2×M (1×M) block.

[0314] Fig. 20 is a diagram illustrating an example of transposing a 2×M block according to an embodiment.

[0315] Fig. 20 (a) is a diagram illustrating that LFNST can be applied by reading 2×M blocks of input data in column-major order, Fig. 20 (b) is a diagram illustrating that LFNST can be applied by reading input data of an M×2(M×1) block in a row-major order. A method of applying LFNST to the top left M×2(M×1) or 2×M(M×1) block is described as follows.

[0316] 1. First, if Fig. 20 As shown in (a) and (b) of FIG. , the input data is arranged to form the input vector of the forward LFNST. For example, referring to Fig.19 , for an M×2 block predicted in mode 2, follow Fig. 20 For a 2×M block predicted in mode 66, the input data is in the order of Fig. 20 The sequential arrangement of (a) can then apply the LFNST set for mode 2.

[0317] 2. For M×2 (M×1) blocks, considering WAIP, determine the LFNST set based on the modified intra prediction mode. As described above, a preset mapping relationship is established between the intra prediction mode and the LFNST set, which can be represented by the mapping table shown in Table 2.

[0318] For a 2×M (1×M) block, a symmetric pattern around the prediction mode (mode 34 in the case of the VVC standard) in a 45-degree diagonal direction downward from the modified intra prediction mode in consideration of WAIP can be obtained, and then the LFNST set is determined based on the corresponding symmetric pattern and the mapping table. The symmetric pattern (y) around mode 34 can be derived by the following formula. The mapping table will be described in more detail below.

[0319] [Formula 11]

[0320] If 2≤x≤66, then y=68-x,

[0321] Otherwise (x≤-1 or x≥67), y=66-x

[0322] 3. When the forward LFNST is applied, the transform coefficient may be derived by multiplying the input data prepared in process 1 by the LFNST kernel. The LFNST kernel may be selected according to the LFNST set determined in process 2 and a predetermined LFNST index.

[0323] For example, when M=8 and a 16×16 matrix is ​​applied as the LFNST kernel, 16 transform coefficients can be generated by multiplying the matrix by 16 input data. The generated transform coefficients can be arranged in the upper left 8×2 or 2×8 area according to the scanning order used in the VVC standard.

[0324] Fig.21 The scanning order of 8×2 or 2×8 areas according to an embodiment is illustrated.

[0325] All regions except the upper left 8×2 or 2×8 region may be filled with zero values ​​(cleared), or existing transform coefficients to which one transform is applied may be left intact. The predetermined LFNST index may be one of the LFNST index values ​​(0, 1, 2) attempted when calculating the RD cost while changing the LFNST index value in a programming process.

[0326] In the case of a configuration that adjusts the worst-case computational complexity to a certain degree or lower (e.g., 8 multiplications / sample), for example, after generating only 8 transform coefficients by multiplying an 8×16 matrix taking only the upper 8 rows of the 16×16 matrix, the transform coefficients may be calculated as Fig.21 The scan order setting of , and zeroing can be applied to the remaining coefficient regions. The worst-case complexity control will be described later.

[0327] 4. When inverse LFNST is applied, a preset number (e.g., 16) of transform coefficients are set as an input vector, and the LFNST set obtained from process 2 and an LFNST kernel (e.g., a 16×16 matrix) derived from the selected parsed LFNST index are selected, and then an output vector can be derived by multiplying the LFNST kernel with the corresponding input vector.

[0328] In the case of M×2(M×1) blocks, the output vector can be expressed as Fig. 20 In the case of 2×M (1×M) blocks, the output vector can be Fig. 20 (a) Column priority setting.

[0329] The remaining areas except the area where the corresponding output vector is set within the upper left M×2 (M×1) or 2×M (M×2) area and the area except the upper left M×2 (M×1) or 2×M (M×2) area in the partition block (M×2 in the partition block) can all be cleared to have zero values, or can be configured to keep the reconstructed transform coefficients as is through residual encoding and inverse quantization processing.

[0330] When constructing the input vector, as in point 3, the input data can be Fig.21 The scanning order arrangement is carried out, and in order to control the computational complexity of the worst case to a certain level or lower, the input vector can be constructed by reducing the number of input data (for example, 8 instead of 16).

[0331] For example, when M=8, if 8 input data are used, only the left 16×8 matrix can be taken from the corresponding 16×16 matrix and multiplied to obtain 16 output data. Worst case complexity control will be described later.

[0332] In the above embodiment, when LFNST is applied, the case where symmetry is applied between M×2 (M×1) blocks and 2×M (1×M) blocks is shown, but according to another example, a different LFNST set may be applied to each of the two block shapes.

[0333] Hereinafter, various examples of a mapping method using an intra prediction mode and a LFNST set configuration of an ISP mode will be described.

[0334] In the case of ISP mode, the LFNST set configuration may be different from the existing LFNST set. In other words, a core different from the existing LFNST core may be applied, and a mapping table different from the mapping table between the intra prediction mode index and the LFNST set applied to the current VVC standard may be applied. The mapping table applied to the current VVC standard may be the same as the mapping table of Table 2.

[0335] In Table 2, the preModeIntra value indicates an intra prediction mode value changed in consideration of WAIP, and the lfnstTrSetIdx value is an index value indicating a specific LFNST set. Each LFNST set is configured with two LFNST cores.

[0336] When the ISP prediction mode is applied, if both the horizontal length and the vertical length of each partition block are equal to or greater than 4, the same kernel as the LFNST kernel applied in the current VVC standard can be applied, and the mapping table can be applied as is. A mapping table and LFNST kernel different from the current VVC standard can be applied.

[0337] When the ISP prediction mode is applied, when the horizontal length or vertical length of each partition block is less than 4, a mapping table and LFNST core different from the current VVC standard may be applied. Hereinafter, Tables 6 to 8 represent mapping tables between intra prediction mode values ​​(intra prediction mode values ​​changed in consideration of WAIP) and LFNST sets, which may be applied to M×2 (M×1) blocks or 2×M (1×M) blocks.

[0338] [Table 6]

[0339] predModeIntra lfnstTrSetIdx predModeIntra<0 1 0<=predModeIntra<=1 0 2<=predModeIntra<=12 1 13<=predModeIntra<=23 2 24<=predModeIntra<=34 3 35<=predModeIntra<=44 4 45<=predModeIntra<=55 5 56<=predModeIntra<=66 6 67<=predModeIntra<=80 6 81<=predModeIntra<=83 0

[0340] [Table 7]

[0341] predModeIntra lfnstTrSetIdx predModeIntra<0 1 0<=predModeIntra<=1 0 2<=predModeIntra<=23 1 24<=predModeIntra<=44 2 45<=predModeIntra<=66 3 67<=predModeIntra<=80 3 81<=predModeIntra<=83 0

[0342] [Table 8]

[0343] predModeIntra lfnstTrSetIdx predModeIntra<0 1 0<=predModeIntra<=1 0 2<=predModeIntra<=80 1 81<=predModeIntra<=83 0

[0344] The first mapping table of Table 6 is configured with seven LFNST sets, the mapping table of Table 7 is configured with four LFNST sets, and the mapping table of Table 8 is configured with two LFNST sets. As another example, when it is configured with one LFNST set, the lfnstTrSetIdx value may be fixed to 0 with respect to the preModeIntra value.

[0345] Hereinafter, a method of maintaining the worst-case computational complexity when applying LFNST to the ISP mode will be described.

[0346] In the case of ISP mode, when LFNST is applied, in order to keep the number of multiplications per sample (or per coefficient, per position) to a certain value or less, the application of LFNST may be limited. Depending on the size of the partition block, the number of multiplications per sample (or per coefficient, per position) can be kept at 8 or less by applying LFNST as follows.

[0347] 1. When both the horizontal length and the vertical length of the partition block are 4 or more, the same method as the worst-case computational complexity control method for LFNST in the current VVC standard may be applied.

