Selection of movement information predictor

Description

1 / 51 “MOTION INFORMATION PREDICTOR SELECTION” CROSS-REFERENCE TO RELATED REQUESTS

[001] This application claims priority over European Application No. 23315124.0, filed on 27 April 2023, which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[002] At least one of the present embodiments refers, in general, to a method and a device for encoding and decoding figures and, more particularly, to a method and a device that allow improving the parallelization of operations in a decoding process. BACKGROUND OF THE TECHNIQUE

[003] To achieve high compression efficiency, video coding schemes generally employ predictions and transforms to take advantage of spatial and temporal redundancies in video content. During encoding, video content figures are divided into pixel blocks, which are then partitioned into one or more sub-blocks, referred to as original sub-blocks. An intra- or inter-figure prediction is then applied to each sub-block to exploit intra- or inter-figure correlations. Whichever prediction method is used (intra- or inter-figure), a predictor sub-block is determined for each original sub-block. Then, a sub-block representing a difference between the original sub-block and the predictor sub-block, often denoted as a prediction error sub-block, a prediction residue sub-block, or simply a residue block, is transformed, quantized, and entropy-encoded to generate an encoded video stream.To reconstruct the video, the compressed data is decoded by inverse processes corresponding to transformation, quantization, and entropic encoding.

[004] In recent video codecs, new methods involving a model Petition 870250085959, dated 09 / 23 / 2025, page 47 / 123 2 / 51 around a current block to infer or reduce the signaling cost have been developed. While in the past, decoding dependencies between successive blocks were avoided as much as possible to allow parallelization of block reconstruction, the models introduced new decoding dependencies preventing parallelization. These new dependencies place a heavy load on the decoding processes, which can no longer parallelize the most computationally intensive stages.

[005] It is desirable to propose solutions that allow for improved parallelization in recent video codecs. BRIEF SUMMARY

[006] In a first aspect, one or more of the present modalities provide a method comprising:

[007] obtain an initial list of candidate motion vector predictors for motion information of a current block;

[008] obtain bi-prediction candidates from the initial list, wherein each bi-prediction candidate comprises two motion vectors, wherein a first motion vector of the two motion vectors indicates a first prediction block in a first reference figure and a second motion vector of the two motion vectors indicates a second prediction block in a second reference figure;

[009] calculate a score for bi-prediction candidates from the initial list; and,

[010] obtain a motion vector predictor for the current block motion information based on the calculated scores;

[011] where each score is a representative value of a difference between a first model and a second model, with the first model comprising Petition 870250085959, dated 09 / 23 / 2025, page 48 / 123 3 / 51 ende samples from the first prediction block and the second model comprises samples from the second prediction block that spatially correspond to the samples from the first model.

[012] In one embodiment, the method comprises reordering the bi-prediction candidates from the initial list based on the scores calculated to obtain a reordered list, wherein the motion vector predictor obtained is derived from the reordered list.

[013] In one embodiment, the first model comprises all samples from the first prediction block and the second model comprises all samples from the second prediction block.

[014] In one embodiment, the first model corresponds to the lower right samples of the first prediction block and the second model corresponds to the lower right samples of the second prediction block.

[015] In one embodiment, the first model corresponds to the upper left samples of the first prediction block and the second model corresponds to the upper left samples of the second prediction block.

[016] In one embodiment, each model has N pixels of thickness, where N is a positive integer value.

[017] In one embodiment, a motion compensation based on integers is applied to identify the first and second models, respectively, in the first and second reference figurations.

[018] In one modality, a motion vector difference is added to at least one bi-prediction candidate from the initial list before calculating the score.

[019] In one embodiment, in response to the initial list comprising uni-prediction candidates, only the bi-prediction candidates are reordered and each uni-prediction candidate retains the same index in the reordered list as in the initial list. Petition 870250085959, dated 09 / 23 / 2025, page 49 / 123 4 / 51

[020] In one modality, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates from the initial list are transformed into bi-prediction candidates before the calculation of scores.

[021] In one embodiment, bi-prediction candidates from the reordered list obtained from a uni-prediction candidate from the initial list are used to predict the movement information of the current block.

[022] In one embodiment, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates are excluded from the reordered list.

[023] In a second aspect, one or more of the present embodiments provide a method for encoding an actual block comprising the method of the first aspect.

[024] In a third aspect, one or more of the present embodiments provide a method for decoding an actual block comprising the method of the first aspect.

[025] In a fourth aspect, one or more of the present embodiments provide a device comprising electronic circuits configured to:

[026] obtain an initial list of candidate motion vector predictors for motion information of a current block;

[027] obtain bi-prediction candidates from the initial list, wherein each bi-prediction candidate comprises two motion vectors, wherein a first motion vector of the two motion vectors indicates a first prediction block in a first reference figure and a second motion vector of the two motion vectors indicates a second prediction block in a second reference figure;

[028] calculate a score for bi-prediction candidates from the initial list; and, Petition 870250085959, dated 09 / 23 / 2025, page 50 / 123 5 / 51

[029] obtain a motion vector predictor for the current block motion information based on the calculated scores;

[030] where each score is a representative value of a difference between a first model and a second model, wherein the first model comprises samples from the first prediction block and the second model comprises samples from the second prediction block that correspond spatially to the samples of the first model.

[031] In one embodiment, the electronic circuit is further configured to reorder the bi-prediction candidates from the initial list based on the scores calculated to obtain a reordered list, wherein the motion vector predictor obtained is obtained from the reordered list.

[032] In one embodiment, the first model comprises all samples from the first prediction block and the second model comprises all samples from the second prediction block.

[033] In one embodiment, the first model corresponds to the lower right samples of the first prediction block and the second model corresponds to the lower right samples of the second prediction block.

[034] In one embodiment, the first model corresponds to the upper left samples of the first prediction block and the second model corresponds to the upper left samples of the second prediction block.

[035] In one embodiment, each model has N pixels of thickness, where N is a positive integer value.

[036] In one embodiment, the electronic circuit is further configured to apply a motion compensation based on integers to identify the first and second models, respectively, in the first and second reference figures. Petition 870250085959, dated 09 / 23 / 2025, page 51 / 123 6 / 51

[037] In one modality, a motion vector difference is added to at least one bi-prediction candidate from the initial list before calculating the score.

[038] In one embodiment, in response to the initial list comprising uni-prediction candidates, only the bi-prediction candidates are reordered and each uni-prediction candidate retains the same index in the reordered list as in the initial list.

[039] In one modality, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates from the initial list are transformed into bi-prediction candidates before the calculation of scores.

[040] In one embodiment, bi-prediction candidates from the reordered list obtained from a uni-prediction candidate from the initial list are used to predict the movement information of the current block.

[041] In one embodiment, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates are excluded from the reordered list.

[042] In a fifth aspect, one or more of the present embodiments provide a device for encoding an actual block comprising the device of the fourth aspect.

[043] In a sixth aspect, one or more of the present embodiments provide a device for decoding a current block comprising the device of the fourth aspect.

[044] In a seventh aspect, one or more of the present embodiments provide a computer program comprising program code instructions for implementing the method of the first, second or third aspect.

[045] In an eighth aspect, one or more of the present embodiments provide a non-transient information storage medium that stores program code instructions for implementing the method of the first, second or third aspect. Petition 870250085959, dated 09 / 23 / 2025, page 52 / 123 7 / 51 BRIEF SUMMARY OF THE DRAWINGS

[046] Figure 1 describes an example of a context in which the following modalities can be implemented.

[047] Figure 2 illustrates an example of partitioning suffered by a pixel image of an original video;

[048] Figure 3 schematically illustrates a method for encoding a video stream performed by an encoding module;

[049] Figure 4 schematically illustrates a method for decoding the encoded video stream;

[050] Figure 5A schematically illustrates an example of a hardware architecture for a processing module capable of implementing an encoding module or a decoding module in which various aspects and modes are implemented;

[051] Figure 5B illustrates a block diagram of an example of a first system in which various aspects and modalities are implemented;

[052] Figure 5C illustrates a block diagram of an example of a second system in which various aspects and modalities are implemented;

[053] Figure 6A schematically illustrates five spatial positions considered for constructing a list of merge candidates;

[054] Figure 6B schematically illustrates positions considered to determine the time-vector predictor of motion;

[055] Figures 7A and 7B illustrate a block-based affine transform motion compensation applied to a block;

[056] Figure 8 illustrates an affine transform prediction based on a subblock;

[057] Figure 9 illustrates a control point motion vector inheritance; Petition 870250085959, dated 09 / 23 / 2025, page 53 / 123 8 / 51

[058] Figure 10 illustrates spatial and temporal neighbors used to derive motion information from control points;

[059] Figure 11 represents spatially neighboring blocks considered in a sub-block temporal motion vector prediction process;

[060] Figure 12 illustrates an example of a process that allows deriving a sub-block temporal motion vector predictor;

[061] Figure 13 illustrates a method based on model matching;

[062] Figure 14A schematically illustrates a predictor reordering process in a predictor list implemented by an encoding module;

[063] Figure 14B schematically illustrates a predictor reordering process in a predictor list implemented by a decoding module; and,

[064] Figure 15 illustrates two temporal predictors used to create the final prediction. DETAILED DESCRIPTION

[065] The following examples of modalities are described in the context of a video format similar to VVC (ISO / IEC 23090-3 - MPEG-1: Versatile Video Coding (VVC) / ITU-T H.266). However, these modalities are not limited to the video encoding / decoding method corresponding to VVC. These modalities are adapted, in particular, to various video formats, including, for example, HEVC (ISO / IEC 23008-2 - MPEG-H Part 2, High Efficiency Video Coding / ITU-T H.265), AVC (ISO / IEC 14496-10), EVC (Essential Video Coding / MPEG-5), AVI, AV2 and VP9.

