Tencent America LLC

Lien-Fei Chen

Shan Liu

Guichun Li

Xiang Li

Xin Zhao

Xiang Yang Li

Ling Li

Xiaozhong Xu

Han Gao

Biao Wang

In a method, a first area and a second area of a current frame is determined when the current frame is a GDR frame. The first area is independently coded and the second area is coded with dependency. When a current block in the current frame is coded by an intraTMP mode or an IBC mode, a search range of the intraTMP mode or the IBC mode is determined such that the search range is only in the first area of the current frame. A reference block is determined from a plurality of candidate reference blocks in the search range. The current block is encoded in a bitstream based on the determined reference block.

Intra template matching prediction in GDR pictures

In some examples, an apparatus for video decoding includes receiving circuitry and processing circuitry. The processing circuitry receives coded information of a current block in a current picture from a coded video bitstream. The coded information includes a syntax element with a value indicative of a merge with motion vector difference (MMVD) mode being applied on the current block. The processing circuitry determines, in response to a motion vector predictor of a merge candidate being a bi-predictor, whether the merge candidate for the current block satisfies a requirement for applying a bi-prediction motion refinement. The processing circuitry can constrain the merge candidate for use in the MMVD mode when the merge candidate for the current block satisfies the requirement. Then, the processing circuitry reconstructs the current block in the MMVD mode with the merge candidate being constrained for use in the MMVD mode.

Merge with motion vector difference

A reference block is determined from a plurality of candidate reference blocks for a current block in a current picture based on template matching (TM) costs of the plurality of candidate reference blocks. The TM costs indicate differences between a template of the current block and respective reference templates of the plurality of candidate reference blocks. Samples of the determined reference block are classified into a plurality of classes of samples. A partitioning pattern of the current block is derived based on the determined reference block from a predetermined plurality of partitioning patterns. The derived partitioning pattern indicates a plurality of partitions of the current block. Each of the plurality of classes of the samples of the determined reference block corresponds to a respective partition of the plurality of partitions of the current block. The current block is reconstructed based on the derived partitioning pattern of the current block.

Template matching based partitioning pattern derivation

Aspects of the disclosure provide a method and an apparatus including processing circuitry that obtaining prediction information indicating whether a current block is coded in a subblock-based temporal motion vector prediction (SbTMVP) mode. If the current block is coded in the SbTMVP mode, whether a subblock in a plurality of subblocks of the current block is bi-predicted is determined. If the subblock is bi-predicted, motion information of the subblock is determined based on the SbTMVP mode. At least one of (i) a bilateral matching (BM)-based motion vector (MV) refinement and (ii) a bi-directional optical flow (BDOF) mode is applied to the subblock to refine the motion information of the subblock. The current block is reconstructed based on refined motion information corresponding to one or more subblocks in the plurality of subblocks. The refined motion information corresponding to the one or more subblocks includes the refined motion information of the subblock.

Decoder-side motion vector refinement and bi-directional optical flow in subblock-based temporal motion vector prediction (SBTMVP)

Processing circuitry determines an initial affine model with bi-prediction for a current block in an affine motion inter prediction mode, the initial affine model predicts the current block based on a first reference picture and a second reference picture. The processing circuitry performs a multi-stage bilateral matching motion refinement with at least a first stage and a second stage on the initial affine model to derive a refined affine model, one of the first stage and the second stage applies a first bilateral matching motion refinement on a translational component of the initial affine model, and the other of the first stage and the second stage applies a second bilateral matching motion refinement on a non-translational component of the initial affine model. The processing circuitry reconstructs the current block using the refined affine model based on the first reference picture and the second reference picture.

Method and apparatus for affine motion refinement

Aspects of the disclosure provide a method and an apparatus including processing circuitry that performs template matching (TM) on motion vector resolution (MVR) pairs of a current block by determining a TM cost corresponding to each one of the MVR pairs. The current block is predicted with an affine adaptive motion vector prediction (ΔMVP) mode. Each MVR pair includes a first MVR in a first plurality of MVRs corresponding to a first reference picture and a second MVR in a second plurality of MVRs corresponding to a second reference picture. Each TM cost is determined based on a current template of the current block and a bi-predictor of the respective MVR pair. The bi-predictor is based on a first reference template in the first reference picture and a second reference template in the second reference picture. An MVR pair is selected based on the determined corresponding TM costs.

Template-matching based adaptive motion vector resolution (AMVR) for bi-prediction and an affine mode

Coded information indicates that an adaptive up-sampling filter is applied to a current block in a current picture. A respective class for each of a plurality of subblocks of the current block is determined. A respective filter coefficient set is determined for each of the plurality of subblocks from a plurality of filter coefficient sets of the adaptive up-sampling filter. The respective filter coefficient set is determined based on at least one class corresponding to the respective subblock and a respective sampling rate of reference pixel resampling (RPR) applied on a reference picture of the current picture. The respective sampling rate is associated with one of a plurality of phases of the RPR. The adaptive up-sampling filter is applied to the current block to generate filtered reconstructed samples of the current block based on the determined respective filter coefficient sets without applying a secondary adaptive loop filter (ALF).