[0348] That is, when the partition block is a 4×4 block, an 8×16 matrix obtained by sampling the upper 8 rows from the 16×16 matrix may be applied in the forward direction instead of the 16×16 matrix, and a 16×8 matrix obtained by sampling the left 8 columns from the 16×16 matrix may be applied in the reverse direction. In addition, when the partition block is an 8×8 block, in the forward direction, instead of the 16×48 matrix, an 8×48 matrix obtained by sampling the upper 8 rows from the 16×48 matrix is ​​applied, and in the reverse direction, instead of the 48×16 matrix, a 48×8 matrix obtained by sampling the left 8 columns from the 48×16 matrix may be applied.

[0349] In the case of a 4×N or N×4 (N>4) block, when forward transform is performed, 16 coefficients generated after applying the 16×16 matrix to only the upper left 4×4 block may be set in the upper left 4×4 region, and the other regions may be filled with a value of 0. In addition, when inverse transform is performed, the 16 coefficients located in the upper left 4×4 block are set in a scanning order to form an input vector, and then 16 output data may be generated by multiplying the 16×16 matrix. The generated output data may be set in the upper left 4×4 region, and the remaining regions except the upper left 4×4 region may be filled with a value of 0.

[0350] In the case of an 8×N or N×8 (N>8) block, when performing forward transform, the generated 16 coefficients after applying the 16×48 matrix to the ROI region within only the upper left 8×8 block (except for the remaining regions except the lower right 4×4 block from the upper left 8×8 block) can be set in the upper left 4×4 region, and all other regions can be filled with a value of 0. In addition, when performing inverse transform, the 16 coefficients located in the upper left 4×4 region are set in a scanning order to form an input vector, and then 48 output data can be generated by multiplying the 48×16 matrix. The generated output data can be filled in the ROI region, and all other regions can be filled with a value of 0.

[0351] 2. When the size of the partition block is N×2 or 2×N and LFNST is applied to the upper left M×2 or 2×M region (M≤N), a matrix sampled according to the N value may be applied.

[0352] In the case of M=8, for a partition block of N=8, i.e., an 8×2 or 2×8 block, in the case of forward transform, an 8×16 matrix obtained by sampling the upper 8 rows from a 16×16 matrix may be applied instead of a 16×16 matrix, and in the case of inverse transform, a 16×8 matrix obtained by sampling the left 8 columns from a 16×16 matrix may be applied instead of a 16×16 matrix.

[0353] When N is greater than 8, in the case of forward transform, 16 output data generated after applying the 16×16 matrix to the upper left 8×2 or 2×8 block are set in the upper left 8×2 or 2×8 block, and the remaining area can be filled with the value 0. In the case of inverse transform, the 16 coefficients located in the upper left 8×2 or 2×8 block are arranged in a scanning order to form an input vector, and then 16 output data can be generated by multiplying the 16×16 matrix. The generated output data can be placed in the upper left 8×2 or 2×8 block, and all the remaining areas can be filled with the value 0.

[0354] 3. When the size of the partition block is N×1 or 1×N and LFNST is applied to the upper left M×1 or 1×M region (M≤N), a matrix sampled according to the N value may be applied.

[0355] When M=16, for a partition block of N=16, i.e., a 16×1 or 1×16 block, in the case of forward transform, an 8×16 matrix obtained by sampling the upper 8 rows from a 16×16 matrix may be applied instead of a 16×16 matrix, and in the case of inverse transform, a 16×8 matrix obtained by sampling the left 8 columns from a 16×16 matrix may be applied instead of a 16×16 matrix.

[0356] When N is greater than 16, in the case of forward transform, 16 output data generated after applying the 16×16 matrix to the upper left 16×1 or 1×16 block can be set in the upper left 16×1 or 1×16 block, and the remaining area can be filled with a value of 0. In the case of inverse transform, the 16 coefficients located in the upper left 16×1 or 1×16 block can be set in a scanning order to form an input vector, and then 16 output data can be generated by multiplying the 16×16 matrix. The generated output data can be set in the upper left 16×1 or 1×16 block, and all remaining areas can be filled with a value of 0.

[0357] As another example, in order to keep the number of multiplications for each sample (or each coefficient, each position) at a certain value or less, the number of multiplications for each sample (or each coefficient, each position) based on the ISP coding unit size rather than the size of the ISP partition block can be kept at 8 or less. When only one block among the ISP partition blocks meets the conditions for applying LFNST, the worst-case complexity calculation of LFNST can be applied based on the corresponding coding unit size rather than the size of the partition block. For example, the luminance coding block of a coding unit (CU) is divided (or partitioned) into four partition blocks, each of which has a size of 4×4. In addition, in this article, among the four partition blocks, if there are no non-zero transform coefficients for two partition blocks, each of the remaining two partition blocks can be configured to generate 16 transform coefficients therein instead of 8 transform coefficients (based on the encoder).

[0358] Hereinafter, a method for signaling a LFNST index in case of the ISP mode will be described.

[0359] As described above, the LFNST index may have values ​​0, 1, 2, where 0 indicates that LFNST is not applied, and where 1 and 2 respectively indicate each of two LFNST kernel matrices included in the selected LFNST set. LFNST is applied based on the LFNST kernel matrix selected by the LFNST index. In the current VVC standard, a method according to which LFNST is transmitted will be described as follows.

[0360] 1. The LFNST index may be sent once for each coding unit (CU), and in the case of a dual tree, the LFNST index may be signaled separately for each of the luma block and the chroma block.

[0361] 2. When the LFNST index is not signaled, the LFNST index is inferred to be 0 as a default value. The case where the LFNST index value is inferred to be 0 will be described below.

[0362] A. When the mode corresponds to a mode in which transform is not applied (e.g., transform skip, BDPCM, lossless coding, etc.)

[0363] B. When the primary transform is not DCT-2 (DST7 or DCT8), that is, when the horizontal transform or the vertical transform is not DCT-2

[0364] C. When the horizontal length or vertical length of the luma block of the coding unit exceeds the maximum luma transform size available for transformation, for example, when the maximum luma transform size available for transformation is equal to 64, and when the size of the luma block of the coding block is equal to 128×16, LFNST cannot be applied.

[0365] In the case of a dual tree, it is determined whether the maximum luma transform size is exceeded for each of the coding units of the luma element and the coding units of the chroma element. That is, it is checked whether the maximum luma transform size available for transform is exceeded for the luma block, and it is checked whether the maximum luma transform size and horizontal / vertical length of the transform of the corresponding luma block available for the color format are exceeded for the chroma block. For example, when the color format is 4:2:0, each of the horizontal / vertical lengths of the corresponding luma block becomes twice the length of the corresponding chroma block, and the transform size of the corresponding luma block becomes twice the size of the corresponding chroma block. As another example, when the color format is 4:4:4, the horizontal / vertical length and transform size of the corresponding luma block are the same as those of the corresponding chroma block.

[0366] The 64-length transform or the 32-length transform indicates a transform applied to a horizontal length or a vertical length of 64 or 32, respectively. In addition, the "transform size" may indicate a corresponding length of 64 or 32.

[0367] In the case of a single tree, after checking whether the horizontal length or vertical length of the luma block exceeds the maximum luma transform block size available for transform, LFNST index signaling may be skipped (or omitted) when the length exceeds the transform block size.

[0368] D. LFNST index may be transmitted only when both the horizontal length and the vertical length of the coding unit are equal to 4 or more.

[0369] In the case of a dual tree, the LFNST index may be signaled only when both the horizontal length and the vertical length of the corresponding element (ie, luma element or chroma element) are equal to 4 or greater.

[0370] In the case of a single tree, when both the horizontal length and the vertical length of the luma element are equal to 4 or greater, the LFNST index may be signaled.

[0371] E. When the last non-zero coefficient position is not the DC position (upper left position in the block), if the block is a dual-tree type luma block and if the last non-zero coefficient position is not the DC position, the LFNST index is sent. If the block is a dual-tree type chroma block and if at least one of the last non-zero coefficient position of Cb and the last non-zero coefficient position of Cr is not the DC position, the corresponding LFNST index is sent.

