Image encoding / decoding method, method for transmitting bitstream and recording medium having bitstream stored therein

Abstract

An image encoding / decoding method, a bit stream transmission method, and a computer-readable recording medium storing a bit stream are provided. An image decoding method according to the present disclosure comprises the steps of: acquiring a source picture timing information (SPTI) supplemental enhancement information (SEI) message; and reconstructing the image based on the SPTI SEI message, where the SPTI SEI message may include information about source picture timing for a picture associated with the SEI.

Description

Image encoding / decoding methods, methods for transmitting bit streams, and recording media containing the bit streams stored therein. Technical Field

[0001] This disclosure relates to an image encoding / decoding method, a method for transmitting a bit stream, and a recording medium having a bit stream stored therein, and more specifically, to an image encoding / decoding method, a method for transmitting a bit stream, and a recording medium having a bit stream stored therein, which are related to the timing of a source image. Background Technology

[0002] Recently, there has been an increasing demand for high-resolution and high-quality images, such as high-definition (HD) and ultra-high-definition (UHD) images, across various fields. As the resolution and quality of image data improve, the amount of information or bits transmitted increases relative to existing image data. This increase in the amount of information or bits transmitted leads to increased transmission and storage costs.

[0003] Therefore, there is a need for efficient image compression techniques to effectively send, store, and reproduce information about high-resolution and high-quality images. Summary of the Invention

[0004] Technical issues

[0005] The purpose of this disclosure is to provide an image encoding / decoding method and apparatus with improved encoding / decoding efficiency.

[0006] Another objective of this disclosure is to provide a method for encoding and processing SPTI SEI messages.

[0007] Another objective of this disclosure is to more clearly specify the source image timing associated with the SPTI SEI message.

[0008] Another objective of this disclosure is to clarify the meaning of image 0 used for timing the export of the source image.

[0009] Another objective of this disclosure is to reduce decoder errors and improve encoding quality and efficiency by clarifying the meaning of information related to the timing of image 0 and the source image.

[0010] Another objective of this disclosure is to improve coding efficiency by clarifying the constraints on image 0.

[0011] Another object of this disclosure is to provide a non-transitory computer-readable recording medium having a bitstream stored therein generated by an image encoding method according to this disclosure.

[0012] Another object of this disclosure is to provide a non-transitory computer-readable recording medium having a bitstream stored therein that is received and decoded by an image decoding apparatus according to this disclosure, the image decoding apparatus being used to reconstruct an image.

[0013] Another object of this disclosure is to provide a method for transmitting a bitstream generated by an image encoding method according to this disclosure.

[0014] The purposes to be achieved by this disclosure are not limited to those mentioned above, and other technical purposes not expressly stated will become apparent to those skilled in the art from the description provided below.

[0015] Technical solution

[0016] The image decoding method according to embodiments of the present disclosure may include: obtaining source image timing information (SPTI) supplemental enhancement information (SEI) messages, and reconstructing an image based on the SPTI SEI messages, wherein the SPTI SEI messages may include information about source image timing for an image related to the SEI.

[0017] The image encoding method according to embodiments of the present disclosure may include: determining a Source Picture Timing Information (SPTI) Supplemental Enhancement Information (SEI) message, and encoding a bitstream including the SPTI SEI message, wherein the SPTI SEI message may include information about the source picture timing for an image associated with the SEI.

[0018] Furthermore, according to this disclosure, a non-transitory computer-readable recording medium having a bitstream generated by an image encoding method stored therein can be provided.

[0019] Furthermore, according to this disclosure, a non-transitory computer-readable recording medium may be provided having a bit stream stored therein that is received and decoded by an image decoding device used in reconstructing an image.

[0020] Furthermore, according to this disclosure, a method for transmitting a bitstream generated by an image encoding method can be provided.

[0021] The features of this disclosure that have been briefly summarized above are merely exemplary aspects described in the following detailed description and are not intended to limit the scope of this disclosure.

[0022] Beneficial effects

[0023] According to this disclosure, an image encoding / decoding method and apparatus with improved encoding / decoding efficiency can be provided.

[0024] Furthermore, according to this disclosure, meaning can be conveyed more clearly and decoder errors can be reduced by modifying the semantics of information in SPTI SEI messages.

[0025] Furthermore, according to this disclosure, encoding quality and efficiency can be improved by clearing the source image timed in conjunction with the SPTI SEI message.

[0026] Furthermore, according to this disclosure, encoding quality and efficiency can be improved by more clearly specifying the meaning of image 0 used for source image timing.

[0027] Furthermore, according to this disclosure, a non-transitory computer-readable recording medium may be provided, wherein a bitstream generated by an image encoding method is stored therein.

[0028] Furthermore, according to this disclosure, a non-transitory computer-readable recording medium may be provided, wherein a bit stream is stored therein, which is received and decoded by an image decoding device used in reconstructing an image.

[0029] Furthermore, according to this disclosure, a method for transmitting a bitstream generated by an image encoding method can be provided.

[0030] The effects available in this disclosure are not limited to those described above, and other effects not mentioned will be readily apparent to those skilled in the art from the following description. Attached Figure Description

[0031] Figure 1 is a view schematically illustrating an example of a video coding system to which embodiments of the present disclosure are applicable.

[0032] Figure 2 is a view schematically illustrating an image encoding apparatus to which embodiments of the present disclosure are applicable.

[0033] Figure 3 is a view schematically illustrating an image decoding apparatus to which embodiments of the present disclosure are applicable.

[0034] Figure 4 is a diagram illustrating the interleaving method used to derive the luminance channel.

[0035] Figure 5 is a flowchart illustrating an image decoding method applicable according to an embodiment of the present disclosure.

[0036] Figure 6 is a flowchart illustrating an image encoding method applicable according to an embodiment of the present disclosure.

[0037] Figure 7 is an exemplary diagram illustrating a content streaming system applicable to an embodiment of the present disclosure. Detailed Implementation

[0038] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings to facilitate implementation by those skilled in the art. However, this disclosure can be implemented in various different forms and is not limited to the embodiments described herein.

[0039] In describing this disclosure, detailed descriptions of relevant known functions or constructions will be omitted if they unnecessarily obscure the scope of the disclosure. In the accompanying drawings, portions irrelevant to the description of this disclosure are omitted, and similar reference numerals are appended to similar portions.

[0040] In this disclosure, when a component is "connected," "coupled," or "linked" to another component, it may include not only direct connections but also indirect connections where intermediate components exist. Furthermore, when a component "comprises" or "has" other components, unless otherwise stated, it means that other components may be further included, rather than excluded.

[0041] In this disclosure, the terms first, second, etc., are used only for the purpose of distinguishing one component from other components and do not limit the order or importance of the components unless otherwise stated. Accordingly, within the scope of this disclosure, a first component in one embodiment may be referred to as a second component in another embodiment, and similarly, a second component in one embodiment may be referred to as a first component in another embodiment.

[0042] In this disclosure, the distinguishing components are intended to clearly describe each feature and do not imply that the components must be separate. That is, multiple components may be integrated and implemented in a single hardware or software unit, or a single component may be distributed and implemented in multiple hardware or software units. Therefore, unless otherwise specified, embodiments in which components are integrated or distributed are included within the scope of this disclosure.

[0043] In this disclosure, the components described in the various embodiments are not necessarily essential components, and some components may be optional. Therefore, embodiments comprising a subset of the components described in the embodiments are also included within the scope of this disclosure. Furthermore, embodiments that include other components besides those described in the various embodiments are also included within the scope of this disclosure.

[0044] This disclosure relates to the encoding and decoding of images, and unless redefined in this disclosure, the terms used herein may have the general meaning commonly used in the art to which this disclosure pertains.

[0045] In this disclosure, "picture" generally refers to the basis of an image within a specific time period, and a slice / tile is a coding unit that constitutes a part of a picture. A picture may consist of one or more slices / tiles. In addition, slices / tiles may include one or more coding tree units (CTUs).

[0046] In this disclosure, "pixel" or "cell" can refer to the smallest unit that constitutes a picture (or image). Additionally, "sample" can be used as a term corresponding to a pixel. A sample can generally represent a pixel or a pixel value, and can represent only the pixel / pixel value of the luminance component or only the pixel / pixel value of the chrominance component.

[0047] In this disclosure, "unit" can refer to a basic unit of image processing. A unit may include at least one of a specific region of an image and information associated with that region. In some cases, the term "unit" may be used interchangeably with terms such as "sample array," "block," or "area." Generally, an M×N block may include a set (or array) of samples (or sample arrays) with M columns and N rows, or a set (or array) of transform coefficients.

[0048] In this disclosure, "current block" can mean one of "current coding block," "current coding unit," "coding target block," "decoding target block," or "processing target block." When performing prediction, "current block" can mean "current prediction block" or "prediction target block." When performing transform (inverse transform) / quantization (dequantization), "current block" can mean "current transform block" or "transform target block." When performing filtering, "current block" can mean "filter target block."

[0049] Additionally, in this disclosure, unless explicitly stated as a chroma block, "current block" may mean a block that includes both luma component blocks and chroma component blocks, or "the luma block of the current block." The luma component block of the current block can be expressed by an explicit description including luma component blocks such as "luma block" or "current luma block." Similarly, "the chroma component block of the current block" can be expressed by an explicit description including chroma component blocks such as "chroma block" or "current chroma block."

[0050] In this disclosure, the terms “ / ” or “,” can be interpreted as indicating “and / or”. For example, the expressions “A / B” and “A, B” can mean “A and / or B”. Furthermore, “A / B / C” and “A, B, C” can mean “at least one of A, B and / or C”.

[0051] In this disclosure, the term "or" should be interpreted to indicate "and / or". For example, expressing "A or B" can include 1) only "A", 2) only "B", or 3) both "A and B". In other words, in this disclosure, the term "or" should be interpreted to indicate "additionally or alternatively".

[0052] Overview of Video Encoding Systems

[0053] Figure 1 is a view schematically illustrating an example of a video coding system to which embodiments of the present disclosure are applicable.

[0054] The video encoding system according to the embodiment may include an encoding device 10 and a decoding device 20. The encoding device 10 may deliver encoded video and / or image information or data to the decoding device 20 in the form of files or streams via a digital storage medium or network.

[0055] The encoding apparatus 10 according to an embodiment may include a video source generator 11, an encoding unit (also referred to as an encoder) 12, and a transmitter 13. The decoding apparatus 20 according to an embodiment may include a receiver 21, a decoding unit (also referred to as a decoder) 22, and a renderer 23. The encoder 12 may be referred to as a video / image encoding apparatus, and the decoding unit 22 may be referred to as a video / image decoding apparatus. The transmitter 13 may be included in the encoder 12. The receiver 21 may be included in the decoding unit 22. The renderer 23 may include a display, and the display may be configured as a separate device or an external component.

[0056] The video source generator 11 can acquire video / images through a process of capturing, compositing, or generating video / images. The video source generator 11 may include a video / image capture device and / or a video / image generation device. The video / image capture device may include, for example, one or more cameras, a video / image archive including previously captured video / images, etc. The video / image generation device may include, for example, a computer, tablet computer, and smartphone, and can generate video / images (electronically). For example, virtual video / images can be generated by a computer, etc. In this case, the video / image capture process can be replaced by a process of generating related data.

[0057] Encoder 12 can encode input video / images. For compression and encoding efficiency, encoder 12 can perform a series of processes such as prediction, transformation, and quantization. Encoder 12 is able to output encoded data (encoded video / image information) in the form of a bitstream.

[0058] Transmitter 13 can acquire encoded video / image information or data output as a bitstream and forward it to receiver 21 of decoding device 20 or another external object via digital storage medium or network as file or streaming data. Digital storage medium can include various storage media such as USB, SD, CD, DVD, Blu-ray, HDD, SSD, etc. Transmitter 13 can include elements for generating media files in a predetermined file format and may include elements for transmission via broadcast / communication networks. Transmitter 13 can be provided as a transmission device separate from encoding device 12, and in this case, the transmission device can include at least one processor that acquires encoded video / image information or data output as a bitstream; and a transmission unit for sending it as file or streaming data. Receiver 21 can extract / receive bitstreams from storage medium or network and send the bitstreams to decoding unit 22.

[0059] The decoding unit 22 can decode video / images by performing a series of processes such as dequantization, inverse transform, and prediction, which correspond to the operations of the encoder 12.

[0060] Renderer 23 can render decoded video / images. The rendered video / images can be displayed on a monitor.

[0061] Overview of image encoding devices

[0062] Figure 2 is a view schematically illustrating an image encoding apparatus to which embodiments of the present disclosure may be applied.

[0063] As shown in Figure 2, the image encoding device 100 may include an image partitioner 110, a subtractor 115, a transformer 120, a quantizer 130, a dequantizer 140, an inverse transformer 150, an adder 155, a filter 160, a memory 170, an inter-frame prediction unit 180, an intra-frame prediction unit 185, and an entropy encoder 190. The inter-frame prediction unit 180 and the intra-frame prediction unit 185 may be collectively referred to as "prediction units". The transformer 120, quantizer 130, dequantizer 140, and inverse transformer 150 may be included in a residual processor. The residual processor may further include a subtractor 115.

[0064] In some embodiments, all or at least some of the components of the image encoding apparatus 100 may be configured by a single hardware component (e.g., an encoder or a processor). Additionally, the memory 170 may include a decoded image buffer (DPB) and may be configured by a digital storage medium.

[0065] Image partitioner 110 can partition an input image (or picture or frame) input to image encoding device 100 into one or more processing units. For example, a processing unit may be referred to as a coding unit (CU). A coding unit can be obtained by recursively partitioning a coding tree unit (CTU) or a maximum coding unit (LCU) according to a quadtree / binary tree / tritree (QT / BT / TT) structure. For example, a coding unit can be partitioned into multiple coding units of greater depth based on a quadtree structure, a binary tree structure, and / or a ternary tree structure. For the partitioning of a coding unit, a quadtree structure can be applied first, and a binary tree structure and / or a ternary tree structure can be applied later. The encoding process according to embodiments of the present disclosure can be performed based on the final coding unit that is no longer partitioned. The maximum coding unit can be used as the final coding unit, or a deeper coding unit obtained by partitioning the maximum coding unit can be used as the final coding unit. Here, the encoding process may include prediction, transformation, and reconstruction processes. As another example, the processing unit of the encoding process may be a prediction unit (PU) or a transformation unit (TU). Prediction and transform units can be segmented or partitioned from the final coding unit. The prediction unit can be a unit for predicting samples, and the transform unit can be a unit for deriving transform coefficients and / or a unit for deriving residual signals from transform coefficients.

[0066] The prediction unit (inter-frame prediction unit 180 or intra-frame prediction unit 185) can perform prediction on the block to be processed (the current block) and generate a prediction block including prediction samples for the current block. The prediction unit can determine whether to apply intra-frame prediction or inter-frame prediction based on the current block or CU. The prediction unit can generate various information related to the prediction of the current block and send the generated information to the entropy encoder 190. The information about the prediction can be encoded in the entropy encoder 190 and output as a bitstream.