[066] Figure 1 describes an example of a context in which the following modalities can be implemented.

[067] In Figure 1, a system 11, which could be a camera, a device Petition 870250085959, dated 09 / 23 / 2025, page 54 / 123 9 / 51 storage, a computer, a server, or any device capable of providing a video stream, transmits a video stream to a system 13 using a communication channel 12. The video stream is encoded and transmitted by system 11 or received and / or stored by system 11 and then transmitted. The communication channel 12 is a wired network link (e.g., Internet or Ethernet) or a wireless network link (e.g., WiFi, 3G, 4G, or 5G).

[068] System 13, which could be, for example, a decoder, receives and decodes the video stream to generate a sequence of decoded figures.

[069] The sequence obtained from decoded figures is then transmitted to a display system 15 using a communication channel 14 which can be a wired or wireless network. The display system 15 then displays said figures.

[070] In one embodiment, system 13 is comprised within display system 15. In this case, system 13 and display 15 are composed of a TV, a computer, a tablet computer, a smartphone, a head-mounted display, etc.

[071] Figures 2, 3 and 4 show an example of a video format.

[072] Figure 2 illustrates an example of partitioning undergone by a pixel figuration 21 of an original video sequence 20. Here, a pixel is considered to be composed of three components: a luminance component and two chrominance components. However, other types of pixels are possible, comprising fewer or more components, such as only a luminance component or an additional depth or transparency component.

[073] A figuration is divided into a plurality of coding entities. First, as represented by reference 23 in Figure 2, a figuration is divided into a grid of blocks called coding tree units (CTUs). A CTU consists of an N x N block of luminance samples joined together. Petition 870250085959, dated 09 / 23 / 2025, page 55 / 123 10 / 51 mind with two corresponding blocks of chrominance samples. TV is generally a power of two with a maximum value of “128”, for example. Second, a figuration is divided into one or more groups of CTUs. For example, it can be divided into one or more tile rows and tile columns, where a tile is a sequence of CTUs that cover a rectangular region of a figuration. In some cases, the block can be divided into one or more bricks, each of which consists of at least one CTU row within the block. Above the concept of tiles and bricks, another coding entity, called a slice, exists, which can contain at least one tile of a figuration or at least one brick of a tile.

[074] In the example in Figure 2, as represented by reference 22, figure 21 is divided into three slices S1, S2 and S3 of the raster scan slice mode, each comprising a plurality of tiles (not shown), each tile comprising only one brick.

[075] As represented by reference 24 in Figure 2, a CTU can be partitioned in the form of a hierarchical tree of one or more sub-blocks called coding units (CU). The CTU is the root (i.e., the parent node) of the hierarchical tree and can be partitioned into a plurality of CUs (i.e., child nodes). Each CU becomes a leaf of the hierarchical tree if it is not subsequently partitioned into smaller CUs, or becomes a parent node of smaller CUs (i.e., child nodes) if it is subsequently partitioned.

[076] In the example in Figure 2, CTU 24 is first partitioned into “4” square CUs using a quaternary tree partitioning. The upper left CU is a leaf of the hierarchical tree, since it is no longer partitioned, that is, it is not a parent node of any other CU. The upper right CU is further partitioned into 4 smaller square CUs again using a quaternary tree partitioning. The lower right CU is vertically partitioned into “2” rectangular CUs using a binary tree partitioning. The lower left CU Petition 870250085959, dated 09 / 23 / 2025, page 56 / 123 11 / 51 is vertically partitioned into three rectangular CUs using a ternary tree partitioning pattern.

[077] During the encoding of a figuration, partitioning is adaptive, with each CTU being partitioned in order to optimize the compaction efficiency of the CTU criterion.

[078] In HEVC, the concepts of prediction unit (PU) and transform unit (TU) emerged. In fact, in HEVC, the encoding entity that is used for prediction (i.e., a PU) and transform (i.e., a TU) can be a subdivision of a CU. For example, as represented in Figure 1, a CU of size 2N x 2N can be divided into PU 2411 of size N x 2N or of size 2N x TV. Furthermore, said CU can be divided into “4” TUs 2412 of size N x N or into “16” TUs of size (-2) x ÇL).

[079] It can be noted that in VVC, except in some particular cases, the edges of TU and PU are aligned on the edges of CU. Consequently, a UC generally comprises a TU and a PU.

[080] In the present application, the term “block” or “figuration block” may be used to refer to any of a CTU, a CU, a PU and a TU. In addition, the term “block” or “figuration block” may be used to refer to a macroblock, a partition and a subblock, as specified in H.264 / AVC or other video coding standards and, more generally, to refer to a sample array of various sizes.

[081] In the present application, the terms “reconstructed” and “decoded” may be used interchangeably, the terms “pixel” and “sample” may be used interchangeably, the terms “image”, “figuration”, “subfiguration”, “slice” and “frame” may be used interchangeably. Normally, but not necessarily, the term “reconstructed” is used on the encoder side, while “decoded” is used on the decoder side. Petition 870250085959, dated 09 / 23 / 2025, page 57 / 123 12 / 51

[082] Figure 3 schematically describes a method for encoding a video stream executed by an encoding module. Variations of this encoding method are contemplated, but the encoding method in Figure 3 is described below for clarity, without describing all expected variations.

[083] Before being encoded, a current original figuration of an original video sequence may undergo preprocessing. For example, in a 301 step, a color transform is applied to the current original figuration (e.g., converting RGB 4:4:4 to YCbCr 4:2:0), or a remapping is applied to the components of the current original figuration to obtain a signal distribution more resilient to compression (e.g., using a histogram equalization of one of the color components). Figurations obtained by preprocessing are referred to as preprocessed figurations hereafter.

[084] The encoding of a pre-processed figure begins with a partitioning of the pre-processed figure during a 302 step, as described in relation to Figure 2. The pre-processed figure is then partitioned into CTU, CU, PU, ​​TU, etc. For each block, the encoding module determines an encoding mode between an intra prediction and an inter prediction.

[085] Intra prediction consists of predicting, according to an intra prediction method, during a 303 step, the pixels of a current block from a prediction block derived from pixels of reconstructed blocks located in a causal adjacency of the current block to be encoded. The result of intra prediction is a prediction mode that indicates which pixels of the adjacent blocks to use and a residual block resulting from the calculation of a difference between the current block and the prediction block.

[086] Inter-prediction consists of predicting the pixels of a current block from a block of pixels, called a reference block, from a previous or subsequent figuration to the current figuration, this figuration being called the reference figuration. During the encoding of a current block according to the inter-prediction method, Petition 870250085959, dated 09 / 23 / 2025, page 58 / 123 13 / 51 A block from the reference figure closest, according to a similarity criterion, to the current block is determined by a motion estimation step 304. During step 304, a motion vector indicating the position of the reference block in the reference figure is determined. The motion vector is used during a motion compensation step 305 during which a residual block is calculated in the form of a difference between the current block and the reference block.

[087] When the current block is encoded according to an inter prediction, a process is applied to encode the movement information.

[088] Two processes are employed to encode motion information: AMVP (Adaptive Motion Vector Prediction) or Blending. In each process, motion information is predicted.

[089] In an AMVP mode implementation, a motion vector predictor (MVP) is selected, and an observed motion vector difference MVd relative to the selected MVP is computed. The MVP is selected from a list of AMVP candidates consisting of “2” candidates. The index of the chosen MVP and the MVd are then encoded by the entropic encoder during step 310, along with the transformed and quantized residual block resulting from the interpretation of the current block.

[090] The AMVP candidate list is constructed by first deriving a first spatial candidate from a left block adjacent to the current block, if this block is available and intercoded. Then, a second spatial candidate is derived from a top block adjacent to the current block, if this block is available and intercoded. Then, a temporal candidate is derived from the reference figuration considered for the current block at a colocalized position to the current block, if an interblock exists at that colocalized position. Each derived MVP candidate is scaled according to a temporal distance between the associated reference figuration. Petition 870250085959, dated 09 / 23 / 2025, page 59 / 123 14 / 51 to this MVP candidate and the reference figuration considered for the current block. A redundancy check is then conducted between the derived spatial candidates and, if there is a duplicate candidate, this candidate is discarded. The final list of AMVP candidates contains the first two derived MVP candidates. If fewer than “2” MVP candidates are obtained through the above process, the AMVP candidate list will be completed with zero motion vectors.

[091] The merge mode consists of deriving motion information from a current block from a selected candidate to predict motion information. The motion information considered here includes all inter-prediction parameters of a block, that is: the type of one-way or two-way temporal prediction, the reference figure index within each list of reference figures, and the motion vectors. The selected motion information predictor candidate (i.e., the merge candidate) is selected from a list of motion information predictor candidates (i.e., from a list of merge candidates). When a block is encoded in merge mode, the index of the selected merge candidate is encoded. If no residual block is encoded for the current block, the current block is considered encoded according to a specific merge mode called skip mode.

[092] In some implementations, the merge candidate list is systematically composed of “5” merge candidates. Up to “5” spatial positions are considered to retrieve some potential candidates for the merge candidate list. Figure 6A schematically illustrates the five spatial positions considered to construct a merge candidate list. These positions are investigated according to the following order:

[093] Left (A1)

[094] Above (B1)

[095] Top right (B0) Petition 870250085959, dated 09 / 23 / 2025, page 60 / 123 15 / 51

[096] Bottom left (A0)

[097] Top left (B2)

[098] Each space candidate is introduced into the merge candidate list, provided that the motion information corresponding to that candidate is not already present in the merge candidate list.