[0372] In case of a single tree type, for any one of the luma element, the Cb element, and the Cr element, if the corresponding last non-zero coefficient position is not the DC position, the LFNST index is transmitted.

[0373] Herein, when a coded block flag (CBF) value indicating the presence or absence of a transform coefficient of one transform block is equal to 0, in order to determine whether to perform LFNST index signaling, the last non-zero coefficient position of the corresponding transform block is not checked. That is, when the corresponding CBF value is equal to 0, since the transform is not applied to the corresponding block, the last non-zero coefficient position may not be considered when checking the condition for LFNST index signaling.

[0374] For example, 1) in the case of a dual tree type and a luma element, if the corresponding CBF value is equal to 0, the LFNST index is not signaled, 2) in the case of a dual tree type and a chroma element, if the CBF value of Cb is equal to 0 and the CBF value of Cr is equal to 1, only the position of the last non-zero coefficient position of Cr is checked in order to send the corresponding LFNST index, and 3) in the case of a single tree type, only the last non-zero coefficient position of the luma element, Cb element, or Cr element, each having a CBF value of 1, is checked.

[0375] F. When verifying that a transform coefficient exists in a position other than a position where an LFNST transform coefficient may exist, LFNST index signaling may be skipped (or omitted). In the case of a 4×4 transform block and an 8×8 transform block, according to the transform coefficient scanning order of the VVC standard, the LFNST transform coefficient may exist in 8 positions starting from the DC position, and all the remaining positions may be filled with 0. Additionally, in the case where the transform block is not a 4×4 transform block and an 8×8 transform block, according to the transform coefficient scanning order of the VVC standard, the LFNST transform coefficient may exist in 16 positions starting from the DC position, and all the remaining positions may be filled with 0.

[0376] Therefore, after performing residual encoding, when a non-zero transform coefficient exists in a region that should be filled with only 0 values, LFNST index signaling may be skipped (or omitted).

[0377] In addition, the ISP mode can be applied only to luma blocks or can be applied to both luma blocks and chroma blocks. As described above, when ISP prediction is applied, prediction is performed after the corresponding coding unit is divided (or partitioned) into 2 or 4 partition blocks, and the transform can also be applied to each of the corresponding partition blocks. Therefore, even when the conditions for signaling the LFNST index are determined by the coding unit, it should be considered that the LFNST can be applied to each of the corresponding partition blocks. Additionally, when the ISP prediction mode is applied only to a specific element (e.g., a luma block), the LFNST index should be signaled based on the fact that the coding unit is divided into partitions only for the corresponding element. The LFNST index signaling method that can be used for the ISP mode will be described below.

[0378] 1. The LFNST index may be sent once for each coding unit (CU), and in the case of a dual tree, the LFNST index may be signaled separately for each of the luma block and the chroma block.

[0379] 2. When the LFNST index is not signaled, the LFNST index is inferred to be 0 as a default value. The case where the LFNST index value is inferred to be 0 will be described below.

[0380] A. When the mode corresponds to a mode in which transform is not applied (e.g., transform skip, BDPCM, lossless coding, etc.)

[0381] B. When the primary transform is not DCT-2 (DST7 or DCT8), that is, when the horizontal transform or the vertical transform is not DCT-2

[0382] Whether to perform LFNST index signaling can be determined based on the size of the partition block rather than the size of the coding unit. That is, when the horizontal length or vertical length of the partition block corresponding to the luminance block exceeds the maximum luminance transform size available for transformation, LFNST index signaling can be skipped (or omitted) and the LFNST index value can be inferred to be 0.

[0383] In the case of a dual tree, for each coding unit or partition block of the luma element and for each coding unit or partition block of the chroma element, it is determined whether the maximum block transform size is exceeded. That is, by comparing each of the horizontal length and vertical length of the coding unit or partition block of the luma element with the maximum luma transform size, and when it is determined that at least one length is greater than the maximum luma transform size, LFNST is not applied. In addition, in the case of the coding unit or partition block of the chroma element, the horizontal / vertical length of the corresponding luma block for the color format is compared with the maximum luma transform size that can be used for the maximum transform. For example, when the color format is 4:2:0, each of the horizontal / vertical lengths of the corresponding luma block becomes twice the length of the corresponding chroma block, and the transform size of the corresponding luma block becomes twice the size of the corresponding chroma block. As another example, when the color format is 4:4:4, the horizontal / vertical length and transform size of the corresponding luma block are the same as those of the corresponding chroma block.

[0384] In the case of a single tree, after checking whether the horizontal length or vertical length of the luma block (coding unit or partition block) exceeds the maximum luma transform block size available for transform, LFNST index signaling can be skipped (or omitted) when the length exceeds the transform block size.

[0385] C. If LFNST included in the current VVC standard is applied, the LFNST index may be transmitted only when both the horizontal length and the vertical length of the partition block are equal to 4 or more.

[0386] In addition to the LFNST included in the current VVC standard, if LFNST for 2×M (1×M) or M×2 (M×1) blocks is applied, the LFNST index may be transmitted only for the case where the partition block size is equal to or greater than 2×M (1×M) or M×2 (M×1) blocks. Here, when a P×Q block is equal to or greater than an R×S block, this means that P≥R and Q≥S.

[0387] In summary, the LFNST index may be signaled only for the case where the partition block size is equal to or greater than the minimum size to which LFNST may be applied. In the case of a dual tree, the LFNST index may be signaled only if the size of the partition block of the luma or chroma element is equal to or greater than the minimum size to which LFNST may be applied. In the case of a single tree, the LFNST index may be signaled only if the size of the partition block of the luma element is equal to or greater than the minimum size to which LFNST may be applied.

[0388] In this specification, when an M×N block is equal to or greater than a K×L block, this means that M is equal to or greater than K and N is equal to or greater than L. When an M×N block is greater than a K×L block, this means that M is equal to or greater than K and N is equal to or greater than L, while M is greater than K or N is greater than L. When an M×N block is less than or equal to a K×L block, this means that M is less than or equal to K and N is less than or equal to L. In addition, when an M×N block is less than a K×L block, this means that M is less than or equal to K and N is less than or equal to L, while M is less than K or N is less than L.

[0389] D. When the last non-zero coefficient position is not the DC position (upper left position in the block), if the block is a dual-tree type luma block, and if the corresponding last non-zero coefficient position of even one partition block among all partition blocks is not the DC position, the LFNST index may be sent. If the block is a dual-tree type chroma block, and if even one of the last non-zero coefficient position of all partition blocks for Cb (when the ISP mode is not applied to chroma elements, it is assumed that the number of partition blocks is equal to 1) and the last non-zero coefficient position of all partition blocks for Cr (when the ISP mode is not applied to chroma elements, it is assumed that the number of partition blocks is equal to 1) is not the DC position, the corresponding LFNST index may be sent.

[0390] In case of a single tree type, for any one of a luma element, a Cb element, and a Cr element, if a corresponding last non-zero coefficient position of even one partition block among all partition blocks is not a DC position, a LFNST index may be transmitted.

[0391] Herein, when a coded block flag (CBF) value indicating the presence or absence of a transform coefficient of each partition block is equal to 0, in order to determine whether to perform LFNST index signaling, the last non-zero coefficient position of the corresponding partition block is not checked. That is, when the corresponding CBF value is equal to 0, since the transform is not applied to the corresponding block, when checking the condition for LFNST index signaling, the last non-zero coefficient position of the corresponding partition block is considered.

[0392] For example, 1) in the case of a dual tree type and a luma element, if the corresponding CBF value of each partition block is equal to 0, the corresponding partition block is excluded when determining whether to perform LFNST index signaling, 2) in the case of a dual tree type and a chroma element, if for each partition block, the CBF value of Cb is equal to 0 and the CBF value of Cr is equal to 1, only the position of the last non-zero coefficient position of Cr is checked to determine whether to perform the corresponding LFNST index signaling, and 3) in the case of a single tree type, only the last non-zero coefficient position of the luma element, Cb element, or Cr element of all partition blocks, each of which has a CBF value of 1, is checked to determine whether to perform LFNST index signaling.