[0067] Intra-prediction unit (intra-predictor) 185 can predict the current block by referencing samples in the current image. Depending on the intra-prediction mode and / or intra-prediction technique, the reference samples may be located among the neighbors of the current block or may be placed separately. Intra-prediction modes may include multiple non-directional modes and multiple directional modes. Non-directional modes may include, for example, DC mode and planar mode. Depending on the level of detail in the prediction direction, directional modes may include, for example, 33 or 65 directional prediction modes. However, this is merely an example, and more or fewer directional prediction modes may be used depending on the settings. Intra-prediction unit 185 can determine the prediction mode to be applied to the current block by using prediction modes applied to neighboring blocks.

[0068] The inter-frame prediction unit (inter-frame predictor) 180 can derive a prediction block for the current block based on a reference block (reference sample array) specified by motion vectors on a reference image. In this case, to reduce the amount of motion information transmitted in the inter-frame prediction mode, motion information can be predicted on a block, sub-block, or sample basis based on the correlation between motion information between neighboring blocks and the current block. Motion information may include motion vectors and reference image indices. Motion information may further include inter-frame prediction direction (L0 prediction, L1 prediction, Bi prediction, etc.) information. In the case of inter-frame prediction, neighboring blocks may include spatially neighboring blocks existing in the current image and temporally neighboring blocks existing in the reference image. The reference image including the reference block and the reference image including the temporally neighboring block may be the same or different. The temporally neighboring block may be referred to as a juxtaposed reference block, a juxtaposed CU (colCU), etc. The reference image including the temporally neighboring block may be referred to as a juxtaposed image (colPic). For example, the inter-frame prediction unit 180 can construct a motion information candidate list based on neighboring blocks and generate information indicating which candidate to use to derive the motion vector and / or reference image index of the current block. Inter-frame prediction can be performed based on various prediction modes. For example, in skip mode and merge mode, the inter-frame prediction unit 180 can use motion information from neighboring blocks as motion information for the current block. In skip mode, unlike merge mode, residual signals may not be transmitted. In motion vector prediction (MVP) mode, motion vectors from neighboring blocks can be used as motion vector predictors, and the motion vector of the current block can be transmitted by signaling an encoded motion vector difference and an indicator for the motion vector predictor. The motion vector difference can refer to the difference between the motion vector of the current block and the motion vector predictor.

[0069] The prediction unit can generate a prediction signal based on various prediction methods and techniques described below. For example, the prediction unit can apply not only intra-frame prediction or inter-frame prediction, but also both intra-frame prediction and inter-frame prediction simultaneously to predict the current block. A prediction method that simultaneously applies intra-frame prediction and inter-frame prediction to predict the current block can be called combined intra-frame and inter-frame prediction (CIIP). Additionally, the prediction unit can perform intra-block copying (IBC) on the prediction of the current block. Intra-block copying can be used for content image / video coding in games, such as Screen Content Coding (SCC). IBC is a method of predicting the current image using a previously reconstructed reference block in the current image at a predetermined distance from the current block. When IBC is applied, the position of the reference block in the current image can be encoded as a vector (block vector) corresponding to the predetermined distance. IBC essentially performs prediction in the current image, but can be performed similarly to inter-frame prediction because the reference block is derived within the current image. That is, IBC can use at least one of the inter-frame prediction techniques described in this disclosure.

[0070] The prediction signal generated by the prediction unit can be used to generate a reconstructed signal or a residual signal. Subtractor 115 generates a residual signal (residual block or residual sample array) by subtracting the prediction signal (prediction block or prediction sample array) output from the prediction unit from the input image signal (original block or original sample array). The generated residual signal can be sent to converter 120.

[0071] Transformer 120 can generate transform coefficients by applying a transform technique to the residual signal. For example, the transform technique may include at least one of Discrete Cosine Transform (DCT), Discrete Sine Transform (DST), Karhunen-Loève Transform (KLT), Graphical Based Transform (GBT), or Conditional Nonlinear Transform (CNT). Here, GBT refers to a transform obtained from a graph when the relationship information between pixels is represented graphically. CNT refers to a transform obtained based on a prediction signal generated using all previously reconstructed pixels. Furthermore, the transform process can be applied to square pixel blocks of the same size or to blocks of variable size instead of square.

[0072] Quantizer 130 can quantize the transform coefficients and send them to entropy encoder 190. Entropy encoder 190 can encode the quantized signal (information about the quantized transform coefficients) and output a bitstream. The information about the quantized transform coefficients can be referred to as residual information. Quantizer 130 can rearrange the block-type quantized transform coefficients into a one-dimensional vector form based on the coefficient scan order, and generate information about the quantized transform coefficients based on the one-dimensional vector form of the quantized transform coefficients.

[0073] The entropy encoder 190 can 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 190 can encode, together or separately, the information necessary for video / image reconstruction, excluding quantized transform coefficients (e.g., values ​​of syntax elements, etc.). The encoded information (e.g., encoded video / image information) can be transmitted or stored in the form of a bitstream at the Network Abstraction Layer (NAL) level. The video / image information may further include information about various parameter sets, such as adaptive parameter sets (APS), picture parameter sets (PPS), sequence parameter sets (SPS), or video parameter sets (VPS). Additionally, the video / image information may further include general constraint information. The information transmitted by signaling, the transmitted information, and / or syntax elements described in this disclosure can be encoded and included in the bitstream through the above encoding process.

[0074] The bitstream can be transmitted over a network or stored in a digital storage medium. The network may include broadcast networks and / or communication networks, 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) for transmitting the signal output from the entropy encoder 190 and / or a storage unit (not shown) for storing the signal may be included as internal / external components of the image encoding apparatus 100. Alternatively, a transmitter may be provided as a component of the entropy encoder 190.

[0075] The quantized transform coefficients output from quantizer 130 can be used to generate residual signals. For example, the residual signal (residual block or residual sample) can be reconstructed by applying dequantization and inverse transform to the quantized transform coefficients via dequantizer 140 and inverse transformer 150.

[0076] Adder 155 adds the reconstructed residual signal to the prediction signal output from inter-frame prediction unit 180 or intra-frame prediction unit 185 to generate a reconstructed signal (reconstructed image, reconstructed block, reconstructed sample array). If there is no residual for the block to be processed, such as in the case of applying a skip mode, the prediction block can be used as the reconstructed block. Adder 155 may be referred to as a reconstructor or reconstructed block generator. The generated reconstructed signal can be used for intra-frame prediction of the next block to be processed in the current image, and can be used for inter-frame prediction of the next image by filtering as described below.

[0077] Filter 160 can improve subjective / objective image quality by applying filtering to the reconstructed signal. For example, filter 160 can generate a modified reconstructed image by applying various filtering methods to the reconstructed image and store the modified reconstructed image in memory 170, specifically in the DPB of memory 170. Various filtering methods may include, for example, deblocking filtering, sample adaptive shifting, adaptive loop filtering, bilateral filtering, etc. Filter 160 can generate various filtering-related information and send the generated information to entropy encoder 190, as described later in the description of each filtering method. The filtering-related information can be encoded by entropy encoder 190 and output as a bitstream.

[0078] The modified reconstructed image sent to memory 170 can be used as a reference image in inter-frame predictor 180. When inter-frame prediction is applied by image coding device 100, prediction mismatch between image coding device 100 and image decoding device can be avoided and coding efficiency can be improved.

[0079] The DPB of memory 170 can store modified reconstructed images for use as reference images in inter-frame predictor 180. Memory 170 can store motion information of blocks from which motion information in the current image is derived (or encoded) and / or motion information of already reconstructed blocks in the image. The stored motion information can be sent to inter-frame predictor 180 and used as motion information for spatially or temporally neighboring blocks. Memory 170 can store reconstructed samples of reconstructed blocks in the current image and can transmit the reconstructed samples to intra-frame predictor 185.

[0080] Overview of an image decoding device

[0081] Figure 3 is a view schematically illustrating an image decoding apparatus to which embodiments of the present disclosure may be applied.

[0082] As shown in Figure 3, the image decoding device 200 may include an entropy decoder 210, a dequantizer 220, an inverse transformer 230, an adder 235, a filter 240, a memory 250, an inter-frame predictor 260, and an intra-frame predictor 265. The inter-frame predictor 260 and the intra-frame predictor 265 may be collectively referred to as "predictors". The dequantizer 220 and the inverse transformer 230 may be included in a residual processor.

[0083] According to an embodiment, all or at least some of the components of the image decoding device 200 may be configured by hardware components (e.g., a decoder or a processor). Additionally, the memory 250 may include a decoded image buffer (DPB) or may be configured by a digital storage medium.

[0084] The image decoding device 200, having received a bitstream including video / image information, can reconstruct the image by executing a process corresponding to that performed by the image encoding device 100 of FIG. 2. For example, the image decoding device 200 can perform decoding using a processing unit applied in the image encoding device. Therefore, the decoding processing unit can be, for example, an encoding unit. The encoding unit can be obtained through a partitioned coding tree unit or a maximum coding unit. The reconstructed image signal decoded and output by the image decoding device 200 can be reproduced by a reproduction device (not shown).

[0085] Image decoding apparatus 200 can receive signals output from the image encoding apparatus of FIG2 in the form of a bitstream. The received signals can be decoded by entropy decoder 210. For example, entropy decoder 210 can parse the bitstream to derive information (e.g., video / image information) necessary for image reconstruction (or picture reconstruction). The video / image information may further include information about various parameter sets, such as adaptive parameter sets (APS), picture parameter sets (PPS), sequence parameter sets (SPS), or video parameter sets (VPS). In addition, the video / image information may further include general constraint information. The image decoding apparatus can further decode the picture based on the information about the parameter sets and / or general constraint information. The information and / or syntax elements transmitted / received by signals described in this disclosure can be decoded and obtained from the bitstream through a decoding process. For example, entropy decoder 210 decodes the information in the bitstream based on encoding methods such as exponential Golomb coding, CAVLC, or CABAC, and outputs the values ​​of syntax elements necessary for image reconstruction and quantized values ​​of transform coefficients for the residuals. More specifically, the CABAC entropy decoding method can receive bins corresponding to each syntax element in the bitstream, determine a context model using information about the target syntax element, decoding information of neighboring blocks and the target block, or information about symbols / bins decoded in previous stages, and perform arithmetic decoding on the bins based on the determined context model by predicting the occurrence probability of the bins, generating symbols corresponding to the value of each syntax element. In this case, the CABAC entropy decoding method can update the context model after determining the context model by using the information of the decoded symbols / bins for the context model of the next symbol / bin. The prediction-related information in the information decoded by the entropy decoder 210 can be provided to the predictors (inter-frame predictor 260 and intra-frame predictor 265), and the residual values ​​of the entropy decoding performed in the entropy decoder 210, i.e., the quantized transform coefficients and related parameter information, can be input to the dequantizer 220. Additionally, the filtering information in the information decoded by the entropy decoder 210 can be provided to the filter 240. The receiver (not shown) for receiving signals output from the image encoding device may be further configured as an internal / external element of the image decoding device 200, or the receiver may be a component of the entropy decoder 210.

[0086] The image decoding apparatus according to this disclosure can be referred to as a video / image / picture decoding apparatus. The image decoding apparatus can be classified as an information decoder (video / image / picture information decoder) and a sample decoder (video / image / picture sample decoder). The information decoder may include an entropy decoder 210. The sample decoder may include at least one of a dequantizer 220, an inverse transformer 230, an adder 235, a filter 240, a memory 250, an inter-frame predictor 260, or an intra-frame predictor 265.

[0087] The dequantizer 220 can dequantize the quantized transform coefficients and output the transform coefficients. The dequantizer 220 can rearrange the quantized transform coefficients in the form of two-dimensional blocks. In this case, the rearrangement can be performed based on the coefficient scan order performed in the image encoding device. The dequantizer 220 can obtain the transform coefficients by performing dequantization on the quantized transform coefficients using quantization parameters (e.g., quantization step size information).

[0088] The inverse transformer 230 can perform inverse transformation on the transformation coefficients to obtain the residual signal (residual block, residual sample array).

[0089] The predictor can perform predictions on the current block and generate a prediction block that includes prediction samples for the current block. The predictor can determine whether to apply intra-frame prediction or inter-frame prediction to the current block based on information about the predictions output from the entropy decoder 210, and can determine a specific intra-frame / inter-frame prediction mode (prediction technique).

[0090] Similar to the predictor described for the image coding device 100, the predictor can generate a prediction signal based on various prediction methods (techniques) described later.

[0091] Intra-predictor 265 can predict the current block by referring to samples in the current image. The description of intra-predictor 185 is also applied to intra-predictor 265.

[0092] Inter-frame predictor 260 can derive a prediction block for the current block based on a reference block (reference sample array) specified by motion vectors on a reference image. In this case, to reduce the amount of motion information transmitted in the inter-frame prediction mode, motion information can be predicted on a block, sub-block, or sample basis based on the correlation between motion information between neighboring blocks and the current block. Motion information may include motion vectors and reference image indices. Motion information may further include inter-frame prediction direction (L0 prediction, L1 prediction, Bi prediction, etc.) information. In the case of inter-frame prediction, neighboring blocks may include spatially neighboring blocks present in the current image and temporally neighboring blocks present in the reference image. For example, inter-frame predictor 260 can configure a motion information candidate list based on neighboring blocks and derive the motion vector and / or reference image index for the current block based on received candidate selection information. Inter-frame prediction can be performed based on various prediction modes, and information about the prediction may include information indicating the inter-frame prediction mode used for the current block.

[0093] Adder 235 generates a reconstructed block by adding the obtained residual signal to the prediction signal (prediction block, prediction sample array) output from the predictor (including inter-frame predictor 260 and / or intra-frame predictor 265). If no residual exists for the block to be processed, such as when a skip mode is applied, the prediction block can be used as the reconstructed block. The description of adder 155 also applies to adder 235. Adder 235 may be referred to as a reconstructor or reconstructed block generator. The generated reconstructed signal can be used for intra-frame prediction of the next block to be processed in the current image, and can be used for inter-frame prediction of the next image by filtering as described below.

[0094] Filter 240 can improve subjective / objective image quality by applying filtering to the reconstructed signal. For example, filter 240 can generate a modified reconstructed image by applying various filtering methods to the reconstructed image, and store the modified reconstructed image in memory 250, specifically in the DPB of memory 250. Various filtering methods may include, for example, deblocking filtering, adaptive sample shifting, adaptive loop filtering, bilateral filtering, etc.

[0095] The (modified) reconstructed image stored in the DPB of memory 250 can be used as a reference image in inter-frame predictor 260. Memory 250 can store motion information of blocks from which motion information in the current image is derived (or decoded) and / or motion information of already reconstructed blocks in the image. The stored motion information can be sent to inter-frame predictor 260 to be utilized as motion information of spatially or temporally neighboring blocks. Memory 250 can store reconstructed samples of reconstructed blocks in the current image and transmit the reconstructed samples to intra-frame predictor 265.

[0096] In this disclosure, the embodiments described with respect to the filter 160, inter-frame predictor 180 and intra-frame predictor 185 of the image coding apparatus 100 can be equally or correspondingly applied to the filter 240, inter-frame predictor 260 and intra-frame predictor 265 of the image decoding apparatus 200.

