Adaptive transform for compound inter-intra prediction modes
The use of composite inter-intra prediction modes with data-driven transform kernels addresses the challenge of efficient redundancy reduction in video coding, enhancing compression efficiency and distortion management.
Patent Information
- Application Number
- JP2024131217
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-09-06
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-06
AI Technical Summary
Existing video coding techniques face challenges in efficiently reducing redundancy and achieving high compression ratios while maintaining acceptable distortion levels, particularly in complex inter-intra prediction modes, which are crucial for handling large volumes of video data.
The implementation of composite inter-intra prediction (CIIP) modes that combine intra-prediction and inter-prediction, utilizing data-driven transform kernels such as LGT, RCT, and pre-trained transformation kernels to optimize video block decoding and encoding processes.
Enhances compression efficiency and reduces distortion by adaptively applying transform kernels based on video block characteristics, leading to improved data utilization and reduced bandwidth requirements.
Smart Images

Figure 2024156935000001_ABST
Abstract
Description
[Technical field]
[0001] This disclosure relates generally to a set of advanced video coding / decoding techniques, and more specifically to transformation techniques and configurations for mixed inter-intra prediction modes. [Background technology]
[0002] The discussion of the background art provided herein is intended to generally present the context of the present disclosure. The inventors' work is not admitted, expressly or impliedly, as prior art to the present disclosure to the extent that that work is described in this Background section, together with aspects of the description that may not otherwise be admitted as prior art at the time of filing of this application.
[0003] Video coding and decoding can be performed using inter-picture prediction with motion compensation. Uncompressed digital video can include a sequence of pictures, each with spatial dimensions of, for example, 1920x1080 luma samples and associated full-sampled or sub-sampled chroma samples. The sequence of pictures can have a fixed or variable picture rate (also called frame rate), for example, 60 pictures per second or 60 frames per second. Uncompressed video has specific bit rate requirements for streaming or data processing. For example, a video with a pixel resolution of 1920x1080, a frame rate of 60 frames / second, and 4:2:0 chroma subsampling with 8 bits per pixel per color channel requires a bandwidth close to 1.5 Gbit / s. One hour of such video requires more than 600 GByte of storage space.
[0004] One goal of video coding and video decoding may be the reduction of redundancy in an uncompressed input video signal through compression. Compression may help reduce the aforementioned bandwidth and / or storage space requirements, in some cases by more than one order of magnitude. Both lossless and lossy compression, and combinations thereof, may be used. Lossless compression refers to techniques where an exact copy of the original signal can be reconstructed from the compressed original signal by the decoding process. Lossy compression refers to a coding / decoding process where the original video information is not fully preserved during coding and cannot be fully recovered during decoding. When using lossy compression, the reconstructed signal may not be identical to the original signal, but the distortion between the original and reconstructed signals is small enough to make the reconstructed signal useful for its intended application, even with some information loss. For video, lossy compression has been widely adopted in many applications. The amount of distortion that can be tolerated depends on the application. For example, a user of a particular consumer video streaming application may tolerate higher distortion than a user of a movie or television broadcast application. The compression ratio achievable by a particular coding algorithm may be selected or adjusted to reflect different distortion tolerances. That is, generally speaking, higher distortion tolerance allows for coding algorithms that result in higher losses and higher compression ratios.
[0005] Video encoders and decoders can utilize techniques from several broad categories and steps, including, for example, motion compensation, Fourier transform, quantization, and entropy coding.
[0006] Video codec techniques can include a technique known as intra-coding. In intra-coding, sample values are represented without reference to samples or other data from previously reconstructed reference pictures. In some video codecs, a picture is spatially subdivided into blocks of samples. If all blocks of samples are coded in intra mode, the picture can be called an intra-picture. Intra-pictures and their derived pictures, such as independent decoder refresh pictures, can be used to reset the decoder state and can therefore be used as the first picture in a coded video bitstream and video session or as a still image. Samples of the intra-predicted block can then be transformed to the frequency domain, and the transform coefficients so generated can be quantized before entropy coding. Intra-prediction refers to a technique that minimizes sample values in the pre-transformed domain. In some cases, the smaller the DC value after transformation and the smaller the AC coefficients, the fewer bits are needed for a given quantization step size to represent the block after entropy coding.
[0007] Conventional intra-coding, for example as known from MPEG-2 generation coding techniques, does not use intra-prediction. However, some newer video compression techniques include techniques that attempt to code / decode a block based on surrounding sample data and / or metadata that precedes the block of data being intra-coded or intra-decoded in decoding order, e.g., obtained during encoding and / or decoding of spatial neighbors. Such techniques are hereafter referred to as "intra-prediction" techniques. It should be noted that in at least some cases, intra-prediction uses reference data only from the current picture being reconstructed, and not from other reference pictures.
[0008] There may be many different forms of intra-prediction. When more than one of such techniques is available in a given video coding technique, the technique used may be referred to as an intra-prediction mode. One or more intra-prediction modes may be provided in a particular codec. In certain cases, a mode may have sub-modes and / or may be associated with various parameters, and the mode / sub-mode information and intra-coding parameters of a block of video may be coded separately or collectively included in the codeword of the mode. Which codeword is used for a given mode, sub-mode, and / or parameter combination may affect the coding efficiency gains via intra-prediction, and thus may also affect the entropy coding technique used to convert the codeword into a bitstream.
[0009] Certain modes of intra prediction were introduced in H.264, improved in H.265, and further refined in newer coding techniques such as the joint exploration model (JEM), versatile video coding (VVC), and benchmark set (BMS). In general, in intra prediction, the predictor block can be formed using neighboring sample values that become available. For example, the available values of a particular set of neighboring samples along a particular direction and / or line can be copied to the predictor block. The reference to the direction used can be coded in the bitstream or can itself be predicted.
[0010] Referring to FIG. 1A, shown at the bottom right is a subset of nine predictor directions specified in the 33 possible intra predictor directions of H.265 (corresponding to the 33 angle modes of the 35 intra modes specified in H.265). The point where the arrows converge (101) represents the sample being predicted. The arrows represent the direction from which neighboring samples are used to predict sample 101. For example, arrow (102) indicates that sample (101) is predicted from one or more neighboring samples to the upper right, at an angle of 45 degrees from the horizontal. Similarly, arrow (103) indicates that sample (101) is predicted from one or more neighboring samples to the lower left of sample (101), at an angle of 22.5 degrees from the horizontal.
[0011] 1A, at the top left is shown a square block (104) of 4×4 samples (indicated by a thick dashed line). The square block (104) contains 16 samples, each labeled with "S", its position in the Y dimension (e.g., row index), and its position in the X dimension (e.g., column index). For example, sample S21 is the second sample (from the top) in the Y dimension and the first sample (from the left) in the X dimension. Similarly, sample S44 is the fourth sample in both the Y and X dimensions in the block (104). Since the size of the block is 4×4 samples, S44 is at the bottom right. Also shown are examples of reference samples that follow a similar numbering scheme. The reference samples are labeled with R, their Y position (e.g., row index) and X position (column index) relative to the block (104). In both H.264 and H.265, predicted samples that neighbor the block being reconstructed are used.
[0012] Intra-picture prediction of block 104 may start by copying reference sample values from adjacent samples according to a signaled prediction direction. For example, assume that the coded video bitstream includes signaling for this block 104 indicating the prediction direction of the arrow (102), i.e., the sample is predicted from one or more prediction samples to the upper right and at an angle of 45 degrees from the horizontal. In such a case, samples S41, S32, S23, S14 are predicted from the same reference sample R05. Then, sample S44 is predicted from reference sample R08.
[0013] In certain cases, values of multiple reference samples may be combined, for example by interpolation, to calculate a reference sample, especially when the orientation is not evenly divisible by 45 degrees.
[0014] The number of possible directions has increased as video coding technology continues to develop. In H.264 (2003), for example, nine different directions are available for intra prediction. This increases to 33 in H.265 (2013), and JEM / VVC / BMS can support up to 65 directions at the time of this disclosure. Experimental studies have been conducted to help identify the most suitable intra prediction directions, and those most suitable directions can be coded with a small number of bits using certain techniques of entropy coding, accepting a certain bit penalty for the direction. Furthermore, the direction itself may be predicted from the neighboring directions used in the intra prediction of the decoded neighboring blocks.
[0015] FIG. 1B shows a schematic diagram (180) showing 65 intra prediction directions according to JEM to illustrate the increasing number of prediction directions in various coding techniques that have evolved over time.
[0016] Methods for mapping bits representing intra-prediction directions in a coded video bitstream to prediction directions can vary across video coding techniques and can range, for example, from simple direct mappings of prediction directions to intra-prediction modes to complex adaptation schemes involving codewords, most-probable modes, and similar techniques. In all cases, however, there may be certain directions of intra-prediction that are statistically less likely to occur in the video content than certain other directions. Because the goal of video compression is to reduce redundancy, in a well-designed video coding technique, these less likely directions may be represented with more bits than the more likely directions.
[0017] Inter-picture prediction, or inter-prediction, may be based on motion compensation, in which sample data from a previously reconstructed picture or part thereof (reference picture) may be used to predict a newly reconstructed picture or picture part (e.g., block) after being spatially shifted in a direction indicated by a motion vector (hereafter MV). In some cases, the reference picture may be the same as the picture currently being reconstructed. The MV may have two dimensions X and Y, or three dimensions, with the third dimension being an indication of the reference picture to be used (similar to the temporal dimension).
[0018] In some video compression techniques, a current MV applicable to a particular area of sample data can be predicted from other MVs, e.g., from other MVs related to other areas of sample data that are spatially adjacent to the area being reconstructed and that precede the current MV in decoding order. Doing so can significantly reduce the overall amount of data required to code the MV by relying on the removal of redundancy in correlated MVs, thereby increasing compression efficiency. MV prediction can work effectively because, for example, when coding an input video signal derived from a camera (known as natural video), areas larger than the area to which a single MV is applicable have a statistical likelihood to move in a similar direction in the video sequence, and therefore, in some cases, can be predicted using similar motion vectors derived from MVs of neighboring areas. As a result, the actual MV of a given area is similar or identical to the MV predicted from the surrounding MVs. Such MVs can further be represented with fewer bits after entropy coding than would be used if the MV was directly coded instead of predicted from the neighboring MV(s). In some cases, MV prediction can be an example of lossless compression of a signal (i.e., an MV) derived from an original signal (i.e., a sample stream). In other cases, the MV prediction itself may be non-lossy, for example due to rounding errors when computing a predictor from several surrounding MVs.
[0019] Various MV prediction mechanisms are described in H.265 / HEVC (ITU-T Rec. H.265, "High Efficiency Video Coding", December 2016). Among the many MV prediction mechanisms specified by H.265, the one described below is a technique that we will refer to as "spatial merging" from here on.
[0020] Specifically, referring to Figure 2, a current block (201) contains samples that were detected by the encoder during the motion search process as predictable from a previous block of the same size but spatially shifted. Instead of coding its MV directly, the MV can be derived from metadata associated with one or more reference pictures, e.g., from the last reference picture (in decoding order), using the MV associated with any one of the five surrounding samples represented as A0, A1, and B0, B1, B2 (202 to 206, respectively). In H.265, MV prediction can use predictors from the same reference picture that neighboring blocks use. Summary of the Invention [Means for solving the problem]
[0021] Aspects of the present disclosure relate generally to a set of advanced video coding / decoding techniques, and more specifically, to transformation techniques and configurations for mixed inter-intra prediction modes.
[0022] In some example implementations, a method for decoding a video block in a video stream is disclosed. The method may include determining that the video block is predicted in a composite inter-intra prediction (CIIP) mode, and the video block is derived as a combination of intra prediction and inter prediction in the CIIP mode. The method may further include determining at least one inter-intra prediction weighting configuration for the video block, identifying a data-driven transform kernel for the video block from a set of transform kernels, generating a residual block of the video block by inverse transforming a set of transform coefficients extracted from a video stream for the video block using at least the data-driven transform kernel, and deriving the video block based on the residual block, the at least one inter-intra prediction weighting configuration, the at least one reconstructed inter prediction block of the video block, and the at least one reconstructed intra prediction reference sample of the video block.
[0023] In such implementations, the data-driven transformation kernels may include pre-trained transformation kernels or derived transformation kernels based on data distribution within the video block.
[0024] In any one of the aforementioned implementations, the data-driven transformation kernel includes one of a data-driven line graph transformation (LGT) kernel, a pre-trained Coronen-Loeve transform kernel, or a pre-trained row-column transformation (RCT) kernel.
[0025] In any one of the aforementioned implementations, the data-driven transformation kernel includes a one-dimensional separable transformation kernel.
[0026] In any one of the aforementioned implementations, the data-driven transformation kernel includes a two-dimensional non-separable transformation kernel.
[0027] In any one of the aforementioned implementations, the method further includes identifying a non-data-driven transform kernel for the video block from among the set of transform kernels, and generating a residual block of the video block may include transforming a set of transform coefficients extracted from the video stream for the video block using a data-driven transform kernel in one of two dimensions of the video block and a non-data-driven transform kernel in the other of the two dimensions of the video block to generate the residual block of the video block.
[0028] In any one of the aforementioned implementations, the method may further include determining a CIIP sub-mode for the video block from among a plurality of CIIP sub-modes, the plurality of CIIP sub-modes may correspond to a subset of intra-prediction modes of the CIIP mode, and the data-driven transform kernel is selected from the transform kernel set based on the CIIP sub-mode.
[0029] In any one of the above implementations, multiple CIIP sub-modes map to the same data-driven transformation kernel.
[0030] In any one of the aforementioned implementations, the method further includes identifying a non-data-driven transform kernel, where the CIIP submode corresponds to a VERTICAL_PRED intra prediction mode, and generating a residual block for the video block includes inverse transforming a set of transform coefficients extracted from the video stream for the video block using a data-driven transform kernel in a vertical dimension of the video block and a non-data-driven transform kernel in a horizontal dimension of the video block to generate the residual block for the video block.
[0031] In any one of the aforementioned implementations, the method further includes identifying a non-data-driven transform kernel, where the CIIP submode corresponds to a HORIZONTAL_PRED intra prediction mode, and generating a residual block for the video block includes inverse transforming a set of transform coefficients extracted from the video stream for the video block using a data-driven transform kernel in a horizontal dimension of the video block and a non-data-driven transform kernel in a vertical dimension of the video block to generate the residual block for the video block.
[0032] In any one of the aforementioned implementations, identifying a data-driven transform kernel for the video block from among the set of transform kernels is based on a size of the video block or a shape of the video block.
[0033] In any one of the above implementations, the data-driven transform kernel includes an LGT kernel, and a self-loop ratio of the LGT kernel is indicated by a size or shape of the video block.
[0034] In any one of the aforementioned implementations, the size of the video block may include the width, height, maximum width and height, minimum width and height, or area of the video block, and the shape of the video block may include the width-height aspect ratio, the height-width aspect ratio, the maximum width-height aspect ratio and height-width aspect ratio, or the minimum width-height aspect ratio and height-width aspect ratio.
[0035] In any one of the aforementioned implementations, the at least one inter-intra prediction weighting configuration indicates whether the inter-intra prediction spatial weights for the video block should be explicitly signaled within the video stream or derived from one of a set of pre-defined inter-intra spatial weighting patterns.
[0036] In any one of the aforementioned implementations, the step of identifying a data-driven transform kernel for the video block from among the transform kernel set is responsive to at least one inter-intra prediction weighting configuration indicating that inter-intra prediction spatial weights for the video block are explicitly signaled.