[099] Next, an observed temporal predictor TMVP is determined. Figure 6B schematically illustrates positions considered to determine the TMVP. The determination of the TMVP consists first of investigating position H and, if no motion information is available at position H, position C is investigated. A scale can be applied to the motion information obtained to obtain the TMVP.

[0100] A final pruning process is then applied to ensure that the set of spatial and temporal candidates does not contain redundant candidates.

[0101] In the case of slice B (slice that allows bi-predicted blocks), candidates of another type, called combined candidates, are introduced into the merge candidate list if this list is not complete.

[0102] Finally, if the merge list is not yet complete, zero motion vectors will be introduced at the end of the merge list until it is complete.

[0103] Recently, the representation of motion information has evolved slightly with the emergence of two main categories of motion representation: whole-block based motion representation and sub-block based motion representation.

[0104] The whole-block motion representation consists of assigning a set of motion information, composed of one or two motion vectors and reference figures associated with an internal block. Thus, the motion information of this block is represented in the form of one or two Petition 870250085959, dated 09 / 23 / 2025, page 61 / 123 16 / 51 motion vectors for the entire block and a reference figure associated with each motion vector.

[0105] The sub-block-based motion encoding mode typically consists of dividing a block into sub-blocks of 4x4 or 8x8 luma samples and assigning an individual set of motion information (one or two pairs of a motion vector and a reference figure) to each sub-block.

[0106] While in previous AMVP implementations only spatial candidates, temporal candidates, and the zero motion vector candidate were considered to build the AMVP candidate list, a new category of candidates, called HMVP (History-Based Motion Vector Prediction) candidates, has been added in AMVP implementations adapted to the whole block-based motion representation.

[0107] One principle of HMVP candidates is to use previously encoded motion vectors as MVPs. These motion vectors are associated with adjacent or non-adjacent blocks relative to a current block. To do this, an HMVP candidate table (i.e., HMVP table) is maintained and dynamically updated as a "first-in, first-out" (FIFO) buffer of MVPs. There are up to five candidates in the HMVP table. After encoding an inter-predicted block, provided that block is not in sub-block mode (including affine mode) or GPM (geometric partition mode), the HMVP table is updated by appending the motion information of the inter-predicted block to the end of the HMVP table as a new HMVP candidate. In addition to the usual FIFO rule, a mechanism to remove redundant HMVP candidates is applied. It may be noted that the HMVP table is reset on each CTU row to allow for parallel processing.

[0108] In recent implementations of the merge mode adapted to the integer-based block representation, the list of merge candidates has been modified and three new merge modes have been introduced. Petition 870250085959, dated 09 / 23 / 2025, page 62 / 123 17 / 51

[0109] The merge candidate list is constructed with the following candidate types:

[0110] Space candidates.

[0111] Temporary candidates.

[0112] HMVP Candidates. Several HMVP candidates are entered into the merge candidate list so that the list reaches the maximum allowed number of merge candidates minus 1.

[0113] Paired average candidates. Up to one paired average candidate is added to the merge candidate list. Paired candidates are computed as follows: The first two merge candidates present in the merge candidate list are considered and their motion vectors are calculated. This average is calculated separately for each reference figuration list. If each of the first two merge candidates is bi-predictive, the motion vectors related to lists L0 and L1 will be averaged. If only one motion vector is present, it will be considered as is to form the paired candidate.

[0114] Candidate with zero motion vector.

[0115] The three new blending modes include MMVD (Motion Vector Difference Blending Mode), GPM (Geometric Partitioning Mode), and CIIP (Combined Intra / Inter Prediction). These new modes are detailed in the document JVET-T2002-v2: Algorithm description for Versatile Video Coding and Test Model 11 (VTM11), Joint Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29, 20th Meeting, by teleconference, October 7-16, 2020, Jianle Chen, Yan Ye, Seung Hwan Kim.

[0116] MMVD can be seen as a type of merge mode in which a merge candidate is refined by aMVd. In MMVD, after a merge candidate is selected, it is subsequently refined by information from Petition 870250085959, dated 09 / 23 / 2025, page 63 / 123 18 / 51 Signaled MVd. The signaling of an MMVD mode comprises a merge candidate flag, an index to specify a motion magnitude, and an index indicating a motion direction. The merge candidate flag is signaled to specify which one is used between the first and second merge candidates. The index specifying a motion magnitude and the index indicating a motion direction allow signaling a limited number of motion vector differences (MVd) in addition to a signaled merge candidate, i.e., “4” vector directions and “8” magnitude values.

[0117] As video compression standards evolve, the family of inter modes has grown significantly and now comprises many different inter modes.

[0118] For example, it has been observed that a translational motion model cannot accurately represent motions such as approach / retreat, rotation, perspective motions, and other irregular motions. To deal with this problem, in some implementations, a block-based affine transform motion compensation prediction has been proposed.

[0119] Figures 7A and 7B illustrate a block-based affine transformation motion compensation applied to a current block Cur. As shown in Figures 6A and 6B, an affine motion field of the current block is described by motion information from two control point motion vectors (CPMVs) (4 parameters) in Figure 7A or three control point motion vectors (6 parameters) in Figure 7B.

[0120] For the 4-parameter affine motion model, a motion vector at the sample location (x, y) in the current block is derived as follows: mv. rw mvlv-mvnv= W iv iv Eq. 1 Petition 870250085959, dated 09 / 23 / 2025, page 64 / 123 19 / 51

[0121] For the 6-parameter affine motion model, the motion vector at the sample location (x, y) in the current block is derived as follows: ____ ........ mv* -------x +-------y + wflJÍιηυ1ν-ιηνονιηΐί,-μ-ίηί'πν 2=— w—x +—H—y+

[0122] Where (mvox. mvoy) is a motion vector at a control point in the upper left corner of the current block, (mvix, mvia) is a motion vector at a control point in the upper right corner of the current block and (mv2x, mv2y) is a motion vector from a control point in the lower left corner of the current block.

[0123] To simplify motion compensation prediction, a sub-block based affine transform prediction was proposed.

[0124] Figure 8 illustrates an affine transform prediction based on subblocks. To derive a motion vector for each 4x4 luma subblock of a current block, a motion vector from a central sample of each subblock is calculated according to equations Eq. 1 or Eq. 2 and rounded to a fractional accuracy of 1 / 16. Then, motion compensation interpolation filters are applied to generate the prediction for each subblock with the derived motion vector. The subblock size of the chroma components is also defined as 4x4. The motion vector of a 4x4 chroma subblock is calculated as the average of the motion vectors of the upper left and lower right luma subblocks in a colocalized 8x8 luma region.

[0125] As with translational inter-motion prediction, there are also two affine inter-motion prediction modes: affine merging mode and affine AMVP mode.

[0126] In affine merge mode, the control point motion vectors (CPMVs) of a current block are generated based on the motion information of neighboring space blocks.

[0127] There may be up to five candidates for predictor of CPMV (CPMVP) and one Petition 870250085959, dated 09 / 23 / 2025, page 65 / 123 The 20 / 51 index is flagged to indicate which one to use for the current block. The following three types of CPVM candidates are used to form an affine merge candidate list:

[0128] Candidates for related mergers inherited from the CPMVs of neighboring related blocks;

[0129] Constructed affine merge candidates CPMVPs that are derived using the translational motion vectors of neighboring blocks;

[0130] Zero motion vectors.

[0131] In some implementations, there are at most two inherited affine candidates that are derived from the affine motion model of neighboring blocks, one from the left neighboring blocks and one from the above neighboring blocks. The candidate blocks are shown in Figure 6A. For the left predictor, the scan order is A0>A1, and for the above predictor, the scan order is B0->B1->B2. Only the first inherited candidate from each side is selected. No pruning check is performed between two inherited candidates. When a neighboring affine block is identified, its CPMVs are used to derive the CPMVP candidate in the current block's affine merge list. As shown in Figure 9, if the lower left neighboring block A is encoded in affine mode, the motion vectors v2, v3, and v4 from the upper left corner, above the right corner, and from the lower left corner of a block containing block A are hit.When block A is encoded with a 4-parameter affine model, the two current block CPMVs are calculated according to v2 and v3. If block A is encoded with a 6-parameter affine model, the three current CU CPMVs are calculated according to v2, v3, and v4.

[0132] Affine constructed candidate means that the candidate is constructed by combining the neighboring translational motion information of each control point. The motion information for the control points is derived from specified spatial and temporal neighbors represented in Figure 10. CPMVk (k=1, Petition 870250085959, dated 09 / 23 / 2025, page 66 / 123 21 / 51 2, 3, 4) represents the k-th control point. For CPMVi, blocks B2->B3->A2 are checked and the motion vector of the first available block is used. For CPMV2, blocks B1->B0 are checked and for CPMVs, blocks A1->A0 are checked. TMVP is used as CPMV4 if available.

[0133] After the motion vectors of four control points are obtained, the affine merge candidates are constructed based on this motion information. The following combinations of control point MVs are used to construct them in order: {CPMV1, CPMV2, CPMV3}, {CPMV1, CPMV2, CPMV4}, {CPMV1, CPMV3, CPMV4}, {CPMV2, CPMV3, CPMV4}, {CPMV1, CPMV2}, {CPMV1, CPMV3}

[0134] The combination of three CPMVs builds a 6-parameter affine merge candidate, and the combination of 2 CPMVs builds a 4-parameter affine merge candidate. To avoid the motion sizing process, if the reference indices of the control points are different, the related combination of control point MVs will be discarded.

[0135] After the inherited affine merge candidates and the constructed affine merge candidates are checked. If the list is not yet full, zero-move vectors will be inserted at the end of the list.