[0097] General post-processing filtering procedure using NNPF

[0098] The input to the post-processing filtering procedure can be a bitstream, BitstreamToFilter. The output of this procedure can be a list of NNPF output images, ListNnpfOutputPics. First, BitstreamToFilter can be decoded, and the list CroppedDecodedPictures can be set as a list of cropped decoded images produced by decoding BitstreamToFilter in output order. Next, the post-processing filtering procedure for an image can be repeatedly called in output order for each cropped decoded image in CroppedDecodedPictures, and one or more NNPFs can be activated for the corresponding image. The order of the images in ListNnpfOutputPics can be the output order. Within ListNnpfOutputPics, there should be no more than one image for any particular output time instance. When multiple NNPFs are activated for any particular image in CroppedDecodedPictures, and only one NNPF can be selected for application (any NNPF can be selected), the above constraints should apply regardless of which NNPF is selected for the particular image.

[0099] Post-processing filtering can be applied to each cropped decoded image, referred to as the current image, within `CroppedDecodedPictures`, and can activate one or more NNPFs. When an NNPF is applied to the current image, the filtered and / or interpolated images can be generated by the NNPF through the NNPF procedure specified in the semantics of the `NNPFC` SEI message. When an NNPF is applied to the current image, the images generated by the NNPF through the applied NNPF procedure are stored in the NNPF output tensor in the output order. When the applied NNPF is the last NNPF applied to the current image, the images generated by the NNPF and output by the NNPF procedure can be included in `ListNnpfOutputPics` in the same order as when the images are stored in the NNPF output tensor.

[0100] Post-filter characteristics of neural networks (NNPFC)

[0101] The combination of Tables 1 to 3 shows the NNPFC syntax structure.

[0102] [Table 1]

[0103] [Table 2]

[0104] [Table 3]

[0105] The NNPFC syntax structures in Tables 1 to 3 can be signaled as Supplemental Enhancement Information (SEI) messages. SEI messages that signal the NNPFC syntax structures in Tables 1 to 3 can be referred to as NNPFC SEI messages.

[0106] The NNPFC SEI message can specify the neural network used as a post-processing filter. The Neural Network Post-Filter Activation (NNPFA) SEI message can be used to indicate the use of a specific neural network post-processing filter for a particular image. Here, the terms "post-processing filter" and "post-filter" can have the same meaning.

[0107] Using these SEI messages requires defining the following variables: - The ability to crop the width and height of the input image in units of luminance samples. Width and height can be represented here by CroppedWidth and CroppedHeight, respectively.

[0108] - The luminance sample array CroppedYPic[idx] and chrominance sample arrays CroppedCbPic[idx] and CroppedCrPic[idx] (if present) of the input image can be used as input to NNPF. Here, the index idx can range from 0 to numInputPics-1.

[0109] - BitDepth Y It can indicate the bit depth of the luminance sample array used for the input image.

[0110] - BitDepth C You can indicate the bit depth of the chroma sample array (if any) used for the input image.

[0111] - ChromaFormatIdc can indicate a chroma format identifier.

[0112] - When nnpfc_auxiliary_inp_idc equals 1, for the input image, the filter strength control value array StrengthControlVal[idx] should contain real numbers in the range of 0 to 1, where the index idx can be in the range of 0 to numInputPics-1.

[0113] An input image with index 0 can correspond to an image whose NNPF is activated by an NNPFC SEI message via an NNPFA SEI message. An input image with index i in the range of 1 to numInputPics-1 (inclusive) can be output before an input image with index i-1.

[0114] `nnpfc_purpose` indicates the purpose of the NNPF as specified in Table 20, where `(nnpfc_purpose & bitMask)` not equal to 0 indicates that the NNPF has a purpose associated with the value of `bitMask` in Table 4. When `nnpfc_purpose` is greater than 0 and `(nnpfc_purpose & bitMask)` equals 0, the purpose associated with the `bitMask` value is not applicable to the NNPF. When `nnpfc_pupose` equals 0, the NNPF can be used as determined by the application. The value of `nnpfc_purpose` can be constrained to be in the range of 0 to 63 in the bitstream. Values ​​of 64 to 65535 used for `nnpfc_purpose` can be reserved for future use. The decoder should ignore NNPFC SEI messages with `nnpfc_purpose` in the range of 64 to 65535.

[0115] [Table 4]

[0116] The variables chromaUpsamplingFlag (indicating whether the purpose of NNPF, as indicated by nnpfc_purpose, is chroma upsampling), resolutionResamplingFlag (indicating whether the purpose of NNPF, as indicated by nnpfc_purpose, is resolution resampling), pictureRateUpsamplingFlag (indicating whether the purpose of NNPF, as indicated by nnpfc_purpose, is image frame rate upsampling), bitDepthUpsamplingFlag (indicating whether the purpose of NNPF, as indicated by nnpfc_purpose, is bit depth upsampling), and colourizationFlag (indicating whether the purpose of NNPF, as indicated by nnpfc_purpose, is colorization) can be derived as shown in Table 5 below.

[0117] [Table 5]

[0118] When ChromaFormatIdc equals 3, chromaUpsampingFlag can be constrained to 0. When ChromaFormatIdc or ChromaUpsampingFlag is not equal to 0, colourizationFlag can be constrained to 0. When pictureRateUpsampingFlag equals 1 and an input picture with index 0 is associated with a frame packing arrangement SEI message with fp_arrangement_type equal to 5, all input pictures can be associated with a frame packing arrangement SEI message with the same value of fp_arrangement_type equal to 5 and fp_current_frame_is_frame0_flag.

[0119] nnpfc_id can contain an identifier that can be used to identify NNPF. The value of nnpfc_id should be between 0 and 2. 32 - Within the range of 2. Between 256 and 511 and 2 31 to 2 32 Values ​​of nnpfc_id in the range of -2 can be reserved for future use. The decoder should ignore values ​​between 256 and 511 or 2. 31 to 2 32 NNPFC SEI messages with nnpfc_id in the range of -2.

[0120] When the NNPFC SEI message is the first NNPFC SEI message in decoding order and has a special value of nnpfc_id within the current CLVS, the following can be applied: - The SEI message can indicate the underlying NNPF.

[0121] - SEI messages can be about the current decoded image of the current layer and all subsequent decoded images in output order, until the current CLVS ends.

[0122] An NNPFC SEI message can be a repetition of a previous NNPFC SEI message in the current CLVS in decoding order, and subsequent semantics can be applied as if this SEI message were the only NNPFC SEI message with the same content in the current CLVS.

[0123] Setting nnpfc_base_flag to 1 allows you to specify the underlying NNPF for the SEI message. Setting nnpf_base_flag to 0 allows you to specify updates related to the underlying NNPF for the SEI message.

[0124] The following constraints can be applied to the value of nnpfc_base_flag.

[0125] - When the NNPFC SEI message is the first NNPFC SEI message in the current CLVS with a special value of nnpfc_id in the decoding order, the value of nnpfc_base_flag should be equal to 1.

[0126] - When NNPFC SEI message nnpfcB is not the first NNPFC SEI message in the current CLVS with a specific value of nnpfc_id in decoding order and a value of nnpfc_base_flag equal to 1, the NNPFC SEI message can correspond to a duplicate of the first NNPFC SEI message nnpfcA with the same value of nnpfc_id in decoding order. That is, the payload content of nnpfcB should be the same as the payload content of nnpfcA.

[0127] When nnpfc_base_flag equals 0, the following constraints apply: - This SEI message can define updates associated with a previous base NNPF in decoding order, having the same value as nnpfc_id. Updates are not cumulative, but rather each update can be applied to a base NNPF, which is the NNPF specified in decoding order by the first NNPFC SEI message, having a specific value of nnpfc_id within the current CLVS. The NNPF defined by this SEI message can be obtained by applying updates associated with the base NNPF having the same value as nnpfc_id, as defined by this SEI message.

[0128] This SEI message can be related to the current decoded image of the current layer and all subsequent decoded images in output order, up to the end of the current CLVS, excluding decoded images that follow the current decoded image in output order within the current CLVS. This SEI message can be associated with subsequent NNPFC SEI messages in decoding order that have nnpfc_base_flag equal to 0 and a specific value of nnpfc_id within the current CLVS (whichever is earlier).

[0129] A `nnpfc_mode_idc` value of 0 indicates that the SEI message contains a bitstream that can specify the underlying NNPF (when `nnpfc_base_flag` equals 1), or an update relative to a underlying NNPF with the same value of `nnpfc_id` (when `nnpfc_base_flag` equals 0). When `nnpfc_base_flag` equals 1, `nnpfc_mode_idc` equals 1 can specify that the underlying NNPF associated with the value of `nnpfc_id` is a neural network identified by a URI, which can be indicated by an `nnpfc_uri` in the format identified by the tag URI `nnpfc_tag_uri`. When `nnpfc_base_flag` equals 0, `nnpfc_mode_idc` equals 1 can specify that the update associated with the underlying NNPF with the same value of `nnpfc_id` is defined by a URI, which can be indicated by an `nnpfc_uri` in the format identified by the tag URI `nnpfc_tag_uri`.

[0130] In the bitstream, the value of nnpfc_mode_idc can be constrained to the range of 0 to 1. Values ​​of 2 to 255 for nnpfc_mode_idc can be reserved for future use and may not exist in the bitstream. The decoder should ignore NNPFC SEI messages with nnPFC_mode_idc values ​​in the range of 2 to 255. Values ​​of nnpfc_mode_idc greater than 255 may not exist in the bitstream and may not be reserved for future use.

[0131] Based on the constraints of the bitstream, nnpfc_reserved_zero_bit_a can be constrained to be equal to 0. The decoder can be constrained to ignore NNPFC SEI messages in which nnpfc_reserved_zero_bit_a is not equal to 0.

[0132] The nnpfc_tag_uri can contain a tag URI that identifies the neural network used as the underlying NNPF, with syntax and semantics as specified in IETF RFC 4151, or an update associated with the underlying NNPF having a value of nnpfc_id specified by the nnpfc_uri. The nnpfc_tag_uri can uniquely identify neural network data in a format specified by the nnrpf_uri without requiring a central registry. An nnpfc_tag_uri equal to “tag:iso.org,2023:15938-17” indicates that the neural network data identified by the nnpfc_uri conforms to ISO / IEC 15938-17.

[0133] nnpfc_uri can contain a URI that identifies the neural network used as the underlying NNPF and has the syntax and semantics specified in IETF Internet Standard 66, or an update associated with the underlying NNPF that has the same value as nnpfc_id.

[0134] A value of 1 for `nnpfc_property_present_flag` indicates the presence of syntax elements related to filter purpose, input formatting, output formatting, and complexity. A value of 0 for `nnpfc_property_present_flag` indicates the absence of syntax elements related to filter purpose, input formatting, output formatting, and complexity. When `nnpfc_base_flag` is 1, `nnpfc_property_present_flag` can be constrained to be 1. When `nnpfc_property_present_flag` is 0, the values ​​of all syntax elements that could exist only when `nnpfc_property_present_flag` is 1 can be inferred to be equal to the values ​​of their respective syntax elements in the NNPFCSEI message that may include this SEI message to provide updates to the underlying NNPF.

[0135] The following constraints may apply when the NNPFC SEI message nnpfcCurr is not the first NNPFC SEI message with a special value of nnpfc_id in the current CLVS in the decoding order and is not a duplicate of the first NNPFC SEI message with a special value of nnpfc_id (i.e., the value of nnpfc_base_flag is equal to 0), and nnpfc_property_present_flag is equal to 1.

[0136] The value of nnpfc_purpose in the NNPFC SEI message should be equal to the value of nnpfc_purpose in the first NNPFC SEI message that has the special value of nnpfc_id in the current CLVS in the decoding order.

[0137] The values ​​of the syntax elements nnpfc_base_flag in the NNPFC SEI message and the previous nnpfc_complex_info_present_flag should be the same as the values ​​of the corresponding syntax elements in the first NNPFC SEI message that have the special value of nnpfc_id in the current CLVS in the order of decoding.

[0138] - In the first NNPFC SEI message in the current CLVS with the special value of nnpfc_id in the decoding order, nnpfc_complexity_info_present_flag should be equal to 0 or nnpfc_complexity_info_present_flag should be equal to 0 or 1, and all of the following applies: (1) nnpfc_parameter_type_idc in nnpfcCurr should be equal to nnpfc_parameter_type_idc in nnpfcBase.

[0139] (2) nnpfc_log2_parameter_bit_length_minus3 in nnpfcCurr (if it exists) should be less than or equal to nnpfc_log2_parameter_bit_length_minus3 in nnpfcBase.

[0140] (3) If nnpfc_num_parameters_idc in nnpfcBase is equal to 0, then nnpfc_num_parameters_idc in nnpfcCurr should be equal to 0.

[0141] (4) Otherwise (if nnpfc_num_parameters_idc in nnpfcBase is greater than 0), nnpfc_num_parameters_idc in nnpfcCurr should be greater than 0 and less than or equal to nnpfc_num_parameters_idc in nnpfcBase.

[0142] (5) If nnpfc_num_kmac_operations_idc in nnpfcBase is equal to 0, then nnpfc_num_kmac_operations_idc in nnpfcCurr should be equal to 0.

[0143] (6) Otherwise (if nnpfc_num_kmac_operations_idc in nnpfcBase is greater than 0), nnpfc_num_kmac_operations_idc in nnpfcCurr should be greater than 0 and less than or equal to nnpfc_num_kmac_operations_idc in nnpfcBase.

[0144] (7) If nnpfc_total_ki0lobyte_size in nnpfcBase is equal to 0, then nnpfc_total_kikbyte_size in nnpfcCurr should be equal to 0.

[0145] (8) Otherwise (if nnpfc_total_kkibyte_size in nnpfcBase is greater than 0), nnpfc_total_kkibyte_size in nnpfcCurr should be greater than 0 and less than or equal to nnpfc_total_kkibyte_size in nnpfcBase.

[0146] `nnpfc_num_input_pics_minus1+1` indicates the number of decoded output images used as input to NNPF. The value of `nnpfc_num_input_pics_minus1` can be constrained to the range of 0 to 63. When `pictureRateUpsampingFlag` equals 1, `nnpfc_num_input_pics_minus1` can be constrained to be greater than 0.

[0147] The variable numInputPics, which specifies the number of images used as input to NNPF, can be derived as shown in Equation 1.

[0148] [Equation 1]

[0149] numInputPics=nnpfc_num_input_pics_minus1+1

[0150] `nnpfc_input_pic_output_flag[i]` equal to 1 indicates that NNPF generates the corresponding output image for the i-th input image. `nnpfc_input_pic_output_flag[i]` equal to 0 indicates that NNPF does not generate the corresponding output image for the i-th input image. When `nnpfc_num_input_pics_minus1` equals 0, `nnpfc_input_pic_output_flag[0]` can be inferred to be equal to 1. When `pictureRateUpsampingFlag` equals 0 and `nnpfc_num_input_pics_minus1` is greater than 0, for at least one value of `i` in the range of 0 to `nnpfc_num_input_pics_minus1`, `nnpfc_input_pic_output_flag[i]` can be constrained to be equal to 1. `nnpfc_input_pic_output_flag[i]` can be referred to as `nnpfc_input_pic_filtering_flag[i]`.

[0151] Setting `nnpfc_absent_input_pic_zero_flag` to 1 indicates that input images not present in the NNPF expected bitstream will be represented by an array of samples with a value of 0. Setting `nnpfc_absent_input_pic_flag` to 0 also indicates that input images not present in the NNPF expected bitstream will be represented by the closest input image in the bitstream in output order.