[0037] In any one of the aforementioned implementations, the step of identifying a data-driven transform kernel for the video block from among the set of transform kernels is responsive to at least one inter-intra prediction weighting configuration indicating that inter-intra prediction spatial weights for the video block are derived from one of a set of predetermined inter-intra spatial weighting patterns.
[0038] In any one of the aforementioned implementations, one of the set of predefined inter-intra spatial weighting patterns is indicated as part of the at least one inter-intra prediction weighting configuration, and a data-driven transform kernel of the set of transform kernels is identified according to one of the set of predefined inter-intra spatial weighting patterns as indicated in the at least one inter-intra prediction weighting configuration.
[0039] In some other implementations, a method for encoding a video block in a video stream is disclosed. The method may include determining that the video block should be predicted in a composite inter-intra prediction (CIIP) mode, where the video block is derived as a combination of intra prediction and inter prediction in the CIIP mode. The method may further include determining at least one inter-intra prediction weighting configuration for the video block, selecting a data-driven transform kernel for the video block from among a set of transform kernels, generating a residual block of the video block based on at least the video block, the at least one inter-intra prediction weighting configuration, the at least one inter-prediction block for the video block, and at least one intra-prediction reference sample for the video block, transforming the residual block using at least the data-driven transform kernel to generate a set of transform coefficients for the video block, and quantizing and entropy coding the set of transform coefficients for inclusion in the video stream with the inter-intra prediction weighting configuration after being encoded.
[0040] In some implementations, a video device is disclosed that can include a memory for storing computer instructions and processing circuitry configured to execute the computer instructions to perform each of the methods described above.
[0041] Aspects of the present disclosure also provide a non-transitory computer-readable medium storing instructions that, when executed by a computer for decoding and / or encoding video, cause the computer to perform any one of the implementations of the above methods for decoding and / or encoding video.
[0042] Further features, nature and various advantages of the disclosed subject matter will become more apparent from the following detailed description and the accompanying drawings. [Brief description of the drawings]
[0043] [Figure 1A] FIG. 13 is a schematic diagram of an example subset of intra-prediction directional modes. [Figure 1B] FIG. 2 illustrates an exemplary intra-prediction direction. [Diagram 2] FIG. 2 is a schematic diagram illustrating a current block and its surrounding spatial merge candidates for motion vector prediction in one example. [Diagram 3] FIG. 3 is a schematic diagram illustrating a simplified block diagram of a communication system (300) according to an example embodiment. [Figure 4] FIG. 4 is a schematic diagram illustrating a simplified block diagram of a communication system (400) according to an example embodiment. [Diagram 5] FIG. 2 is a schematic diagram illustrating a simplified block diagram of a video decoder according to an example embodiment. [Figure 6] FIG. 1 is a schematic diagram illustrating a simplified block diagram of a video encoder according to an example embodiment. [Figure 7] FIG. 2 is a block diagram illustrating a video encoder according to another example embodiment. [Figure 8] FIG. 2 is a block diagram illustrating a video decoder according to another example embodiment. [Figure 9] FIG. 2 illustrates a coding block partitioning scheme according to an exemplary embodiment of the present disclosure. [Figure 10] FIG. 13 illustrates another scheme for coding block partitioning according to an exemplary embodiment of the present disclosure. [Figure 11] FIG. 13 illustrates another scheme for coding block partitioning according to an exemplary embodiment of the present disclosure. [Figure 12] FIG. 13 illustrates another scheme for coding block partitioning according to an exemplary embodiment of the present disclosure. [Figure 13] 2 illustrates a scheme for splitting a coding block into multiple transform blocks and the coding order of the transform blocks according to an exemplary embodiment of the present disclosure. [Figure 14]4A-4C are diagrams illustrating another manner of dividing a coding block into multiple transform blocks and the coding order of the transform blocks according to an exemplary embodiment of the present disclosure. [Figure 15] FIG. 2 illustrates another scheme for splitting a coding block into multiple transform blocks, according to an exemplary embodiment of the present disclosure. [Figure 16] FIG. 2 illustrates various baseline-based intra-prediction schemes, according to an exemplary embodiment of the present disclosure. [Figure 17] Indicates the top, left, and top-left positions for the PAETH mode in the block. [Figure 18] 1 illustrates an exemplary recursive intra-filtering mode. [Figure 19] 1 illustrates a planar rotation transform according to an exemplary embodiment of the present disclosure. [Figure 20] 1 illustrates various DCT-2, DCT-4 partial butterfly lookup tables according to an exemplary embodiment of the present disclosure. [Figure 21] 1 illustrates a DST-7 partial butterfly lookup table according to an exemplary embodiment of the present disclosure. [Figure 22] 1 illustrates a line graph transformation according to an exemplary embodiment of the present disclosure. [Figure 23] 1 illustrates a flowchart of a method according to an exemplary embodiment of the present disclosure. [Figure 24] 1 illustrates a flowchart of a method according to an exemplary embodiment of the present disclosure. [Diagram 25] FIG. 1 illustrates a schematic diagram of a computer system according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0044] FIG. 3 illustrates a simplified block diagram of a communication system (300) according to one embodiment of the present disclosure. The communication system (300) includes a plurality of terminal devices that can communicate with each other, for example, via a network (350). For example, the communication system (300) includes a first pair of terminal devices (310) and (320) interconnected via the network (350). In the example of FIG. 3, the first pair of terminal devices (310) and (320) may perform unidirectional transmission of data. For example, the terminal device (310) may code video data (e.g., a stream of video pictures captured by the terminal device (310)) for transmission to the other terminal device (320) via the network (350). The encoded video data may be transmitted in the form of one or more coded video bitstreams. The terminal device (320) may receive the coded video data from the network (350), decode the coded video data to reconstruct the video pictures, and display the video pictures according to the reconstructed video data. Unidirectional data transmission may be implemented, for example, in media serving applications.
[0045] In another example, the communication system (300) includes a second pair of terminal devices (330) and (340) performing bidirectional transmission of coded video data, which may be implemented, for example, during video conferencing applications. For the bidirectional transmission of data, in one example, each of the terminal devices (330) and (340) may code video data (e.g., of a stream of video pictures captured by that terminal device) for transmission to the other of the terminal devices (330) and (340) over the network (350). Each of the terminal devices (330) and (340) may also receive coded video data transmitted by the other of the terminal devices (330) and (340), decode the coded video data to recover the video pictures, and display the video pictures on an accessible display device according to the recovered video data.
[0046] In the example of FIG. 3, the terminal devices (310), (320), (330), and (340) may be embodied as a server, a personal computer, and a smartphone, although the applicability of the principles underlying the present disclosure is not so limited. The embodiments of the present disclosure may be implemented in desktop computers, laptop computers, tablet computers, media players, wearable computers, dedicated video conferencing equipment, and the like. The network (350) represents any number and type of network that conveys coded video data between the terminal devices (310), (320), (330), and (340), including, for example, wired and / or wireless communication networks. The communication network (350) 9 may exchange data over circuit-switched channels, packet-switched channels, and / or other types of channels. Exemplary networks include telecommunications networks, local area networks, wide area networks, and / or the Internet. For purposes of this discussion, the architecture and topology of the network (350) may not be important to the operation of the present disclosure unless explicitly described herein.
[0047] 4 illustrates an arrangement of video encoders and video decoders in a video streaming environment as an example of an application of the disclosed subject matter. The disclosed subject matter may be equally applied to other video-enabled applications including, for example, video conferencing, digital television broadcasting, gaming, virtual reality, storage of compressed video on digital media including CDs, DVDs, memory sticks, etc.
[0048] The video streaming system may include a video source (401), such as a video capture subsystem (413), which may include a digital camera, for creating a stream of uncompressed video pictures or images (402). In one example, the stream of video pictures (402) includes samples recorded by the digital camera of the video source 401. The stream of video pictures (402), shown in bold to emphasize its high amount of data compared to the encoded video data (404) (or coded video bitstream), may be processed by an electronic device (420) including a video encoder (403) coupled to the video source (401). The video encoder (403) may include hardware, software, or a combination thereof to enable or implement aspects of the disclosed subject matter as described in more detail below. The encoded video data (404) (or encoded video bitstream (404)), shown with thin lines to emphasize its low amount of data compared to the stream of uncompressed video pictures (402), may be stored directly on the streaming server (405) or on a downstream video device (not shown) for future use. One or more streaming client subsystems, such as the client subsystems (406) and (408) of FIG. 4, may access the streaming server (405) to obtain copies (407) and (409) of the encoded video data (404). The client subsystem (406) may include a video decoder (410), for example, within the electronic device (430). The video decoder (410) decodes an input copy (407) of the encoded video data and creates an output stream (411) of video pictures that is uncompressed and can be rendered on a display (412) (e.g., a display screen) or other rendering device (not shown). The video decoder 410 may be configured to perform some or all of the various functions described in this disclosure.In some streaming systems, the encoded video data (404), (407), and (409) (e.g., video bitstreams) may be encoded according to a particular video coding / compression standard. Examples of such standards include ITU-T Recommendation H.265. In one example, a video coding standard under development is informally known as Versatile Video Coding (VVC). The disclosed subject matter may be used in the context of VVC, as well as other video coding standards.
[0049] It should be noted that the electronic devices (420) and (430) can include other components (not shown). For example, the electronic device (420) can include a video decoder (not shown), and the electronic device (430) can also include a video encoder (not shown).
[0050] 5 shows a block diagram of a video decoder (510) according to any of the following embodiments of the present disclosure. The video decoder (510) can be included in an electronic device (530). The electronic device (530) can include a receiver (531) (e.g., a receiving circuit). The video decoder (510) can be used in place of the video decoder (410) in the example of FIG. 4.
[0051] The receiver (531) may receive one or more coded video sequences to be decoded by the video decoder (510). In the same or another embodiment, one coded video sequence may be decoded at a time, with the decoding of each coded video sequence being independent of the other coded video sequences. Each video sequence may be associated with multiple video frames or video images. The coded video sequences may be received from a channel (501), which may be a hardware / software link to a storage device that stores the coded video data, or a streaming source that transmits the coded video data. The receiver (531) may receive the coded video data along with other data, such as coded audio data and / or auxiliary data streams, which may be forwarded to respective processing circuits (not shown). The receiver (531) may separate the coded video sequences from other data. To combat network jitter, a buffer memory (515) may be located between the receiver (531) and the entropy decoder / parser (520) (hereafter "parser (520)"). In certain applications, the buffer memory (515) may be implemented as part of the video decoder (510). In other applications, the buffer memory (515) may be separate and external to the video decoder (510) (not shown). In still other applications, there may be a buffer memory (not shown) external to the video decoder (510), for example to combat network jitter, and there may be another buffer memory (515) internal to the video decoder (510), for example to handle playback timing. When the receiver (531) is receiving data from a storage / forwarding device with sufficient bandwidth and controllability, or from an isosynchronous network, the buffer memory (515) may be unnecessary or may be small. For use with best-effort packet networks such as the Internet, a buffer memory (515) of sufficient size may be required, and its size may be relatively large.Such a buffer memory may be implemented with an adaptive size and may be implemented, at least in part, in an operating system or similar element (not shown) external to the video decoder (510).
[0052] The video decoder (510) may include a parser (520) to recover symbols (521) from the coded video sequence. These categories of symbols include information used to manage the operation of the video decoder (510) and potentially information for controlling a rendering device such as a display (512) (e.g., a display screen) that may or may not be an integral part of the electronic device (530) but may be coupled to the electronic device (530) as shown in FIG. 5. The control information for the rendering device(s) may be in the form of a supplemental enhancement information (SEI message) or a video usability information (VUI) parameter set fragment (not shown). The parser (520) may parse / entropy decode the coded video sequence received by the parser (520). The entropy coding of the coded video sequence may be according to a video coding technique or standard and according to various principles including variable length coding, Huffman coding, arithmetic coding with or without context sensitivity, etc. The parser (520) may extract from the coded video sequence a set of subgroup parameters for at least one of the subgroups of pixels in the video decoder based on at least one parameter corresponding to the subgroup. The subgroups may include groups of pictures (GOPs), pictures, tiles, slices, macroblocks, coding units (CUs), blocks, transform units (TUs), prediction units (PUs), etc. The parser (520) may also extract information from the coded video sequence, such as transform coefficients (e.g., Fourier transform coefficients), quantization parameter values, motion vectors, etc.
[0053] The parser (520) can perform entropy decoding / parsing operations on the video sequence received from the buffer memory (515) to produce symbols (521).
[0054] The reconstruction of the symbols (521) may involve a number of different processing or functional units, depending on the type of video picture or portion thereof being coded (inter-picture and intra-picture, inter-block and intra-block, etc.), as well as other factors. The units that are included and how they are included may be controlled by subgroup control information parsed from the coded video sequence by the parser (520). The flow of such subgroup control information between the parser (520) and the following processing or functional units is not shown for the sake of simplicity.
[0055] Besides the functional blocks already mentioned, the video decoder (510) can be conceptually subdivided into several functional units, as described below. In an actual implementation operating under commercial constraints, many of these functional units will interact closely with each other and may be integrated, at least in part, with each other. However, to clearly explain the various functions of the disclosed subject matter, the following disclosure adopts a conceptual subdivision into functional units.
[0056] The first unit may include a scalar / inverse transform unit (551), which may receive quantized transform coefficients as well as control information from the parser (520) including information indicating which type of inverse transform to use, block size, quantization coefficients / parameters, quantization scaling matrices, etc. The scalar / inverse transform unit (551) may output a block including sample values that may be input to an aggregator (555).
[0057] In some cases, the output samples of the scaler / inverse transform (551) may relate to intra-coded blocks, i.e., blocks that do not use prediction information from a previously reconstructed picture, but may use prediction information from a previously reconstructed portion of the current picture. Such prediction information may be provided by an intra-picture prediction unit (552). In some cases, the intra-picture prediction unit (552) may generate a block of the same size and shape as the block being reconstructed using information of surrounding blocks that have already been reconstructed and stored in a current picture buffer (558). The current picture buffer (558) may, for example, buffer a partially reconstructed and / or a fully reconstructed current picture. In some implementations, the aggregator (555) may add, on a sample-by-sample basis, the prediction information generated by the intra-prediction unit (552) to the output sample information provided by the scaler / inverse transform unit (551).
[0058] In other cases, the output samples of the scalar / inverse transform unit (551) may relate to an inter-coded, potentially motion-compensated block. In such cases, the motion compensated prediction unit (553) may access the reference picture memory (557) to fetch samples used for inter-picture prediction. After motion compensating the fetched samples according to the symbols (521) related to the block, these samples may be added to the output of the scalar / inverse transform unit (551) by the aggregator (555) to generate output sample information (the output of unit 551 may be referred to as a residual sample or residual signal). The address in the reference picture memory (557) from which the motion compensated prediction unit (553) fetches the prediction sample may be controlled by a motion vector available to the motion compensated prediction unit (553) in the form of a symbol (521) that may have, for example, an X component, a Y component (shift), and a reference picture component (time). Motion compensation may also include interpolation of sample values fetched from a reference picture memory (557) when sub-sample accurate motion vectors are used, and may be associated with a motion vector prediction mechanism, etc.