[0393] In the case of the ISP mode, the image information may be configured such that the last non-zero coefficient position is not checked, and the corresponding implementation will be described below.

[0394] i. In the case of ISP mode, the check of the last non-zero coefficient position of both the luminance block and the chrominance block is skipped, and LFNST index signaling can be authorized. That is, even if the last non-zero coefficient position of all partition blocks is the DC position or has a corresponding CBF value of 0, the corresponding LFNST index signaling can be authorized.

[0395] ii. In the case of ISP mode, only the check of the last non-zero coefficient position of the luma block is skipped, and for the chroma block, the check of the last non-zero coefficient position according to the above method may be performed. For example, in the case of a dual tree type and a luma block, the check of the last non-zero coefficient position is not performed, and LFNST index signaling may be authorized. In addition, in the case of a dual tree type and a chroma block, the presence or absence of a DC position corresponding to the last non-zero coefficient position is checked according to the above method to determine whether to perform signaling of the corresponding LFNST index.

[0396] iii. In the case of ISP mode and single tree type, method number i and method number ii may be applied. That is, in the case of applying method number i to ISP mode and single tree type, checking of the last non-zero coefficient position of both the luminance block and the chrominance block may be skipped, and LFNST index signaling may be authorized. Alternatively, by applying method number ii, checking of the last non-zero coefficient position of the partition block of the luminance element may be skipped, and checking of the last non-zero coefficient position of the partition block of the chrominance element may be performed according to the above method (when the ISP mode is not applied to the chrominance element, it may be assumed that the number of partition blocks is equal to 1) in order to determine whether to perform the corresponding LFNST index signaling.

[0397] E. When it is verified that a transform coefficient exists in a position other than a position where a LFNST transform coefficient may exist even for one partition block among all partition blocks, LFNST index signaling may be skipped (or omitted).

[0398] For example, in the case of a 4×4 partition block and an 8×8 partition block, according to the transform coefficient scanning order of the VVC standard, the LFNST transform coefficient may be present in 8 positions starting from the DC position, and all the remaining positions may be filled with 0. Additionally, in the case where the partition block is equal to or greater than 4×4, and in the case where the partition block is not a 4×4 partition block and an 8×8 partition block, according to the transform coefficient scanning order of the VVC standard, the LFNST transform coefficient may be present in 16 positions starting from the DC position, and all the remaining positions may be filled with 0.

[0399] Therefore, after performing residual encoding, when a non-zero transform coefficient exists in a region that should be filled with only 0 values, LFNST index signaling may be skipped (or omitted).

[0400] If LFNST can be applied even for a case where the partition block size is equal to 2×M (1×M) or M×2 (M×1), an area where the LFNST transform coefficient can be located can be specified as described below. An area outside the area where the LFNST transform coefficient can be located can be filled with 0. In addition, when it is assumed that LFNST has been applied, if there is a non-zero transform coefficient in an area that should be filled with 0, LFNST index signaling can be skipped.

[0401] i. When LFNST can be applied to a 2×M or M×2 block, and when M=8, only 8 LFNST transform coefficients can be generated for a 2×8 or 8×2 partition block. Fig. 20 When arranging the transform coefficients in the scan order shown, 8 transform coefficients are arranged in a scan order starting from the DC position, and the remaining 8 positions may be filled with zeros.

[0402] 16 LFNST transform coefficients can be generated for 2×N or N×2 (N>8) partition blocks. Fig. 20 When arranging the transform coefficients in the scan order shown, 16 transform coefficients are arranged in a scan order starting from the DC position, and the remaining area can be filled with 0. That is, in a 2×N or N×2 (N>8) partition block, the area other than the upper left 2×8 or 8×2 block can be filled with 0. Instead of 8 LFNST transform coefficients, 16 coefficient blocks can also be generated for a 2×8 or 8×2 partition block, and in this case, there is no area that needs to be filled with 0. As described above, when LFNST is applied, when a non-zero transform coefficient is detected as existing in an area designated to be filled with 0 (even in one partition block), LFNST index signaling can be skipped and the LFNST index can be inferred to be 0.

[0403] ii. When LFNST can be applied to a 1×M or M×1 block, and when M=16, 8 LFNST transform coefficients can be generated only for the 1×16 or 16×1 partition block. When the transform coefficients are arranged in a scanning order from left to right or from top to bottom, the 8 transform coefficients are arranged in a corresponding scanning order starting from the DC position, and the remaining 8 positions can be filled with 0.

[0404] 16 LFNST transform coefficients may be generated for a 1×N or N×1 (N>16) partition block. When the transform coefficients are arranged in a scan order from left to right or from top to bottom, the 16 transform coefficients are arranged in a corresponding scan order starting from the DC position, and the remaining area may be filled with 0. That is, in a 1×N or N×1 (N>16) partition block, an area other than the upper left 1×16 or 16×1 block may be filled with 0.

[0405] Instead of 8 LFNST transform coefficients, 16 coefficient blocks may also be generated for a 1×16 or 16×1 partition block, and in this case, there is no area that needs to be filled with 0. As described above, when LFNST is applied, when a non-zero transform coefficient is detected as existing in an area designated to be filled with 0 (even in one partition block), LFNST index signaling may be skipped and the LFNST index may be inferred to be 0.

[0406] In addition, in the case of ISP mode, in the current VVC standard, DST-7 is applied instead of DCT-2 by independently (or separately) referring to the length conditions in the horizontal direction and the vertical direction, and signaling for the MTS index is not performed. Depending on whether the horizontal length or the vertical length is equal to or greater than 4 and less than or equal to 16, a transform kernel is determined. Therefore, in the case of ISP mode, and when LFNST can be applied, the following transform combination can be configured as described below.

[0407] 1. For the case where the LFNST index is 0 (including the case where the LFNST index is inferred to be 0), the conditions for determining a primary transformation corresponding to the ISP mode included in the current VVC standard may be followed. That is, by independently (or separately) checking whether the length conditions in the horizontal direction and the vertical direction (i.e., the condition that the length is equal to or greater than 4 and less than or equal to 16) are satisfied, if the length condition is satisfied, DST-7 is applied for the primary transformation instead of DCT-2. In addition, if the length condition is not satisfied, DCT-2 may be applied.

[0408] 2. For the case where the LFNST index is greater than 0, the following two configurations are possible for one transformation.

[0409] A. DCT-2 can be applied to both horizontal and vertical directions.

[0410] B. The conditions for determining a transformation corresponding to the ISP mode included in the current VVC standard may be followed. That is, by independently (or separately) checking whether the length conditions in the horizontal direction and the vertical direction (i.e., the condition that the length is equal to or greater than 4 and less than or equal to 16) are satisfied, if the length condition is satisfied, DST-7 is applied instead of DCT-2. In addition, if the length condition is not satisfied, DCT-2 may be applied.

[0411] In the case of the ISP mode, the image information may be configured so that the LFNST index may be transmitted for each partition block instead of for each coding unit. In this case, the above-mentioned LFNST index signaling method assumes that only one partition block exists within the unit through which the LFNST index is transmitted, and whether to perform LFNST index signaling may be determined.

[0412] The above-mentioned implementation of applying LFNST in ISP mode can be summarized as follows.

[0413] (1) When LFNST is applied in ISP mode, the split (or partitioned) transform unit should have a minimum size of 4×4 or larger.

[0414] (2) A LFNST core may be used, which is the same as a conventional LFNST core applied to a coding unit to which the ISP mode is not applied.

[0415] (3) All transform units shall satisfy the maximum (or highest) last position value condition (ie, the position condition of the last non-zero significant coefficient). If one or more transform units do not satisfy the maximum last position condition, LFNST is not used and the LFNST index is not parsed.

[0416] (4) When the ISP mode is applied, a configuration that allows the LFNST to be applied when a significant coefficient exists in a position other than the DC position may be discarded (or ignored).