[0152] When `chromaUpsampingFlag` equals 0, `nnpfc_out_sub_c_flag` can specify the values ​​of variables `outSubWidthC` and `outSubHeightC`. `nnpfc_out_sub_c_flag` equal to 1 specifies that both `outSubWidthC` and `outSubHeightC` are 1. `nnpfc_out_sub_c_flag` equal to 0 specifies that both `outSubWidthC` and `outSubHeightC` are 2. When `ChromaFormatIdc` equals 2 and `nnpfc_out_sub_c_flag` exists, its value should be 1.

[0153] When `colourizationFlag` equals 1, `nnpfc_out_colour_format_idc` specifies the color format of the NNPF output, and correspondingly specifies the values ​​of the variables `outSubWidthC` and `outSubHeightC`. `nnpfc_out_colour_format_idc` equal to 1 specifies the NNPF output color format as 4:2:0, with both `outSubWidthC` and `outSubHeightC` equal to 2. `nnpfc_out_colour_format_idc` equal to 2 specifies the NNPF output color format as 4:2:2, with `outSubWidthC` equal to 2 and `outSubHeightC` equal to 1. `nnpfc_out_colour_format_idc` equal to 3 specifies the NNPF output color format as 4:4:4, with both `outSubWidthC` and `outSubHeightC` equal to 1. The value of `nnpfc_out_colour_format_idc` can be constrained to be non-zero. When both chromaUpsampingFlag and colourizationFlag are equal to 0, outSubWidthC and outSubHeightC can be inferred to be equal to SubWidthC and SubHeightC, respectively.

[0154] `nnpfc_pic_width_num_minus1+1` and `nnpfc_pic_width_deom_minus1+1` can specify the numerator and denominator of the resampling ratio used for the NNPF output image width relative to `CroppedWidth`, respectively. The value of `(nnpfc_pic_width_num_minus1+1)` divided by `(nnpfc_pic_width_deom_minus1+1)` will be in the range of 1 / 16 to 16 (inclusive). When `nnpfc_pic_width_num_minus1` and `nnpfc_pic_width_deom_minus1` do not exist, both `nnpfc_pic_width_num_minus1` and `nnpfc_pic_width_deom_minus1` can be inferred to be equal to 0.

[0155] The variable nnpfcOutputPicWidth, which indicates the width of the brightness sample array of the image produced by applying the NNPF identified by nnpfc_id to the input image, can be derived as shown in Equation 2.

[0156] [Equation 2]

[0157] The remainder when nnpfcOutputPicWidth is divided by outSubWidthC should be 0.

[0158] `nnpfc_pic_height_num_minus1+1` and `nnpfc_pic_height_deom_minus1+1` can specify the numerator and denominator of the resampling ratio of the NNPF output image height relative to `CroppedHeight`, respectively. The value of `(nnpfc_pic_height_num_minus1+1)` divided by `(nnpfc_pic_height_deom_minus1+1)` should be in the range of 1 / 16 to 16 (inclusive). When `nnpfc_pic_height_num_minus1` and `nnpfc_pic_height_deom_minus1` do not exist, both `nnpfc_pic_height_num_minus1` and `nnpfc_pic_height_deom_minus1` can be inferred to be equal to 0.

[0159] The variable nnpfcOutputPicHeight, which indicates the height of the luminance sample array of the resulting image due to the application of the NNPF identified by nnpfc_id to the input image, can be derived as shown in Equation 3.

[0160] [Equation 3]

[0161] The remainder when nnpfcOutputPicHeight is divided by outSubHeightC should be 0.

[0162] When nnpfc_pic_width_num_minus1, nnpfc_pic_width_deom_minus1, nnpfc_pic_height_num_minus1, and nnpfc_pic_height_deom_minus1 exist, at least one of the following constraints can be true: - The value of nnpfcOutputPicWidth is not equal to CroppedWidth.

[0163] The value of nnpfcOutputPicHeight is not equal to CroppedHeight.

[0164] `nnpfc_interpolated_pics[i]` can specify the number of interpolated pictures generated by NNPF between the i-th picture and the (i+1)-th picture used as input to NNPF. The value of `nnpfc_interpolated_pics[i]` can be constrained to be in the range of 0 to 63 (inclusive). For at least one value of `i` in the range of 0 to `nnpfc_num_input_pics_minus1-1` (inclusive), the value of `nnpfc_interpolated_pics[i]` can be constrained to be greater than 0.

[0165] The following variables can be exported: NumInpPicsInOutputTensor, which specifies the number of images with corresponding input images that exist in the NNPF output tensor; InpIdx[idx], which specifies the input image index of the idx-th image that exists in the NNPF output tensor with corresponding input images; and numOutputPics, which specifies the total number of images that exist in the NNPF output tensor, as shown in Table 6.

[0166] [Table 6]

[0167] `nnpfc_component_last_flag` equal to 1 indicates that the last dimension in both the NNPF input tensor and the NNPF-generated output tensor is used for the current channel. `nnpfc_component_last_flag` equal to 0 indicates that the third dimension in both the NNPF input tensor and the NNPF-generated output tensor is used for the current channel. The first dimension in both the input and output tensors can be used as a batch index, which is a common practice in some neural network frameworks. While the formula in the semantics of this SEI message uses a batch size corresponding to a batch index equal to 0, this depends on the post-processing implementation that determines the batch size used as input for neural network inference. For example, when `nnpfc_inp_order_idc` equals 3 and `nnpfc_auxiliary_inp_idc` equals 1, there can be 7 channels in the input tensor, including four luminance matrices, two chrominance matrices, and one auxiliary input matrix. In this case, the procedure DeriveInputTensors() can derive each of the seven channels of the input tensor one by one. When processing a particular channel among these channels, that channel can be referred to as the current channel during the procedure.

[0168] `nnpfc_inp_format_idc` can indicate the method for converting sample values ​​of the input image to input values ​​for NNPF. When `nnpfc_inp_format_idc` equals 1, the input values ​​to NNPF can be real numbers, and the functions `InpY()` and `InpC()` can be specified as shown in Equation 4.

[0169] [Equation 4]

[0170] When nnpfc_inp_format_idc equals 1, the input value to NNPF can be an unsigned integer, and the functions InpY() and InpC() can be derived as shown in Table 7.

[0171] [Table 7]

[0172] variable inpTensorBitDepth Y It can be exported from the syntax element nnpfc_inp_tensor_luma_bitdepth_minus8 specified below. The variable inpTensorBitDepth C The nnpfc_inp_tensior_chroma_bitdepth_minus8 syntax element can be derived from the syntax element specified below. Values ​​of nnpfc_inp_format_idc greater than 1 can be reserved for future use and should not be present in the bitstream. The decoder should ignore NNPFC SEI messages containing reserved values ​​of nnpfc_inp_format_idc.

[0173] A value greater than 0 for `nnpfc_auxiliary_inp_idc` indicates that auxiliary input data exists in the input tensor of the NNPF. A value equal to 0 for `nnpfc_auxiliary_inp_idc` indicates that auxiliary input data does not exist in the input tensor. A value equal to 1 for `nnpfc_auxiliary_inp_idc` indicates that the auxiliary input data was derived using the methods specified in Tables 10 to 12. In the bitstream, the value of `nnpfc_auxiliary_inp_idc` should be in the range of 0 to 1 (inclusive). Values ​​of 2 to 255 (inclusive) for `nnpfc_auxiliary_inp_idc` can be reserved for future use and should not appear in the bitstream. The decoder should ignore NNPFC SEI messages with `nnpfc_auxiliary_inp_idc` in the range of 2 to 255 (inclusive). Values ​​greater than 255 in nnpfc_auxiliary_inp_idc should not appear in the bitstream and are not reserved for future use.

[0174] `nnpfc_inp_order_idc` indicates the method by which the sample array of the input image is ordered to form the input tensor to NNPF. In the bitstream, the value of `nnpfc_inp_order_idc` should be in the range of 0 to 3 (inclusive). Values ​​of 4 to 255 (inclusive) for `nnpfc_inp_order_idc` can be reserved for future use and should not exist in the bitstream. The decoder should ignore NNPFC SEI messages with `nnpfc_inp_order_idc` in the range of 4 to 255 (inclusive). Values ​​of `nnpfc_inp_order_idc` greater than 255 should not exist in the bitstream and are not reserved for future use. `nnpfc_inp_order_idc` should not be equal to 3 when `ChromaFormatIdc` is not equal to 1. When `ChromaFormatIdc` is equal to 0, `nnpfc_inp_order_idc` should be equal to 0. When chromaUpsampingFlag equals 1, nnpfc_inp_order_idc should not equal 0.

[0175] Table 8 contains descriptions of the values ​​of nnpfc_inp_order_idc.

[0176] [Table 8]

[0177] `nnpfc_inp_tensor_luma_bitdepth_minus8+8` specifies the bit depth of the luminance sample values ​​in the input integer tensor. `inpTensorBitDepth` Y The value of can be derived as shown in Equation 5.

[0178] [Equation 5]

[0179] The value of nnpfc_inp_tensor_luma_bitdepth_minus8 can be constrained to be in the range of 0 to 24 (inclusive).

[0180] `nnpfc_inp_tensor_chroma_bitdepth_minus8+8` can specify the bit depth of the chroma sample values ​​in the input integer tensor. The value of `inpTensorBitDepthC` can be derived as shown in Equation 6.

[0181] [Equation 6]

[0182] The value of nnpfc_inp_tensior_chroma_bitdepth_minus8 can be constrained to be in the range of 0 to 24 (inclusive).

[0183] When nnpfc_auxiliary_inp_idc equals 1, the variable strength-ControlScaledVal can be exported as shown in Table 9.

[0184] [Table 9]

[0185] A tile can be a rectangular array of samples from the components of an image (e.g., luminance or chrominance components).

[0186] The procedure DevrivInputTensor(), used to derive an input tensor with given vertical sample coordinates cTop and horizontal sample coordinates cLeft for the top-left sample position of a patch of samples included in a specified input tensor, can be represented as a combination of Tables 10 to 12.

[0187] [Table 10]

[0188] [Table 11]

[0189] [Table 12]

[0190] When nnpfc_out_format_idc is equal to 0, it indicates that the sample values output by NNPF are real numbers, where the value range from 0 to 1 (including 0 and 1) is linearly mapped to the unsigned integer value range from 0 to (1<<bitDepth)-1 (including 0 and (1<<bitDepth)-1)) for any desired bit depth bitDepth for subsequent post-processing or display. When nnpfc_out_format_idc is equal to 1, it indicates that the sample values output by NNPF are unsigned integer values within the range of 0 to (1<<outTensorBitDepth Y )-1 (including 0 and (1<<outTensorBitDepth Y )-1), and the sample values output by NNPF are unsigned integer values within the range of 0 to (1<<outTensorBitDepth C )-1 (including 0 and (1<<outTensorBitDepth C )-1). Values of nnpfc_out_format_idc greater than 1 can be reserved for future use and should not be present in the bitstream. The decoder should ignore NNPFC SEI messages containing reserved values of nnpfc_out_format_idc.

[0191] nnpfc_out_order_idc can indicate the output order of the samples output from NNPF. In the bitstream, the value of nnpfc_out_order_idc should be within the range of 0 to 3 (including 0 and 3). Values from 4 to 255 (including 4 and 255) for nnpfc_out_order_idc can be reserved for future use and should not be present in the bitstream. The decoder should ignore NNPFC SEI messages having nnpfc_out_order_idc within the range of 4 to 255 (including 4 to 255). Values of nnpfc_out_order_idc greater than 255 should not be present in the bitstream and are not reserved for future use. When chromaUpsampingFlag is equal to 1, nnpfc_out_order_idc should not be equal to 0 or 1. When colourizationFlag is equal to 1, nnpfc_out_order_idc should not be equal to 0.

[0192] Table 13 contains the description of the values of nnpfc_out_order_idc.

[0193] [Table 13]

[0194] `nnpfc_out_tensor_luma_bitdepth_minus8+8` specifies the bit depth of the luminance sample values ​​in the output integer tensor. The value of `nnpfc_out_tensor_luma_bitdepth_minus8` should be in the range of 0 to 24 (inclusive). `outTensorBitDepth` Y The value of can be derived as shown in Equation 7.

[0195] [Equation 7]

[0196] `nnpfc_out_tensor_chroma_bitdepth_minus8+8` specifies the bit depth of the chroma sample values ​​in the output integer tensor. The value of `nnpfc_out_tensor_chroma_bitdepth_minus8` should be in the range of 0 to 24 (inclusive). `outTensorBitDepth` C The value of can be derived as shown in Equation 8.

[0197] [Equation 8]

[0198] When bitDepthUpsampingFlag equals 1, the value of nnpfc_out_format_idc should be equal to 1, and at least one of the following conditions can be true: - nnpfc_out_tensor_luma_bitdepth_minus8 exists, and outTensorBitDepth Y Greater than BitDepth Y .

[0199] -nnpfc_out_tensor_chroma_bitdepth_minus8 exists, and outTensorBitDepth C Greater than BitDepth C .

[0200] When nnpfc_inp_tensor_luma_bitdepth_minus8, nnpfc_inp_tensor_chroma_bitdepth_minus8, nnpfc_out_tensor_luma_bitdepth_minus8, and nnpfc_out_tensor_chroma_bitdepth_minus8 exist and outTensorBitDepth Y Greater than inpTensorBitDepth Y At that time, outTensorBitDepth C Will be no less than inpTensorBitDepth C When nnpfc_inp_tensor_luma_bitdepth_minus8, nnpfc_inp_tensor_chroma_bitdepth_minus8, nnpfc_out_tensor_luma_bitdepth_minus8, and nnpfc_out_tensor_chroma_bitdepth_minus8 exist and outTensorBitDepth C Greater than inpTensorBitDepth C At that time, outTensorBitDepth Y Will be no less than inpTensorBitDepth Y .

[0201] The procedure StoreOutputTensers(), used to derive sample values ​​from the filtered output sample arrays FilteredYPic, FilteredCbPic, and FilteredCrPic from the output tensor outputTensor for a given vertical sample coordinate cTop and horizontal sample coordinate cLeft, can be represented as a combination of Tables 14 and 15, where the given vertical sample coordinate cTop and horizontal sample coordinate cLeft specify the top-left sample position of the block of samples included in the input tensor.

[0202] [Table 14]

[0203] [Table 15]

[0204] A value of 1 for `nnpfc_separate_colour_description_present_flag` indicates that a different combination of color primaries, transformation properties, matrix coefficients, and scaling and offset values ​​is specified in the SEI message syntax structure for the image generated by NNPF. A value of 0 for `nnpfc_separate_colour_description_present_flag` indicates that the combination of color primaries, transformation properties, matrix coefficients, and scaling and offset values ​​used for the image generated by NNPF is the same as specified in the VUI parameters used for CLVS.

[0205] nnpfc_colour_primaries can have the same semantics as those defined for the vui_colour_primaries syntax element, with the following exceptions.

[0206] - nnpfc_colour_primaries can indicate the color primaries of the image produced by the NNPF specified by the application in the SEI message, instead of the color primaries used in CLVS.

[0207] - When nnpfc_colour_primaries does not exist in the NNPFC SEI message, the value of nnpfc_colour_primaries can be inferred to be equal to vui_colour_primaries.