[0059] The output samples of the aggregator (555) may be subjected to various loop filtering techniques in the loop filter unit (556). Video compression techniques may include in-loop filter techniques that are controlled by parameters contained in the coded video sequence (also called the coded video bitstream) and made available to the loop filter unit (556) as symbols (521) from the parser (520), but may also be responsive to previously reconstructed and loop filtered sample values as well as meta-information obtained during decoding of previous portions (in decoding order) of the coded picture or coded video sequence. As described in more detail below, several types of loop filters may be included as part of the loop filter unit 556 in various orders.
[0060] The output of the loop filter unit (556) may be a sample stream that can be output to a rendering device (512) and also stored in a reference picture memory (557) for use in future inter-picture prediction.
[0061] Once a particular coded picture is fully reconstructed, it may be used as a reference picture for future inter-picture prediction. For example, once a coded picture corresponding to a current picture is fully reconstructed and the coded picture is identified (e.g., by the parser (520)) as a reference picture, the current picture buffer (558) can become part of the reference picture memory (557), and any unused current picture buffer can be reallocated before beginning reconstruction of the next coded picture.
[0062] The video decoder (510) may perform decoding operations according to a given video compression technique adopted in a standard such as ITU-T Rec. H.265. The coded video sequence may conform to a syntax specified by the video compression technique or standard being used, in the sense that the coded video sequence adheres to both the syntax of the video compression technique or standard and to a profile documented in the video compression technique or standard. In particular, a profile may select certain tools from among all tools available in the video compression technique or standard as tools that are dedicated to use only under that profile. To comply with a standard, the complexity of the coded video sequence may be within a range defined by a level of the video compression technique or standard. In some cases, the level limits the maximum picture size, maximum frame rate, maximum reconstruction sample rate (e.g., measured in megasamples per second), maximum reference picture size, etc. The limits set by the level may be further limited in some cases by the specification of a hypothetical reference decoder (HRD) and metadata for HRD buffer management signaled within the coded video sequence.
[0063] In some example embodiments, the receiver (531) may receive additional (redundant) data along with the encoded video. The additional data may be included as part of the coded video sequence(s). The additional data may be used by the video decoder (510) to properly decode the data and / or to more accurately recover the original video data. The additional data may be in the form of, for example, temporal, spatial, or signal-to-noise ratio (SNR) enhancement layers, redundant slices, redundant pictures, forward error correction codes, etc.
[0064] 6 shows a block diagram of a video encoder (603) according to an exemplary embodiment of the present disclosure. The video encoder (603) may be included in an electronic device (620). The electronic device (620) may further include a transmitter (640) (e.g., a transmission circuit). The video encoder (603) may be used in place of the video encoder (403) of the example of FIG.
[0065] The video encoder (603) may receive video samples from a video source (601) (which is not part of the electronic device (620) in the example of FIG. 6) that may capture video image(s) to be coded by the video encoder (603). In another example, the video source (601) may be implemented as part of the electronic device (620).
[0066] The video source (601) may provide a source video sequence to be coded by the video encoder (603) in the form of a digital video sample stream that may be of any suitable bit depth (e.g., 8-bit, 10-bit, 12-bit, ...), any color space (e.g., BT.601 Y CrCb, RGB, XYZ...), and any suitable sampling structure (e.g., Y CrCb 4:2:0, Y CrCb 4:4:4). In a media serving system, the video source (601) may be a storage device that can store previously prepared video. In a video conferencing system, the video source (601) may be a camera that captures local image information as a video sequence. The video data may be provided as a number of separate pictures or images that give motion when viewed in sequence. The picture itself may be organized as a spatial array of pixels, each pixel may contain one or more samples, depending on the sampling structure, color space, etc. used. Those skilled in the art can easily understand the relationship between pixels and samples. The following description focuses on samples.
[0067] According to some example embodiments, the video encoder (603) may code and compress pictures of a source video sequence into a coded video sequence (643) in real time or under any other time constraint required by the application. Enforcing an appropriate coding rate constitutes one function of the controller (650). In some embodiments, the controller (650) may be operatively coupled to and control other functional units, as described below. For simplicity, coupling is not shown. Parameters set by the controller (650) may include rate control related parameters (picture skip, quantizer, lambda value of rate distortion optimization technique, etc.), picture size, group of pictures (GOP) layout, maximum motion vector search range, etc. The controller (650) may be configured to have other appropriate functions associated with the video encoder (603) optimized for a particular system design.
[0068] In some example embodiments, the video encoder (603) may be configured to operate in a coding loop. As an oversimplified explanation, in one example, the coding loop may include a source coder (630) (e.g., responsible for creating symbols, such as a symbol stream, based on an input picture to be coded and a reference picture(s)) and a (local) decoder (633) embedded in the video encoder (603). The decoder (633) reconstructs the symbols to create sample data in a similar manner as a (remote) decoder would create them, even if the embedded decoder 633 processes a video stream coded by the source coder 630 without entropy coding (because in the video compression techniques considered in the disclosed subject matter, any compression between the symbols in the entropy coding and the coded video bitstream may be lossless). The reconstructed sample stream (sample data) is input to a reference picture memory (634). Since decoding of the symbol stream leads to bit-exact results regardless of the location of the decoder (local or remote), the contents in the reference picture memory (634) are also bit-exact between the local and remote encoders. In other words, the predictive part of the encoder "sees" exactly the same sample values as the reference picture samples that the decoder will "see" when using the prediction during decoding. This basic principle of reference picture synchrony (and the resulting drift if synchrony cannot be maintained, e.g., due to channel errors) is used to improve coding quality.
[0069] The operation of the "local" decoder (633) may be the same as that of a "remote" decoder, such as the video decoder (510), already described in detail above in conjunction with Figure 5. Referring also briefly to Figure 5, however, because symbols are available and the encoding / decoding of symbols into a coded video sequence by the entropy coder (645) and parser (520) may be lossless, the entropy decoding portion of the video decoder (510), including the buffer memory (515) and parser (520), may not be fully implemented in the local decoder (633) within the encoder.
[0070] At this point, it can be said that any decoder technology, except for parsing / entropy decoding, which may only exist in the decoder, may also necessarily need to exist in the corresponding encoder in substantially the same functional form. For this reason, the subject matter of the disclosure may focus on the decoder operation, which is similar to the decoding part of the encoder. Thus, the description of the encoder technology may be omitted, since it is the reverse of the decoder technology described in general. Only in certain areas or aspects, a more detailed description of the encoder is given below.
[0071] In operation, in some example implementations, the source coder (630) may perform motion-compensated predictive coding, which predictively codes an input picture with reference to one or more previously coded pictures from a video sequence designated as "reference pictures." In this manner, the coding engine (632) codes color channel differences (or residuals) between pixel blocks of the input picture and pixel blocks of the reference picture(s) that may be selected as predictive reference(s) to the input picture. The terms "residue" and its adjective form "residual" may be used interchangeably.
[0072] The local video decoder (633) may decode the coded video data of pictures that may be designated as reference pictures based on the symbols created by the source coder (630). The operation of the coding engine (632) may advantageously be a lossy process. When the coded video data may be decoded in a video decoder (not shown in FIG. 6), the reconstructed video sequence may be a replica of the source video sequence, usually with some errors. The local video decoder (633) may replicate the decoding process that may be performed by the video decoder on the reference pictures and store the reconstructed reference pictures in a reference picture cache (634). In this way, the video encoder (603) may locally store (without transmission errors) copies of reconstructed reference pictures that have common content with reconstructed reference pictures obtained by a far-end (remote) video decoder.
[0073] The predictor (635) may perform a prediction search for the coding engine (632). That is, for a new picture to be coded, the predictor (635) may search the reference picture memory (634) for sample data (as candidate reference pixel blocks) or specific metadata such as reference picture motion vectors, block shapes, etc., that may serve as suitable prediction references for the new picture. The predictor (635) may operate on sample blocks on a pixel block by pixel block basis to find suitable prediction references. In some cases, as determined by the search results obtained by the predictor (635), the input picture may have prediction references drawn from multiple reference pictures stored in the reference picture memory (634).
[0074] The controller (650) can manage the coding operations of the source coder (630), including, for example, setting the parameters and subgroup parameters used to encode the video data.
[0075] The output of all the aforementioned functional units may undergo entropy coding in an entropy coder (645), which converts the symbols produced by the various functional units into a coded video sequence by lossless compression of the symbols according to techniques such as Huffman coding, variable length coding, arithmetic coding, etc.
[0076] The transmitter (640) can buffer the coded video sequence created by the entropy coder (645) and prepare it for transmission over a communication channel (660), which may be a hardware / software link to a storage device that stores the encoded video data. The transmitter (640) can merge the coded video data from the video coder (603) with other data to be transmitted, such as coded audio data and / or auxiliary data streams (sources not shown).
[0077] The controller (650) can manage the operation of the video encoder (603). During coding, the controller (650) can assign a particular coded picture type to each coded picture, which may affect the coding techniques that may be applied to the respective picture. For example, pictures may often be assigned as one of the following picture types:
[0078] An intra picture (I-picture) may be a picture that can be coded and decoded without using any other picture in a sequence as a source of prediction. Some video codecs allow different types of intra pictures, including, for example, independent decoder refresh ("IDR") pictures. Those skilled in the art are aware of these variations of I-pictures and their respective uses and characteristics.
[0079] A predictive picture (P picture) may be a picture that can be coded and decoded using intra- or inter-prediction, which uses at most one motion vector and reference index to predict sample values for each block.
[0080] A bidirectionally predicted picture (B-picture) may be a picture that can be coded and decoded using intra- or inter-prediction, which uses at most two motion vectors and reference indices to predict the sample values of each block. Similarly, a multi-prediction picture may use more than two reference pictures and associated metadata for the reconstruction of a single block.
[0081] A source picture may generally be spatially subdivided into multiple sample coding blocks (e.g., blocks of 4x4, 8x8, 4x8, or 16x16 samples, respectively) and coded block by block. Blocks may be predictively coded with reference to other (already coded) blocks as determined by the coding assignment applied to the block's respective picture. For example, blocks of an I picture may be non-predictively coded or predictively coded with reference to already coded blocks of the same picture (spatial or intra prediction). Pixel blocks of a P picture may be predictively coded via spatial prediction with reference to one previously coded reference picture or via temporal prediction. Blocks of a B picture may be predictively coded by spatial prediction or by temporal prediction with reference to one or two previously coded reference pictures. Source pictures or intermediate processed pictures may be subdivided into other types of blocks for other purposes. The division of coding blocks and other types of blocks may or may not follow the same method, as described in more detail below.
[0082] The video encoder (603) may perform coding operations in accordance with a given video coding technique or standard, such as ITU-T Rec. H.265. In its operations, the video encoder (603) may perform various compression operations, including predictive coding operations that exploit temporal and spatial redundancy in the input video sequence. Thus, the coded video data may conform to a syntax specified by the video coding technique or standard being used.
[0083] In some example embodiments, the transmitter (640) may transmit additional data along with the encoded video. The source coder (630) may include such data as part of the coded video sequence. The additional data may include temporal / spatial / SNR enhancement layers, other forms of redundant data such as redundant pictures or slices, SEI messages, VUI parameter set fragments, etc.
[0084] Video may be captured in time sequence as multiple source pictures (video pictures). Intra-picture prediction (often abbreviated as intra-prediction) exploits spatial correlation in a given picture, while inter-picture prediction exploits temporal or other correlation between pictures. For example, a particular picture being coded / decoded, called the current picture, may be divided into blocks. If a block in the current picture resembles a reference block in a previously coded yet buffered reference picture in the video, it may be coded by a vector, called a motion vector. A motion vector points to a reference block in a reference picture and may have a third dimension that identifies the reference picture if multiple reference pictures are used.
[0085] In some exemplary embodiments, bi-prediction techniques can be used for inter-picture prediction. According to such bi-prediction techniques, two reference pictures, such as a first reference picture and a second reference picture, are used, both of which advance the current picture in the video in decoding order (but may be in the past or future, respectively, in display order). A block in the current picture can be coded by a first motion vector that points to a first reference block in the first reference picture and a second motion vector that points to a second reference block in the second reference picture. A block can be jointly predicted by a combination of the first and second reference blocks.
[0086] Furthermore, merge mode techniques may be used to improve coding efficiency in inter-picture prediction.
[0087] According to some exemplary embodiments of the present disclosure, predictions such as inter-picture prediction and intra-picture prediction are performed on a block-by-block basis. For example, a picture in a sequence of video pictures is divided into coding tree units (CTUs) for compression, and the CTUs in a picture may have the same size, such as 64×64 pixels, 32×32 pixels, or 16×16 pixels. In general, a CTU may include three parallel coding tree blocks (CTBs), namely, one luma CTB and two chroma CTBs. Each CTU may be recursively quad-tree partitioned into one or more coding units (CUs). For example, a CTU of 64×64 pixels may be partitioned into one CU of 64×64 pixels or four CUs of 32×32 pixels. Each of one or more of the 32×32 blocks may be further partitioned into four CUs of 16×16 pixels. In some example embodiments, each CU may be analyzed during encoding to determine its prediction type among various prediction types, such as inter prediction type and intra prediction type. A CU may be divided into one or more prediction units (PUs) according to temporal and / or spatial predictability. In general, each PU includes one luma prediction block (PB) and two chroma PBs. In one embodiment, prediction operations in coding (encoding / decoding) are performed on a prediction block basis. The division of a CU into PUs (or PBs of different color channels) may be performed in various spatial patterns. A luma PB or a chroma PB may include a matrix of sample values (e.g., luma values), such as, for example, 8×8 pixels, 16×16 pixels, 8×16 pixels, 16×8 pixels, etc.
[0088] 7 shows a diagram of a video encoder (703) according to another exemplary embodiment of this disclosure. The video encoder (703) is configured to receive a processed block (e.g., a predictive block) of sample values in a current video picture in a sequence of video pictures and to encode the processed block into a coded picture that is part of a coded video sequence. The exemplary video encoder (703) may be used in place of the example video encoder (403) of FIG. 4.
[0089] For example, the video encoder (703) receives a matrix of sample values for a processing block, such as a prediction block of 8x8 samples. The video encoder (703) then determines, for example using rate distortion optimization (RDO), whether the processing block is best coded using intra-mode, inter-mode, or bi-predictive mode. If it is determined that the processing block is coded in intra-mode, the video encoder (703) may use intra-prediction techniques to encode the processing block into a coded picture, and if it is determined that the processing block is coded in inter-mode or bi-predictive mode, the video encoder (703) may use inter-prediction techniques or bi-prediction techniques to encode the processing block into a coded picture, respectively. In some exemplary embodiments, a merge mode may be used as a sub-mode of inter-picture prediction, in which motion vectors are derived from one or more motion vector predictors without the benefit of coded motion vector components outside the predictors. In some other exemplary embodiments, there may be motion vector components applicable to the current block. Accordingly, the video encoder (703) may include components not explicitly shown in FIG. 7, such as a mode decision module, to determine the prediction mode of a processing block.
[0090] In the example of FIG. 7, the video encoder (703) includes an inter-encoder (730), an intra-encoder (722), a residual calculator (723), a switch (726), a residual encoder (724), a general controller (721), and an entropy encoder (725), coupled together as shown in the exemplary configuration of FIG.
[0091] The inter encoder (730) is configured to receive samples of a current block (e.g., a processing block), compare the block to one or more reference blocks in a reference picture (e.g., blocks in previous and subsequent pictures in display order), generate inter prediction information (e.g., a description of redundancy information, motion vectors, merge mode information according to an inter coding technique), and calculate an inter prediction result (e.g., a predicted block) based on the inter prediction information using any suitable technique. In some examples, the reference picture is a decoded reference picture decoded based on the encoded video information using a decoding unit 633 incorporated in the example encoder 620 of FIG. 6 (shown as a residual decoder 728 of FIG. 7, as described in more detail below).