[0417] (5) If LFNST is applied, DCT-2 is used for one transform of the transform unit to which the ISP mode is applied.

[0418] Table 9 shown below indicates syntax elements including the details described above.

[0419] [Table 9]

[0420]

[0421] Table 9 configures the width and height of the area where LFNST is applied according to the tree type and indicates the condition for transmitting the LFNST index. The syntax elements of Table 9 can be signaled at the coding unit (CU) level. In the case of a dual tree, a separate LFNST index can be signaled for each of the luma block and the chroma block.

[0422] First, if the tree type of a coding unit is dual-tree chroma, the width of the region to which LFNST is applied (lfnstWidth) may be set to reflect the width of the color format according to the width of the coding unit ((treeType==DUAL_TREE_CHROMA)?cbWidth / SubWidthC).

[0423] On the contrary, if the tree type of the coding unit is not dual-tree chroma, that is, in the case of dual-tree luminance or single tree, the width (lfnstWidth) of the area to which LFNST is applied can be set to a value obtained by dividing the coding unit by the number of sub-partitions, or the width of the coding unit ((IntraSubPartitionsSplitType==ISP_VER_SPLIT)?cbWidth / NumIntraSubPartitions:cbWidth) according to whether the coding unit is divided (or partitioned) by ISP. That is, if the coding unit is divided (or partitioned) in the vertical direction by ISP (IntraSubPartitionsSplitType==ISP_VER_SPLIT), the width of the area to which LFNST is applied can be set to a value obtained by dividing the coding unit by the number of sub-partitions (cb Width / NumIntraSubPartitions), and if the coding unit is not divided (or partitioned), the width of the corresponding area can be set to the width of the coding unit (cbWidth).

[0424] Similarly, if the tree type of the coding unit is dual-tree chroma, the height of the area to which LFNST is applied (lfnstHeight) can be set to reflect the height of the color format according to the height of the coding unit ((treeType==DUAL_TREE_CHROMA)?cbHeight / SubHeightC).

[0425] On the contrary, if the tree type of the coding unit is not dual-tree chroma, that is, in the case of dual-tree luminance or single tree, the height of the area to which LFNST is applied (lfnstHeight) can be set to a value obtained by dividing the coding unit by the number of sub-partitions, or the height of the coding unit ((IntraSubPartitionsSplitType==ISP_HOR_SPLIT)?cbHeight / NumIntraSubPartitions:cbHeight) according to whether the coding unit is divided (or partitioned) by ISP. That is, if the coding unit is divided (or partitioned) in the horizontal direction by ISP (IntraSubPartitionsSplitType==ISP_HOR_SPLIT), the height of the area to which LFNST is applied can be set to a value obtained by dividing the coding unit by the number of sub-partitions (cbHeight / NumIntraSubPartitions), and if the coding unit is not divided (or partitioned), the height of the corresponding area can be set to the height of the coding unit (cbHeight).

[0426] In order to allow LFNST to be applied as described above, the width and height of the region to which LFNST is applied should each be equal to or greater than 4 (Min(lfnstWidth, lfnstHeight)>=4). That is, in the case of encoding a dual tree, the LFNST index may be signaled only when both the horizontal length and the vertical length of the corresponding element (i.e., luma or chroma element) are equal to or greater than 4. In addition, in the case of a single tree, the LFNST index may be signaled when both the horizontal length and the vertical length of the luma element are equal to or greater than 4.

[0427] In case that ISP is applied to a coding unit, a LFNST index may be transmitted only if both the horizontal length and the vertical length of a partition block are equal to or greater than 4.

[0428] Additionally, if the horizontal length or vertical length of the luma block of the coding unit exceeds the maximum luma transform size available for transform (when the condition of Max(cbWidth, cbHeight)<=MaxTbSizeY is not satisfied), LFNST cannot be applied and the LFNST index is not sent.

[0429] Additionally, the LFNST index may be signaled only when the last non-zero coefficient position is not the DC position (upper left position in the block).

[0430] In the case of a dual-tree type luma block, if the last non-zero coefficient position is not the DC position, the LFNST index is transmitted. In the case of a dual-tree chroma block, even if one of the last non-zero coefficient position of Cb and the last non-zero coefficient position of Cr is not the DC position, the corresponding LFNST index is transmitted. In the case of a single tree type, if the corresponding last non-zero coefficient position of even one of the luma element, Cb element, and Cr element is not the DC position, the LFNST index may be transmitted.

[0431] In addition, when ISP is applied to a coding unit, the LFNST index can be signaled without checking the last non-zero coefficient position (IntraSubPartitionsSplitType!=ISP_NO_SPLIT||LfnstDcOnly==0). That is, even if the last non-zero coefficient position of all partition blocks is at the DC position, LFNST index signaling can be authorized.

[0432] Finally, when it is verified that a transform coefficient exists in a position other than a position where a LFNST transform coefficient may exist, LFNST index signaling may be skipped (Lfnst Zero Out Sig Coeff Flag==1).

[0433] When ISP is applied to a coding unit, if it is verified that a transform coefficient exists in a position other than a position where a LFNST transform coefficient may exist (even for one partition block among all partition blocks), LFNST index signaling may be skipped.

[0434] Hereinafter, when the ISP mode is applied, an example of using a 16×36 matrix when an 8×8 LFNST is used will be described.

[0435] In the case of ISP mode, and when LFNST is applied, when both the horizontal length and the vertical length of the partition block are equal to or greater than 8, a 16×36 matrix (based on forward LFNST) may be used instead of a 16×48 matrix (based on forward LFNST).

[0436] As described above, when the inverse 8×8 LFNST is applied, 48 coefficients can be derived by applying the 48×16 LFNST kernel to the 16 coefficients. The derived 48 coefficients can be arranged in the transform unit as Figure 8 shown.

[0437] Furthermore, according to an embodiment, a 36×16 LFNST kernel that derives 36 coefficients instead of 48 coefficients may be proposed.

[0438] Fig. 22 is a diagram showing an arrangement of 36 coefficients according to a 16×36 LFNST kernel according to an example.

[0439] The positions of the 36 coefficients can be determined according to the LFNST index. For example, if the LFNST index is 1, the 36 coefficients can be as follows Fig. 22 , and if the LFNST index is 2, the 36 coefficients can be arranged as shown in (a) of Fig. 22 Arrange as shown in (b).

[0440] The following figures are provided to describe specific examples of the present disclosure. Since the specific names of the devices shown in the figures or the names of the specific signals / messages / fields are provided for illustration, the technical features of the present disclosure are not limited to the specific names used in the following figures.

[0441] Fig.23 is a flowchart illustrating the operation of a video decoding device according to an embodiment of this document.

[0442] Fig.23 Each step disclosed in Figures 4 to 22 Therefore, some of the contents described in the above will be omitted or simplified Figures 3 to 22 Detailed descriptions of those described in .

[0443] The decoding apparatus 300 according to the present embodiment may receive residual information from a bitstream ( S2310 ).

[0444] More specifically, the decoding device 300 can decode information about the quantized transform coefficient of the target block from the bitstream, and can derive the quantized transform coefficient of the current block based on the information about the quantized transform coefficient of the current block. The information about the quantized transform coefficient of the target block may be included in a sequence parameter set (SPS) or a slice header, and may include information about whether a reduced transform (RST) is applied, information about a simplification factor, information about a minimum transform size for applying a reduced transform, information about a maximum transform size for applying a reduced transform, a reduced inverse transform size, and at least one of information about a transform index indicating any one transform kernel matrix included in a transform set.

[0445] Additionally, the decoding device may also receive information about the intra prediction mode of the current block and information about whether the ISP is applied to the current block. By receiving and parsing the flag information indicating whether the ISP code or ISP mode is applied, the decoding device may deduce whether the current block is divided (or split or partitioned) into a predetermined number of sub-partition transform blocks. In this article, the current block may be a coding block. In addition, the decoding device may deduce the size and number of the divided sub-partition blocks by indicating the flag information along which direction the current block is to be divided (partitioned).