[0208] nnpfc_transfer_characteristics can have the same semantics as those defined for the vui_transfer_characteristics syntax element, with the following exceptions.

[0209] - nnpfc_transfer_characteristics can indicate the transformation characteristics of the image generated by the NNPF specified in the SEI message, rather than the transformation characteristics used for CLVS.

[0210] - When nnpfc_transfer_characteristics does not exist in the NNPFC SEI message, the value of nnpfc_transfer_characteristics can be inferred to be equal to vui_transfer_characteristics.

[0211] `nnpfc_matrix_coeffs` describes the equations used in deriving luminance and chrominance signals from green, blue, and red, or Y, Z, and X primary colors. The semantics of `nnpfc_matrix_coeffs` can be applied to images generated from the NNPF specified in this SEI message. For example, `BitDepth` is used for `MatrixCoefficients`. Y and BitDepth C They can be equal to outTensorBitDepth respectively. Y and outTensorBitDepth C When nnpfc_matrix_coeffs is not present in the NNFPFC SEI message, the value of nnpfc_matrix_coeffs can be inferred to be equal to vui_matrix_coeffs.

[0212] nnpfc_matrix_coeffs should not be equal to 0 unless both of the following conditions are true.

[0213] - nnpfc_out_tensor_chroma_bitdepth_minus8 is equal to nnpfc_out_tensor_luma_bitdepth_minus8.

[0214] - nnpfc_out_order_idc equals 2, outSubHeightC equals 1, and outSubWidthC equals 1.

[0215] nnpfc_matrix_coeffs should not be equal to 8 unless one of the following conditions is true.

[0216] - nnpfc_out_tensor_chroma_bitdepth_minus8 is equal to nnpfc_out_tensor_luma_bitdepth_minus8.

[0217] - nnpfc_out_tensor_chroma_bitdepth_minus8 equals nnpfc_out_tensor_luma_bitdepth_minus8+1, nnpfc_out_order_idc equals 2, outSubHeightC equals 2, and outSubWidthC equals 1.

[0218] `nnpfc_full_range_flag` indicates the scaling and offset values ​​applied in association with the matrix coefficients, as specified by `nnpfc_matrix_coeffs`. The semantics of `nnpfc_full_range_flag` can be as specified for `VideoFullRangeFlag`. When `nnpfc_full_range_flag` is not present, its value can be inferred to be 0.

[0219] A value of 1 for `nnpfc_chroma_loc_info_present_flag` indicates the presence of the `nnpfc_chroma_sample_loc_type_frame` syntax element in the NNFPFC SEI message. A value of 0 for `nnpfc_chroma_loc_info_present_flag` indicates the absence of the `nnpfc_chroma_sample_loc_type_frame` syntax element in the NNFPFC SEI message. The value of `nnpfc_chroma_loc_info_present_flag` can be constrained to 0 when `colourizationFlag` is 0 or `nnpfc_out_colour_format_idc` is not 1.

[0220] When `nnpfc_chroma_sample_loc_type_frame` is not equal to 6 and `nnpfc_out_colour_format_idc` is equal to 1, `nnpfc_chroma_sample_loc_type_frame` can indicate the location of the chroma samples in the output image. `nnpfc_chroma_sample_loc_type_frame` being equal to 6 and `nnpfc_out_colour_format_idc` being equal to 1 indicates that the location of the chroma samples is unknown, unspecified, or has been specified by other means. The value of `nnpfc_chroma_sample_loc_type_frame` should be in the range of 0 to 6 (inclusive).

[0221] nnpfc_overlap indicates the horizontal and vertical sample counts of overlap between adjacent input tensors in NNPF. The value of nnpfc_overlap should be in the range of 0 to 16383 (inclusive).

[0222] Setting `nnpfc_constant_patch_size_flag` to 1 instructs NNPF to accept the exact patch size specified by `nnpfc_patch_width_minus1` and `nnpfc_patch_height_minus1` as input. Setting `nnpfc_constant_patch_size_flag` to 0 instructs NNPF to accept any patch size with a width `inpPatchWidth` and a height `inpPatchHeight` as input. Here, `inpPatchWidth + 2` is used. The width of the extended patch of nnpfc_overlap (i.e., the patch plus the overlapping region) is nnpfc_extended_patch_width_cd_delta_minus1+1+2 A positive integer multiple of nnpfc_overlap, and equal to inpPatchHeight + 2 The height of the extended patch in nnpfc_overlap is nnpfc_extended_patch_delta_minus1+1+2 A positive integer multiple of nnpfc_overlap.

[0223] When nnpfc_constant_patch_size_flag equals 1, npfc_patch_width_minus1+1 indicates the level sample count of the patch size required for the NNPF input. The value of nnpfc_patch_width_minus1 should be in the range of 0 to Min(32766, CroppedWidth-1) (inclusive).

[0224] When nnpfc_constant_patch_size_flag equals 1, npfc_patch_height_minus1+1 indicates the vertical sample count of the patch size required for the NNPF input. The value of nnpfc_patch_height_minus1 should be in the range of 0 to Min(32766, CroppedHeight-1) (inclusive).

[0225] When nnpfc_constant_patch_size_flag equals 0, nnpfc_extended_patch_width_cd_delta_minus1+1+2 `nnpfc_overlap` can indicate the common factor of all allowed values ​​for the width of the extended patch required for the NNPF input. The value of `nnpfc_extended_patch_width_cd_delta_minus1` should be in the range of 0 to Min(32766, CroppedWidth-1) (inclusive).

[0226] When nnpfc_constant_patch_size_flag equals 0, nnpfc_extended_patch_height_cd_delta_minus1+1+2 nnpfc_overlap can indicate the common factor of all allowed values ​​for the height of the extended patch required for the NNPF input. The value of nnpfc_extended_patch_height_cd_delta_minus1 should be in the range of 0 to Min(32766, CroppedHeight-1) (inclusive).

[0227] The variables inpPatchWidth and inpPatchHeight can be set to the tile size width and tile size height, respectively.

[0228] If nnpfc_constant_patch_size_flag equals 0, then the following can be applied.

[0229] The values ​​of inpPatchWidth and inpPatchHeight can be provided by an external device or set by a post-processor.

[0230] - inpPatchWidth+2 The value of nnpfc_overlap should be nnpfc_extended_patch_width_cd_delta_minus1+1+2 The value should be a positive integer multiple of nnpfc_overlap, and inpPatchWidth should be less than or equal to CroppedWidth. inpPatchHeight + 2 The value of nnpfc_overlap should be nnpfc_extended_patch_height_cd_delta_minus1+1+2 The value should be a positive integer multiple of nnpfc_overlap, and inpPatchHeight should be less than or equal to CroppedHeight.

[0231] Otherwise (if nnpfc_constant_patch_size_flag equals 1), the value of inpPatchWidth can be set to equal nnpfc_patch_width_minus1+1, and the value of inpPatchHeight can be set to equal nnpfc_patch_height_minus1+1.

[0232] The variables outPatchWidth, outPatchHeight, horCScaling, verCScaling, outPatchCWidth, and outPatchCHeight can be derived as shown in Table 16.

[0233] [Table 16]

[0234] outPatchWidth CroppedWidth should be equal to nnpfcOutputPicWidth inpPatchWidth and outPatchHeight CroppedHeight should be equal to nnpfcOutputPicHeight inpPatchHeight.

[0235] `nnpfc_padding_type` indicates the padding process when referencing sample locations outside the boundaries of the input image, as described in Table 17. The value of `nnpfc_padding_type` should be in the range of 0 to 4 (inclusive). Values ​​of 5 to 15 (inclusive) for `nnpfc_padding_type` can be reserved for future use and should not exist in the bitstream. The decoder should ignore NNPFC SEI messages with `nnpfc_padding_type` in the range of 5 to 15 (inclusive). Values ​​of `nnpfc_padding_type` greater than 15 should not exist in the bitstream and are not reserved for future use.

[0236] [Table 17]

[0237] nnpfc_luma_padding_val can indicate the luma value to be used for padding when nnpfc_padding_type is equal to 4. The value of nnpfc_luma_padding_val should be in the range of 0 to (1<<BitDepthY)-1 (including 0 and (1<<BitDepthY)-1).

[0238] nnpfc_cb_padding_val can indicate the Cb value to be used for padding when nnpfc_padding_type is equal to 4. The value of nnpfc_cb_padding_val should be in the range of 0 to (1<<BitDepthC) (including 0 and (1<<BitDepthC)).

[0239] nnpfc_cr_padding_val can indicate the Cr value to be used for padding when nnpfc_padding_type is equal to 4. The value of nnpfc_cr_padding_val should be in the range of 0 to (1<<BitDepthC)-1 (including 0 and (1<<BitDepthC)-1).

[0240] The function InpSampleVal(y, x, picHeight, picWidth, croppedPic, cIdx), with inputs of vertical sample position y, horizontal sample position x, picture height picHeight, picture width picWidth, sample array croppedPic, and component index cIdx (equal to 0 for luma, 1 for Cb, and 2 for Cr), can return the value of the derived sampleVal as shown in Table 18.

[0241] [Table 18]

[0242] NNPF PostProcessingFilter() is the target NNPF derived as in the semantics of the NNPFA SEI message.

[0243] The procedure in Table 19 can be used with NNPF PostProcessingFilter() to generate a filtered and / or interpolated picture in a tile-by-tile manner. The filtered and / or interpolated picture can contain the Y, Cb, and Cr sample arrays FilteredYPic, FilteredCbPic, and FilteredCrPic respectively, as indicated by nnpfc_out_order_idc.

[0244] [Table 19]

[0245] An image generated by NNPF with index i can contain sample arrays FilteredYPic[i], FilteredCbPic[i], and FilteredCrPic[i] (if they exist). The image generated by NNPF may not include overlapping regions.

[0246] The NNPF procedure may include outputting NNPF-generated images in ascending order of index according to the procedure defined by Table 19, wherein all NNPF-generated images interpolated by NNPF may be output, and all NNPF-generated images corresponding to any input image to NNPF may be output, as specified in the semantics of the NNPFA SEI message.

[0247] A value of 1 for nnpfc_complex_info_present_flag indicates the presence of one or more syntax elements that indicate the complexity of the NNPF associated with nnpfc_id. A value of 0 for nnpfc_complex_info_present_flag indicates the absence of one or more syntax elements that indicate the complexity of the NNPF associated with nnpfc_id.

[0248] An `nnpfc_parameter_type_idc` value of 0 indicates that the neural network uses only integer parameters. An `nnpfc_parameter_type_flag` value of 1 indicates that the neural network can use either floating-point or integer parameters. An `nnpfc_parameter_type_idc` value of 2 indicates that the neural network uses only binary parameters. An `nnpfc_parameter_type_idc` value of 3 can be reserved for future use and should not exist in the bitstream. The decoder should ignore NNPFC SEI messages with an `nnpfc_parameter_type_idc` value of 3.

[0249] The values ​​0, 1, 2, and 3 for nnpfc_log2_parameter_bit_length_minus3 indicate that the neural network should not use parameters with bit lengths greater than 8, 16, 32, and 64, respectively. When nnpfc_parameter_type_idc exists and nnpfc_log2_parameter_bit_length_minus3 does not exist, the neural network can avoid using parameters with a bit length greater than 1.

[0250] `nnpfc_num_parameters_idc` indicates the maximum number of neural network parameters used in NNPF, in powers of 2048. `nnpfc_num_parameters_idc` equal to 0 indicates that the maximum number of neural network parameters is unknown. The value of `nnpfc_num_parameters_idc` should be in the range of 0 to 52 (inclusive). Values ​​of `nnpfc_num_parameters_idc` greater than 52 are reserved for future use and should not exist in the bitstream. The decoder should ignore NNPFC SEI messages with `nnpfc_num_parameters_idc` greater than 52.

[0251] If the value of nnpfc_num_parameters_idc is greater than 0, then the variable maxNumParameters can be derived as shown in Equation 9.

[0252] [Equation 9]

[0253] The number of neural network parameters in NNPF can be constrained to be less than or equal to maxNumParameters.

[0254] A value greater than 0 for nnpfc_num_kmac_operations_idc indicates that the maximum number of multiplication-accumulation operations per sample in NNPF is less than or equal to nnpfc_num_kmac_operations_idc. 1000. An nnpfc_num_kmac_operations_idc value of 0 indicates that the maximum number of multiply-accumulate operations in the network is unknown. The value of nnpfc_num_kmac_operations_idc should be between 0 and 2. 32 -2 (including 0 and 2) 32 Within the range of -2).

[0255] A value greater than 0 for `nnpfc_total_kilobyte_size` indicates the total size, in kilobytes, required to store the uncompressed parameters for the neural network. The total size, in bits, can be greater than or equal to the sum of the bits used to store each parameter. `nnpfc_total_kilobyte_size` can be the total size (in bits) divided by 8000 and rounded up. A value of 0 for `nnpfc_total_kilobyte_size` indicates that the total size required to store the neural network parameters is unknown. The value of `nnpfc_total_kilobyte_size` should be between 0 and 2.32 - 2 (including 0 and 2) 32 - 2) within the range.

[0256] A value of 0 for `nnpfc_metadata_extension_num_bits` indicates that `nnpfc_reserved_metadata_extension` does not exist. A value greater than 0 for `nnpfc_metadata_extension_num_bits` indicates the length (in bits) of `nnpfc_reserved_metadata_extension`. `nnpfc_metadata_extension_num_bits` should be 0. Values ​​in the range 0 to 2048 (inclusive) for `nnpfc_metadata_extension_num_bits` are reserved for future use and should not exist in the bitstream. The decoder may allow any value for `nnpfc_metadata_extension_num_bits` in the range 0 to 2048 (inclusive). Values ​​greater than 2048 for `nnpfc_metadata_extension_num_bits` should not exist in the bitstream and are not reserved for future use.

[0257] The `nnpfc_reserved_metadata_extension` should not exist in the bitstream. However, the decoder will ignore the presence and value of `nnpfc_reserved_metadata_extension`. When `nnpfc_reserved_metadata_extension` exists, its length can be equal to `nnpfc_metadata_extension_num_bits`.

[0258] nnpfc_reserved_zero_bit_b should be equal to 0 in the bitstream. The decoder should ignore NNPFC SEI messages where nnpfc_reserved_zero_bit_b is not equal to 0.

[0259] nnpfc_payload_byte[i] may contain the i-th byte of the bitstream. The sequence of bytes used for all current values ​​of i, nnpfc_payload_byte[i], should be a complete bitstream conforming to ISO / IEC 15938-17.

[0260] Post-Filter Activation (NFFPA) of Neural Networks

[0261] The syntax structure for NNPFA is shown in Table 20.

[0262] [Table 20]

[0263] The NNPFA syntax structure in Table 20 can be sent as a signal in the form of an SEI message. A SEI message sent using a signal based on the NNPFA syntax structure in Table 20 can be called an NNPFA SEI message.

[0264] The NNPFA SEI message can activate or deactivate the possible use of a target neural network post-processing filter (NNPF), identified by nnpfa_target_id and nnpfa_target_base_flag, for post-processing filtering of a set of images. For a specific image where the NNPF is activated, the target NNPF can be derived as follows.

[0265] - If nnpfa_target_base_flag equals 1, then the target NNPF can be the base NNPF with nnpfc_id equal to nnpfa_target_id.