[0136] In addition, some implementations have proposed a sub-block merging mode. The sub-block merging mode uses a sub-block temporal motion vector prediction to generate a sub-block temporal motion predictor (SbTMVP). The SbTMVP differs from a regular TMVP (RTMVP), as described above, in the following two main aspects:

[0137] RTMVP provides movement at the block level, while SbTMVP provides movement at the sub-block level;

[0138] while the RTMVP is derived from a block placed in the placed figuration, the position of the actual block is first shifted before deriving the SbTMVP from Petition 870250085959, dated 09 / 23 / 2025, page 67 / 123 22 / 51 a block placed with the position offset from the current block in the placed figure. The change, called the displacement of movement hereafter, is obtained from a motion vector of a block spatially adjacent to the current block.

[0139] Figure 12 illustrates an example of a process that allows deriving the temporal motion vector predictor of the sub-block.

[0140] The sub-block motion vector prediction predicts the motion vectors of sub-blocks within a current block 1110 of a current configuration 111 in two steps:

[0141] In the first stage, the blocks spatially adjacent to the current block 1110 are examined. Figure 11 represents the spatially adjacent blocks considered in the process of predicting the temporal motion vector of the sub-block. As can be seen in Figure 11, four blocks are considered: two blocks A1 and A0 located in the lower left corner of the current block 1110, and two blocks B1 and B0 located in the upper right corner of the current block 1110. The spatially adjacent blocks are examined in the order A1, B1, B0, and A0. In this order, as soon as a spatially adjacent block with a motion vector pointing towards the placed figure 110 is identified, that motion vector is selected to be the motion change to be applied. If no motion vector is identified in the spatially adjacent blocks A1, B1, B0, and A0, then the motion change is defined as (0, 0), that is, no motion.

[0142] In the second step, the motion change identified in the first step is applied to the position of the current block 1110 (i.e., added to the coordinates of the current block 1110). Then, sub-block level motion data (motion vectors and reference indices) are derived from a current block 1100 of the placed figure 110 placed with the displaced position of the current block 1110. In the example in Figure 12, the motion change is assumed to be defined by the motion of block Al. For each sub-block of the current block 1110, the motion data of its Petition 870250085959, dated 09 / 23 / 2025, page 68 / 123 The corresponding 23 / 51 sub-block (the smallest motion grid covering the central sample) in block 1100 is used to derive motion data for that sub-block from the current block 1110. The SbTMVP derivation is then finalized by applying a temporal motion vector scale to the derived motion vectors for each sub-block to align the reference figures of these derived motion vectors with those of the current block 1110. For each sub-block, the scaled motion vector is used as a motion vector for the sub-block.

[0143] The sub-block size used in SbTMVP is generally 8x8. In this case, the SbTMVP mode is only applicable to blocks with a width and height greater than or equal to “8”.

[0144] In some implementations, a combined sub-block-based merge list containing SbTMVP candidates and related merge candidates is used for signaling the sub-block-based merge mode. The SbTMVP mode is enabled / disabled by a sequence parameter set (SPS) flag. If the SbTMVP mode is enabled, the SbTMVP predictor will be added as the first entry in the sub-block-based merge candidate list, followed by related merge candidates. The size of the sub-block-based merge list is signaled in the SPS, and the maximum allowed size of the sub-block-based merge list is usually “5”.

[0145] To further improve the efficiency of motion information compression, some implementations reduce the bit rate of the motion information, allowing decoders to determine a portion of this motion information. For example, it is proposed to let a decoder refine the motion information. To maintain consistency between the encoder and the decoder, the processes applied on the decoder side are identically replicated on the encoder side. Petition 870250085959, dated 09 / 23 / 2025, page 69 / 123 24 / 51

[0146] Some methods that allow refining motion information are based on model matching.

[0147] Model matching (TM) is a decoder-side motion vector derivation method for refining the motion information of a current block by finding a closest match between a model (i.e., a set of reconstructed samples) neighboring the current block in the current figure and a reference model in a reference figure. Figure 13 illustrates a model matching-based method. As illustrated in Figure 13, a better motion vector is searched around an initial motion vector of a current block Curr within a [-8, 8]-pel search window.

[0148] In AMVP mode, a motion vector predictor candidate is determined based on the model matching error to select the one that achieves the minimum difference between the current block model and the reference block model. Then, TM is run only for that specific motion vector predictor candidate for motion vector refinement. TM refines this motion vector predictor candidate, starting from the accuracy of the full pel motion vector difference within a [-8, 8]-pel search window using iterative diamond search. The resulting motion vector predictor candidate can be further refined using cross-search with full pel motion vector difference accuracy, followed sequentially by half pel and quarter pel.In the search process, if the difference between the previous minimum cost and the current minimum cost in the iteration is less than a limit equal to the block area, the search process will be terminated.

[0149] In merge mode, a similar search method is applied to the merge candidate indicated by a merge index.

[0150] TM can also be used to adaptively rearrange the Petition 870250085959, dated 09 / 23 / 2025, page 70 / 123 25 / 51 merge candidates proposed in the Adaptive Reordering of Merge Candidates with Model Matching (ARMC-TM) method described in section 2.7 of the document JVET-X2025-V2: Algorithm description of Enhanced Compression Model 3 (ECM 3), Muhammed Coban, Joint Video Experts Team (JVET), of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29, 23rd Meeting, by teleconference, July 7-16, 2021.

[0151] After a list of merge candidates is constructed, the merge candidates are divided into several subgroups. The merge candidates in each subgroup are reordered in ascending order.

[0152] according to cost values ​​based on model matching. For simplicity, merge candidates in the last, but not the first subgroup, are not reordered.

[0153] The model matching cost of a merge candidate is measured by the sum of absolute differences (SAD) between samples of a model from the current block and their corresponding reference samples. The model comprises a set of reconstructed samples neighboring the current block. Reference samples of the model are located by the move information of the merge candidate.

[0154] For merge candidates based on sub-blocks with sub-block size equal to Wsub x Hsub, a model above a current block comprising the sub-block is split into multiple submodels with size Wsub x 1, and a model to the left of the current block is split into multiple submodels with size 1 x Hsub. The movement information of the sub-blocks in the first row and first column of the current block is used to derive the reference samples of each submodel.

[0155] During a selection step 306, the prediction mode that optimizes compaction performance, according to an optimization criterion of Petition 870250085959, dated 09 / 23 / 2025, page 71 / 123 26 / 51 rate / distortion (i.e., RDO criterion), among the prediction modes tested (intra-prediction modes, inter-prediction modes), is selected by the encoding module.

[0156] When the prediction mode is selected, the residue block is transformed during a 307 step. In some implementations, a plurality of transform types can be applied to a transformed residue block. In fact, in addition to DCT-II, a Multiple Transform Selection (MTS) scheme is used for inter- and intra-predicted blocks. This uses multiple selected transforms from DCT-VIII / DST-VII.

[0157] The transformed block is then quantized during a 309 step.

[0158] Note that the encoding module can skip the transformation and apply quantization directly to the untransformed residual signal.

[0159] The quantized residual block determined for the current block during an inter or intra prediction is encoded by an entropic encoder during a step 310. Note that the encoding module can ignore transformation and quantization, i.e., entropic encoding is applied to the residual without applying transformation or quantization processes. The result of the input encoding is inserted into the video data 311.

[0160] When the current block is encoded according to an intra prediction mode, the intra prediction mode is encoded by the entropic encoder during step 310 in the video data 311.

[0161] When the current block is encoded according to an inter prediction, the inter mode and the result of the process applied to encode the motion information are then encoded by the entropic encoder during step 310 in the video data 311.

[0162] Metadata such as SEI (Supplemental Enhancement Information) messages can be appended to the 311 encoded video stream. An SEI message, as defined, for example, in standards such as AVC, HEVC, or VVC, is a Petition 870250085959, dated 09 / 23 / 2025, page 72 / 123 27 / 51 data container associated with a video stream and comprises metadata that provides information relating to the video stream.

[0163] After quantization step 309, the current block is reconstructed so that the pixels corresponding to that block can be used for future predictions. This reconstruction phase is also called the prediction loop. An inverse quantization is therefore applied to the transformed and quantized residue block during step 312, and an inverse transformation is applied during step 313. According to the prediction mode used for the block obtained during step 314, the block's prediction block is reconstructed. If the current block is encoded according to an inter-prediction mode, the encoding module applies, when appropriate, during step 316, a motion compensation using the motion information of the current block to identify each reference block of the current block. If the current block is encoded according to an intra-prediction mode, during step 315, the intra-prediction mode is used to reconstruct the prediction block of the current block.The prediction block and the reconstructed residue block are added together to obtain the reconstructed current block.

[0164] After reconstruction, a loop filtering intended to reduce encoding artifacts is applied, during a step 317, to the reconstructed block. This filtering is called loop filtering, since this filtering occurs in the prediction process to obtain, in the decoder, the same reference figures as the encoder and, therefore, avoid a departure between the encoding and decoding processes. Loop filtering tools include unlock filtering, SAO (Adaptive Sample Shift) and ALF (Adaptive Loop Filtering).

[0165] When a block is reconstructed, it is inserted during a step 318 into a reconstructed picture stored in a memory 319 of reconstructed pictures, usually called the Decoded Picture Buffer (DPB). The figures Petition 870250085959, dated 09 / 23 / 2025, page 73 / 123 28 / 51 reconstructed rations stored in this way can then serve as reference figurations for other figurations to be coded.