[0446] For example, Fig.17 As shown, when the size (width×height) of the current block is 8×4, the current block can be vertically divided and divided into two sub-blocks, and when the size (width×height) of the current block is 4×8, the current block can be horizontally divided and divided into two sub-blocks. Alternatively, as Fig.18 As shown, when the size (width × height) of the current block is greater than 4 × 8 or 8 × 4, that is, when the size of the current block is: 1) 4 × N or N × 4 (N ≥ 16) or 2) M × N (M ≥ 8, N ≥ 8), the current block can be divided into 4 sub-blocks in the horizontal direction or the vertical direction.

[0447] The same intra prediction mode is applied to the sub-partition blocks divided (or partitioned) from the current block, and the decoding device can derive the prediction samples of each sub-partition block. That is, according to the division format of the sub-partition blocks, for example, the decoding device performs intra prediction in sequence horizontally or vertically, from left to right or from top to bottom. For the leftmost sub-block or the rightmost sub-block, the reconstructed pixels of the coded block that has been encoded are referenced, as in the general intra prediction method. Additionally, when the previous sub-partition block is not adjacent to each side of the subsequent internal sub-partition block, in order to derive the reference pixels adjacent to the corresponding side, the reconstructed pixels of the coded adjacent coding block are referenced, as in the general intra prediction method.

[0448] The decoding apparatus 300 may induce residual information of the current block, that is, induce a transform coefficient by performing dequantization on the quantized transform coefficient ( S2320 ).

[0449] The derived transform coefficients may be arranged (or aligned) according to an inverse diagonal scan order in a 4×4 block unit, and the transform coefficients within a 4×4 block may also be arranged according to an inverse diagonal scan order. That is, the transform coefficients processed using dequantization may be arranged according to an inverse scan order applied in a video codec such as VVC or HEVC.

[0450] As described above, the transform coefficient derived based on such residual information may be a dequantized transform coefficient. That is, whether quantization is performed or not, the transform coefficient will be sufficient to be data that can check whether the coefficient is non-zero data in the current block.

[0451] According to an example, the decoding apparatus may derive modified (or corrected) transform coefficients by applying LFNST to the transform coefficients.

[0452] Unlike a single transform that transforms the coefficients of the transform target separately in the vertical or horizontal direction, LFNST is an inseparable transform that applies the transform without separating the coefficients in a specific direction. Such an inseparable transform may be a low-frequency inseparable transform that applies the forward transform only in the low-frequency region rather than the entire block region.

[0453] The LFNST index information is received as syntax information, and the syntax information may be received as a binarized bin string including 0 and 1.

[0454] The syntax element of the LFNST index according to the present embodiment may indicate whether to apply any one of the inverse LFNST or the inverse inseparable transform and the transform kernel matrix included in the transform set.

[0455] That is, according to an embodiment, the syntax element value for the LFNST index may include 0, 1, and 2, 0 indicating a case where inverse LFNST is not applied to the target block, 1 indicating the first transform core matrix of the two transform core matrices, and 2 indicating the second transform core matrix of the two transform core matrices.

[0456] Intra prediction mode information and LFNST index information may be signaled at a coding unit level.

[0457] Additionally, the decoding apparatus according to the embodiment may also determine whether the transform coefficient exists in the second region other than the upper left first region of the current block. In addition, if the transform coefficient does not exist in the second region, the LFNST index may be parsed.

[0458] By deriving a variable indicating whether a significant coefficient exists in a second region other than the upper left first region of the current block, it may be checked whether clearing of the second region has been performed.

[0459] The variable may be a variable LfnstZeroOutSigCoeffFlag, which may indicate that zeroing has been performed when LFNST is applied. The variable LfnstZeroOutSigCoeffFlag may be initially set to 1, and when there are significant coefficients in the second region, the variable LfnstZeroOutSigCoeffFlag may be changed to 0.

[0460] The variable LfnstZeroOutSigCoeffFlag may be derived to be 0 when the index of the sub-block in which the last non-zero coefficient exists is greater than 0 and both the width and height of the transform block are equal to or greater than 4, or when the last position of the non-zero coefficient within the sub-block in which the last non-zero coefficient exists is greater than 7, and when the transform block size is equal to 4×4 or 8×8.

[0461] That is, when a non-zero coefficient is derived from an area in the transform block other than the upper left area where the LFNST transform coefficient may exist, or when a non-zero coefficient exists outside the eighth position of the scanning order of a 4×4 block and an 8×8 block, the variable LfnstZeroOutSigCoeffFlag is set to 0.

[0462] According to an example, when ISP is applied to a coding unit, clearing of the second region for LFNST may also be applied to all sub-partition blocks. At this time, if it is verified that a transform coefficient exists in a position other than a position where a LFNST transform coefficient may exist for even one sub-partition block, LFNST index signaling may be skipped.

[0463] The decoding device can derive residual samples of the current block based on an inverse transform of the transform coefficients for correction (or modification) or the transform coefficients derived from the residual information, and when the current block is divided into sub-partition blocks, the decoding device can derive a transform kernel for an inverse transform applied to the sub-partition block based on the horizontal length or vertical length of the sub-partition block (S2330).

[0464] As described above, when the current block is divided into sub-partition blocks, for inverse primary transform, multiple MTS indexes may not be signaled. In this case, an implicit MTS may be applied to the sub-partition blocks, and a transform kernel may be derived based on the horizontal length or vertical length of the sub-partition blocks.

[0465] According to an example, the transform kernel may be derived based on whether the horizontal length or the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16.

[0466] If each of the horizontal length and the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, the transform kernel can be derived as DST-7. In addition, in other cases, that is, if each of the horizontal length and the vertical length of the sub-partition block does not satisfy the condition of being equal to or greater than 4 and less than or equal to 16, the transform kernel can be derived as DCT-2.

[0467] Additionally, the decoding device may configure the size of a non-cleared block to which the transform kernel is applied and indicating an area within the sub-partition block where a valid transform coefficient may exist, and the decoding device may derive residual samples based on the size of such a non-cleared block.

[0468] Since each of the horizontal length and the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, if DST-7 is used as the transform kernel, the width and height of the non-cleared block may be set to 16.

[0469] Furthermore, if each of the horizontal length and vertical length of the sub-partition block does not satisfy the condition of being equal to or greater than 4 and less than or equal to 16, and therefore if DCT-2 is used as the transform kernel, the width and height of the non-cleared block can be set to up to 32 or less.

[0470] That is, according to an example, when multi-transform selection is applied (when DST-7 or DCT-8 can be applied), the decoding device may set the width or height of the non-cleared block to 16. In addition, when multi-transform selection is not applied, the decoding device may set the width or height of the non-cleared block to 32 or less.

[0471] As described above, when the ISP is applied to the current block, the MTS index indicating the transform core may not be signaled.

[0472] Furthermore, the above-described LFNST may be applied to the current block, and in this case, clearing of the second area corresponding to the LFNST may be performed on the sub-partitioned block.

[0473] When the LFNST index is equal to 0, or when it is assumed that the LFNST index is equal to 0 because it cannot be received, DCT-2 can be used as the transform kernel for the inverse primary transform in the sub-partition block, and the width or height of the non-cleared block can be set to 32 or less.

[0474] If LFNST is performed by assigning a value of 1 or 2 instead of 0 to the LFNST index, DCT-2 may be applied to the modified transform coefficient on which zeroing of the second region is performed. Also in this case, the width or height of the non-zeroed block may be set to 32 or less. That is, the residual sample may be derived by reflecting zero according to one transform.

[0475] The decoding apparatus may induce residual samples of the current block from the transform coefficients by applying the derived transform kernel ( S2340 ).

[0476] Subsequently, the decoding apparatus 300 may generate a reconstructed sample based on the residual sample of the current block and the prediction sample of the current block ( S2350 ).

[0477] The following figures are provided to describe specific examples of the present disclosure. Since the specific names of the devices shown in the figures or the names of the specific signals / messages / fields are provided for illustration, the technical features of the present disclosure are not limited to the specific names used in the following figures.