[0266] - Otherwise (if nnpfa_target_base_flag equals 0), the target NNPF can be the NNPF specified by the last NNPFC SEI message with nnpfc_id equal to nnpfa_target_id. Here, the last NNPFC SEI message can precede the first VCL NAL unit of the current image in the decoding order, and may not correspond to a repetition of the NNPFC SEI message containing the base NNPF.

[0267] There can be several NNPFA SEI messages for the same image, for example, when NNPF is intended for different purposes or for filtering different color components.

[0268] `nnpfa_target_id` can indicate the target NNPF specified by one or more NNPFC SEI messages about the current image and having `nnpfc_id` equal to `nnpfa_target_id`. The value of `nnpfa_target_id` should be between 0 and 2. 32 -2 (including 0 and 2) 32 - 2) within the range.

[0269] NNPFA SEI messages with the special value nnpfa_target_id should not exist in the current PU unless one or both of the following conditions are true.

[0270] - Within the current CLVS, there exists an NNPFC SEI message with a special value nnpfc_id equal to the nnpfa_target_id that exists in the PU that precedes the current PU in the decoding order.

[0271] - There exists an NNPFC SEI message with a special value nnpfc_id equal to nnpfa_target_id in the current PU.

[0272] When a PU contains both an NNPFC SEI message with a special value of nnpfc_id and an NNPFA SEI message with a special value of nnpfa_target_id equal to nnpfc_id, the NNPFC SEI message should be decoded before the NNPFASEI message.

[0273] A `nnpfa_cancel_flag` value of 1 indicates that the persistence of a target NNPF established by any previous NNPFA SEI message with the same `nnpfa_target_id` as the current SEI message is canceled. In other words, the target NNPF is no longer used unless it is activated by another NNPFA SEI message with the same `nnpfa_target_id` as the current SEI message and a `nnpfa_cancel_flag` value of 0. A `nnpfa_cancel_flag` value of 0 can also indicate that `nnpfa_target_base_flag`, `nnpfa_persistence_flag`, and `nnpfa_num_output_entries` follow.

[0274] A `nnpfa_target_base_flag` value of 1 indicates that the target NNPF is the base NNPF with `nnpfc_id` equal to `nnpfa_target_id`. A `nnpfa_target_base_flag` value of 0 indicates that the target NNPF is the NNPF specified by the last NNPFC SEI message with `nnpfc_id` equal to `nnpfa_target_id`. Here, the last NNPFC SEI message may precede the first VCL NAL unit of the current image in the decoding order and may not correspond to a repetition of the NNPFC SEI message containing the base NNPF.

[0275] `nnpfa_persistence_flag` indicates the persistence of the target NNPF used for the current layer. `nnpfa_persistence_flag` equal to 0 indicates that the target NNPF can be used only for post-processing filtering of the current image. `nnpfa_persistence_flag` equal to 1 indicates that the target NNPF can be used for post-processing filtering of the current image and all subsequent images in the current layer in output order, until one or more of the following conditions are true.

[0276] - A new CLVS begins in the current layer.

[0277] - End of bitstream.

[0278] - Output the images in the current layer that are associated with the NNPFA SEI message that has the same nnpfa_target_id as the current SEI message and whose nnpfa_cancel_flag is equal to 1, following the current images in the output order.

[0279] For subsequent images in the current layer associated with an NNPFA SEI message that has the same nnpfa_target_id as the current SEI message and nnpfa_cancel_flag equal to 1, no target NNPF is applied.

[0280] `nnpfcTargetPictures` can be a collection of pictures associated with the last NNPFA SEI message that precedes the current NNPFA SEI message in decoding order and has `nnpfc_id` equal to `nnpfa_target_id`. `nnpfaTargetPictures` can also be a collection of pictures that activate the target NNPF through the current NNPFA SEI message. Any pictures included in `nnpfTargetPictures` should also be included in `nnpfcTargetPictures`.

[0281] nnpfa_num_output_entry can indicate the number of nnpfa_output_flag[i] syntax elements present in the NNPFA SEI message. The value of nnpfa_num_output_entry should be in the range of 0 to NumInpPicsInOutputTensor (inclusive).

[0282] An NNpfa_output_flag[i] equal to 1 indicates that the NNPF-generated image corresponding to the input image with index InpIdx[i] is output by the NNPF process activated by the NNPPFA SEI message. Here, the NNPF process can be specified in the semantics of the NNPFC SEI message. An NNpfa_output_flag[i] equal to 0 indicates that no NNPF-generated image corresponding to the input image with index InpIdx[i] is output by the NNPF process activated by the NNPFASEI message. When nnpfa_num_output_entry is less than NumInpPicsInOutputTensor, for each value of i in the range from nnpfa_num_output_entry to NumInpPicsInOutputTensor-1 (inclusive), nnpfa_output_flag[i] can be inferred to be equal to 1.

[0283] Post-filter bank characteristics of neural networks (NNPFGC)

[0284] Table 21 shows the syntax structure used for NNFGGC.

[0285] [Table 21]

[0286] The NNPFGC syntax structure in Table 21 can be sent as a signal in the form of an SEI message. Sending the SEI message of the NNPFGC syntax structure in Table 21 as a signal can be called an NNPFGC SEI message.

[0287] The NNPFGC SEI message can specify the post-filter bank of a neural network. The SEI message can indicate whether an NNPF group defines an NNPF cascade or defines NNPF groups or NNPF cascades that are alternatives to each other. The use of an NNPF group for a specific image's NNPF cascade can be indicated using the Neural Network Post-Filter Bank Activation (NNPFGA) SEI message.

[0288] nnpfgc_id can contain an identifier number that can be used to identify NNPF groups. The value of nnpfgc_id should be between 0 and 2. 32 - 2 (including 0 and 2) 32 - 2) range. From 256 to 511 (inclusive) and from 2 31 to 2 32 -2 (including 2) 31 and 232 The value of nnpfgc_id (-2) can be reserved for future use. The decoder can be constrained to ignore values ​​in the range of 256 to 511 (inclusive) or greater than 2... 31 Up to 2 32 -2 (including 2) 31 and 2 32 -2) NNPFGC SEI messages with nnpfgc_id within the range of nnpfgc_id. The value of nnpfgc_id should not be equal to the value of nnpfgc_id in any NNPFC SEI message existing in the same CLVS. When the value of nnpfgc_id of NNPFGC SEI message nnpfgcSeiA is equal to the value of nnpfgc_id of another NNPFGC SEI message nnpfgcSeiB existing in the same CLVS, nnpfgcSeiA and nnpfgcSeiB should be the same.

[0289] `nnpfgc_grouping_type` equal to 0 indicates that this SEI message can specify a set of cascaded NNPFs. `nnpfgc_grouping_type` equal to 1 indicates that the NNPF or NNPF group identified by `nnpfgc_member_id[i]` is an alternative to each other that the postprocessor should select only to be applied. `nnpfgc_grouping_type` equal to 2 indicates that this SEI message can specify a set of NNPFs intended to be used in combination or activated alternately, such that at most one NNPF is activated for any picture. `nnpfgc_grouping_type` equal to 3 indicates that the NNPF or NNPF group identified by `nnpfgc_member_id[i]` is intended to be used in parallel. `nnpfgc_grouping_type` equal to 4 indicates that the NNPF or NNPF group identified by `nnpfgc_member_id[i]` is optional, i.e., they may or may not be applied by the postprocessor. The value of `nnpfgc_grouping_type` should be in the range of 0 to 255 (inclusive). Values ​​of `nnpfgc_grouping_type` in the range of 5 to 255 (inclusive) can be reserved for future use and can be constrained to not exist in the bitstream. The decoder can be constrained to ignore NNPFGC SEI messages with `nnpfgc_grouping_type` in the range of 5 to 255 (inclusive).

[0290] nnpfgc_purpose can have the semantics of nnpfc_purpose, except for specifying semantics for NNPF groups defined by the NNPFGC SEI message, rather than for NNPF groups defined by the NNPFC SEI message.

[0291] nnpfgc_num_memers_minus2+2 can indicate the number of NNPFs or NNPF groups in the NNPF group defined by the NNPFGC SEI message.

[0292] nnpfgc_member_id[i] can indicate the i-th member in the NNPF group defined by the NNPFGC SEI message as follows.

[0293] If there exists an NNPF defined in CLVS with an nnpfc_id equal to nnpfgc_member_id[i], then the i-th member in the NNPF group defined by the NNPFGC SEI message can be an NNPF with an nnpfc_id equal to nnpfgc_member_id[i].

[0294] - Otherwise (if there is no NNPF defined in CLVS with an nnpfc_id equal to nnpfgc_member_id[i]), the i-th member of the NNPF group defined by the NNPFGC SEI message can be an NNPF group with an nnpfgc_id equal to nnpfgc_member_id[i].

[0295] When the value of nnpfgc_member_id[i] references the value of nnpfgc_id in the NNPFGC SEI message nnpfgcSei, the NNPFGC SEI message nnpfgcSei can be constrained to have an nnpfgc_grouping_type equal to 0. When nnpfgc_grouping_type equals 0 or 2, an NNPF defined in CLVS with an nnpfc_id equal to nnpfgc_member_id[i] should exist. When nnpfgc_grouping_type equals 1, 3, or 4, an NNPF defined in CLVS with an nnpfc_id equal to nnpfgc_member_id[i] or an NNPF with an nnpfgc_id equal to nnpfgc_member_id[i] should exist. When nnpfgc_grouping_type equals 0, NNPFs with nnpfc_id equal to nnpfgc_member_id[i] can be cascaded in ascending order of i, as activated by an NNPFGASEI message with nnpfga_target_id equal to nnpfgc_id. The semantics of nnpfgc_complex_info_present_flag, nnpfgc_parameter_type_idc, nnpfgc_log2_parameter_bit_length_minus3, nnpfgc_num_parameters_idc, nnpfgc_num_kmac_operations_idc, and nnpfgc_total_kkibyte_size are respectively nnpfc_complexity_info_present_flag, nnpfc_parameter_type_idc, nnpfc_log2_parameter_bit_length_minus3, nnpfc_num_parameters_idc, nnpfc_num_kmac_operations_idc, and nnpfc_total_kkibyte_size, but except that they specify semantics for the NNPF group defined by this SEI message rather than the NNPF defined by the NNPFC SEI message. When nnpfgc_grouping_type equals 1, nnpfgc_complex_info_present_flag should equal 0.

[0296] Neural Network Post-Filter Bank Activation (NNPFGA)

[0297] The syntax structure for NNAFPGA is shown in Table 22.

[0298] [Table 22]

[0299] The NNPFGA syntax structure in Table 22 can be sent as a signal in the form of an SEI message. A SEI message sent using a signal to represent the NNPFGA syntax structure in Table 22 can be called an NNPFGA SEI message.

[0300] The NNPFGA SEI message can activate or deactivate the potential use of a target NNPFGC identified by the nnpfga_target_id of the NNPF group for post-processing filtering of a collection of pictures. The nnpfgc_grouping_type used for the identified NNPF group should be equal to 0 (cascaded) or 1 (alternative). When nnpfgc_grouping_type equals 1, each member of the group will have the same number of input pictures and NNPF output pictures. For a specific picture that activates an NNPFGC, the target NNPFG can be the NNPFG specified by the last NNPFGC SEI message of the first VCL NAL unit that has an nnpfgc_id equal to nnpfga_target_id and is decoded before the current picture. The NNPF of the target NNPFG can be defined by an NNPFC SEI message with an nnpfgc_id equal to the nnpfgc_member_id[i] of the target NNPFG. They can exist in the current picture unit or be decoded before the current picture.

[0301] The use of NNPFGC SEI messages may require the following definition.

[0302] - The input image width and height in units of brightness samples, InitCroppedWidth[idx] and InitCroppedHeight[idx], have an index idx in the range of 0 to numCandInputPics-1 (inclusive), and can be used as input for NNPFGC.

[0303] - A luminance sample array InitCroppedYPic[idx] and chrominance sample arrays InitCroppedCbPic[idx] and InitCroppedCrPic[idx] of an input image with an index idx in the range of 0 to numCandInputPics-1 (inclusive) can be used as input for NNPFGC.

[0304] - BitDepth of the luminance sample array used for candidate input images Y .

[0305] - BitDepth of the chroma sample array used for candidate input images C .

[0306] - Chroma format indicator, here represented by ChromaFormatIdc.

[0307] - When nnpfc_assistary_inp_idc equals 1, the filter strength control value array StrengthControlVal[idx] should contain real numbers in the range 0 to 1 (inclusive) of the input image with an index idx in the range of 0 to numCandInputPics-1 (inclusive).

[0308] A candidate input image with index 0 can represent the image used to activate NNPFGC by the NNPFGA SEI message. An input image with index idx in the range of 1 to numCandInputPics-1 (inclusive) can precede the candidate input image with index i-1 in the output order. candInputPicList[0] can be a list of candidate input images in reverse output order.

[0309] `nnpfga_target_id` can indicate the target NNPGFG specified by the NNPFGC SEI message about the current image, and can have an `nnpfgc_id` equal to `nnpfga_target_id`. The value of `nnpfga_target_id` should be between 0 and 2. 32 -2 (including 0 and 2) 32 Within the range of -2). NNPFGA SEI messages with the special value nnpfga_target_id should not exist in the current PU, unless an NNPFGC SEI message with the special value nnpfgc_id equal to nnpfga_target_id and nnpfgc_grouping_type equal to 0 exists in the current PU or in a PU that is decoded before the current PU. When a PU contains both an NNPFGC SEI message with the special value nnpfgc_id and an NNPFGA SEI message with the special value nnpfga_target_id equal to nnpfgc_id, the NNPFGC SEI message should be decoded before the NNPFGA SEI message.

[0310] A `nnpfga_cancel_flag` value of 1 indicates that the persistence of a target NPFG established by any previous NNPFGA SEI message with the same `nnpfga_target_id` as the current SEI message is cancelled. That is, the target NPFG may no longer be used unless it is activated by another NNPFGA SEI message with the same `nnpfga_target_id` as the current SEI message and a `nnpfga_cancel_flag` value of 0. A `nnpfga_cancel_flag` value of 0 indicates that the target NPFG is activated for use.

[0311] `nnpfga_persistence_flag` indicates the persistence of the target NPFG used for the current layer. `nnpfga_persistence_flag` equal to 0 indicates that the target NPFG can be used only for post-processing filtering of the current image. `nnpfga_persistence_flag` equal to 1 indicates that the target NPFG can be used for post-processing filtering of the current image and all subsequent images in the current layer in output order, until one of the following conditions is true: - A new CLVS for the current layer begins.

[0312] - End of bitstream.

[0313] - Includes images in the current layer that follow the current image in output order and have the same nnpfga_target_id as the NNPFGA SEI message.

[0314] The target NNPPG is not applied to subsequent pictures in the current layer associated with the NNPPG SEI message, which has the same nnpfga_target_id as the current SEI message. nnpfgcTargetPictures is a set of pictures corresponding to the last NNPPG SEI message that precedes the current NNPPG SEI message in decoding order and has an nnpfgc_id equal to nnpfga_target_id. nnpfgaTargetPictures can be a set of pictures whose target NNPPG is activated by the current NNPPG SEI message. Any pictures included in nnpfgaTargetPictures should also be included in nnpfgcTargetPictures.