[0092] The intra encoder (722) is configured to receive samples of a current block (e.g., a processing block), compare the block to already coded blocks in the same picture, generate transformed quantized coefficients, and possibly also generate intra prediction information (e.g., intra prediction direction information according to one or more intra encoding techniques). The intra encoder (722) may calculate an intra prediction result (e.g., a predicted block) based on the intra prediction information and a reference block in the same picture.
[0093] The generic controller (721) may be configured to determine generic control data and control other components of the video encoder (703) based on the generic control data. In one example, the generic controller (721) determines a prediction mode of a block and provides a control signal to the switch (726) based on the prediction mode. For example, if the prediction mode is an intra mode, the generic controller (721) controls the switch (726) to select an intra mode result for use by the residual calculator (723) and controls the entropy encoder (725) to select intra prediction information and include the intra prediction information in the bitstream, and if the prediction mode of the block is an inter mode, the generic controller (721) controls the switch (726) to select an inter prediction result for use by the residual calculator (723) and controls the entropy encoder (725) to select inter prediction information and include the inter prediction information in the bitstream.
[0094] The residual calculator (723) may be configured to calculate a difference (residual data) between a received block and a prediction result for a block selected from the intra-encoder (722) or the inter-encoder (730). The residual encoder (724) may be configured to encode the residual data to generate transform coefficients. For example, the residual encoder (724) may be configured to transform the residual data from the spatial domain to the frequency domain to generate transform coefficients. The transform coefficients then undergo a quantization process to obtain quantized transform coefficients. In various exemplary embodiments, the video encoder (703) also includes a residual decoder (728). The residual decoder (728) is configured to perform an inverse transform and generate decoded residual data. The decoded residual data may be used by the intra-encoder (722) and the inter-encoder (730) as appropriate. For example, the inter-encoder (730) can generate decoded blocks based on the decoded residual data and the inter-prediction information, and the intra-encoder (722) can generate decoded blocks based on the decoded residual data and the intra-prediction information. The decoded blocks are appropriately processed to generate decoded pictures, which can be buffered in a memory circuit (not shown) and used as reference pictures.
[0095] The entropy encoder (725) may be configured to format a bitstream to include the encoded blocks and to perform entropy coding. The entropy encoder (725) may be configured to include various information in the bitstream. For example, the entropy encoder (725) may be configured to include general control data, selected prediction information (e.g., intra-prediction information or inter-prediction information), residual information, and other suitable information in the bitstream. Residual information may not be present when coding a block in a merged sub-mode of either an inter mode or a bi-prediction mode.
[0096] 8 shows a diagram of an example video decoder (810) according to another embodiment of the present disclosure. The video decoder (810) is configured to receive coded pictures that are part of a coded video sequence and decode the coded pictures to generate reconstructed pictures. In one example, the video decoder (810) may be used in place of the example video decoder (410) of FIG. 4.
[0097] In the example of FIG. 8, the video decoder (810) includes an entropy decoder (871), an inter-decoder (880), a residual decoder (873), a reconstruction module (874), and an intra-decoder (872), all coupled together as shown in the exemplary configuration of FIG. 8.
[0098] The entropy decoder (871) may be configured to recover from the coded picture certain symbols that represent syntax elements of which the coded picture is composed. Such symbols may include, for example, prediction information (e.g., intra-mode, inter-mode, bi-predictive mode, merged sub-mode, or another sub-mode) that may identify the mode in which the block is coded, certain samples or metadata used for prediction by the intra-decoder (872) or inter-decoder (880), residual information, for example in the form of quantized transform coefficients, etc. In one example, if the prediction mode is an inter-mode or bi-predictive mode, the inter-prediction information is provided to the inter-decoder (880), and if the prediction type is an intra-prediction type, the intra-prediction information is provided to the intra-decoder (872). The residual information may undergo inverse quantization and is provided to the residual decoder (873).
[0099] The inter decoder (880) may be configured to receive the inter prediction information and generate inter prediction results based on the inter prediction information.
[0100] The intra decoder (872) may be configured to receive intra prediction information and to generate a prediction result based on the intra prediction information.
[0101] The residual decoder (873) may be configured to perform inverse quantization to extract inverse quantized transform coefficients and process the inverse quantized transform coefficients to transform the residual from the frequency domain to the spatial domain. The residual decoder (873) may also utilize certain control information (to include quantization parameters (QPs)), which may be provided by the entropy decoder (871) (datapath not shown as this may be only a small amount of control information).
[0102] The reconstruction module (874) may be configured to combine, in the spatial domain, the residual as output by the residual decoder (873) and the prediction result (possibly as output by an inter-prediction module or an intra-prediction module) to form a reconstructed block that forms part of a reconstructed picture as part of the reconstructed video. It should be noted that other suitable operations, such as deblocking operations, may also be performed to improve visual quality.
[0103] It should be noted that the video encoders (403), (603), and (703) and the video decoders (410), (510), and (810) may be implemented using any suitable technique. In some exemplary embodiments, the video encoders (403), (603), and (703) and the video decoders (410), (510), and (810) may be implemented using one or more integrated circuits. In another embodiment, the video encoders (403), (603), and (603) and the video decoders (410), (510), and (810) may be implemented using one or more processors executing software instructions.
[0104] Looking at the coding block partitioning, in some example implementations, a predefined pattern may be applied. As shown in FIG. 9, an example four-way partitioning tree may be used starting from a first predefined level (e.g., 64×64 block level) to a second predefined level (e.g., 4×4 level). For example, the base block may follow four partitioning options shown at 902, 904, 906, and 908, and the partitions represented by R may be recursively partitioned in that the same partitioning tree shown in FIG. 9 may be repeated at lower scales down to the lowest level (e.g., 4×4 level). In some implementations, additional restrictions may be applied to the partitioning scheme of FIG. 9. In the implementation of FIG. 9, rectangular partitions (e.g., 1:2 / 2:1 rectangular partitions) may be used but not repeatedly, whereas square partitions may be used repeatedly. Subsequent partitioning of FIG. 9 by recursion generates a final set of coding blocks, if necessary. Such a scheme may be applied to one or more of the color channels.
[0105] FIG. 10 illustrates another exemplary predefined partitioning pattern that allows forming a partitioning tree by recursive partitioning. As illustrated in FIG. 10, an exemplary ten-way partitioning structure or pattern may be predefined. The root block may start from a predefined level (e.g., from the 128×128 level or the 64×64 level). The exemplary partitioning structure in FIG. 10 includes various 2:1 / 1:2 and 4:1 / 1:4 rectangular partitions. A partition type having three subpartitions, shown as 1002, 1004, 1006, and 1008 in the second column of FIG. 10, may be referred to as a "T-type" partition. The "T-type" partitions 1002, 1004, 1006, and 1008 may be referred to as a left T-type, an upper T-type, a right T-type, and a lower T-type. In some implementations, none of the rectangular partitions in FIG. 10 may be further subdivided. A coding tree depth may be further defined to indicate the partitioning depth from the root node or root block. For example, the coding tree depth of the root node or root black of a 128x128 block may be set to 0, and the coding tree depth increases by 1 after the root block is split one more time following FIG. 10. In some implementations, only the all-square partitions of 1010 may allow recursive splitting to the next level of the split tree following the pattern of FIG. 10. In other words, recursive splitting is not possible for square partitions of patterns 1002, 1004, 1006, and 1006. Subsequent splitting of FIG. 10 by recursion generates a final set of coding blocks, if necessary. Such a scheme may be applied to one or more of the color channels.
[0106] After partitioning or splitting the base block according to any of the above partitioning procedures or other procedures, a final set of partitions or coding blocks may still be obtained. Each of these partitions may be at one of various partitioning levels. Each partition may be called a coding block (CB). In the various exemplary partitioning implementations above, each resulting CB may be of any of the allowed sizes and partitioning levels. They are called coding blocks because they may form a unit for which some basic coding / decoding decisions may be made and coding / decoding parameters may be optimized, determined, and signaled in the encoded video bitstream. The highest level in the final partition represents the depth of the coding block partitioning tree. The coding block may be a luma coding block or a chroma coding block.
[0107] In some other example implementations, a quadtree structure may be used to recursively split the base luma blocks and base chroma blocks into coding units. Such a split structure may be called a coding tree unit (CTU), and the CTU is split into coding units (CUs) by adapting the split to various local characteristics of the base CTU using the quadtree structure. In such implementations, an implicit quadtree split may be performed at the picture boundary, such that the block continues the quadtree split until its size fits into the picture boundary. The term CU is used to collectively refer to the units of luma coding block (CB) and chroma coding block (CB).
[0108] In some implementations, the CB may be further divided. For example, the CB may be further divided into multiple prediction blocks (PBs) for the purpose of intra-frame or inter-frame prediction during the coding and decoding processes. In other words, the CB may be further partitioned into different sub-partitions, where individual prediction decisions / configurations may be made. In parallel, the CB may be further divided into multiple transform blocks (TBs) for the purpose of describing the level at which the transformation or inverse transformation of the video data is performed. The division scheme of the CB into PBs and TBs may be the same or different. For example, each division scheme may be performed using a unique procedure based on, for example, various characteristics of the video data. The division schemes of the PBs and TBs may be independent in some exemplary implementations. The division schemes and boundaries of the PBs and TBs may be correlated in some other exemplary implementations. In some implementations, for example, the TBs may be divided after the PB division, and in particular, each PB may be determined following the division of the coding block, and then further divided into one or more TBs. For example, in some implementations, the PB may be divided into one, two, four, or some other number of TBs.
[0109] In some implementations, luma and chroma channels may be processed differently to divide base blocks into coding blocks and further into prediction and / or transform blocks. For example, some implementations may allow division of coding blocks into prediction and / or transform blocks for a luma channel, but not for a chroma channel(s). In such implementations, transformation and / or prediction of luma blocks may thus be performed only at the coding block level. In another example, the minimum transform block size of the luma channel and the chroma channel(s) may be different, e.g., a coding block of a luma channel may be allowed to be divided into smaller transform and / or predictive blocks than a chroma channel. In yet another example, the maximum depth of division of coding blocks into transform and / or predictive blocks may be different between the luma and chroma channels, e.g., a coding block of a luma channel may be allowed to be divided into deeper transform and / or predictive blocks than a chroma channel(s). As a specific example, a luma coding block may be divided into transform blocks of multiple sizes that can be represented by a recursive partitioning down by up to two levels, and transform block shapes such as square, 2:1 / 1:2, 4:1 / 1:4, etc., and transform block sizes from 4×4 to 64×64 may be allowed. However, for chroma blocks, only the largest possible transform block designated for the luma block may be allowed.
[0110] In some example implementations for partitioning a coding block into PBs, the depth, shape, and / or other characteristics of the PB partition may depend on whether the PB is intra-coded or inter-coded.
[0111] The division of a coding block (or a prediction block) into transform blocks may be performed in various exemplary manners, including but not limited to quadtree division and predetermined pattern division, recursively or non-recursively, further considering the transform blocks at the boundaries of the coding block or the prediction block. In general, the resulting transform blocks may be at different division levels, may not be the same size, and may not be square in shape (e.g., they may be rectangular with some allowed sizes and aspect ratios).
[0112] In some implementations, a coding partition tree scheme or structure may be used. The coding partition tree schemes used for the luma channel and the chroma channel may not be the same. In other words, the luma channel and the chroma channel may have separate coding tree structures. Furthermore, whether the luma channel and the chroma channel use the same coding partition tree structure or different coding partition tree structures, and the actual coding partition tree structure to be used, may depend on whether the slice being coded is a P slice, a B slice, or an I slice. For example, for an I slice, the chroma channel and the luma channel may have separate coding partition tree structures or coding partition tree structure modes, while for a P slice or a B slice, the luma channel and the chroma channel may share the same coding partition tree scheme. When separate coding partition tree structures or modes are applied, the luma channel may be partitioned into CB by one coding partition tree structure, and the chroma channel may be partitioned into chroma CB by another coding partition tree structure.
[0113] A specific exemplary implementation of the division of coding blocks and transform blocks is described below. In one such exemplary implementation, a base coding block may be divided into coding blocks using the recursive quadtree division described above. At each level, whether to continue further quadtree division of a particular partition may be determined by local video data characteristics. The resulting CBs may be at various quadtree division levels with various sizes. The decision of whether to code a picture area using inter-picture (temporal) prediction or intra-picture (spatial) prediction may be made at the CB level (or at the CU level in case of three color channels). Each CB may be further divided into one, two, four, or other number of PBs according to the PB division type. Within one PB, the same prediction process may be applied, and related information is sent to the decoder on a PB basis. After obtaining the residual block by applying the prediction process based on the PB division type, the CB may be divided into TBs according to another quadtree structure similar to the coding tree of the CB. In this specific implementation, the CB or TB may not be limited to a square shape. Further, in this particular example, the PB may be square or rectangular in shape for inter prediction, and only square in intra prediction. The coding block may be further divided, for example, into four square-shaped TBs. Each TB may be further divided recursively (using quad-tree partitioning) into smaller TBs called Residual Quad-Tree (RQT).
[0114] Another specific example for splitting a base coding block into CB and other PBs and / or TBs is described below. For example, instead of using multiple partition unit types as shown in FIG. 10, a quadtree with nested multi-type trees using bipartition and tripartition segmentation structures may be used. Separation of the concepts of CB, PB, and TB (i.e., splitting CB into PB and / or TB, and splitting PB into TB) may be abandoned except when necessary for CBs with sizes too large for the maximum transform length, which may require further splitting. This exemplary splitting scheme may be designed to support greater flexibility of CB splitting shapes, so that both prediction and transformation can be performed at the CB level without further splitting. In such a coding tree structure, the CB may have either a square or rectangular shape. Specifically, a coding tree block (CTB) may be first split by a quadtree structure. Then, the leaf nodes of the quadtree may be further split by a multi-type tree structure. An example of a multi-type tree structure is shown in FIG. 11. Specifically, the exemplary multi-type tree structure of FIG. 11 includes four split types: vertical bisection (SPLIT_BT_VER) (1102), horizontal bisection (SPLIT_BT_HOR) (1104), vertical trisection (SPLIT_TT_VER) (1106), and horizontal trisection (SPLIT_TT_HOR) (1108). CB then corresponds to a leaf of the multi-type tree. In this exemplary implementation, as long as CB is not too large relative to the maximum transform length, this segmentation is used for both prediction and transform processing without further splitting. This means that in most cases, CB, PB, and TB have the same block size in a quadtree with a nested multi-type tree coding block structure. An exception occurs when the maximum supported transform length is smaller than the width or height of the color components of CB.