[0478] Fig.24 is a flowchart illustrating the operation of a video encoding device according to an embodiment of this document.

[0479] Fig.24 Each step disclosed in Figures 4 to 22 Therefore, some of the contents described in the above will be omitted or simplified Figure 2 and Figures 4 to 22 Detailed descriptions of those described in .

[0480] The encoding apparatus 200 according to an embodiment may first derive a prediction sample of the current block based on an intra prediction mode applied to the current block.

[0481] When applying the ISP to the current block, the encoding apparatus may perform prediction per sub-partitioned transform block.

[0482] The encoding device can determine whether ISP encoding or ISP mode is applied to the current block (ie, the encoding block), determine the direction in which the current block will be divided according to the determination result, and derive the size and number of divided sub-blocks.

[0483] For example, when the size (width × height) of the current block is 8 × 4, Fig.17 As shown, the current block can be vertically divided and divided into two sub-blocks, and when the size (width×height) of the current block is 4×8, the current block can be horizontally divided and divided into two sub-blocks. Fig.18 As shown, when the size (width × height) of the current block is greater than 4 × 8 or 8 × 4, that is, when the size of the current block is: 1) 4 × N or N × 4 (N ≥ 16) or 2) M × N (M ≥ 8, N ≥ 8), the current block can be divided into 4 sub-blocks in the horizontal direction or the vertical direction.

[0484] The same intra prediction mode can be applied to the sub-partition transform blocks divided from the current block, and the encoding device can derive the prediction samples of each sub-partition transform block. That is, the encoding device performs intra prediction in sequence from left to right or from top to bottom, for example, horizontally or vertically, according to the division form of the sub-partition transform blocks. For the leftmost or topmost sub-block, the reconstructed pixels of the already encoded coding block are referenced, as in the traditional intra prediction method. In addition, for each side of the subsequent internal sub-partition transform block, when it is not adjacent to the previous sub-partition transform block, in order to derive the reference pixels adjacent to the corresponding side, the reconstructed pixels of the already encoded adjacent coding block are referenced, as in the traditional intra prediction method.

[0485] The encoding apparatus 200 may induce a residual sample of the current block based on the prediction sample ( S2410 ).

[0486] Additionally, the encoding device 200 may derive a transform coefficient of the current block based on a primary transform of the residual sample. More specifically, when the current block is divided into sub-partition blocks, the encoding device 200 derives a transform kernel (S2420) of a primary transform applied to the sub-partition block based on a horizontal length or a vertical length of the sub-partition block. Furthermore, then, the encoding device 200 may derive a transform coefficient (S2430) based on the transform kernel.

[0487] One transform may be performed by a plurality of transform kernels, and in this case, the transform kernel may be selected based on the intra prediction mode.

[0488] According to an example, when the current block is divided into sub-partition blocks, an implicit MTS may be applied to the sub-partition blocks, and a transform kernel may be derived based on a horizontal length or a vertical length of the sub-partition blocks.

[0489] According to an example, the transform kernel may be derived based on whether the horizontal length or the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16.

[0490] If the horizontal length or the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, the transform kernel may be derived as DST-7. In addition, in other cases, that is, if the horizontal length or the vertical length of the sub-partition block does not satisfy the condition of being equal to or greater than 4 and less than or equal to 16, the transform kernel may be derived as DCT-2.

[0491] Additionally, the encoding device may configure the size of a non-cleared block to which the transform kernel is applied and indicating an area within the sub-partition block where a valid transform coefficient may exist, and the encoding device may derive the transform coefficient based on the size of such a non-cleared block.

[0492] Since each of the horizontal length and the vertical length of the sub-partition block is equal to or greater than 4 and less than or equal to 16, if DST-7 is used as the transform kernel, the width and height of the non-cleared block may be set to 16.

[0493] Furthermore, if each of the horizontal length and vertical length of the sub-partition block does not satisfy the condition of being equal to or greater than 4 and less than or equal to 16, and therefore if DCT-2 is used as the transform kernel, the width and height of the non-cleared block can be set to up to 32 or less.

[0494] That is, according to an example, when multi-transform selection is applied (when DST-7 or DCT-8 can be applied), the encoding device may set the width or height of the non-cleared block to 16. In addition, when multi-transform selection is not applied, the encoding device may set the width or height of the non-cleared block to 32 or less.

[0495] As described above, when the ISP is applied to the current block, the MTS index indicating the transform core may not be signaled.

[0496] Furthermore, the above-described LFNST may be applied to the current block, and in this case, clearing of the second area corresponding to the LFNST may be performed on the sub-partitioned block.

[0497] When LFNST is not performed, DCT-2 may be used as a transform kernel for one transform in a sub-partition block, and the width or height of a non-zero block may be set to 32 or less.

[0498] When performing LFNST, if LFNST is applied to transform coefficients that perform zeroing according to a primary transform, the second region may be processed again with zeroing after applying LFNST. Also in this case, DCT-2 may be used as a transform kernel for a primary transform in a sub-partition block, and the width or height of a non-zeroed block may be set to 32 or less.

[0499] The encoding apparatus 200 may determine whether to perform a secondary transform or an inseparable transform (more specifically, LFNST) on a transform coefficient of a current block, and may derive a modified transform coefficient by applying LFNST to the transform coefficient.

[0500] Unlike a single transform that transforms the coefficients of the transform target separately in the vertical or horizontal direction, LFNST is an inseparable transform that applies the transform without separating the coefficients in a specific direction. Such an inseparable transform may be a low-frequency inseparable transform that applies the forward transform only in the low-frequency region rather than the entire block region.

[0501] In addition, according to an example, the encoding device may clear the second area except the upper left first area of ​​the sub-partition block or the current block, such as Fig.14 and Fig.15 shown.

[0502] The encoding apparatus may induce a modified transform coefficient of the current block or induce a quantized transform coefficient by performing quantization based on the transform coefficient, and may encode image information including information about the quantized transform coefficient ( S2440 ).

[0503] That is, the encoding device may generate residual information including information about the quantized transform coefficients. The residual information may include the above-mentioned transform-related information / syntax elements. The encoding device may encode the image / video information including the residual information and output the encoded image / video information in the form of a bitstream.

[0504] More specifically, the encoding apparatus 200 may generate information about a quantized transform coefficient and encode the information about the generated quantized transform coefficient.

[0505] When the image information can apply LFNST, the image information may include a LFNST index indicating a LFNST matrix.

[0506] The syntax element of the LFNST index according to the present embodiment may indicate whether to apply (inverse) LFNST and any one LFNST matrix included in the LFNST set, and when the LFNST set includes two transform kernel matrices, the syntax element of the LFNST index may have three values.

[0507] According to an embodiment, when the partition tree structure of the current block is a dual tree type, a LFNST index may be encoded for each of the luma block and the chroma block.

[0508] According to an embodiment, the syntax element value of the transform index may be derived as 0, 1, and 2, 0 indicating that (inverse) LFNST is not applied to the current block, 1 indicating the first LFNST matrix in the LFNST matrix, and 2 indicating the second LFNST matrix in the LFNST matrix.

[0509] In the present disclosure, at least one of quantization / dequantization and / or transform / inverse transform may be omitted. When quantization / dequantization is omitted, the quantized transform coefficient may be referred to as a transform coefficient. When transform / inverse transform is omitted, the transform coefficient may be referred to as a coefficient or a residual coefficient, or may still be referred to as a transform coefficient for consistency of expression.

[0510] In addition, in the present disclosure, the quantized transform coefficient and the transform coefficient may be referred to as a transform coefficient and a scaled transform coefficient, respectively. In this case, the residual information may include information about the transform coefficient, and the information about the transform coefficient may be signaled by the residual coding syntax. The transform coefficient may be derived based on the residual information (or information about the transform coefficient), and the scaled transform coefficient may be derived by an inverse transform (scaling) of the transform coefficient. The residual sample may be derived based on an inverse transform (transform) of the scaled transform coefficient. These details may also be applied / expressed in other parts of the present disclosure.