[0315] Increasing 2 by nnpfga_num_filters_minus2 indicates the number of NNPFs activated in the NNPFGA SEI message within the NNPFG. The value of nnpfga_num_filter_minus2 should be equal to the value of nnpfgc_num_minus2 in the NNPFGC SEI message with nnpfgc_id equal to nnpfga_target_id.

[0316] An nnpfga_target_base_flag[i] equal to 1 indicates that the i-th NNPF in the target NNPFGC is the base NNPF with an nnpfgc_id equal to nnpfgc_member_id[i] in the NNPFGC SEI message with an nnpfgc_id equal to nnpfga_target_id. An nnpfga_target_base_flag[i] equal to 0 indicates that the i-th NNPF in the target NNPFGC is the NNPF specified by the last NNPFC SEI message with an nnpfgc_id equal to nnpfgc_member_id[i] in the NNPFGC SEI message with an nnpfgc_id equal to nnpfga_target_id. This message may precede the first VCL NAL unit of the current image in the decoding order and may not be a duplicate of the NNPFC SEI message containing the base NNPF.

[0317] `nnpfga_input_all_pics_flag[i]` equal to 1 indicates that the i-th NNPF input image should be selected from the list of candidate input images `candInputPicList[i]` without skipping any. `nnpfga_input_all_pics_flag[i]` equal to 0 indicates that the i-th NNPF input image should be selected from `candInputPicList[i]` while skipping some candidate input images.

[0318] nnpfga_num_input_pics_minus1[i] indicates the number of input images for the i-th NNPF in the target NNPFG. When nnpfga_num_input_pics_minus1[i] exists, for an NNPF with an nnpfc_id equal to nnpfgc_member_id[i] in the NNPFGC SEI message with nnpfgc_id equal to nnpfga_target_id, nnpfga_num_input_pics_minus1[i] should be equal to nnpfc_num_input_pics_minus1. When nnpfga_num_input_pics_minus1[i] does not exist, for an NNPF GC SEI message with nnpfgc_id equal to nnpfga_target_id and nnpfgc_num_input_pics_minus1[i], nnpfga_num_input_pics_minus1[i] can be inferred to be equal to nnpfc_num_input_pics_minus1.

[0319] `nnpfga_input_pic_skip_count[i][j]` indicates the count of the j-th picture skipped in the list of candidate input pictures `candInputPicList[i]` when selecting an input picture for the NNPF activated by the i-th loop entry. When `nnpfga_input_pic_skip_count[i][j]` is not present, for all values ​​of `j` in the range 0 to `nnpfga_num_input_pics_minus1[i]` (inclusive), `nnpfga_input_pic_skip_count[i][j]` can be inferred to be equal to 0. The variable `numCandInputPics` can be derived as shown in Table 23, which indicates the number of candidate input pictures to the NNFFG.

[0320] [Table 23]

[0321] candInputPicList[m] (where m is in the range of 1 to nnpfga_num_filters_minus2+1 (inclusive)) can be a list of images in reverse output order. It can initially be empty and formed in descending order of n in the range of 0 to m-1 (inclusive). This list can be configured by including each image output by the NNPF process from the nth loop entry that is not yet present in candInputPicList[m] and finally including each image present in candInputPicList[0] that is not yet present in candInputPicList[m].

[0322] When the candidate input image candInputPicList[m][idx] is the NNPF output image of the nth NNPF process (where the value of n is less than the value of m), for any value of m in the range from 1 to nnpfga_num_filters_minus2+1 (inclusive), the width and height of the candidate input image can be equal to the nnpfcOutputPicWidth and nnpfcOutputPicHeight of the NNPF output image, respectively.

[0323] As shown in Table 24, a list of input images for NNPF, inputPicList[m], can be exported to the m-th loop entry.

[0324] [Table 24]

[0325] The number of images in `candIdx` should not exceed the number of images in `candInputPicList[m]`. Images present in `inputPicList[m]` for any value of `m` in the range 1 to `nnpfga_num_filter_minus2+1` (inclusive) should have the same width, height, bit depth, and chroma format. To interpret an NNPFC SEI message with an `nnpfc_id` equal to `nnpfga_target_id` in an NNPFGC SEI message with an `nnpfgc_id` equal to `nnpfga_member_id[i]`, the following variables can be specified for the `i`th loop entry.

[0326] - You can use the variables BitDepthY, BitDepthC, and ChromaFormatIdc, as provided for the interpretation of NNPFGA SEI messages.

[0327] - CroppedWidth and CroppedHeight can be set to equal to the width and height (in brightness samples) of the image in inputPicList[i], respectively.

[0328] - For each input image k in the range of 0 to nnpfga_num_input_pics_minus1[i] (inclusive of 0 and nnpfga_num_input_pics_minus1[i]), the following can be applied.

[0329] (1) CroppedYPic[k], CroppedCbPic[k] and CroppedCrPic[k] (if they exist) can be set to the corresponding sample arrays equal to inputPicList[i][k].

[0330] (2) When nnpfc_auxiliary_inp_idc equals 1 for an NNPF with nnpfc_id equal to nnpfgc_member_id[i] in the NNPFGC SEI message (with nnpfgc_id equal to nnpfga_target_id), the following can be applied.

[0331] (2.1) For any idx value in the range of 0 to numCandInputPics-1 (inclusive), inputPicList[i][k] should be the same as candInputPicList[0][idx].

[0332] (2.2) StrengthControlVal[k] can be set to equal InitStrengthControlVal[idx].

[0333] `nnpfga_num_output_entry[i]` indicates the number of `nnpfga_output_flag[i][j]` syntax elements present in the NNPFGA SEI message. For an NNPF with an `nnpfgc_id` equal to `nnpfga_target_id` in an NNPFGC SEI message, the value of `nnpfga_num_output_entries[i]` should be in the range of 0 to `NumInpPicsInOutputTensor` (inclusive).

[0334] An NNpfga_output_flag[i][j] equal to 1 indicates that the NNPF-generated image corresponding to the input image with the index InpIdx[j] derived from the i-th NNPF for the target NNPPFG is output through the NNPF process activated by this loop entry. Here, the NNPF process can be specified in the semantics of the NNPFC SEI message. An NNpfga_output_flag[i][j] equal to 0 indicates that the NNPF-generated image corresponding to the input image with the index InpIdx[j] derived from the i-th NNPF for the target NNPPFG is not output through the NNPF process activated by this loop entry. When nnpfga_num_output_entries[i] is less than NumInpPicsInOutputTenser derived for the i-th NNPF of the target NNPFG, for each value of i in the range from nnpfga_num_output_entries[i] to NumInpPicsInOutputTenser-1 (inclusive), nnpfga_num_output_entries[i] and NumInpPicsInOutputTenser-1 can be inferred to be equal to 1.

[0335] NnpfgaOutputPicList, which is a list of images output in the output order by the NNPF process of NNPFG, can be initially empty and is formed by including each image output by the NNPF process of the nth loop entry that is not yet present in NnpfgaOutputPicList in descending order of n in the range of 0 to nnpfga_num_filter_minus2+1 (inclusive).

[0336] Post-filter hints

[0337] The syntax structure for post-filter hints is shown in Table 25.

[0338] [Table 25]

[0339] The syntax structure of the post-filter hints in Table 25 can be signaled as an SEI message. An SEI message that signals the syntax structure of the post-filter hints in Table 25 can be called a post-filter hint SEI message.

[0340] The post-filter hint SEI message can provide the coefficients of the post-filter or relevant information for the design of the post-filter, for potential use in the post-processing of the image set after they are decoded and output to obtain improved display quality.

[0341] A filter_hint_cancel_flag value of 1 indicates that the SEI message cancels the persistence of any previous post-filter hint SEI messages applied to the current layer in output order. A filter_hint_cancel_flag value of 0 indicates that a post-filter hint message should follow.

[0342] `filter_hint_persistence_flag` indicates the persistence of the post-filter hint SEI message used for the current layer. `filter_hint_persistence_flag` equal to 0 indicates that the post-filter hint is applied only to the currently decoded image. `filter_hint_persistence_flag` equal to 1 indicates that the post-filter hint SEI message is applied to the currently decoded image and continues to be applied to all subsequent images in the current layer in output order until one or more of the following conditions are true.

[0343] - A new CLVS begins in the current layer.

[0344] - End of bitstream.

[0345] - The image in the current layer of the AU associated with the post-filter hint SEI message is output, following the current image in the output order.

[0346] The filter_hint_size_y parameter indicates the vertical size of the filter coefficients or the correlation array. The value of filter_hint_size_y should be in the range of 1 to 15 (inclusive).

[0347] `filter_hint_size_x` can indicate the horizontal size of the filter coefficients or correlation array. The value of `filter_hint_size_x` should be in the range of 1 to 15 (inclusive).

[0348] `filter_hint_type` indicates the type of filter hint sent, as specified in Table 26. The value of `filter_hint_type` should be in the range of 0 to 2 (inclusive). A value of `filter_hint_type` equal to 3 can be reserved for future use and should not exist in the bitstream. The decoder should ignore post-filter hint SEI messages with `filter_hint_type` equal to 3.

[0349] [Table 26]

[0350] A filter_hint_chroma_coeff_present_flag value of 1 indicates the presence of filter coefficients for chroma. A filter_hint_chroma_coeff_present_flag value of 0 indicates the absence of filter coefficients for chroma.

[0351] `filter_hint_value[ cIdx ][ cy ][ cx ]` indicates the filter coefficients or elements of the cross-correlation matrix between the original and decoded signals with 16-bit precision. The value of `filter_hint_value[ cIdx ][ cy ][ cx ]` should be in the range of -2. 31 + 1 to 2 31 - 1 (including -2) 31 + 1 and 2 31 - 1) within the range. cIdx can represent the associated color component, cy can represent the vertical counter, and cx can represent the horizontal counter. The following can be applied depending on the value of filter_hint_type.

[0352] - If filter_hint_type equals 0, then a message with a size of filter_hint_size_y can be sent. The coefficients of a two-dimensional finite impulse response (FIR) filter are represented by filter_hint_size_x.

[0353] Otherwise, if `filter_hint_type` equals 1, the filter coefficients of two one-dimensional FIR filters can be sent. In this case, `filter_hint_size_y` should equal 2. An index `cy` of 0 indicates the filter coefficients of the horizontal filter, and `cy` of 1 indicates the filter coefficients of the vertical filter. During filtering, the horizontal filter can be applied first, and the result can be filtered by the vertical filter.

[0354] - Otherwise (if filter_hint_type equals 2), the hint sent can specify the cross-correlation matrix between the original signal s and the decoded signal s'.

[0355] It can be defined as follows for a filter with a size of filter_hint_size_y The normalized cross-correlation matrix of the relevant color components identified by cIdx for filter_hint_size_x.

[0356] [Equation 10]

[0357] In Equation 10, s represents the sample array of the color component cIdx of the original image, s' represents the corresponding array of the decoded image, h represents the vertical height of the correlated color component, w represents the horizontal width of the correlated color component, and bitDepth represents the bit depth of the color component. OffsetY equals (filter_hint_size_y >> 1), OffsetX equals (filter_hint_size_x >> 1), 0 <= cy < filter_hint_size_y and 0 <= cx < filter_hint_size_x.

[0358] The decoder can derive the Wiener post-filter from the cross-correlation matrix of the original signal and the decoded signal, as well as the autocorrelation matrix of the decoded signal.

[0359] Source Image Timing Information (SPTI)

[0360] The syntax structure used for SPTI is shown in Table 27.

[0361] [Table 27]

[0362] The SPTI syntax structure in Table 27 can be sent as a signal in the form of an SEI message. Sending an SEI message using the SPTI structure in Table 27 as a signal can be called an SPTI SEI message.

[0363] SPTI SEI messages can indicate the temporal distance between source images associated with the corresponding decoded output image prior to encoding. For example, for content captured by a camera, the temporal distance between source images can represent the difference between the time it took for the image sensor to be exposed to produce the source image associated with the current decoded image and the time it took for the image sensor to be exposed to produce the source images associated with the previously decoded images in output order.

[0364] A `spti_cancel_flag` value of 1 indicates that an SPTI SEI message cancels the persistence of any previous SPTI SEI messages applied to the current layer. A `spti_cancel_flag` value of 0 indicates that the source image timing information should follow.

[0365] `spti_persistence_flag` specifies the persistence of SPTI SEI messages used for the current layer. `spti_persistence_flag` equal to 0 specifies that SPTI SEI messages apply only to the currently decoded picture. `spti_persistence_flag` equal to 1 specifies that SPTI SEI messages apply to the currently decoded picture and persist for all subsequent pictures of the current layer in output order until one or more of the following conditions are true.

[0366] - A new CLVS begins for the current layer.

[0367] - End of bitstream.

[0368] - Images in the AU, including SPTI SEI messages, are output, following the current image in the output order.

[0369] As shown in Table 28, `spti_source_picture_timing_type` indicates the timing relationship between the source picture and the corresponding decoded output picture. A non-zero value for `(spti_source_picture_timing_type & bitMask)` indicates that the timing relationship has an interpretation associated with the value of `bitMask`. When `spti_source_picture_timing_type` is greater than 0 and `(spti_source_picture_timing_type & bitMask)` is equal to 0, the interpretation associated with the value of `bitMask` does not apply to SPTI SEI messages. When `spti_source_picture_timing_type` is equal to 0, the timing relationship can be specified by application.

[0370] The value of spti_source_picture_timing_type should be in the range of 0 to 127 (inclusive). Values ​​of 128 to 255 (inclusive) for spti_source_picture_timing_type can be reserved for future use and should not exist in the bitstream. The decoder should ignore SPTI SEI messages with spti_source_picture_timing_type in the range of 128 to 255 (inclusive).

[0371] [Table 28]

[0372] The variable `temporalReversalFlag` can be equal to `(spti_source_picture_timing_type & 0x10 ) ? 1 : 0`. `spti_source_timing_equals_output_timing_flag` equal to 1 indicates that the timing of the source image is the same as the timing of the corresponding decoded output image. `spti_source_timing_equals_output_timing_flag` equal to 0 indicates that the timing of the source image is different from the timing of the corresponding decoded output image. When `spti_source_timing_equals_output_timing_flag` equals 1 and a picture timing SEI message exists for the current image, the timing of the source image can be determined from the information transmitted in the picture timing SEI message.

[0373] spti_time_scale indicates the number of time units that elapse in one second. The value of spti_time_scale should not be equal to 0. For example, a time coordinate system using a 27MHz clock has a spti_time_scale of 27,000,000.

[0374] `spti_num_units_in_elementary_source_picture_interval` indicates the number of clock time units operating at a frequency corresponding to the `spti_time_scale` Hz of consecutive element source picture intervals in CLVS in output order. The indicated element source picture interval, i.e., the interval indicated by the variable `ElementalSourcePictureInterval`, can be in seconds and can be equal to `spti_num_units_in_element_source_picture_interval` divided by `spti_time_scale`. For example, when the element source picture interval is 0.04 seconds, `spti_time_scale` can be set to 27,000,000, and `spti_num_units_in_element_source_picture_interval` can be set to 1,080,000.

[0375] spti_max_sublayers_minus_1+1 indicates the maximum number of time sublayers that can exist in CLVS.