[0166] Figure 4 schematically represents a method for decoding the encoded video stream 311 encoded according to the method described in relation to Figure 3, executed by a decoding module. Variations of this decoding method are contemplated, but the decoding method of Figure 4 is described below for clarity, without describing all expected variations.

[0167] Decoding is done block by block. For a current block, it starts with an entropic decoding of the current block during a 410 step. Entropic decoding allows obtaining, at least, the prediction mode of the block.

[0168] If the current block has been encoded according to an inter prediction mode, entropic decoding allows obtaining, when appropriate, representative information of a movement of the current block and of a residue block. During a 408 step, the movement information is reconstructed for the current block using the decoded information representative of the movement information.

[0169] If the applied inter-block prediction mode involves a refinement of the decoding lateral motion vector, the specified refinement process will be applied to the motion information of the current block. When the decoding lateral motion vector refinement process is based on models involving samples of reconstructed blocks neighboring the current block, dependencies are created between the neighboring blocks and the current block. These dependencies can prevent parallel processing of the current and neighboring blocks. As can be seen, model-based inter-blocks introduce decoding latencies compared to the non-model-based inter-block mode.

[0170] If the block was encoded according to an intra-prediction mode, entropic decoding allows obtaining the intra-prediction mode and a residue block. Steps 412, 413, 414, 415, 416 and 417 implemented by the decoding module Petition 870250085959, dated 09 / 23 / 2025, page 74 / 123 29 / 51 are identical in all respects to steps 412, 413, 414, 415, 416 and 417 implemented by the coding module, respectively.

[0171] The decoded blocks are saved in decoded figures, and the decoded figures are stored in a DPB 419 in a step 418. When the decoding module decodes a given figure, the figures stored in DPB 419 are identical to the figures stored in DPB 319 by the encoding module during the encoding of said provided image. The decoded figure can also be output by the decoding module, for example, to be displayed.

[0172] Post-processing step 421 may comprise an inverse color transformation (e.g., converting YCbCr 4:2:0 to RGB 4:4:4), an inverse mapping performing the reverse of the remapping process performed in pre-processing step 301, and post-filtering to improve the reconstructed figures based, for example, on filter parameters provided in an SEI message.

[0173] Figure 5A schematically illustrates an example of the hardware architecture of a processing module 500 capable of implementing an encoding module or a decoding module capable of implementing, respectively, a method for encoding as shown in Figure 3 and a method for decoding as shown in Figure 4, modified according to different aspects and modalities. The encoding module is, for example, included in system 11 when this device is responsible for encoding the video stream. The decoding module is, for example, included in system 13. The processing module 500 comprises, connected by a communication bus 5005: a processor or CPU (central processing unit) 5000 encompassing one or more microprocessors, general-purpose computers, special-purpose computers, and basic processors. Petition 870250085959, dated 09 / 23 / 2025, page 75 / 123 30 / 51 seated in a multi-core architecture, as non-limiting examples; a random access memory (RAM) 5001; a read-only memory (ROM) 5002; a storage unit 5003, which may include non-volatile memory and / or volatile memory, including, but not limited to, electrically erasable programmable read-only memory (EEPROM), read-only memory (ROM), programmable read-only memory (PROM), random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), flash, magnetic disk drive and / or optical disk drive, or a storage media reader, such as an SD (Secure Digital) card reader and / or a hard disk drive (HDD) and / or a network accessible storage device; at least one communication interface 5004 for exchanging data with other modules, devices or equipment.The 5004 communication interface may include, but is not limited to, a transceiver configured to transmit and receive data over a communication channel. The 5004 communication interface may include, but is not limited to, a modem or network card.

[0174] If processing module 500 implements a decoding module, communication interface 5004 allows, for example, processing module 500 to receive encoded video streams and provide a sequence of decoded figures. If processing module 500 implements an encoding module, communication interface 5004 allows, for example, processing module 500 to receive a sequence of original figure data to encode and provide an encoded video stream.

[0175] Processor 5000 is capable of executing instructions loaded into RAM 5001 from ROM 5002, from external memory (not shown), from a storage medium, or from a communication network. When processing module 500 is powered on, processor 5000 is able to read instructions from RAM 5001 and execute them. These instructions form a computer program that causes, by Petition 870250085959, dated 09 / 23 / 2025, page 76 / 123 31 / 51 example, the implementation by processor 5000 of a decoding method as described in relation to Figure 4, an encoding method described in relation to Figure 3 and methods described in relation to Figures 14A or 14B, these methods comprising various aspects and modalities described below in this document.

[0176] All or some of the algorithms and steps of the methods in Figures 3, 4, 14A and 14B can be implemented in software form by executing a set of instructions by a programmable machine, such as a DSP (digital signal processor) or a microcontroller, or implemented in hardware form by a dedicated machine or component, such as an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

[0177] As can be observed, microprocessors, general-purpose computers, special-purpose computers, processors based on or not on multi-core architecture, DSP, microcontroller, FPGA and ASIC are electronic circuits adapted to implement at least partially the methods of Figures 3, 4, 14A and 14B.

[0178] Figure 5C illustrates a block diagram of an example of System 13 in which various aspects and modalities are implemented. System 13 can be embodied as a device that includes the various components described below and is configured to perform one or more of the aspects and modalities described in this document. Examples of such devices include, but are not limited to, various electronic devices such as personal computers, laptops, smartphones, tablets, digital multimedia set-top boxes, digital television receivers, personal video recording systems, connected household appliances, and head-mounted displays. Elements of System 13, individually or in combination, can be embodied in a single integrated circuit (IC), multiple ICs, and / or discrete components. For example, in at least one embodiment, System 13 comprises a module of Petition 870250085959, dated 09 / 23 / 2025, page 77 / 123 32 / 51 processing 500 that implements a decoding module. In various embodiments, system 13 is communicatively coupled to one or more other systems, or other electronic devices, by means of, for example, a communications bus or by means of dedicated input and / or output ports. In various embodiments, system 13 is configured to implement one or more of the aspects described in this document.

[0179] Input to the 500 processing module may be provided by means of various input modules, as indicated in block 531. These input modules include, but are not limited to, (i) a radio frequency (RF) module that receives an RF signal transmitted, for example, through the air by a transmitter, (ii) a component input module (COMP) (or a set of COMP input modules), (iii) a Universal Serial Bus (USB) input module, and / or (iv) a High Definition Multimedia Interface (HDMI) input module. Other examples, not shown in Figure 5D, include composite video.

[0180] In various embodiments, the 531 block input modules have respective associated input processing elements, as known in the art. For example, the RF module may be associated with elements suitable for (i) selecting a desired frequency (also said to be selecting a signal or limiting the bandwidth of a signal to a frequency band), (ii) downconverting the selected signal, (iii) further band-limiting to a narrower frequency band to select (e.g.) a signal frequency band which may be referred to as a channel in certain embodiments, (iv) demodulating the downconverted and band-limited signal, (v) performing error correction, and (vi) demultiplexing to select the desired stream of data packets. The RF module of various embodiments includes one or more elements to perform these functions, for example, frequency selectors, signal selectors, bandwidth limiters, channel selectors, Petition 870250085959, dated 09 / 23 / 2025, p. 78 / 123 33 / 51 filters, down-converters, demodulators, error correctors, and demultiplexers. The RF portion may include a tuner that performs several of these functions, including, for example, down-converting the received signal to a lower frequency (e.g., an intermediate frequency or a frequency close to baseband) or to baseband. In a decoder embodiment, the RF module and its associated input processing element receive an RF signal transmitted via a wired medium (e.g., cable) and perform frequency selection by means of filtering, down-converting, and filtering again to a desired frequency band. Various embodiments rearrange the order of the elements described above (and others), remove some of these elements, and / or add other elements that perform similar or different functions.Adding elements can include inserting elements between existing elements, such as inserting amplifiers and an analog-to-digital converter. In several configurations, the RF module includes an antenna.

[0181] In addition, the USB and / or HDMI modules may include respective interface processors for connecting the system 13 to other electronic devices via USB and / or HDMI connections. It should be understood that various aspects of input processing, for example, Reed-Solomon error correction, may be implemented, for example, within a separate input processing IC or within the processing module 500, as required. Similarly, aspects of USB or HDMI interface processing may be implemented in separate interface ICs or within the processing module 500, as required. The demodulated, error-corrected, and demultiplexed stream is provided to the processing module 500.

[0182] Several elements of system 13 can be provided within an integrated housing. Within the integrated housing, the various elements can be interconnected and transmit data to each other using suitable connection arrangements. Petition 870250085959, dated 09 / 23 / 2025, page 79 / 123 34 / 51 for example, an internal bus as known in the art, including the Inter-IC (I2C) bus, wiring and printed circuit boards. For example, in system 13, processing module 500 is interconnected to other elements of said system 13 by bus 5005.

[0183] The communication interface 5004 of the processing module 500 allows the system 13 to communicate on the communication channel 52. As mentioned above, the communication channel 52 can be implemented, for example, within a wired and / or wireless medium.

[0184] Data is transmitted, or otherwise provided, to system 13 in various modes using a wireless network, such as a Wi-Fi network, for example IEEE 802.11 (IEEE refers to the Institute of Electrical and Electronics Engineers). The Wi-Fi signal of these modes is received by communication channel 52 and communication interface 5004, which are adapted for Wi-Fi communications. Communication channel 52 of these modes is typically connected to an access point or router that provides access to external networks, including the Internet, to enable streaming applications and other over-the-top communications. Other modes provide transmitted data to system 13 using the RF connection of input block 531. As indicated above, several modes provide data in a non-streaming manner. In addition, several modes use wireless networks other than Wi-Fi, for example, a cellular network or a Bluetooth network.