Adaptive up-sampling filter for luma and chroma with reference picture resampling (RPR)

In a method, coded information of a current block and neighboring blocks of the current block in a current picture is received from a coded video bitstream. A list of merge candidates of the current block is generated based on the neighboring blocks of the current block. The list of merge candidates of the current block is divided into a plurality of subgroups. Each of the plurality of subgroups includes one or more merge candidates. The one or more merge candidates are ordered within each subgroup by a respective template matching (TM) cost associated with each of the one or more merge candidates. The current block is reconstructed based on a merge candidate selected from the list of merge candidates of the current block.

Grouping based adaptive reordering of merge candidate

In the method, coded information of a coding unit (CU) in a picture of a video is received from a coded video bitstream. The CU is partitioned into a first partition and a second partition based on a geometric partition mode (GPM). The first partition and the second partition are rectangular partitions that are adjacent to each other and divided by a straight partition line. A first intra prediction mode for the first partition of the CU and a second intra prediction mode for the second partition of the CU are determined. The first intra prediction mode is different from the second intra prediction mode. The first partition of the CU is reconstructed based on the first intra prediction mode and the second partition of the CU is reconstructed based on the second intra prediction mode.

Geometric partition based intra prediction

Coded information of a first color component and a second color component associated with a current block is determined. A first prediction mode is determined to be selected from an intra prediction mode, an inter prediction mode, or a combined intra-inter prediction mode for the first color component associated with the current block. A second prediction mode is determined to be selected from the intra prediction mode, the inter prediction mode, or the combined intra-inter prediction mode for the second color component associated with the current block. The second prediction mode is different from the first prediction mode. The current block is reconstructed based on predicting the first color component of the current block in the first prediction mode and predicting the second color component of the current block in the second prediction mode.

Component dependent intra and inter prediction signaling

Aspects of the disclosure provide a method and an apparatus for video decoding. The apparatus includes processing circuitry that receives a coded bitstream comprising a first block in a current picture. The processing circuitry obtains prediction information indicating whether the first block is coded in an intra template matching prediction (IntraTMP) mode. If the IntraTMP mode is applied to the first block, the first block is reconstructed based on a prediction block within a reconstructed search area in the current picture. A reference template of the prediction block is matched to a current template of the first block in the IntraTMP mode. An IntraTMP-based block vector BVof the first block is stored. The IntraTMP-based block vector indicates a position displacement between the current template of the first block and the reference template of the prediction block. A second block is reconstructed based on the stored IntraTMP-based block vector.

Intra block copy (IBC) candidate list construction with motion information of intra template-matching prediction

Aspects of the disclosure provide a method and an apparatus including processing circuitry that decodes a value of a syntax element that indicates whether a current block is coded with a geometric partition mode (GPM) along a partition edge intersecting the current block. If the value of the syntax element indicates that the current block is coded with the GPM and a condition for adaptive blending is satisfied, width candidates in a width candidate list are reordered using template matching (TM) that is based on a current template of the current block and reference templates that correspond to the respective width candidates. A width candidate is selected from the reordered width candidate list. A width of a blending area defined by boundaries on both sides of the partition edge is determined based on the selected width candidate. The blending area is determined based on the width of the blending area.

Adaptive blending for geometric partition mode (GPM)

Aspects of the disclose provide methods and an apparatus for video and/or image coding. The apparatus includes processing circuitry that receives a bitstream including a plurality of pictures. The processing circuitry determines an initial motion vector (MV) associated with a motion compensation padding (MCP) block according to motion information of a boundary block located at a picture boundary of a picture in the plurality of pictures. The boundary block is located in the picture. The MCP block is located in an MCP area outside of the picture and adjacent to the picture boundary. The initial MV has a fractional precision. The processing circuitry generates a final MV associated with the MCP block based on the initial MV. The final MV has an integer precision. The processing circuitry reconstructs, according to the generated final MV having the integer precision, at least a sample in the MCP block.

Full-pixel based and group-based motion compensated picture boundary padding

In a method, coded information of a current block in a current picture of a video is received from a coded video bitstream. The coded information indicates that the current block is partitioned into at least a first partition and a second partition based on a geometric partition mode (GPM). The first partition is associated with an intra block copy (IBC) prediction mode, and the second partition is associated with an intra prediction mode. Prediction mode information for a first subblock of a plurality of subblocks of the current block is stored. The prediction mode information includes one of a DC mode, a Planar mode, and an intra prediction mode based on whether the first subblock is associated with one of the IBC prediction mode, the intra prediction mode, and both the IBC prediction mode and the intra prediction mode.

Storage of intra mode information in geometric partition mode

Aspects of the disclosure provide a method and an apparatus for video decoding. The apparatus includes processing circuitry receiving prediction information of a plurality of blocks in a current picture from a coded video bitstream. The processing circuitry determines, based on the prediction information, whether at least one of a decoder-side motion vector refinement (DMVR) mode or a bi-directional optical flow (BDOF) mode is allowed for the plurality of blocks. If the DMVR mode or the BDOF mode is allowed for the plurality of blocks, the processing circuitry determines that a plurality of merge with motion vector difference (MMVD) merge flags indicating whether a MMVD mode is applied to the plurality of blocks, respectively, is inferred as false for the plurality of blocks. If the plurality of MMVD merge flags is inferred as false, the processing circuitry reconstructs each block in the plurality of blocks without applying the MMVD mode.

Growing community of inventors

Taipei, Taiwan

Lien-Fei Chen