[0511] In the above-mentioned embodiments, the method is explained based on the flowchart with the aid of a series of steps or boxes, but the present disclosure is not limited to the order of the steps, and a certain step may be performed in an order or step different from the above order or step, or a certain step may be performed concurrently with other steps. In addition, it can be understood by those of ordinary skill in the art that the steps shown in the flowchart are not exclusive, and another step may be incorporated or one or more steps in the flowchart may be deleted without affecting the scope of the present disclosure.

[0512] The above method according to the present disclosure may be implemented in a software form, and the encoding device and / or decoding device according to the present disclosure may be included in a device for image processing such as a television, a computer, a smart phone, a set-top box, a display device, and the like.

[0513] When the embodiments in the present disclosure are implemented by software, the above methods can be implemented as modules (steps, functions, etc.) for performing the above functions. These modules can be stored in a memory and can be executed by a processor. The memory can be inside or outside the processor and can be connected to the processor in various well-known ways. The processor may include an application-specific integrated circuit (ASIC), other chipsets, logic circuits and / or data processing devices. The memory may include a read-only memory (ROM), a random access memory (RAM), a flash memory, a memory card, a storage medium and / or other storage devices. In other words, the embodiments described in the present disclosure can be implemented and executed on a processor, a microprocessor, a controller or a chip. For example, the functional units shown in each of the accompanying drawings can be implemented and executed on a computer, a processor, a microprocessor, a controller or a chip.

[0514] In addition, the decoding device and encoding device of the present disclosure can be included in a multimedia broadcast transceiver, a mobile communication terminal, a home theater video device, a digital theater video device, a surveillance camera, a video chat device, a real-time communication device (such as video communication), a mobile streaming device, a storage medium, a camera, a video on demand (VoD) service provider, an over-the-top (OTT) video device, an Internet streaming service provider, a three-dimensional (3D) video device, a video phone video device, and a medical video device, and can be used to process a video signal or a data signal. For example, an over-the-top (OTT) video device can include a game console, a Blu-ray player, an Internet access TV, a home theater system, a smart phone, a tablet PC, a digital video recorder (DVR), etc.

[0515] In addition, the processing method of the present disclosure can be produced in the form of a program executed by a computer and can be stored in a computer-readable recording medium. Multimedia data with a data structure according to the present disclosure can also be stored in a computer-readable recording medium. Computer-readable recording media include various storage devices and distributed storage devices that store computer-readable data. Computer-readable recording media may include, for example, a Blu-ray disc (BD), a universal serial bus (USB), a ROM, a PROM, an EPROM, an EEPROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device. In addition, the computer-readable recording medium includes a medium implemented in the form of a carrier (e.g., transmission on the Internet). In addition, the bit stream generated by the encoding method can be stored in a computer-readable recording medium or transmitted via a wired or wireless communication network. In addition, the embodiments of the present disclosure can be implemented as a computer program product through a program code, and the program code can be executed on a computer according to the embodiments of the present disclosure. The program code can be stored on a computer-readable carrier.

[0516] The claims disclosed herein may be combined in various ways. For example, the technical features of the method claims of the present disclosure may be combined to be implemented or performed in a device, and the technical features of the device claims may be combined to be implemented or performed in a method. In addition, the technical features of the method claims and the device claims may be combined to be implemented or performed in a device, and the technical features of the method claims and the device claims may be combined to be implemented or performed in a method.

Claims

1. An image decoding method performed by a decoding device, the image decoding method comprising the following steps: Obtaining residual information and inseparable transform indices from a bitstream; deriving transform coefficients for a current block based on the residual information; Deriving residual samples for the current block by applying an inverse transform to the transform coefficients using a vertical transform kernel and a horizontal transform kernel; as well as Generate a reconstructed picture based on the residual samples, wherein, based on the current block being divided into sub-partition blocks as the intra-sub-partition ISP is applied and the inseparable transform index being equal to 0, the vertical transform kernel is determined by checking the vertical length of the sub-partition block divided from the current block, and the horizontal transform kernel is determined by checking the horizontal length of the sub-partition block, wherein, for the sub-partition block whose vertical length is checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DST-7, and for the sub-partition block whose vertical length is not checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DCT-2, wherein, for the sub-partition block whose horizontal length is checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DST-7, and for the sub-partition block whose horizontal length is not checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DCT-2, and Wherein, based on the current block being divided into the sub-partition blocks as the ISP is applied and the inseparable transform index being greater than 0, the vertical transform kernel and the horizontal transform kernel are respectively determined to be DCT-2.

2. An image encoding method performed by an encoding device, the image encoding method comprising the following steps: Derive residual samples for the current block; deriving transform coefficients by applying a transform to the residual samples using a vertical transform kernel and a horizontal transform kernel; as well as encoding residual information including information about the transform coefficients and an inseparable transform index, wherein, based on the current block being divided into sub-partition blocks as the intra-sub-partition ISP is applied and the inseparable transform index being equal to 0, the vertical transform kernel is determined by checking the vertical length of the sub-partition block divided from the current block, and the horizontal transform kernel is determined by checking the horizontal length of the sub-partition block, wherein, for the sub-partition block whose vertical length is checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DST-7, and for the sub-partition block whose vertical length is not checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DCT-2, wherein, for the sub-partition block whose horizontal length is checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DST-7, and for the sub-partition block whose horizontal length is not checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DCT-2, and Wherein, based on the current block being divided into the sub-partition blocks as the ISP is applied and the inseparable transform index being greater than 0, the vertical transform kernel and the horizontal transform kernel are respectively determined to be DCT-2.

3. A non-transitory computer-readable digital storage medium storing a bit stream generated by an image encoding method, the image encoding method comprising the following steps: Derive residual samples for the current block; deriving transform coefficients by applying a transform to the residual samples using a vertical transform kernel and a horizontal transform kernel; as well as encoding residual information including information about the transform coefficients and an inseparable transform index, wherein, based on the current block being divided into sub-partition blocks as the intra-sub-partition ISP is applied and the inseparable transform index being equal to 0, the vertical transform kernel is determined by checking the vertical length of the sub-partition block divided from the current block, and the horizontal transform kernel is determined by checking the horizontal length of the sub-partition block, wherein, for the sub-partition block whose vertical length is checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DST-7, and for the sub-partition block whose vertical length is not checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DCT-2, wherein, for the sub-partition block whose horizontal length is checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DST-7, and for the sub-partition block whose horizontal length is not checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DCT-2, and Wherein, based on the current block being divided into the sub-partition blocks as the ISP is applied and the inseparable transform index being greater than 0, the vertical transform kernel and the horizontal transform kernel are respectively determined to be DCT-2.

4. A method for sending image data, the sending method comprising the following steps: obtaining a bitstream for the image, wherein the bitstream is generated based on deriving residual samples for a current block, deriving transform coefficients by applying a transform to the residual samples using a vertical transform kernel and a horizontal transform kernel, and encoding residual information including information about the transform coefficients and an inseparable transform index; and sending said data comprising said bit stream, wherein, based on the current block being divided into sub-partition blocks as the intra-sub-partition ISP is applied and the inseparable transform index being equal to 0, the vertical transform kernel is determined by checking the vertical length of the sub-partition block divided from the current block, and the horizontal transform kernel is determined by checking the horizontal length of the sub-partition block, wherein, for the sub-partition block whose vertical length is checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DST-7, and for the sub-partition block whose vertical length is not checked to be equal to or greater than 4 and less than or equal to 16, the vertical direction transform kernel is determined to be DCT-2, wherein, for the sub-partition block whose horizontal length is checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DST-7, and for the sub-partition block whose horizontal length is not checked to be equal to or greater than 4 and less than or equal to 16, the horizontal direction transform kernel is determined to be DCT-2, and Wherein, based on the current block being divided into the sub-partition blocks as the ISP is applied and the inseparable transform index being greater than 0, the vertical transform kernel and the horizontal transform kernel are respectively determined to be DCT-2.

Images