[0115] An example of a quadtree with nested multi-type tree coding block structure of block division of one CTB is shown in FIG. 12. More specifically, FIG. 12 shows that a CTB 1200 is quadtree divided into four square partitions 1202, 1204, 1206, and 1208. A decision to further use the multi-type tree structure of FIG. 11 for division is made for each of the quadtree divided partitions. In the example of FIG. 12, partition 1204 is not further divided. Partition 1202 and partition 1208 adopt another quadtree division, respectively. In partition 1202, the second level quadtree divided upper left partition, upper right partition, lower left partition, and lower right partition adopt the third level division of quadtree, 1104 in FIG. 11, non-partition, and 1108 in FIG. 11, respectively. Partition 1208 adopts another quadtree partitioning, and the second level quadtree partitioned top-left partition, top-right partition, bottom-left partition, and bottom-right partition adopt the third level partitioning of 1106 in FIG. 11, not partitioned, not partitioned, and 1104 in FIG. 11, respectively. Two of the subpartitions of the top-left partition of the third level of 1208 are further partitioned according to 1104 and 1108. Partition 1206 adopts the second level partitioning pattern according to 1102 in FIG. 11 into two partitions, and the two partitions are further partitioned at the third level according to 1108 and 1102 in FIG. 11. A fourth level partitioning is further applied to one of them according to 1104 in FIG. 11.
[0116] In the above specific example, the maximum luma transform size may be 64 x 64, and the maximum supported chroma transform size may also be different from the luma, e.g., 32 x 32. If the width or height of a luma coding block or a chroma coding block is larger than the maximum transform width or height, the luma coding block or the chroma coding block may be automatically split horizontally and / or vertically to meet the horizontal and / or vertical transform size restrictions.
[0117] In the specific example for splitting the base coding block into CBs above, the coding tree scheme may support the ability for luma and chroma to have separate block tree structures. For example, for P slices and B slices, the luma CTB and chroma CTB in one CTU may share the same coding tree structure. For I slices, for example, luma and chroma may have separate coding block tree structures. When the separate block tree mode is applied, the luma CTB may be split into luma CBs by one coding tree structure, and the chroma CTB is split into chroma CBs by another coding tree structure. This means that a CU in an I slice may consist of a coding block of a luma component or a coding block of two chroma components, and a CU in a P slice or B slice always consists of coding blocks of all three color components unless the video is monochrome.
[0118] Exemplary implementations for splitting coding or prediction blocks into transform blocks and the coding order of the transform blocks are described in further detail below. In some exemplary implementations, the transform splitting may support transform blocks of multiple shapes, e.g., 1:1 (square), 1:2 / 2:1, and 1:4 / 4:1, with transform block sizes ranging from, e.g., 4×4 to 64×64. In some implementations, if the coding block is 64×64 or smaller, the transform block splitting may be applied only to the luma component, such that for chroma blocks, the transform block size is the same as the coding block size. Otherwise, if the width or height of the coding block is greater than 64, both the luma coding block and the chroma coding block may be implicitly split into multiples of transform blocks of min(W,64)×min(H,64) and min(W,32)×min(H,32), respectively.
[0119] In some example implementations, for both intra-coded and inter-coded blocks, the coding block may be further divided into multiple transform blocks with a division depth up to a predetermined number of levels (e.g., two levels). The division depth and size of the transform block may be related. An example mapping from the transform size of the current depth to the transform size of the next depth is shown below in Table 1.
[0120] [Table 1]
[0121] According to the example mapping of Table 1, for a 1:1 square block, the next level transform split may create four 1:1 square sub-transform blocks. The transform split may stop at, for example, 4×4. Thus, a transform size of the current depth of 4×4 corresponds to the same size of 4×4 at the next depth. In the example of Table 1, for a 1:2 / 2:1 non-square block, the next level transform split creates two 1:1 square sub-transform blocks, and for a 1:4 / 4:1 non-square block, the next level transform split creates two 1:2 / 2:1 sub-transform blocks.
[0122] In some example implementations, further restrictions may be applied to the luma components of intra-coded blocks. For example, for each level of transform partitioning, all sub-transform blocks may be restricted to have equal size. For example, for a 32×16 coding block, the level 1 transform partitioning creates two 16×16 sub-transform blocks, and the level 2 transform partitioning creates eight 8×8 sub-transform blocks. In other words, to keep the transform units equal in size, the second level partitioning must be applied to all first level sub-blocks. An example of transform block partitioning for an intra-coded square block according to Table 1 is shown in FIG. 13 with the coding order indicated by the arrows. Specifically, 1302 shows a square coding block. The first level partitioning according to Table 1 into four equal-sized transform blocks is shown in 1304 with the coding order indicated by the arrows. The second level partitioning of all first level equal-sized blocks according to Table 1 into 16 equal-sized transform blocks is shown in 1306 with the coding order indicated by the arrows.
[0123] In some example implementations, the above restrictions on intra-coding may not apply to the luma components of an inter-coded block. For example, after the first level of transform partitioning, any one of the sub-transform blocks may be further partitioned independently at another level. Thus, the resulting transform blocks may or may not be of the same size. An example partitioning of an inter-coded block into transform blocks with coding order is shown in FIG. 14. In the example of FIG. 14, an inter-coded block 1402 is partitioned into transform blocks at two levels according to Table 1. At the first level, the inter-coded block is partitioned into four transform blocks of equal size. Then, only one of the four transform blocks (but not all of them) is further partitioned into four sub-transform blocks, resulting in a total of seven transform blocks with two different sizes, as shown at 1404. An example coding order of these seven transform blocks is indicated by an arrow at 1404 in FIG. 14.
[0124] In some example implementations, for the chroma component(s), some additional restrictions on the transform blocks may be applied, e.g., for the chroma component(s), the transform block size may be as large as the coding block size, but may not be smaller than a certain size, e.g., 8×8.
[0125] In some other example implementations, for coding blocks whose width (W) or height (H) is greater than 64, both the luma coding block and the chroma coding block may be implicitly divided into multiples of min(W,64)×min(H,64) and min(W,32)×min(H,32) transform units, respectively.
[0126] Figure 15 further illustrates another alternative exemplary scheme for splitting a coding block or a predictive block into transform blocks. As illustrated in Figure 15, instead of using recursive transform partitioning, a set of predefined partition types may be applied to a coding block according to the transform type of the coding block. In the particular example illustrated in Figure 15, one of six exemplary partition types may be applied to split the coding block into various numbers of transform blocks. Such a scheme may be applied to either a coding block or a predictive block.
[0127] More specifically, the partitioning scheme of FIG. 15 provides up to six partition types for any given transform type, as shown in FIG. 15. In this scheme, every coding block or predictive block may be assigned a transform type based on, for example, a rate-distortion cost. In one example, the partition type assigned to a coding block or predictive block may be determined based on the transform partition type of the coding block or predictive block. As shown by the four partition types illustrated in FIG. 15, a particular partition type may correspond to the partition size and pattern (or partition type) of the transform block. The correspondence between various transform types and various partition types may be predefined. An exemplary correspondence is shown below with capitalized labels indicating the transform types that may be assigned to a coding block or predictive block based on the rate-distortion cost.
[0128] ·PARTITION_NONE: Allocate the transformation size equal to the block size.
[0129] ·PARTITION_SPLIT: Allocates a transformation size that is 1 / 2 the width and 1 / 2 the height of the block size.
[0130] ·PARTITION_HORZ: Allocates a transformation size with the same width as the block size and 1 / 2 the height of the block size.
[0131] ·PARTITION_VERT: Allocates a transformation size that is half the width of the block size and the same height as the block size.
[0132] ·PARTITION_HORZ4: Allocates a transformation size with the same width as the block size and 1 / 4 of the block size in height.
[0133] ·PARTITION_VERT4: Allocates a transformation size that is 1 / 4 the width of the block size and the same height as the block size.
[0134] In the above example, all the partition types shown in Figure 15 include uniform transform sizes for the partitioned transform blocks. This is not a limitation but merely an example. In some other implementations, mixed transform block sizes may be used for the partitioned transform blocks in a particular partition type (or pattern).
[0135] The PB (or CB, also called PB if not further divided into predictive blocks) resulting from any of the above partitioning schemes can become individual blocks for coding via either intra-prediction or inter-prediction. For inter-prediction on the current PB, the residual between the current block and the predictive block can be generated, coded, and included in the coded bitstream.
[0136] Returning to the intra prediction process, samples in a block (e.g., a luma or chroma prediction block, or a coding block if not further divided into prediction blocks) are predicted by samples from neighbors, next neighbors, or one or more other lines, or a combination thereof, to generate a prediction block. The residual between the actual block being coded and the prediction block may then be processed by a transform after quantization. Various intra prediction modes may be made available, and parameters related to the selection of the intra mode and other parameters may be signaled in the bitstream. The various intra prediction modes may relate, for example, to one or more line positions used to predict the samples, the direction to go in selecting a prediction sample from one or more prediction lines, and other special intra prediction modes.
[0137] For example, the set of intra-prediction modes (also referred to as "intra modes") may include a predefined number of directional intra-prediction modes. As described above with respect to the implementation of FIG. 1, these intra-prediction modes may correspond to a predefined number of directions to follow when selecting samples outside a particular block as destinations for samples being predicted within the particular block. In another particular implementation, eight primary directional modes corresponding to angles from 45 to 207 degrees relative to the horizontal axis may be the supported predefined modes.
[0138] In other implementations of intra prediction, the directional intra modes may be further extended to a finer set of angles to further exploit more types of spatial redundancy in the directional texture. For example, as shown in FIG. 16, the above eight angle implementation may be configured to provide eight named angles (referred to as V_PRED, H_PRED, D45_PRED, D135_PRED, D113_PRED, D157_PRED, D203_PRED, and D67_PRED), and for each named angle, a predefined number (e.g., seven) of smaller angle increments may be added. This extension results in a large total number of directional angles (e.g., 56 in this example) that can be used for intra prediction, which correspond to the same number of predefined directional intra modes. The prediction angles may be represented by named intra angles and associated angle increments (angle deltas). In the particular example above where there are seven angular directions with small increments for each named angle, the angular increments may be from -3 to 3, with increasing step sizes in increments of 3 degrees.
[0139] In some implementations, in addition to or instead of the above directional intra modes, a predefined number of non-directional intra prediction modes may also be predefined and available. For example, five non-directional intra modes, called smooth intra prediction modes, may be specified. These non-directional intra mode prediction modes may be specifically called DC intra mode, PAETH intra mode, SMOOTH intra mode, SMOOTH_V intra mode, and SMOOTH_H intra mode. Prediction of samples of a particular block using examples of these non-directional modes is shown in FIG. 17. For example, FIG. 17 shows how a 4×4 block 2002 is predicted by samples obtained from the neighboring line above and / or the neighboring line to the left. A particular sample 1710 in block 1702 may correspond to sample 1704 directly above sample 1710 in the top neighborhood of block 1702, sample 1706 located above and to the left of sample 1710 as an intersection of the top neighborhood and the left neighborhood, and sample 1708 located directly to the left of sample 1710 in the left neighborhood of block 1702. In an example DC intra prediction mode, the average value of left neighborhood sample 1708 and top neighborhood sample 1704 may be used as the predicted value of sample 1710. In an example PAETH intra prediction mode, the top, left and top left reference samples 1704, 1708 and 1706 may be taken, and then any value between these three reference samples that is closest to (top+left-top left) may be set as the predicted value of sample 1710. In an example SMOOTH_V intra prediction mode, sample 1710 may be predicted by vertical quadratic interpolation of top-left neighbor sample 1706 and left neighbor sample 1708. In an example SMOOTH_H intra prediction mode, sample 1710 may be predicted by horizontal quadratic interpolation of top-left neighbor sample 1706 and above neighbor sample 1704. In an example SMOOTH intra prediction mode, sample 1710 may be predicted by an average of vertical and horizontal quadratic interpolations.The above non-directional intra mode implementations are shown as non-limiting examples only: other neighboring lines, other non-directional sample selection methods, and combinations of prediction samples to predict a particular sample within a prediction block are also possible.
[0140] The bitstream may signal which intra prediction mode is selected by the encoder from the above directional and non-directional modes for various coding scales (picture, slice, block, unit, etc.). In some implementations, eight typical named directional modes may be signaled first along with five smooth modes without angles (total of 13 options). Then, if the signaled mode is one of the eight intra modes with named angles, an index indicating the selected angle step for the corresponding signaled named angle is further signaled. In other implementations, all intra prediction modes (e.g., 56 directional modes plus five non-directional modes to obtain 61 intra prediction modes) may be indexed together for signaling.
[0141] In some implementations, the 56 or other number of example directional intra-prediction modes may be implemented using a unified directional predictor that projects each sample of a block to a reference sub-sample position and interpolates the reference sample using a two-tap bilinear filter.
[0142] In some implementations, a further filter mode, called FILTER INTRA mode, may be designed to capture the reduced spatial correlation with the references on the edges. In this mode, samples predicted within the block may be used as intra prediction reference samples for some patches within the block in addition to samples outside the block. These modes may be, for example, predefined and made available for intra prediction of at least the luma block (or only the luma block). A predefined number (e.g., 5) of filter intra modes may be predesigned, each of which is represented by a set of n-tap filters (e.g., 7-tap filters) that reflect the correlation between a sample in a 4×2 patch and its n neighbors. In other words, the weight coefficients of the n-tap filters may be position-dependent. As shown in FIG. 18, using an 8×8 block, a 4×2 patch, and 7-tap filtering as an example, the 8×8 block 2002 may be divided into eight 4×2 patches. In FIG. 18, these patches are indicated as B0, B1, B1, B3, B4, B5, B6, and B7. For each patch, its seven neighbors (denoted R0-R7 in FIG. 18) may be used to predict samples in the target patch. For patch B0, all neighbors may already be reconstructed, while other patches may not be reconstructed because some of their neighbors are within the target block, in which case the predicted values of their close neighbors are used as reference. For example, patch B7 shown in FIG. 18 does not have all its neighbors reconstructed, so the predicted samples of its neighbors are used instead.
[0143] In some implementations of intra prediction, one color component may be predicted using one or more other color components. The color components may be in any one of the YCrCb, RGB, XYZ, etc. color spaces. For example, a prediction Chroma from Luma (CfL) may be performed to predict a chroma component (e.g., a chroma block) from a luma component (e.g., a luma reference sample). In some implementations, cross-color prediction is mostly only allowed from luma to chroma. For example, chroma samples in a chroma block may be modeled as a linear function of the corresponding reconstructed luma samples. CfL prediction may be performed as follows: CfL(α)=α×L AC +DC (1)
[0144] Here, L AC denotes the AC contribution of the luma component, α denotes a parameter of the linear model, and DC denotes the DC contribution of the chroma components. For example, the AC components are obtained for each sample of the block, while the DC components are obtained for the whole block. Moreover, the reconstructed luma samples may be subsampled to obtain the chroma resolution, and then the average luma value (DC of luma) may be subtracted from each luma value to generate the luma AC contribution. The luma AC contribution is then used in the linear mode of equation (1) to predict the AC values of the chroma components. Instead of requiring the decoder to calculate the scaling parameters to approximate or predict the chroma AC components from the luma AC contribution, a CfL implementation may determine the parameter α based on the original chroma samples and signal this in the bitstream. This reduces the decoder complexity and provides a more accurate prediction. The DC contribution of the chroma components may be calculated using an intra DC mode in the chroma components in some implementations.
[0145] Instead of intra prediction, the PB may be inter predicted in either single reference or mixed reference inter prediction modes. In particular, in the inter prediction modes, a video block may be predicted by one or more other reference blocks or inter prediction blocks from one or more other frames via either single reference or mixed reference inter prediction. To implement inter prediction, a reference block may be specified by its frame identifier (temporal location of the reference block) and a motion vector (spatial location of the reference block) indicating a spatial offset between the current block being coded or decoded and the reference block. The reference frame identification and the motion vector may be signaled in the bitstream. The motion vector as a spatial block offset may be directly signaled or may itself be predicted by another reference motion vector or a predictor motion vector. For example, the current motion vector may be predicted directly by a reference motion vector (e.g., a candidate neighboring block) or by a combination of the reference motion vector and a motion vector difference (MVD) between the current motion vector and the reference motion vector. The latter is sometimes referred to as merge mode with motion vector difference (MMVD). A reference motion vector may be identified in the bitstream, for example, as a pointer to a spatially adjacent block of the current block or to a temporally adjacent but spatially co-located block.