[0376] spti_sublayer_source_picture_interval_scale_factor[i] (if present) can indicate the scaling factor used to determine the source picture intervals of consecutive pictures in CLVS with TemporalId equal to i and in output order. For spti_sublayer_source_picture_interval_scale_factor[i], a value of 0 can be used to indicate that the source picture corresponding to the current decoded output picture is the same as the source picture corresponding to the previous decoded output picture.

[0377] The source image interval associated with the output image having a TemporalId less than or equal to I can be represented by the variable SourcePictureInterval[i] in seconds and can be derived as shown in Equation 11.

[0378] [Equation 11]

[0379] SourcePictureInterval[i]=ElementalSourcePictureInterval SPTI sublayer source image interval scaling factor [i]

[0380] When image n is an output image with a TemporalId less than or equal to 1 and is not the first image in the bitstream in output order, the value of the variable SourcePictureTime[n] can be derived as follows: - If temporalReversalFlag equals 0, then SourcePictureTime[n] = SourcePictureTime[previous PicInOutputOrder] + SourcePictureInterval[i]; - Otherwise, if temporalReversalFlag equals 1, then SourcePictureTime[n] = SourcePictureTime[previous PicInOutputOrder] - SourcePictureInterval[i], where previous PicInOutputOrder can represent the last output image with a TemporalId less than or equal to i that precedes image n in output order (if such an image exists). If the value of SourcePictureTime[0] is not provided by an external device, the value of SourcePictureTime[0] can be inferred to be equal to 0.

[0381] When spti_sublayer_syntheed_picture_flag[i] exists, spti_sublayer_synthesized_picture_flag[i] being equal to 0 indicates that the decoded output image belonging to the i-th temporal sublayer was synthesized and does not correspond to the unmodified original source image. A spti_sublayer_syntheed_picture_flag[i] being equal to 0 may not provide such an indication. When spti_sublayer_syntheed_picture_flag[i] does not exist, the value of spti_sublayer_syntheed_picture_flag[i] can be inferred to be equal to 0.

[0382] Problems of traditional technologies

[0383] The image associated with the source image provided by the SPTI SEI message (e.g., it can be represented as image 0, anchor image, first source timing image, or standard image, but it is always referred to as image 0 below) can be assumed to be an image under specific conditions (e.g., the first output image). However, in this case, it is unclear in the current design which image will be the first output image, i.e., which image will be referred to as image 0.

[0384] In other words, due to the lack of a precise definition for picture 0, picture 0 can be classified as one or more pictures. For example, picture 0 could be classified as the first picture in the bitstream, the first picture in CVS / CLVS, or the first picture associated with an SEI message (e.g., in relation to the SEI message). Therefore, the meaning of picture 0 becomes ambiguous.

[0385] Additionally, the current design includes features regarding whether the image associated with the SPTI SEI message is a composite image. Image 0 could be constrained to not be associated with a composite image because a composite image is not associated with a source image; that is, it does not have an associated source image. However, the current design lacks any relevant constraints, and therefore may cause problems.

[0386] Overview of the Implementation Examples

[0387] This disclosure presents various embodiments capable of solving the problems of conventional designs described above. The following embodiments can be used alone, or in combination with or in conjunction with the embodiments described above, and should also be included in this disclosure.

[0388] 1. The image 0 used for the calculation of the specific syntax of the SPTI SEI message (as an example, the syntax associated with the source image timing can be included, for example, sourcepictureTime[0]) can be specified as the image in the same access unit (AU) containing the SEI message.

[0389] 2. It can be specified that image 0 should not be a constraint for the composite image because it does not have a corresponding source image.

[0390] 3. The time identifier (TemporalId) of image 0 may be constrained to a specific value (e.g., 0).

[0391] In the following description, various embodiments including the above embodiments are described in more detail, and improvements to the images in the Neural Network Post-Filter (NNPF) SEI message for the encoded video bitstream are presented. Although the embodiments described below are based on standard video codecs (e.g., VVC (Video Coding Universal) and the VSEI (Very General Supplemental Enhancement Information Message for Encoded Video Bitstreams), it is apparent that they can be applied to other video coding techniques, which should also be included in this disclosure.

[0392] In describing the embodiments below, the Neural Network Post-Filter (NNPF) SEI message or the NNPF-related SEI message may include the Neural Network Post-Filter Feature (NNPFC) SEI message and / or the Neural Network Post-Filter Activation (NNPFA) SEI message.

[0393] Since the names of the syntaxes used to describe the following embodiments are arbitrarily assigned for clarity of description, it is obvious that the names of the syntaxes can be changed, and even if the names of the syntaxes are changed, they should still be included in this disclosure.

[0394] In the following, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0395] Example

[0396] In the following embodiments, embodiments of items 1, 2, and 3 described in the overview of the embodiments will be described in detail. VSEI message syntax and semantics will be described below.

[0397] As an example, the syntax and semantics of SEI messages (e.g., NNPFASEI messages, etc.) related to neural network post-filters (NNPF) can be modified in part.

[0398] As an example, assume that picture 0 is the picture of the same access unit as the SPTI SEI message, and picture N is the Nth picture from picture 0, where N is greater than 0. As an example, when the TEMPORALID of picture N is less than or equal to any value i and is not the first picture of the bitstream, the source picture timing information can be derived as follows (e.g., SourcePictureTime[n]).

[0399] When specific information (e.g., temporalReversalFlag) equals a specific value (e.g., 0), the source picture timing information SourcePictureTime[n] can be derived based on SourcePictureTime[previous PicInOutputOrder] and information about the source picture interval. For example, SourcePictureTime[n] can be derived based on the addition operation of SourcePictureTime[previousPicInOutputOrder] and SourcePictureInterval[i] (Source Picture Time[previousPicInOutputOrder] + SourcePictureInterval[i]).

[0400] When specific information (e.g., temporalReversalFlag) equals a specific value (e.g., 1), the source picture timing information SourcePictureTime[n] can be derived based on SourcePictureTime[previous PicInOutputOrder] and information about the source picture interval. For example, SourcePictureTime[n] can be derived based on the subtraction operation between SourcePictureTime[previous PicInOutputOrder] and SourcePictureInterval[i] (Source Picture Time[previous PicInOutputOrder] - SourcePictureInterval[i]).

[0401] Here, as an example, `previous PicInOutputOrder` can be a time identifier (Temporal ID) that is less than or equal to a specific threshold (e.g., denoted as `i`) and is the last output image before image `n` in the output order. When a value for the source picture timing information (Source Picture Time[0]) is not provided based on other syntax or other information, `SourcePictureTime[0]` can be inferred to be equal to a specific value (e.g., 0). Furthermore, as an example, the time identifier for image 0 (e.g., anchor image, reference image, first source timing image) can be constrained to a specific value (e.g., 0).

[0402] As an example, when present, a synthesized image-related information spti_sublayer_syntheed_picture_flag[i] equal to 1 can indicate that the decoded output image belonging to the i-th temporal sublayer does not correspond to the unmodified original source image. Conversely, information equal to 0 may not provide such an indication. Furthermore, when the syntax does not exist, the value of the syntax can be inferred to be equal to a specific value (e.g., 0).

[0403] Image 0 (e.g., reference image, anchor image, first source timing image, etc.) can be constrained to be an image that does not belong to the composition time layer.

[0404] The problems of the conventional technology described in this disclosure can be solved. Furthermore, by diversifying the variables, syntax, and / or semantics of the information, clarifying the constraints on the information, or clarifying the semantics of the information, decoder errors can be reduced and encoding quality and efficiency can be improved.

[0405] Examples of image decoding and encoding methods

[0406] Hereinafter, image encoding and image decoding methods according to various embodiments of this application will be described. The image decoding method of FIG5 can be performed by image decoding device 200, and the image encoding method of FIG6 can be performed by image encoding device 100. Furthermore, the image decoding and encoding methods of FIG5 and FIG6 can be based on the above embodiments and the examples described below, respectively.

[0407] Figure 5 illustrates an image decoding method that can be performed by an image decoding device according to an embodiment of the present disclosure.

[0408] First, the source image timing information (SPTI) supplemental enhancement information (SEI) message can be obtained (S510). As an example, the SPTI SEI message can be sent as a bitstream using a signal, and the image can be reconstructed based on the obtained SPTI SEI message (S520).

[0409] As an example, an SPTI SEI message may include information about the timing of the source image for an image associated with the SEI. The timing of the source image can be derived from the anchor image. As an example, the anchor image may be represented as image 0, the standard image, the first source timing image, etc., and the value of the anchor image's time identifier may be equal to a specific value (e.g., 0). Furthermore, as an example, the anchor image may not belong to the composition timing layer; it may be a constraint. Additionally, the anchor image may be included in the same access unit as the access unit that includes the SPTI SEI message. The timing of the source image for a specific image can be derived based on the order relative to the anchor images. As an example, the order may include the output order relative to the anchor images. Other descriptions are the same as those given above, and therefore redundant descriptions are omitted.

[0410] Since the image decoding method in Figure 5 corresponds to an embodiment of this disclosure, it will be apparent that certain steps can be changed, the order of steps can be changed, or some steps can be added or removed, and such modifications are also within the scope of this disclosure.

[0411] Figure 6 illustrates an image encoding method that can be performed by an image encoding apparatus according to an embodiment of the present disclosure. The image encoding method of Figure 6 can also be performed based on the following description and the above embodiments. The description based on the decoder above can also be applied to the image encoding method, as long as it does not conflict with the operation of the encoder.

[0412] As an example, a Source Picture Timing Information (SPTI) Supplemental Enhancement Information (SEI) message can be determined (S610). Determining the SPTISEI message may include determining the information to be included in the SPTI SEI message. Subsequently, the bitstream including the SPTI SEI message can be encoded (S620). In other words, the SPTI SEI message can be encoded, and the SPTI SEI message can be transmitted using signals.

[0413] As described above, as an example, the SPTI SEI message can include information about the timing of the source image for the image associated with the SEI. The timing of the source image can be determined based on the anchor image. As an example, the anchor image can be represented as image 0, the standard image, the first source timing image, etc., and the value of the anchor image's time identifier can be equal to a specific value (e.g., 0). Furthermore, as an example, the anchor image may not belong to the composition timing layer; it can be a constraint. Additionally, the anchor image can be included in the same access unit as the access unit that includes the SPTI SEI message. The timing of the source image for a specific image can be determined based on the order relative to the anchor images. As an example, the order can include the output order relative to the anchor images. Other descriptions are the same as those given above, and therefore redundant descriptions are omitted.

[0414] Subsequently, although not shown in the attached diagram, the image can be reconstructed based on the output image information.

[0415] Furthermore, as an example, a computer-readable medium on which a bitstream generated by an image encoding method is recorded can be provided, and a method for sending the bitstream generated by the image encoding method can be provided.

[0416] Since the image encoding method in Figure 6 corresponds to an embodiment of this disclosure, it will be apparent that certain steps can be changed, the order of steps can be changed, or some steps can be added or removed, and such modifications are also within the scope of this disclosure.

[0417] According to this disclosure, the meaning of information that can be included in the VSEI can be clarified, thereby reducing decoder errors and representing more accurate scenarios. Therefore, encoding quality can be improved. Furthermore, according to this disclosure, clear processing can be achieved even when a corresponding output image is generated but an input image is not available. Therefore, encoding efficiency can be improved.

[0418] Figure 7 is a diagram illustrating an exemplary content streaming system to which embodiments of the present disclosure are applicable.

[0419] As shown in Figure 7, the content streaming system using embodiments of this disclosure may mainly include an encoding server, a streaming server, a web server, media storage, user equipment, and a multimedia input device.

[0420] An encoding server compresses content input from multimedia input devices such as smartphones, cameras, and camcorders into digital data to generate a bitstream, and then sends the bitstream to a streaming server. As another example, when multimedia input devices such as smartphones, cameras, and camcorders generate bitstreams directly, the encoding server can be omitted.

[0421] The bitstream can be generated by the image encoding method or image encoding apparatus 100 applying the embodiments of this disclosure, and the streaming server can temporarily store the bitstream during the sending or receiving of the bitstream.

[0422] A streaming server sends multimedia data to a user's device based on a user's request via a web server, and the web server acts as a medium for notifying the user of services. When a user requests a desired service from the web server, the web server can deliver it to the streaming server, and the streaming server can send the multimedia data to the user. In this scenario, the content streaming system may include a separate control server. In this case, the control server acts as a command / response controller between devices in the content streaming system.

[0423] A streaming server can receive content from media storage and / or encoding servers. For example, when content is received from an encoding server, it can be received in real time. In this case, to provide a smooth streaming service, the streaming server can store the bitstream for a predetermined period of time.

[0424] Examples of user equipment may include mobile phones, smartphones, laptops, digital broadcasting terminals, personal digital assistants (PDAs), portable multimedia players (PMPs), navigation devices, tablet PCs, tablet computers, ultrabooks, wearable devices (e.g., smartwatches, smart glasses, head-mounted displays), digital televisions, desktop computers, digital signage, etc.

[0425] In a content streaming system, each server can operate as a distributed server, in which case the data received from each server can be distributed.

[0426] The scope of this disclosure includes software or machine-executable instructions (e.g., operating systems, applications, firmware, programs, etc.) for enabling the operation of methods according to various embodiments to be executed on a device or computer, and non-transitory computer-readable media having such software or instructions stored thereon and executable on a device or computer.

[0427] Industrial applicability

[0428] The embodiments disclosed herein can be used to encode or decode images.

Claims

1. An image decoding method, comprising: Obtain Source Image Timing Information (SPTI) Supplemental Enhancement Information (SEI) messages, and reconstruct an image based on the SPTI SEI messages, wherein the SPTI SEI messages include information about source image timing for an image related to the SEI.

2. The image decoding method according to claim 1, wherein, The source image is exported based on the anchor image at a set time.

3. The image decoding method according to claim 2, wherein, The time identifier of the anchor image has a value of 0.

4. The image decoding method according to claim 2, wherein, The anchor image is constrained to not belong to the composite time layer.

5. The image decoding method according to claim 2, wherein, The anchor image is included in the same access unit as the access unit that includes the SPTISEI message.

6. The image decoding method according to claim 2, wherein, The source image is exported at a specific time based on the order relative to the anchor images.

7. An image encoding method, comprising: Determine the Source Image Timing Information (SPTI) Supplemental Enhancement Information (SEI) message, and encode the bitstream including the SPTI SEI message, wherein the SPTI SEI message includes information about the source image timing for the image associated with the SEI.

8. The image encoding method according to claim 7, wherein, The source image is exported based on the anchor image at a set time.

9. The image encoding method according to claim 8, wherein, The time identifier of the anchor image has a value of 0.

10. A non-transitory computer-readable medium comprising a bitstream generated by an image encoding method, the image encoding method comprising: Determine the Source Image Timing Information (SPTI) Supplemental Enhancement Information (SEI) message, and encode the bitstream including the SPTI SEI message, wherein the SPTI SEI message includes information about the source image timing for the image associated with the SEI.

11. A method for transmitting a bitstream generated by an image encoding method, the method comprising: The bitstream is transmitted, wherein the image encoding method includes: determining a Source Picture Timing Information (SPTI) Supplemental Enhancement Information (SEI) message, and encoding the bitstream including the SPTI SEI message, wherein the SPTI SEI message includes information about the source picture timing for an image related to the SEI.

Images