[0185] System 13 can provide an output signal to various output devices, including the display system 55, loudspeakers 56, and other peripheral devices 57. The display system 55 of various embodiments includes one or more of the following: for example, a touch screen, an organic light-emitting diode (OLED) screen, a curved screen, and / or a foldable screen. The display 55 can be for a television, a tablet, a laptop, a mobile phone (cell phone), a mounted display. Petition 870250085959, dated 09 / 23 / 2025, page 80 / 123 35 / 51 on the head or other devices. The display system 55 may also be integrated into other components (e.g., as in a smartphone) or separate (e.g., an external monitor for a laptop). Other peripheral devices 57 include, in various embodiments, one or more stand-alone digital video discs (or digital versatile discs) (DVR, for both terms), a disc player, a stereo system, and / or a lighting system. Various embodiments use one or more peripheral devices 57 that provide a function based on the output of the system 13. For example, a disc player performs the function of playing back an output of the system 13.

[0186] In various embodiments, control signals are communicated between system 13 and display system 55, loudspeakers 56, or other peripheral devices 57 using signaling such as AV.Link, Consumer Electronics Control (CEC), or other communication protocols that allow device-to-device control with or without user intervention. Output devices can be communicatively coupled to system 13 via dedicated connections through their respective interfaces 532, 533, and 534. Alternatively, output devices can be connected to system 13 using communication channel 52 via communication interface 5004 or a dedicated communication channel corresponding to communication channel 54 in Figure 5A via communication interface 5004.The display system 55 and the speakers 56 can be integrated into a single unit with the other system components 13 in an electronic device, such as, for example, a television. In several embodiments, the display interface 532 includes a display driver, such as, for example, a timing controller chip (T Con).

[0187] The display system 55 and the loudspeaker 56 may alternatively be separated from one or more of the other components. In various embodiments in which the display system 55 and the loudspeakers 56 are external components, the Petition 870250085959, dated 09 / 23 / 2025, page 81 / 123 36 / 51 output signal can be provided via dedicated output connections, including, for example, HDMI ports, USB ports or COMP outputs.

[0188] Figure 5B illustrates a block diagram of an example of System 51 in which various aspects and modalities are implemented. System 51 is very similar to System 13. System 51 can be embodied as a device that includes the various components described below and is configured to perform one or more of the aspects and modalities described in this document. Examples of such devices include, but are not limited to, various electronic devices such as personal computers, laptops, smartphones, tablets, a camera, and a server. Elements of System 51, individually or in combination, can be embodied in a single integrated circuit (IC), multiple ICs, and / or discrete components. For example, in at least one embodiment, System 51 comprises a processing module 500 that implements an encoding module.In various embodiments, system 51 is communicatively coupled to one or more other systems, or other electronic devices, by means of, for example, a communications bus or by means of dedicated input and / or output ports. In various embodiments, system 51 is configured to implement one or more of the aspects described in this document.

[0189] Input for processing module 500 can be provided via various input modules, as indicated in block 531 already described in relation to Figure 5D.

[0190] Several elements of system 51 can be provided within an integrated housing. Within the integrated housing, the various elements can be interconnected and transmit data to each other using suitable connection arrangements, for example, an internal bus as known in the art, including the Inter-IC (I2C) bus, wiring, and printed circuit boards. For example, in the system Petition 870250085959, dated 09 / 23 / 2025, page 82 / 123 37 / 51 51, processing module 500 is interconnected to other elements of said system 51 by bus 5005.

[0191] The 5004 communication interface of the 500 processing module allows the 500 system to communicate on communication channel 52.

[0192] Data is transmitted, or otherwise provided, to system 51 in various modes using a wireless network, such as a Wi-Fi network, for example IEEE 802.11 (IEEE refers to the Institute of Electrical and Electronics Engineers). The Wi-Fi signal of these modes is received by communication channel 52 and communication interface 5004, which are adapted for Wi-Fi communications. Communication channel 52 of these modes is typically connected to an access point or router that provides access to external networks, including the Internet, to enable streaming applications and other over-the-top communications. Other modes provide data transmitted to system 51 using the RF connection of input block 531.

[0193] As indicated above, several modalities provide data in a non-streaming manner. In addition, several modalities use wireless networks other than Wi-Fi, for example, a cellular network or a Bluetooth network.

[0194] The data provided to system 51 can be provided in different formats. In several modes, this data is encoded and compatible with a known video compression format, such as AVI, VP9, ​​VVC, HEVC, AVC, etc. In several modes, this data is raw data provided, for example, by a figure and / or audio acquisition module connected to or included in system 51. In this case, the processing module is responsible for encoding this data.

[0195] System 51 can provide an output signal to various output devices capable of storing and / or decoding the output signal, such as system 13. Petition 870250085959, dated 09 / 23 / 2025, page 83 / 123 38 / 51

[0196] Several implementations involve decoding. “Decoding”, as used in this application, may encompass all or part of the processes performed, for example, on a received encoded video stream, in order to produce a final output suitable for display.

[0197] In several embodiments, such processes include one or more of the processes normally performed by a decoder, for example, entropy decoding, inverse quantization, inverse transformation, and prediction. In several embodiments, such processes also, or alternatively, include processes performed by a decoder of various implementations described in this application, for example, to reorder a list of motion vector predictors.

[0198] Whether the phrase “decoding process” is intended to refer specifically to a subset of operations or, more generally, to the broader decoding process will become clear from the context of the specific descriptions and is believed to be well understood by those skilled in the art.

[0199] Various implementations involve encoding. Analogous to the discussion above about “decoding”, “encoding”, as used in this application, can encompass all or part of the processes performed, for example, on an input video sequence to produce an encoded video stream. In various embodiments, such processes include one or more of the processes normally performed by an encoder, for example, partitioning, prediction, transformation, quantization, and entropy encoding. In various embodiments, such processes also, or alternatively, include processes performed by an encoder of various implementations described in this application, for example, to reorder a list of motion vector predictors.

[0200] Whether the phrase “coding process” is intended to refer specifically to a subset of operations or, more generally, to the broader coding process will become clear based on the context of the specific descriptions and is believed Petition 870250085959, dated 09 / 23 / 2025, page 84 / 123 39 / 51 so that it is well understood by those versed in the technique.

[0201] Note that the names of the syntax elements used here are descriptive terms. As such, they do not preclude the use of other names for syntax elements.

[0202] When a figure is presented as a flow diagram, it should be understood that it also provides a block diagram of a corresponding device. Similarly, when a figure is presented as a block diagram, it should be understood that it also provides a flow diagram of a corresponding method / process.

[0203] Several approaches relate to rate distortion optimization. In particular, during the encoding process, the balance or trade-off between a rate and distortion is usually considered. Rate distortion optimization is generally formulated as the minimization of a rate distortion function, which is a weighted sum of the rate and distortion. There are different approaches to solving the rate distortion optimization problem. For example, approaches may be based on extensive testing of all encoding options, including all considered modes or encoding parameter values, with a thorough evaluation of the encoding cost and related distortion of a reconstructed signal after encoding and decoding.Faster approaches can also be used to save coding complexity, particularly by calculating an approximate distortion based on a prediction or a residual prediction signal, not the reconstructed one. A combination of these two approaches can also be used, such as using an approximate distortion for only some of the possible coding options and a full distortion for other coding options. Other approaches evaluate only a subset of the possible coding options. More generally, many approaches employ a variety of techniques to perform optimization, but optimization is not necessarily the same. Petition 870250085959, dated 09 / 23 / 2025, page 85 / 123 40 / 51 a complete assessment of the cost of coding and related distortion.

[0204] The implementations and aspects described in this document can be implemented in, for example, a method or a process, a device, a software program, a data stream, or a signal. Even if discussed only in the context of a single implementation form (for example, discussed only as a method), the implementation of the discussed features can also be implemented in other forms (for example, a device or program). A device can be implemented, for example, in appropriate hardware, software, and firmware. Methods can be implemented, for example, in a processor, which refers to processing devices in general, including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device.Processors also include communication devices, such as computers, mobile phones, personal / portable digital assistants (“PDAs”), and other devices that facilitate the communication of information between end users.

[0205] The reference to “a modality” or “a modality” or “an implementation” or “an implementation”, as well as other variations thereof, means that a specific characteristic, structure, feature, and so forth described in connection with the modality are included in at least one modality. Thus, occurrences of the phrase “in a (1) modality” or “in a modality” or “in an implementation” or “in a (1) implementation”, as well as any other variations, which appear in various places throughout this application are not necessarily referring to the same modality.

[0206] In addition, this application may refer to the “determination” of various information. The determination of information may include one or more of the following: for example, estimating information, calculating information, predicting information, retrieving information from memory, or obtaining information, for example, from another device, module, or the user. Petition 870250085959, dated 09 / 23 / 2025, page 86 / 123 41 / 51

[0207] In addition, this request may refer to “access” to various information. Access to information may include one or more of the following: for example, receiving information, retrieving information (e.g., from memory), storing information, moving information, copying information, calculating information, determining information, predicting information, or estimating information.

[0208] Furthermore, this request may refer to the “receiving” of various information. Receiving, as well as “accessing,” is intended to be a broad term. Receiving information may include one or more of the following: for example, accessing the information or retrieving it (e.g., from memory). In addition, “receiving” is normally involved, in one form or another, during operations such as, for example, storing the information, processing the information, transmitting the information, moving the information, copying the information, deleting the information, calculating the information, determining the information, predicting the information, or estimating the information.