[0146] In some implementations, a composite inter-intra prediction (CIIP) mode may be implemented. In the CIIP mode, a prediction block may be derived as a combination of an intra prediction (or intra predictor) block and an inter prediction (intra predictor) block. The inter prediction block in CIIP may be derived using a single reference inter prediction with a translation operation corresponding to a motion vector, while the intra prediction block in CIIP may be determined from neighboring samples based on a subset of the intra prediction modes mentioned above. In some example implementations, the spatial samples in the intra prediction block of a current block being predicted in a CIIP mode may be derived from intra reference line samples according to one of the subsets of intra prediction modes including the DC_PRED, V_PRED, H_PRED, and SMOOTH modes mentioned above. The use of each of this subset of intra prediction modes to derive an intra prediction block may correspond to a CIIP sub-mode index, as shown in Table 2.
[0147] [Table 2]
[0148] The composite inter-intra prediction (or predictor) block in the current block may be generated as a sample-level weighted sum of the intra prediction block and the inter prediction block at the sample level, as derived according to the above description. Thus, the relative weight between the intra prediction block and the inter prediction block may be represented by a weight matrix. Various exemplary aspects of the weight matrix determination may be implemented corresponding to various inter-intra weighting modes of the CIIP.
[0149] In one exemplary implementation of the weighting mode of CIIP, commonly referred to as CIIP, the elements of the inter-intra weighting matrix corresponding to samples in the current block may follow a deterministic relationship with the position of the sample. Such a deterministic relationship may depend on the intra prediction mode being used. In one particular example, the weighting applied to the intra prediction sample P0(x,y) (where x,y represent the sample position in the block) may be derived as follows:
number
[0150] [Table 3]
[0151] In the above example of normal CIIP, the weighting in intra prediction generally decreases as the sample moves away from the upper left corner of the block (or away from the intra prediction reference sample), except for DC_PRED mode, and the inter prediction weighting is independent of the sample position.In other words, the embodiment illustrated in Table 3 reflects a scheme in which the inter prediction weight increases as the sample position moves away from the intra reference sample.
[0152] In some other exemplary inter-intra weighting implementations in CIIP, called wedge CIIP, a set of weight patterns may be defined, and one of the set of weight patterns may be selected by the encoder for the current block. An index of the selected pattern in the set of weight patterns may be signaled in the bitstream. Such a pattern may be applied to the current block to determine a particular weight matrix used to combine / add the intra-predicted and inter-predicted blocks for each sample of the block. For example, 16 different patterns (also called wedge patterns) may be predefined and represented by indexes 0-15. In wedge CIIP, once the index of the pattern is specified, the entire weight matrix of the block is derived using the predefined lookup table in Table 3, rather than being signaled sample-by-sample as in the normal CIIP approach.
[0153] In any CIIP, once the above weighting matrix is obtained and applied to combine the intra-predicted block and the inter-predicted block, such combined block can be used as the actual predicted block in the current block to obtain a residual block. The residual block can then undergo a primary transformation, and optionally a second transformation, as well as the remainder of the quantization and entropy coding process, from the encoder's perspective. In the case of a decoder, the bitstream is parsed / decoded and inversely transformed to obtain the residual block. If the decoder determines from the bitstream that a CIIP is used for the current block, it can further obtain the weighting matrix of the CIIP based on the information extracted from the bitstream (either the lookup index signaled above, or the wedge pattern index signaled above). The predicted block can then be derived from the weighting matrix and the corresponding intra-predicted block and inter-prediction from already reconstructed samples in the current frame or reference frame. The original block can then be restored from the residual block and the predicted block.
[0154] Turning to the primary transform, an exemplary 2D transform process may involve the use of hybrid transform kernels (which may be composed of, for example, a different 1-D transform for each dimension of the coded residual block) in addition to using the same transform kernel for both dimensions. Exemplary primary 1-D transform kernels may include, but are not limited to, a) 4-point (4 p), 8-point (8 p), 16-point (16 p), 32-point (32 p) and 64-point (64 p) DCT-2, b) 4-point, 8-point, 16-point asymmetric DST and their inverted versions, c) 4-point, 8-point, 16-point or 32-point identity transform (DST stands for discrete sine transform). Thus, a 2D transform process may involve the use of hybrid transforms or transform kernels (different transforms for each dimension of the coded residual block), and the selection of the transform or transform kernel to be used for each dimension may be based on a rate-distortion (RD) criterion. The term transform kernel may alternatively be referred to as a transform basis function. For example, Table 4 lists basis functions for 1D DCT-2, DST-4 and DST-7, which can be implemented as a hybrid of 2D transforms (where DCT stands for discrete cosine transform).
[0155] [Table 4]
[0156] For example, DCT-2 (4 p - 64 p), DST-4 (8 p, 16 p), and DST-7 (4 p) transforms exhibit symmetric / anti-symmetric properties, and therefore some exemplary implementations may support a "partial butterfly" implementation to reduce the (multiply, add / part, shift) operation count. The partial butterfly implementation may involve plane rotations using trigonometric cosine and sine functions at various angles, as described in FIG. 19. Exemplary 12-bit lookup tables are shown in FIG. 20 and FIG. 21, which may be utilized to generate the trigonometric function values.
[0157] In some example implementations, a line graph transform (LGT) or transform kernel can be used instead of or in combination with the 1-D DST or DCT and other transform kernels mentioned above. Essentially, as shown in FIG. 22, a graph is a general mathematical structure consisting of a set of vertices and edges that can be used to model affinity relationships between objects of interest. In practice, a weighted graph (used to assign a set of weights to the edges and potentially to the vertices) can provide a sparse representation for robust modeling of signals / data. LGT can improve coding efficiency by providing better adaptation for diverse block statistics. A separable LGT (for each 1-D transform) uses a line graph (with self-loop weights v , v , v ) from the training data to model the underlying row- and column-wise statistics of the residual block signal. c1 , v c2 and edge weight w c (characterized by:
[0158] Given a weighted graph G(W,V), the GGL matrix is defined as follows: L c =D-W+V (4) where W is the non-negative edge weight w cis the adjacency matrix consisting of, D is the diagonal matrix, and V is the weighted self-loop v c1 , v c2 The matrix L is a diagonal matrix that represents c can be expressed as follows:
number
[0159] Then, change LGT to GGL L c It can be derived by eigenvalue decomposition of L c =UΦU T (6) where the columns of the orthogonal matrix U are the basis vectors of the LGT, and Φ is the diagonal eigenvalue matrix. In fact, the GGL can be considered as a general scheme for deriving transforms including DCT and DST. For example, DCT-2, DCT-8 and DST-7 are specific forms of the LGT derived from the GGL. v c1 DCT-2 is derived by setting =0 (7) v c1 =w c DST-7 is derived by setting (8) v c2 =w c The DCT-8 is derived by setting (9) v c1 =2w c DST-4 is derived by setting (10) v c2 =2w c The DCT-4 is derived by setting (11)
[0160] In more general use, the self-loop weights v c1 and v c2Different LGT cores can be generated by setting v to specific values. Furthermore, the basis vectors of the LGT cores can also be adjusted for orthogonality by adjusting the individual weights of the basis vectors over a specific dynamic range to generate an optimal LGT core. The encoder derives the self-loop weights v from the block statistics. c1 and v c2 , optimize and derive LGT basis vectors for use in performing a transform of the current coding block residual, and signal LGT transform related information in the bitstream for the decoder to derive the corresponding inverse LGT.
[0161] In some other implementations, one or more other offline trained linear transform kernels can be used, including, but not limited to, Korhonen-Loeve Transform (KLT) kernels and matrix transform (RCT) kernels. These types of kernels are data-driven and are pre-trained to learn block, column and row statistics and correlations in various types of videos. In practical applications, these offline trained transform kernels or basis functions can form at least a part of a set of candidate transform functions, called a transform set. The encoder can perform an analysis of a particular transform block and determine whether to use an offline trained transform kernel and, if so, which of the candidate offline trained transform kernels. Relevant information is signaled in the bitstream for the decoder to identify or derive the selected kernel to perform the inverse transform.
[0162] The above various transform kernels may be designed as separable 1-D transform kernels or non-separable 2-D transform kernels. For a separable 1-D transform, each of the two dimensions in a 2-D image block can be processed separately by independently applying a 1-D transform kernel (e.g., a separable DCT kernel, a separable DST kernel, a separable LGT kernel, a separable KLT kernel, or a separable RCT kernel, etc.). For a non-separable 2-D transform, the two dimensions of the 2-D block can be jointly transformed based on a non-separable transform kernel (e.g., a non-separable DCT kernel, a non-separable DST kernel, a non-separable LGT kernel, a non-separable KLT kernel, or a non-separable RCT kernel, etc.).
[0163] In some further exemplary implementations, the CIIP above can use any one of the primary transform kernels above to transform the residual of any prediction block. The candidate transform kernels can form a kernel set. The set of candidate kernels can include transform kernels based on any combination of DCT, DST, LGT, KLT, RCT kernels, and other derived or pre-trained kernels. For example, the set of candidate kernels can include multiple transform kernels selected from LGT, KLT, RCT, or other derived or pre-trained transform kernels. The transform kernels derived by the encoder, such as LGT, KLT, RCT kernels, are based on learned block statistics from pre-training and / or data characteristics of a particular input video during the encoding process. These transform kernels may be referred to as data-driven transform kernels, in contrast to deterministic and rule-based transform kernels, such as DCT, DST, ADST, FLIPADST (inverted ADST), IDTX (Identify Transform), etc.
[0164] In some implementations, the type of transform kernel selected from the candidate kernels for a particular prediction block may be based on the CIIP submode, such as those described above in Table 2 for the intra prediction modes used in the CIIP scheme. For example, the type of transform kernel (e.g., LGT, KLT, RCT, or other derived or pre-trained kernel type) or the particular transform kernel in the set of transform kernel candidates may be predefined according to the CIIP submode. In a particular example, according to the general block statistics of a particular type of video being coded, a higher coding gain may be obtained for a particular CIIP submode when a particular type of LGT, KLT, RCT, or other derived or pre-trained transform kernel is used. A smaller set of kernel candidate kernels may be required for a particular CIIP submode (rather than a larger set that requires more signaling overhead and includes kernels that are rarely used). The smaller set of kernels may include kernels of LGT, KLT, RCT, or other derived or pre-trained types to provide a better coding gain.
[0165] In some example implementations, the correspondence between CIIP submodes and types of transform kernels or specific transform kernels need not be a one-to-one correspondence. In certain examples, multiple CIIP submodes may be specified as being associated with one or more of the same type of transform kernel or one or more of the same specific transform kernel. By way of example only, CIIP submodes 0-2 in Table 2 may be mapped to use a KLT or RCT transform kernel, and CIIP submode 3 may be specified to use an LGT kernel.
[0166] In some example implementations, the LGT, KLT, or RCT, and other derived or pre-trained kernels used in one or more CIIP submodes may be separate 1D transform kernels. In some other example implementations, the LGT, KLT, or RCT, and other derived or pre-trained kernels used in one or more CIIP submodes may be non-separate 2D transform kernels.
[0167] In some exemplary implementations of the separable 1-D transform, the two dimensions of the transform block may employ different sets of transform kernels or types of transform kernels. For example, the transform in each dimension in the CIIP may select any kernel, such as LGT, KLT, RCT, or other derived or pre-trained kernel types, or regular predefined DCT, DST, ADST, FLIPADST, IDTX kernel types, etc. The set of candidate kernels or kernel types may further depend on the CIIP submode, such as the exemplary submodes shown in Table 2 above. For example, for CIIP submode 1, II_V_PRED mode in Table 2, intra prediction using weighted inter-intra prediction may be V_PRED submode. In such a situation, for example, the set of vertical transform kernels for the CIIP residual may include one or more LGT, KLT, RCT, or other derived or pre-trained kernels, while the set of horizontal transform kernels may include regular DCT, DST, ADST, FLIPADST, IDTX kernels, etc. Similarly, for CIIP submode 2, II_H_PRED mode in Table 2, intra prediction using weighted inter-intra prediction may be H_PRED submode. In such a situation, for example, the set of horizontal transform kernels of the CIIP residual may include one or more LGT, KLT, RCT, or other derived or pre-trained kernels, while the set of vertical transform kernels may include regular DCT, DST, ADST, FLIPADST, IDTX kernels, etc. For CIIP submode 3 in Table 2, in another example, both the vertical and horizontal transform kernels may include only one or more LGT, KLT, RCT, or other derived or pre-trained kernels. For CIIP submode 0, in yet another example, both the vertical and horizontal transform kernels may include only one or more regular DCT, DST, ADST, FLIPADST, IDTX kernels, etc. The various above implementations of CIIP submode-dependent use of regular and derived / pre-trained transform kernels are merely examples. Various other combinations are also possible.
[0168] In some example implementations, parameters of a set of LGT, KLT, RCT, or other derived or pre-trained kernel candidates may be signaled in the video stream. Various subsets of these kernels may also be signaled in the bitstream, for example, as subgroups of an index of the set of candidate kernels. Each of these subgroups may be identified, for example, by a group index. The selection of a kernel for a particular block may be determined by the group index and then an index within the group. In some implementations, the selection of a kernel from among the LGT, KLT, or RCT, or other derived or pre-trained kernels from the transform group may be signaled at the block level. In other words, the determination and selection of the transform kernel may vary from block to block.
[0169] In some exemplary implementations, the selection of a transform kernel from among the LGT, KLT or RCT, or other derived or pre-trained kernels per block may depend on the block size of the block being coded. The term block size may refer to either the width or height of the block, or the maximum of the width and height, or the minimum of the width and height, or the area size (width*height). In other words, which of the LGT, KLT or RCT, or other derived or pre-trained kernels is used for a particular transform block and / or which kernel is used for a particular transform block may be determined from the width, height, or a combination thereof of the block (e.g., the maximum of the width and height, or the minimum of the width and height, or the area of the block). The mapping between the block size and the kernel used may be pre-determined based on a general analysis of coding statistics. A range of block sizes (e.g., a range of the width, height, maximum of the width and height, minimum of the width and height, or area of the block) may be mapped to a particular LGT, KLT or RCT, or other derived or pre-trained kernel, or a particular type of derived or pre-trained kernel. For example, two or more block size ranges may be predefined, each of which may correspond to one or more predefined transform kernels. In such implementations, the LGT, KLT, or RCT, or other derived or pre-trained kernel that is selected to transform a particular block may be indicated (or implicitly signaled) via the block size and thus may not need to be explicitly signaled in the bitstream.
[0170] In some alternative implementations, the selection of a transform kernel from among LGT, KLT or RCT, or other derived or pre-trained kernels per block may depend on the block shape of the block being coded. The term block shape may refer to the width-height aspect ratio, height-width aspect ratio, maximum of the width-height aspect ratio and height-width aspect ratio, or minimum of the width-height aspect ratio and height-width aspect ratio of the block. In other words, whether and / or which LGT, KLT or RCT, or other derived or pre-trained kernel is used for a particular transform block may be determined from the width-height ratio, height-width aspect ratio, maximum of the width-height aspect ratio and height-width aspect ratio, or minimum of the width-height aspect ratio and height-width aspect ratio of the block. The mapping between block shapes and kernels used may be pre-determined based on a general analysis of coding statistics. A range of block shapes may be mapped to a particular LGT, KLT or RCT, or other derived or pre-trained kernel, or to a particular type of derived or pre-trained kernel. For example, two or more shape ranges may be predefined, with each of the ranges corresponding to one or more pre-defined transform kernels. In such an implementation, the LGT, KLT or RCT, or other derived or pre-trained kernel that is selected to transform a particular block may be indicated (or implicitly signaled) via the block shape and thus may not need to be explicitly signaled in the bitstream.
[0171] In some example implementations, a combination of block size and block shape may be mapped to a particular one or set of LGT, KLT, or RCT, or other derived or pre-trained kernels, or may be mapped to a particular one or more types of derived or pre-trained transform kernels. Again, in such implementations, the LGT, KLT, or RCT, or other derived or pre-trained kernel that is selected to transform a particular block may be indicated (or implicitly signaled) by the block size and shape combination, and thus may not need to be explicitly signaled in the bitstream.
[0172] In some example implementations, when LGT is used, the self-loop weight (self-loop weight v c1 and v c2 ) and / or the self-loop weight ratio (self-loop weight v c1 and v c2The selection of the self-loop weight and / or the self-loop weight ratio of the GTL used to transform the coding block may be based on the size of the prediction block (width, height, maximum width and height, minimum width and height, or area size (width*height)). In other words, the selection of the self-loop weight and / or the self-loop weight ratio of the GTL used to transform the coding block may be determined from the width, height, or a combination thereof of the block (e.g., maximum width and height, or minimum width and height, or area of the block). The mapping between the block size and the self-loop weight or the self-loop weight ratio used for the LGT may be pre-determined based on a general analysis of coding statistics. A range of block sizes (e.g., a range of width, height, maximum width and height, minimum width and height, or area of the block) may be mapped to a specific set of self-loop weights or self-loop weight ratios. For example, two or more block size ranges may be pre-defined, each of which corresponds to a set of self-loop weights or self-loop weight ratios when the GTL is selected to transform the residual of the block. In such an implementation, the LGT self-loop weight or self-loop weight ratio selected for a block may be indicated (or implicitly signaled) by the block size and therefore may not need to be explicitly signaled in the bitstream.
[0173] In some example implementations, when LGT is used, the self-loop weight and / or the self-loop weight ratio (self-loop weight v c1 and v c2The selection of the self-loop weight and / or the self-loop weight ratio may be based on the shape of the prediction block (e.g., the width-height aspect ratio, the height-width aspect ratio, the maximum of the width-height aspect ratio and the height-width aspect ratio, or the minimum of the width-height aspect ratio and the height-width aspect ratio). In other words, the selection of the self-loop weight and / or the self-loop weight ratio may be determined from the maximum of the width-height aspect ratio, the height-width aspect ratio, the width-height aspect ratio and the height-width aspect ratio, or the minimum of the width-height aspect ratio and the height-width aspect ratio of the block. The mapping between the block shapes used for LGT and the self-loop weight or the self-loop weight ratio may be pre-determined based on a general analysis of coding statistics. A range of block shapes may be mapped to a specific set of self-loop weights or self-loop weight ratios. For example, two or more block shape ranges may be pre-defined, each of the ranges corresponding to a set of self-loop weights or self-loop weight ratios when the GTL is selected to transform the residual of the block. For example, a square block may be mapped to the GTL using a particular self-loop weight ratio or a particular set of self-loop weight ratios, while a non-square block may be mapped to the GTL using a different self-loop weight ratio or a different set of self-loop weight ratios. In such an implementation, the LGT self-loop weight or self-loop weight ratio selected for a block may be indicated (or implicitly signaled) via the block shape and therefore may not need to be explicitly signaled in the bitstream.
[0174] In some example implementations, a selection between normal CIIP and wedge CIIP can be made at the block level and any other coding level. Such a selection can be signaled in the bitstream at the block level or other coding level. In some further implementations, whether a block is predicted under a normal CIIP or wedge CIIP weighting scheme not only affects the spatial weighting between intra-predicted and inter-predicted blocks in generating the actual overall predicted block, but may also be used to determine the transform kernel or type of transform kernel for transforming the corresponding residual block.
[0175] For example, the various kernel selection approaches described above regarding the use of LGT, KLT or RCT, or other derived or pre-trained kernels, may only be applicable when the block is predicted under a regular CIIP, whereas when the block is predicted under a wedge CIIP (e.g., a given DCT, ADST, DST transform kernel, etc.), a given kernel or set of given kernels is used.
[0176] Similarly, in some other alternative embodiments, the above various kernel selection techniques regarding the use of LGT, KLT, or RCT, or other derived or pre-trained kernels may be applied only when the block is predicted under a wedged CIIP, whereas a predetermined kernel or a predetermined set of kernels is used when the block is predicted under a normal CIIP (e.g., a predetermined DCT, ADST, DST transform kernel, etc.). In some further implementations of the wedged CIIP, the selection of the LGT, KLT, RCT, or other derived or pre-trained kernel to be used from the transform set for transforming the CIIP residual may depend on the wedge pattern used for the CIIP. In other words, the mapping between the wedge pattern and the transform kernel may be predefined. In such an embodiment, the selection of the LGT, KLT, RCT, or other derived or pre-trained kernel to be used from the transform set may be implicit by the wedge pattern selected for use in the wedged CIIP, and thus does not need to be explicitly signaled (or only needs to be indirectly signaled via the signaling for the wedge pattern selection).
[0177] The above various implementations of CIIP are described in the context of first-order transformations. The underlying principles are also applicable to second-order transformations with LGT, KLT, RCT and other derived or pre-trained kernels, as well as other types of kernel choices, in either separable 1D or non-separable 2D implementations.
[0178] An exemplary decoder incorporating some of the above implementations is provided below. In this example, the decoder may receive a video bitstream and extract a current coding block. The encoder may further extract configurations related to entropy coding of the current block, perform entropy decoding, and then perform inverse quantization of the coding block to generate transform coefficients of a residual block of the current block. The decoder may further extract various coding modes of the current block, including whether the current block is inter-predicted, intra-predicted, or predicted under a CIIP mode, from signaling in the bitstream. If the decoder determines that a CIIP mode is used, the decoder may further determine which CIIP weighting sub-mode and which spatial weighting pattern mode of a plurality of sub-modes (e.g., the CIIP sub-modes shown in Table 2 above) is used for the CIIP (e.g., either a square CIIP or a wedge CIIP). For example, such a determination may be based on signaling information extracted from the bitstream. The inter-intra prediction weighting sub-mode and spatial weighting pattern mode used for the current block may be signaled separately or together in the bitstream. In the case of spatial weighting pattern mode, additional signaling regarding weights may be extracted from the bitstream (e.g., an index of each spatial sample of the current block with reference to an inter-intra prediction weighting lookup table such as Table 3 above, or an index corresponding to a wedge weighting pattern among a plurality of predefined wedge weighting patterns). Thus, the decoder can determine the inter-intra weighting matrix of the current block. The encoder may further determine the primary transform kernel (and the second transform kernel, if it is determined that a secondary transform is also used in coding the current block) to be used for the current block. The transform kernels may be in a transform kernel set. The transform kernel set may be determined by the decoder from information / signaling extracted from the bitstream.The decoder may use additional signaling in the bitstream to determine which of the transform kernels in the transform kernel set to use for the current block. The transform kernel set may include one or more of LGT, KLT, RCT, and other derived / pre-trained transform kernels in addition to predefined DST, DCT, ADST, FLIPADST, IDTX kernels, etc. The encoder may further determine whether a separable 1-D transform or a non-separable 2-D transform is used for the current block and determine an appropriate transform kernel for each or both of the two dimensions. The actual kernel used for the current block may be explicitly signaled (e.g., using an index in the transform kernel set) or may be implicitly derived based on other parameters. For example, the transform kernel(s) may be identified by determining the already extracted CIIP weighting sub-mode, CIIP weighting pattern mode, and / or CIIP weighting pattern index in conjunction with the mapping relationship between these already extracted parameters and the transform kernel(s), as described in more detail above. Once the transform kernel is determined and obtained, the decoder can then perform an inverse transform of the transform coefficients of the residual block (generated through the above decoding process) to generate a residual block. The decoder can then combine the residual block with the already reconstructed inter-predicted and intra-predicted samples according to the inter-intra weighting matrix derived above to generate a block.
[0179] A corresponding example of an encoder follows the above inverse processing steps to determine various CIIP modes, weighting matrices, transform kernel sets, and transform kernels for the block being coded to generate encoded blocks and include various signaling information in the encoded bitstream.
[0180] In the above embodiments, the term block may be interpreted as a coding block, a prediction block, or a transformation block. The block size may refer to either the width or height of the block, or the maximum value of the width and height, or the minimum value of the width and height, or the area size of the block (width*height), or the aspect ratio (width:height, or height:width). Furthermore, the term chroma channel may include both Cb and Cr color components, or both U and V color components. The term luma channel may include a luma color component or a Y color component.
[0181] FIG. 23 shows a flowchart 2300 of an example video decoding method according to the principles underlying the implementation above. The example method flow 2300 starts at S 2301. At S 2310, it is determined that a video block is predicted in a composite inter-intra prediction (CIIP) mode. At S 2320, at least one inter-intra prediction weighting configuration for the video block is determined. At S 2330, a data-driven transform kernel for the video block is identified from among a set of transform kernels. At S 2340, a residual block of the video block is generated by inverse transforming a set of transform coefficients extracted from a video stream for the video block using at least the data-driven transform kernel. At S 2350, a video block is derived based on the residual block, the at least one inter-intra prediction weighting configuration, at least one reconstructed inter-prediction block of the video block, and at least one reconstructed intra-prediction reference sample of the video block. The example method flow 2300 ends at S 2399.
[0182] FIG. 24 illustrates a flowchart 2400 of an exemplary video encoding method according to the principles underlying the above implementations. The exemplary method flow 2400 starts at 2401. At 2410, it is determined that a video block should be predicted in a composite inter-intra prediction (CIIP) mode. At 2420, at least one inter-intra prediction weighting configuration for the video block is determined. At 2430, a data-driven transform kernel for the video block is selected from among a set of transform kernels. At 2440, a residual block of the video block is generated based on at least the video block, the at least one inter-intra prediction weighting configuration, at least one inter-predicted block of the video block, and at least one intra-predicted reference sample of the video block. At 2450, the residual block is transformed using at least the data-driven transform kernel to generate a set of transform coefficients for the video block. At 2460, the set of transform coefficients is quantized and entropy coded for inclusion in a video stream together with the inter-intra prediction weighting configuration after encoding. The example method flow 2400 ends at S2499.
[0183] The embodiments of the present disclosure may be used separately or combined in any order. Furthermore, each of the methods (or embodiments), the encoder, and the decoder may be implemented by a processing circuit (e.g., one or more processors or one or more integrated circuits). In one example, the one or more processors execute a program stored on a non-transitory computer-readable medium. The embodiments of the present disclosure may be applied to luma blocks or chroma blocks.
[0184] The techniques described above may be implemented as computer software using computer-readable instructions and physically stored on one or more computer-readable media. For example, FIG. 25 illustrates a computer system (2500) suitable for implementing certain embodiments of the disclosed subject matter.
[0185] The computer software may be coded using any suitable machine code or computer language that may undergo assembly, compilation, linking, or similar mechanisms to generate code including instructions that may be executed, such as by one or more computer central processing units (CPUs) and graphics processing units (GPUs), either directly, or through interpretation and execution of microcode.
[0186] The instructions may be executed in various types of computers or components thereof, including, for example, a personal computer, a tablet computer, a server, a smart phone, a gaming device, an Internet of Things device, etc.
[0187] 25 with respect to the computer system (2500) are exemplary in nature and are not intended to suggest any limitation as to the scope of use or functionality of the computer software implementing the embodiments of the present disclosure, nor should the arrangement of components be interpreted as having any dependency or requirement relating to any one or combination of components illustrated in the exemplary embodiment of the computer system (2500).
[0188] The computer system (2500) may include certain human interface input devices. Such human interface input devices may respond to input by one or more human users through, for example, tactile input (keystrokes, swipes, data glove movements, etc.), audio input (voice, clapping, etc.), visual input (gestures, etc.), and olfactory input (not shown). The human interface devices may be used to capture certain media that are not necessarily directly related to conscious human input, such as sound (speech, music, ambient sounds, etc.), images (scanned images, photographic images obtained from a still image camera, etc.), and video (two-dimensional video, three-dimensional video including stereoscopic video, etc.).
[0189] The input human interface devices may include one or more of a keyboard (2501), a mouse (2502), a trackpad (2503), a touch screen (2510), a data glove (not shown), a joystick (2505), a microphone (2506), a scanner (2507), and a camera (2508) (only one of each is depicted).
[0190] The computer system (2500) may also include certain human interface output devices. Such human interface output devices may stimulate one or more of the human user's senses, for example, through haptic output, sound, light, and smell / taste. Such human interface output devices may include haptic output devices (e.g., haptic feedback via a touch screen (2510), data gloves (not shown), or joystick (2505), although there may also be haptic feedback devices that do not function as input devices), audio output devices (such as speakers (2509), headphones (not shown)), visual output devices (such as screens (2510), including CRT screens, LCD screens, plasma screens, OLED screens, each with or without touch screen input capability and with or without haptic feedback capability, some of which may output two-dimensional visual output or three or more dimensional output by means of stereoscopic output, virtual reality glasses (not shown), holographic displays, smoke tanks (not shown)), and printers (not shown).
[0191] The computer system (2500) may also include human accessible storage and associated media, such as optical media, including CD / DVD ROM / RW (2520) with media (2521) such as CDs / DVDs, thumb drives (2522), removable hard drives or solid state drives (2523), legacy magnetic media such as tapes and floppy disks (not shown), and specialized ROM / ASIC / PLD based devices such as security dongles (not shown).
[0192] Those skilled in the art will also understand that the term "computer-readable medium" as used in connection with the presently disclosed subject matter does not encompass transmission media, carrier waves, or other transitory signals.
[0193] The computer system (2500) may also include an interface (2554) to one or more communication networks (2555). The networks may be, for example, wireless, wired, optical. The networks may further be local, wide area, metropolitan, vehicular and industrial, real-time, delay tolerant, etc. Examples of networks include local area networks such as Ethernet, cellular networks including wireless LAN, GSM, 3G, 4G, 5G, LTE, etc., television wired or wireless wide area digital networks including cable television, satellite television and terrestrial television, vehicular and industrial including CAN bus, etc. Certain networks typically require an external network interface adapter connected to a particular general-purpose data port or peripheral bus (2549) (e.g., a USB port on the computer system (2500)), while others are generally integrated into the core of the computer system (2500) by connecting to a system bus, as described below (e.g., an Ethernet interface for a PC computer system, or a cellular network interface for a smartphone computer system). Using any of these networks, the computer system (2500) may communicate with other entities. Such communication may be unidirectional, receive only (e.g., television broadcast), unidirectional transmit only (e.g., from the CANbus to a particular CANbus device), or bidirectional, e.g., to other computer systems using local or wide area digital networks. Specific protocols and protocol stacks may be used in each of these networks and network interfaces, as described above.