[0209] It should be understood that the use of any of the following “ / ”, “and / or” and “at least one of”, “one or more of”, for example, in the cases of “A / B”, “A and / or B” and “at least one of A and B”, “one or more of A and B” is intended to cover the selection of the first option listed (A) only, or the selection of the second option listed (B) only, or the selection of both options (A and B). As a further example, in the cases of “A, B and / or C” and “at least one of A, B and C”, “one or more of A, B and C”, such formulation intends to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of the third listed option (C) only, or the selection of the first and second listed options (A and B) only, or the selection of the first and third listed options (A and C) only, or the selection of the second and third listed options (B and C) only, or the selection of all three options (A, B and C). This can be extended, as is evident, to Petition 870250085959, dated 09 / 23 / 2025, page 87 / 123 42 / 51 someone with average skill in this technique and related techniques, for as many items as are listed.

[0210] Furthermore, as used in this document, the word “signal” refers, among other things, to indicating something to a corresponding decoder. For example, in certain embodiments, the encoder signals the use of some encoding tools. In this way, in one embodiment, the same parameters can be used on both the encoder and decoder sides. Thus, for example, an encoder can transmit (explicit signaling) a particular parameter to the decoder so that

[0211] the decoder can use the same specific parameter. On the other hand, if the decoder already has the specific parameter, as well as others, signaling can be used without transmission (implicit signaling) to simply allow the decoder to know and select the specific parameter. By avoiding the transmission of any actual functions, a saving of bits is achieved in several modes. It is important to emphasize that signaling can be performed in several ways. For example, one or more syntax elements, flags, and so on are used to signal information to a corresponding decoder in several modes. Although the preceding relates to the verbal form of the word “signal”, the word “signal” can also be used here as a noun.

[0212] As will be evident to anyone with average skill in the field, implementations can produce a variety of formatted signals to carry information that can, for example, be stored or transmitted. The information may include, for example, instructions to execute a method or data produced by one of the described implementations. For example, a signal may be formatted to carry the encoded video stream and SEI messages of a described modality. This signal may be formatted, for example, as an electromagnetic wave (e.g., using a radio frequency portion of the spectrum) or Petition 870250085959, dated 09 / 23 / 2025, page 88 / 123 43 / 51 as a baseband signal. Formatting may include, for example, encoding a video stream and modulating a carrier with the video stream. The information the signal carries may be, for example, analog or digital information. The signal may be transmitted over a variety of wired or wireless links, as is known. The signal may be stored on a processor-readable medium.

[0213] Next, methods are proposed to increase the possibilities of parallelizing operations in a decoding process. A low-complexity reordering criterion is introduced for bi-prediction motion candidates. The use of this criterion allows avoiding the latency problem caused by reordering methods based on matching known patterns.

[0214] Figure 14A schematically illustrates a predictor reordering process in a predictor list implemented by an encoding module.

[0215] The process in Figure 14A is, for example, implemented by processing module 500 of system 11 when system 11 implements an encoding module that implements, for example, the encoding method in Figure 3. The process in Figure 14A is executed, for example, during step 304 if the encoding of motion information is taken into account in the mode selection of step 306. In any case, the process in Figure 14A is executed during step 308 for the encoding of motion information.

[0216] In a step 1401, processing module 500 obtains a list of candidates for motion vector predictors. The list is, for example, a list of merge candidates, a list of affine merge candidates, etc.

[0217] In a 1402 step, processing module 500 obtains bi-prediction candidates from the list.

[0218] In a step 1403, processing module 500 initializes a Petition 870250085959, dated 09 / 23 / 2025, page 89 / 123 44 / 51 variable / to zero.

[0219] In a step 1404, processing module 500 determines if the variable / is less than a number of bi-prediction candidates in the NumBíPred list.

[0220] If so, processing module 500 continues with step 1405.

[0221] In step 1405, processing module 500 calculates a score[i] for a / -th bi-prediction candidate MVP_BíPred[i]. Similar to TM-based methods, the bi-prediction candidate score is a representative value of a difference between two models. However, to avoid latencies observed in the traditional TM-based method and to allow parallelization of the processing of consecutive interpreted blocks, the model is located within each of the two predictions, which does not depend on the reconstruction of the current slice that makes up the current block.

[0222] For example, in Figure 15, for a current block C in the current configuration, two temporal predictors are used to create the final prediction: P0 the block C is displaced in the reference configuration refO and P1 the displaced block in the reference configuration ref1. The prediction P0 or P1 are not necessarily created by applying a uniform motion across the entire block, but may be motion models based on sub-blocks (e.g., affine or SbTMVP).

[0223] For each prediction (i.e., P0 and P1), a model T is created containing the motion-compensated samples. An error metric representing a difference between the model T° extracted from prediction P0 and the model V extracted from prediction P1 is used to calculate the candidate score MVP_BíPred[í], for example, as follows: D= Σ'^^I

[0224] Here, the score is calculated as a SAD (sum of absolute differences) between the two models. However, other metrics can be used, such as Petition 870250085959, dated 09 / 23 / 2025, pp. 90 / 123 45 / 51 is the sum of squared differences (SSD) or a weighted sum of absolute differences, where the weights depend on the sample position.

[0225] The main advantage of using samples within predictions is that it does not introduce latency in the pipeline, since the samples are already available when decoding the current Block C (whereas for traditional TM-based methods, it is necessary to wait for the neighboring block to be reconstructed).

[0226] In one variant, instead of using a model corresponding to an entire prediction block to calculate the score, the score is calculated on a subpart of the prediction blocks. In Figure 15, a model is used in the lower right corner. The main advantage of this location is that if the candidate has inconsistent movement between the “2” predictors, it is likely to be more different in the lower right location than in the upper left location. In fact, candidates usually come from neighboring blocks, so movement near the upper left edge is likely to be correlated with movement from the upper or left blocks. By using a model far from these boundaries, we increase the probability of rejecting poor candidates.

[0227] In a variant, on the other hand, to have a behavior similar to that of the current matching model, one can choose to use reconstructed samples in the upper left corner. Favoring some samples in the difference calculation can also be done using sample-based difference weighting, with higher weights for the area to be favored.

[0228] In one variant, the model size is chosen to be only N pixels thick (e.g. JV=1) to minimize complexity.

[0229] In one variant, the model's motion compensation is simplified and an integer-based motion compensation (pixel copy) based on the nearest integer motion is computed. In another variant, Petition 870250085959, dated 09 / 23 / 2025, pp. 91 / 123 46 / 51 A low-complexity filtering (e.g., bilinear filter) is used to perform motion compensation of the model samples.

[0230] It can be noted that in the case of motion vector predictor candidates where a motion correction (mvd) is added before motion compensation (e.g. in AMVP mode or MMVD candidates), then the motion vector difference mvd is added to each candidate before calculating the score.

[0231] It may also be noted that, in some cases, a current block may be temporally predicted from more than two reference blocks (that is, a current block may be predicted from more than two forecasts). In this case, it is no longer a bi-predicted block, but rather a multi-predicted block. When more than two forecasts are available, the score is the normalized score (by the number of forecast pairs) between each pair of forecasts or an average or a weighted average of the scores between each pair of forecasts.

[0232] In a step 1406, the processing module 500 increments i by one unit.

[0233] If during step 1404, i is equal to the number of biprediction candidates in the NumBiPred list, step 1404 is followed by step 1407.

[0234] In step 1407, processing module 500 reorders the bi-prediction candidates in the list based on the computed scores to obtain a reordered list. The bi-prediction candidates are reordered from lowest to highest score.

[0235] In a step 1408, processing module 500 signals an index of the motion vector predictor from the selected reordered list to predict the motion information of the current block. Note that no index is signaled if the selected motion vector predictor is in the first position in the reordered list. Petition 870250085959, dated 09 / 23 / 2025, pp. 92 / 123 47 / 51

[0236] Figure 14B schematically illustrates a predictor reordering process in a predictor list implemented by a decoding module.

[0237] The process in Figure 14B is, for example, implemented by processing module 500 of system 13 when system 13 implements a decoding module that implements, for example, the decoding method in Figure 4. The process in Figure 14B is executed, for example, during step 408.

[0238] Steps 1401 to 1407 described in relation to Figure 14A are applied identically during the process in Figure 14B.

[0239] Stage 1408 is replaced by stage 1408Bis.

[0240] During step 1408Bis, the index of the motion vector predictor in the reordered list is decoded by processing module 500. If no index is signaled, processing module 500 has determined that the selected motion vector predictor is in the first position in the reordered list.

[0241] In a first variant, at step 1407, if the list also contains uni-prediction candidates, only the bi-prediction candidates are reordered and the uni-prediction candidates retain the same index in the reordered list as in the “initial” list. index Candidate 0 Uni0 1 Bi0 2 Bi1 3 Uni1 4 Bi2 5 Bi3 Table TAB1 index Candidate 0 Uni0 1 Bi3 2 Bi0 Petition 870250085959, dated 09 / 23 / 2025, pp. 93 / 123 48 / 51 3 Uni1 4 Bi2 5 Bi1 Table TAB2

[0242] Table TAB1 illustrates an “initial” list (before reordering). Table TAB2 illustrates a list reordered according to the first variant of step 1407.

[0243] In a second variant, between steps 1401 and 1402, the uni-prediction candidates from the “initial” list are transformed into bi-prediction candidates using a symmetrization of the motion vector with respect to the current figure comprising the current block. For example, if the current frame has POC (Figure Order Count) equal to “8” and a uni-prediction candidate has a motion mv indicating a reference figure with POC equal to “12”, this is transformed, when possible, into a bi-prediction candidate in which the second prediction uses a motion -mv and a reference with POC equal to “4” (i.e., symmetrical to “12” with respect to “8”). The resulting bi-prediction candidates then replace the uni-prediction candidates in the “initial” list as replacement bi-prediction candidates before continuing the processes in Figures 14A or 14B with steps 1402 to 1408 (respectively 1408Bis).These replacement bi-prediction candidates are used for reordering to obtain the reordered list, but once reordered, the replacement bi-prediction candidates are replaced again by the uni-prediction candidates that were replaced in the reordered list for predicting the movement information of the current block.