[0194] The aforementioned human interface devices, human accessible storage devices, and network interfaces may be attached to the core (2540) of the computer system (2500).
[0195] The cores (2540) may include one or more central processing units (CPUs) (2541), graphics processing units (GPUs) (2542), dedicated programmable processing units in the form of field programmable gate areas (FPGAs) (2543), hardware accelerators for specific tasks (2544), and graphics adapters (2550). These devices may be connected via a system bus (2548), along with read-only memory (ROM) (2545), random access memory (2546), and internal mass storage (2547) such as an internal hard drive or SSD that is not accessible to the user. In some computer systems, the system bus (2548) may be accessed in the form of one or more physical plugs to allow expansion with additional CPUs, GPUs, and the like. Peripheral devices may be connected directly to the core's system bus (2548) or through a peripheral bus (2549). In one example, a screen (2510) may be connected to the graphics adapter (2550). Peripheral bus architectures include PCI, USB, and the like.
[0196] The CPU (2541), GPU (2542), FPGA (2543), and accelerator (2544) may execute certain instructions that may combine to constitute the aforementioned computer code. The computer code may be stored in a ROM (2545) or a RAM (2546). Transient data may also be stored in the RAM (2546), and persistent data may be stored, for example, in internal mass storage (2547). Rapid storage and retrieval from any of the memory devices may be made possible through the use of cache memories closely associated with one or more of the CPU (2541), GPU (2542), mass storage (2547), ROM (2545), RAM (2546), and the like.
[0197] The computer-readable medium may bear computer code for performing various computer-implemented operations. The medium and computer code may be those specially designed and constructed for the purposes of the present disclosure, or they may be of the kind well known and available to those skilled in the computer software arts.
[0198] As a non-limiting example, the computer system (2500) having the architecture, and in particular the core (2540), may provide functionality as a result of the processor(s) (including CPU, GPU, FPGA, accelerator, etc.) executing software embodied in one or more tangible computer-readable media. Such computer-readable media may be media associated with the user-accessible mass storage devices introduced above, as well as with specific storage of the core (2540) of a non-transitory nature, such as the core internal mass storage (2547) or ROM (2545). Software implementing various embodiments of the present disclosure may be stored in such devices and executed by the core (2540). The computer-readable media may include one or more memory devices or chips, depending on the particular needs. The software may cause the core (2540) and in particular the processors therein (including CPU, GPU, FPGA, etc.) to perform certain processes, or certain parts of certain processes, described herein, including defining data structures stored in RAM (2546) and modifying such data structures according to the processes defined by the software. Additionally, or alternatively, the computer system may provide functionality as a result of logic hardwired or otherwise embodied in circuitry (e.g., accelerator (2544)), which may operate in place of or in conjunction with software to perform particular processes or portions of particular processes described herein. Where appropriate, references to software may encompass logic, and vice versa. Where appropriate, references to computer-readable media may encompass circuitry (such as an integrated circuit (IC)) that stores software for execution, circuitry embodying logic for execution, or both. The present disclosure encompasses any suitable combination of hardware and software.
[0199] While this disclosure has described several exemplary embodiments, there are modifications, substitutions, and various alternative equivalents which are within the scope of this disclosure. Thus, it will be appreciated that those skilled in the art will be able to devise numerous systems and methods which, although not explicitly shown or described herein, embody the principles of the present disclosure and are thus within the spirit and scope of the present disclosure. Appendix A: Acronyms JEM: Joint Exploration Model VVC: Versatile Video Coding BMS: Benchmark Set MV: Motion Vector HEVC: High Efficiency Video Coding SEI:Supplementary Enhancement Information VUI: Video Usability Information GOP: Group of Pictures TU: conversion unit PU: Prediction Unit CTU: Coding Tree Unit CTB: coding tree block PB: Predicted block HRD: Hypothetical Reference Decoder SNR: Signal to Noise Ratio CPU: Central Processing Unit GPU: Graphics Processing Unit CRT: cathode ray tube LCD: Liquid crystal display OLED: Organic Light Emitting Diode CD:Compact Disc DVD: Digital Video Disc ROM: Read-Only Memory RAM: Random Access Memory ASIC: Application Specific Integrated Circuit PLD: Programmable Logic Device LAN: Local Area Network GSM: Global System for Mobile Communications LTE: Long Term Evolution CANBus: Controller Area Network Bus USB: Universal Serial Bus PCI: Peripheral Component Interconnect FPGA: Field Programmable Gate Area SSD: Solid State Drive IC: Integrated Circuit HDR: High Dynamic Range SDR: Standard Dynamic Range JVET: Joint Video Exploration Team MPM: Most Probable Mode WAIP: Wide-angle Intra Prediction CU: coding unit PU: Prediction Unit TU: conversion unit CTU: Coding Tree Unit PDPC: Position-dependent prediction combination ISP: Intra Subpartition SPS: Sequence Parameter Settings PPS: Picture Parameter Set APS: Adaptive Parameter Set VPS: Video Parameter Set DPS: Decoding Parameter Set ALF: Adaptive Loop Filter SAO: Sample Adaptive Offset CC-ALF: Cross-component adaptive loop filter CDEF: Constrained Directional Enhancement Filter CCSO: Cross component sample offset LSO: Local Sample Offset LR: Loop recovery filter AV1:AOMedia Video 1 AV2:AOMedia Video 2 [Explanation of symbols]
[0200] 101 Samples 102 Arrow 103 Arrow 104 Square Block 180 Schematic diagram 201 Current Block 202 Surrounding Samples 203 Surrounding Samples 204 Surrounding Samples 205 Surrounding Samples 206 Surrounding Samples 300 Communication Systems 310 Terminal Devices 320 Terminal Devices 330 Terminal Devices 340 Terminal Devices 350 Network 400 Communication Systems 401 Video Source 402 Video Picture Stream 403 Video Encoder 404 Video Data 405 Streaming Server 406 Client Subsystem 407 Copy video data 408 Client Subsystem 409 Copy of video data 410 Video Decoder 411 Video Picture Output Stream 412 Display 413 Video Capture Subsystem 420 Electronic Devices 430 Electronic Devices 501 Channel 510 Video Decoder 512 Rendering Devices, Displays 515 Buffer Memory 520 Parser 521 Symbols 530 Electronic Devices 531 Receiver 551 Scaler / Descaler Unit 552 Intra-picture prediction unit 553 Motion Compensation Prediction Unit 555 Aggregator 556 Loop Filter Unit 557 Reference Picture Memory 558 Current Picture Buffer 601 Video Sources 603 Video Encoder 620 Electronic Devices 630 Source Coder 632 Coding Engine 633 Local Decoder 634 Reference Picture Memory 635 Predictor 640 Transmitter 643 coded video sequence 645 Entropy Coder 650 Controller 660 Communication Channels 703 Video Encoder 721 General-purpose controller 722 Intra Encoder 723 Residual Calculator 724 Residual Encoder 725 Entropy Encoder 726 Switch 728 Residual Decoder 730 InterEncoder 810 Video Decoder 871 Entropy Decoder 872 Intra Decoder 873 Residual Decoder 874 Reconstruction Module 880 Interdecoder 902 Split Options 904 Split Options 906 Split Options 908 Split Options 1002 Left T-shaped partition 1004 Upper T-shaped partition 1006 Right T-shaped partition 1008 Lower T-shaped partition 1010 All square partitions 1102 Vertical bisection (SPLIT_BT_VER) 1104 Horizontal split into two (SPLIT_BT_HOR) 1106 Vertical third division (SPLIT_TT_VER) 1108 Horizontal third division (SPLIT_TT_HOR) 1200 Coding Tree Block (CTB) 1202 Square Partition 1204 Square Partition 1206 Square Partition 1208 Square Partition 1302 Square coding block 1304 First Level Split 1402 Inter-coded Blocks 1404 Conversion Block 1702 Block 1704 Sample 1706 Sample 1708 Sample 1710 Sample 2002 Block 2010 Sample 2300 Flowchart 2400 Flowchart 2500 Computer System 2501 Keyboard 2502 Mouse 2503 Trackpad 2505 Joystick 2506 Microphone 2507 Scanner 2508 Camera 2509 Speaker 2510 Touch Screen 2520 CD / DVD ROM / RW 2521 CD / DVD and other media 2522 Thumb Drive 2523 Removable Hard Drive or Solid State Drive 2540 cores 2541 Central Processing Unit (CPU) 2542 Graphics Processing Unit (GPU) 2543 Field Programmable Gate Area (FPGA) 2544 Hardware Accelerator 2545 Read-Only Memory (ROM) 2546 Random Access Memory (RAM) 2547 Internal Mass Storage 2548 System Bus 2549 Surrounding Bus 2550 Graphics Adapter 2554 Network Interface 2555 Network
Claims
1. 1. A method for decoding a video block in a video stream, comprising: determining that the video block is predicted in a combined inter-intra prediction (CIIP) mode, the video block being derived as a combination of intra prediction and inter prediction in the CIIP mode; determining at least one inter-intra prediction weighting configuration for the video block; identifying a data driven transform kernel for the video block from a set of transform kernels; generating a residual block for the video block by inverse transforming a set of transform coefficients extracted from the video stream for the video block using at least the data-driven transform kernel; deriving the video block based on the residual block, the at least one inter-intra prediction weighting configuration, at least one reconstructed inter-predicted block of the video block, and at least one reconstructed intra-predicted reference sample of the video block. The method includes:
2. The method of claim 1 , wherein the data-driven transformation kernels include pre-trained transformation kernels or derived transformation kernels based on data distribution within the video block.
3. 3. The method of claim 2, wherein the data-driven transformation kernel comprises one of a data-driven line graph transformation (LGT) kernel, a pre-trained Coronen-Loeve transform kernel, or a pre-trained row-column transformation (RCT) kernel.
4. The method of claim 3 , wherein the data-driven transformation kernel comprises a one-dimensional separable transformation kernel.
5. The method of claim 3 , wherein the data-driven transformation kernel comprises a two-dimensional non-separable transformation kernel.
6. The method further includes identifying a non-data driven transform kernel for the video block from among the set of transform kernels, wherein generating the residual block for the video block comprises: transforming the set of transform coefficients extracted from the video stream for the video block using the data-driven transform kernel in one of two dimensions of the video block and the non-data-driven transform kernel in the other of the two dimensions of the video block to generate the residual block for the video block.
2. The method of claim 1, comprising:
7. 2. The method of claim 1, further comprising determining a CIIP submode for the video block from among a plurality of CIIP submodes, the plurality of CIIP submodes corresponding to a subset of intra-prediction modes of the CIIP mode, and the data-driven transform kernel being selected from the transform kernel set based on the CIIP submode.
8. The method of claim 7 , wherein multiple CIIP sub-modes map to the same data-driven transformation kernel.
9. further comprising identifying a non-data driven transformation kernel; the CIIP sub-mode corresponds to a V_PRED intra prediction mode; 8. The method of claim 7, wherein generating the residual block for the video block comprises inverse transforming the set of transform coefficients extracted from the video stream for the video block using the data-driven transform kernel in a vertical dimension of the video block and the non-data-driven transform kernel in a horizontal dimension of the video block to generate the residual block for the video block.
10. further comprising identifying a non-data driven transformation kernel; the CIIP sub-mode corresponds to a HORIZONTAL_PRED intra prediction mode; 8. The method of claim 7, wherein generating the residual block for the video block comprises inverse transforming the set of transform coefficients extracted from the video stream for the video block using the data-driven transform kernel in a horizontal dimension of the video block and the non-data-driven transform kernel in a vertical dimension of the video block to generate the residual block for the video block.
11. The method of claim 1 , wherein identifying the data-driven transform kernel for the video block from among the set of transform kernels is based on a size of the video block or a shape of the video block.
12. the size of the video block includes a width, a height, a maximum of the width and the height, a minimum of the width and the height, or an area of the video block; 12. The method of claim 11 , wherein the shape of the video block includes a width-to-height aspect ratio, a height-to-width aspect ratio, a maximum of the width-to-height aspect ratio and the height-to-width aspect ratio, or a minimum of the width-to-height aspect ratio and the height-to-width aspect ratio.
13. the data-driven transformation kernel comprises an LGT kernel; The method of claim 1 , wherein a self-loop ratio of the LGT kernel is dictated by a size or shape of the video block.
14. the size of the video block includes a width, a height, a maximum of the width and the height, a minimum of the width and the height, or an area of the video block; 14. The method of claim 13, wherein the shape of the video block includes a width-height aspect ratio, a height-to-width aspect ratio, a maximum of the width-height aspect ratio and the height-to-width aspect ratio, or a minimum of the width-height aspect ratio and the height-to-width aspect ratio.
15. 2. The method of claim 1, wherein the at least one inter-intra prediction weighting configuration indicates whether inter-intra prediction spatial weights for the video block should be explicitly signaled in the video stream or derived from one of a set of predetermined inter-intra spatial weighting patterns.
16. 16. The method of claim 15, wherein identifying the data-driven transform kernel for the video block from among the set of transform kernels is responsive to the at least one inter-intra prediction weighting configuration indicating that the inter-intra prediction spatial weight for the video block is explicitly signaled.
17. 16. The method of claim 15, wherein identifying the data-driven transform kernel for the video block from among the set of transform kernels is responsive to the at least one inter intra prediction weighting configuration indicating that the inter intra prediction spatial weights for the video block are derived from one of a set of predetermined inter intra spatial weighting patterns.
18. the one of a set of predefined inter-intra spatial weighting patterns is indicated as part of the at least one inter-intra prediction weighting configuration; 20. The method of claim 17, wherein the data-driven transform kernels of the set of transform kernels are identified according to any of the set of predefined inter-intra spatial weighting patterns as indicated in the at least one inter-intra prediction weighting configuration.
19. 1. A device for decoding video blocks in a video stream, comprising: a memory for storing instructions; and a processing circuit, the processing circuit executing the instructions to: determining that the video block is predicted in a combined inter-intra prediction (CIIP) mode, the video block being derived as a combination of intra prediction and inter prediction in the CIIP mode; determining at least one inter-intra prediction weighting configuration for the video block; identifying a data driven transform kernel for the video block from a set of transform kernels; generating a residual block for the video block by inverse transforming a set of transform coefficients extracted from the video stream for the video block using at least the data-driven transform kernel; deriving the video block based on the residual block, the at least one inter-intra prediction weighting configuration, at least one reconstructed inter-predicted block of the video block, and at least one reconstructed intra-predicted reference sample of the video block. The device is configured to:
20. 1. A method for encoding a video block in a video stream, comprising: determining that the video block should be predicted in a combined inter-intra prediction (CIIP) mode, the video block being derived as a combination of intra prediction and inter prediction in the CIIP mode; determining at least one inter-intra prediction weighting configuration for the video block; selecting a data driven transform kernel for the video block from a set of transform kernels; generating a residual block for the video block based at least on the video block, the at least one inter-intra prediction weighting configuration, at least one inter predicted block for the video block, and at least one intra predicted reference sample for the video block; transforming the residual block using at least the data-driven transform kernel to generate a set of transform coefficients for the video block; quantizing and entropy coding the set of transform coefficients for inclusion in the video stream together with the inter-intra prediction weighting structure after encoding; The method includes:
Citation Information
Patent Citations
Method and apparatus for video coding
US20200389671A1
Adaptive weight in multi-hypothesis prediction in video coding
WO2020182187A1
Transform-based image coding method, and apparatus therefor
WO2021066598A1