[0244] In a third variant, when the bi-prediction reordering mode is used, uni-prediction candidates are never inserted into the reordered list.

[0245] In a fourth variant, when the bi-prediction reordering mode is used, the uni-prediction candidates are transformed into bi-prediction candidates (using the method above) and the bi-prediction information is also used to perform the prediction of the current block movement information (and not Petition 870250085959, dated 09 / 23 / 2025, pp. 94 / 123 49 / 51 only for reordering the list).

[0246] In a fifth variant, when it is not possible to convert a uni-prediction into a bi-prediction because no reference figuration is available with the desired Picture Order Count, then the figuration closest to the current figuration is chosen, and the motion vector pointing to that closest figuration is computed with an ad hoc scale of the uni-prediction motion vector, taking into account the respective temporal distance to the two reference figurations considered.

[0247] In a sixth variant, the uni-prediction candidates are replaced by bi-prediction candidates as in the second variant, but the reordering step 1407 is not applied. A minimum average distortion minDist per sample can be signaled in the video data 311. The error metric D, representative of the difference between model T° and model T1, is computed sequentially for each candidate in the “initial” list (following the order of the “initial” list). As soon as a candidate has an error metric D value less than or equal to minDist x N, where N is the number of samples in the model, then the candidate is selected to calculate the prediction and the other candidates are not tested. This has the advantage of reducing complexity with minimal performance reduction. The “minDist” value can be signaled by slice, figure, or CTU, for example. Note that the error metric can also be calculated on a subset of samples from models T° and T1.

[0248] In a seventh variant, uni-prediction candidates are replaced by bi-prediction candidates as in the second variant, but the 1407 reordering step is not applied. In this seventh variant, a maximum distortion maxDist per sample is used to filter candidates with a difference above a threshold: only candidates with an error metric value D less than or equal to maxDist x N are considered in the initial list. The value of “maxDist” can be fixed or signaled by slice, figure, or CTU, for example. Petition 870250085959, dated 09 / 23 / 2025, pages 95 / 123 50 / 51

[0249] It may be noted that an approach similar to the sixth or seventh variant can be applied to the usual approach based on TM. However, in this case, there is no need to replace uni-prediction candidates with bi-prediction candidates. In fact, since the models are external to the blocks (i.e., external to the current block and external to the reference blocks), the distortion D for a uni-prediction candidate can be calculated between a model neighboring the current block and a model neighboring the reference block pointed to by the uni-prediction candidate.

[0250] In an eighth variant, a high-level syntax, for example, at the SPS level, PPS level, figure header level, or in a slice header, indicates that the predictor reordering process for Figure 14A is applied by the encoding module and that the predictor reordering process for Figure 14B should be applied by the decoding module.

[0251] In a ninth variant, a syntax element, for example, at the CTU or block level, indicates that the predictor reordering process of Figure 14A is applied by the encoding module for a corresponding CTU or block and that the predictor reordering process of Figure 14B should be applied by the decoding module for the same CTU or block.

[0252] In addition, the modalities may include one or more of the following attributes, devices or aspects, individually or in any combination, in various categories and claim types:

[0253] Insert, in a signaling syntax, elements that allow the decoder to identify the motion vector predictor to be used;

[0254] A bit stream or signal that includes syntax that transmits information generated in accordance with any of the modes described;

[0255] Insert into the signaling syntax elements that allow the decoder to adapt the prediction of the motion vector in a manner corresponding to that used by an encoder; Petition 870250085959, dated 09 / 23 / 2025, pp. 96 / 123 51 / 51

[0256] Create and / or transmit and / or receive and / or decode a bit stream or signal that includes one or more of the described syntax elements, or variations thereof.

[0257] A method, process, apparatus, instruction storage medium, data storage medium or signal according to any of the embodiments described;

[0258] A TV, set-top box, mobile phone, tablet computer or other electronic device that performs motion vector prediction according to any of the modes described;

[0259] A TV, signal decoder, mobile phone, tablet computer or other electronic device that performs motion vector prediction according to any of the modes described and that displays (for example, using a monitor, screen or other type of display) a resulting image;

[0260] A TV, signal decoder, mobile phone, tablet computer or other electronic device that selects (for example, using a tuner) a channel to receive a signal including an encoded image and perform motion vector prediction according to any of the modes described.

[0261] A TV, signal decoder, cell phone, tablet computer or other electronic device that receives (for example, using an antenna) a signal over the air that includes an encoded image and performs motion vector prediction according to any of the modes described. Petition 870250085959, dated 09 / 23 / 2025, pp. 97 / 123

Claims

1 / 4 CLAIMS 1. A method CHARACTERIZED in that it comprises: obtaining an initial list of candidate motion vector predictors for motion information of a current block; obtaining bi-prediction candidates from the initial list, each bi-prediction candidate comprising two motion vectors, wherein a first motion vector of the two motion vectors indicates a first prediction block in a first reference figure and a second motion vector of the two motion vectors indicates a second prediction block in a second reference figure; calculating a score for the bi-prediction candidates from the initial list; and, obtaining a motion vector predictor for the motion information of the current block based on the calculated scores;where each score is a representative value of a difference between a first model and a second model, the first model corresponds to the lower right samples of the first prediction block and the second model corresponds to the lower right samples of the second prediction block that are spatially corresponding to the samples of the first model.

2. Method, according to claim 1, CHARACTERIZED in that it comprises reordering the bi-prediction candidates from the initial list based on the calculated scores to obtain a reordered list, wherein the resulting motion vector predictor is obtained from the reordered list.

3. Method, according to claim 1, CHARACTERIZED in that each model has N pixels of thickness, where N is a positive integer value. Petition 870250085959, dated 09 / 23 / 2025, pp. 120 / 123 2 / 4 4. Method, according to claim 1, CHARACTERIZED in that a motion compensation based on integers is applied to identify the first and second models, respectively, in the first and second reference figurations.

5. Method, according to claim 1, CHARACTERIZED in that a motion vector difference is added to at least one biprediction candidate from the initial list before calculating the score.

6. Method, according to claim 2, CHARACTERIZED in that, in response to the initial list comprising uni-prediction candidates, only the bi-prediction candidates are reordered and each uni-prediction candidate maintains the same index in the reordered list as in the initial list.

7. Method, according to claim 1, CHARACTERIZED in that, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates from the initial list are transformed into bi-prediction candidates before the calculation of scores.

8. Method, according to claim 7, CHARACTERIZED in that the bi-prediction candidates of the reordered list obtained from a uni-prediction candidate of the initial list are used for a prediction of the movement information of the current block.

9. Method, according to claim 2, CHARACTERIZED in that, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates are excluded from the reordered list.

10. Method CHARACTERIZED by the fact that it is for encoding an actual block comprising the method, as defined in claim 1.

11. Method CHARACTERIZED by the fact that it is for decoding a current block Petition 870250085959, dated 09 / 23 / 2025, page 121 / 123 3 / 4 comprising the method, as defined in claim 1.

12. Device CHARACTERIZED by the fact that it comprises electronic circuits configured to: obtain an initial list of candidate motion vector predictors for motion information of a current block; obtain bi-prediction candidates from the initial list, each bi-prediction candidate comprising two motion vectors, wherein a first motion vector of the two motion vectors indicates a first prediction block in a first reference figure and a second motion vector of the two motion vectors indicates a second prediction block in a second reference figure; calculate a score for the bi-prediction candidates from the initial list; and, obtain a motion vector predictor for the motion information of the current block based on the calculated scores;where each score is a representative value of a difference between a first model and a second model, the first model corresponds to the lower right samples of the first prediction block and the second model corresponds to the lower right samples of the second prediction block that are spatially corresponding to the samples of the first model.

13. Device according to claim 12, CHARACTERIZED in that the electronic circuit is further configured to reorder the bi-prediction candidates from the initial list based on the scores calculated to obtain a reordered list, wherein the motion vector predictor obtained is derived from the reordered list.

14. Device, according to claim 12, CHARACTERIZED by Petition 870250085959, dated 09 / 23 / 2025, pp. 122 / 123 4 / 4 in the fact that each model has N pixels of thickness, where N is a positive integer value.

15. Device according to claim 12, CHARACTERIZED in that the electronic circuit is further configured to apply a motion compensation based on integers to identify the first and second models, respectively, in the first and second reference figurations.

16. Device, according to claim 12, CHARACTERIZED in that a motion vector difference is added to at least one bi-prediction candidate from the initial list before calculating the score.

17. Device, according to claim 13, CHARACTERIZED in that, in response to the initial list comprising uni-prediction candidates, only the bi-prediction candidates are reordered and each uni-prediction candidate maintains the same index in the reordered list as in the initial list.

18. Device according to claim 12, CHARACTERIZED in that, in response to the initial list comprising uni-prediction candidates, the uni-prediction candidates from the initial list are transformed into bi-prediction candidates before the calculation of scores.

19. Device according to claim 18, CHARACTERIZED in that bi-prediction candidates from the reordered list obtained from a uni-prediction candidate from the initial list are used for a prediction of the movement information of the current block.

20. Device, according to claim 13, CHARACTERIZED in that, in response to the initial list comprising single-prediction candidates, the single-prediction candidates are excluded from the reordered list. Petition 870250085959, dated 09 / 23 / 2025, pp. 123 / 123

Images