Abstract: Some aspects of the disclosure provide a method of video decoding. For example, bitstream is received. The bitstream includes coded information of a sequence of pictures in a video. Training data for an online transform kernel derivation for a current block is collected based on one or more reconstructed blocks in the sequence of pictures. A transform kernel for the current block is derived based on the training data. The current block is reconstructed based on the transform kernel.

Abstract: A method for video encoding includes adjusting an orientation of a first picture from an original orientation to an adjusted orientation. The method includes encoding, in a bitstream, the adjusted first picture in the adjusted orientation. The method includes generating unified intermediate coding information of the first picture based on a unified orientation of a plurality of pictures in a picture buffer. The plurality of pictures in the picture buffer is oriented in the unified orientation. The unified intermediate coding information is stored in one or more intermediate coding information buffers associated with the first picture. The unified orientation is one of (i) the original orientation and (ii) the adjusted orientation. The method includes encoding, in the bitstream, a second picture based on the unified intermediate coding information of the first picture.

Abstract: A method for video decoding includes selecting a set of filter shapes from a plurality of sets of filter shapes associated with extrapolated filter-based intra prediction based on reconstructed information of a current block. The plurality of sets of filter shapes includes a first set of filter shapes and a second set of filter shapes. The method includes predicting, based on the selected set of filter shapes, a current sample in the current block using the extrapolated filter-based intra prediction. The first set of filter shapes only includes above-left positions that are above and to the left of a current position of the current sample. The second set of filter shapes includes one or more of (i) above-right positions that are above and to the right of the current position and (ii) below-left positions that are below and to the left of the current position.

Abstract: Some aspects of the disclosure provide a method of video decoding. For example, a bitstream is received, the bitstream includes coded information of a sequence of pictures in a video. Further, to apply a coding tool on a current block in a current picture in the sequence of pictures is determined, the coding tool is based on boundary gradient computation. Direction information of the current block is determined, and at least a boundary gradient associated with a sample of the current block is computed based on the direction information of the current block. A reconstructed block of the current block is generated by applying the coding tool that is based on at least the boundary gradient associated with the sample.

Abstract: A video bitstream is received. The video bitstream includes coded information of a current block in a current picture and of a plurality of reference pictures of the current block in a reference list. The coded information indicates that the current block is coded in an affine mode. A first control point motion vector (CPMV) of a first control point of the current block is determined. The first CPMV is associated with a sum of a plurality of intermediate vectors. The plurality of intermediate vectors includes an initial CPMV and an intermediate motion vector (MV). The initial CPMV is between the current picture and an initial reference picture of the plurality of reference pictures. The intermediate MV is between two respective reference pictures of the plurality of reference pictures. The current block is reconstructed based on the first CPMV of the first control point of the current block.

Abstract: The various implementations described herein include methods and systems for coding video. In one aspect, a method includes receiving a video bitstream comprising a plurality of blocks that includes a current block; identifying a block vector for the current block; determining a minimum-cost position in a predetermined area of a location indicated by the block vector; deriving a sub-pixel position based on the minimum-cost position; and reconstructing the current block using the sub-pixel position.

Abstract: An example method of video decoding includes receiving a video bitstream comprising a plurality of blocks, including a current block. The method includes obtaining a reference block for the current block. The method includes obtaining a set of reference template samples for the reference block. The method includes determining a direction for the current block based on the set of reference template samples. The method further includes decoding the current block using the direction.

Abstract: A non-transitory computer-readable storage medium storing instructions which when executed by a processor cause the processor to perform a method of encoding a bitstream. A first collocated block in a first collocated picture is determined based on a first motion shift candidate of a current block in a current picture. First motion information of a current template of the current block is determined based on one or more pieces of motion information of the first collocated block. The first motion information indicates one of a first reference template and a first subblock reference template in a first reference picture. The one of the first reference template and the first subblock reference template is determined based on the first motion information. A first template matching cost is determined based on the current template and the one of the first reference template and the first subblock reference template.

Abstract: Aspects of the disclosure provide a method and an apparatus for video encoding/decoding. The apparatus includes processing circuitry for: receiving prediction information of a current coding block in a current picture from a coded video bitstream, the prediction information indicating that the current coding block is coded using a subblock-based temporal motion vector prediction (SbTMVP) mode; deriving multiple displacement vector (DV) candidates by applying multiple DV offset candidates to a fixed DV predictor of the current coding block; comparing a template of the current coding block with each of multiple templates, each template of the multiple templates being located at a position specified by a corresponding one of the multiple DV candidates; calculating a cost value associated with each one of the multiple DV offset candidates based on the comparing; and reordering DV offset indices of the multiple DV offset candidates based on their calculated cost values.

Abstract: Example methods for decoding and encoding using classifiers for determining transformation information are described herein. An example method of video decoding includes receiving a video bitstream comprising a plurality of blocks that includes a current block. The method also includes applying a classifier to coded information for the current block to determine a transform kernel for the current block. The method further includes reconstructing the current block by applying the transform kernel to perform an inverse transform for the current block.

Abstract: A method includes receiving a bitstream indicating an adjustment factor to a linear formula used in a formula-based inter prediction of a current block in a current picture based on a reference block in a reference picture. The formula-based inter prediction generates a prediction sample of the current block based on the linear formula applied to one or more reconstructed samples in the reference block. The adjustment factor is applied on an initial slope parameter to determine an adjusted slope parameter. An adjusted offset parameter is calculated according to the adjusted slope parameter and a first weighted average of a plurality of first samples in a reference template of the reference block. An adjusted linear formula is generated based at least on the adjusted slope parameter and the adjusted offset parameter. At least a reconstructed sample of the current block is determined according to the adjusted linear formula.

Abstract: A method of video decoding includes receiving motion vector (MV) offset information of a current block in a current picture from a coded video bitstream. The current block includes a plurality of subblocks to be coded with a subblock-based temporal motion vector prediction (SbTMVP) mode. Motion information of a subblock in the plurality of subblocks is determined based on motion information of a corresponding subblock in a collocated block in a collocated reference picture. The collocated block is collocated with the current block and is indicated by a displacement vector (DV) of the current block. Updated motion information of the subblock in the plurality of subblocks is determined based on the motion information of the subblock in the plurality of subblocks and an MV offset of the current block indicated by the MV offset information. The subblock in the plurality of subblocks is reconstructed based on the updated motion information.

Abstract: In method of video processing in an encoder, a candidate list that comprises a plurality of motion vector (MV) candidates for predicting a current block is determined. The plurality of MV candidates in the candidate list is modified according to a motion information based diversity assessment of the plurality of MV candidates. The modifying the plurality of MV candidates prioritizes a subset of MV candidates with a larger difference in motion information. An index indicating a selected MV candidate from the modified candidate list is encoded in a video bitstream. The current block is encoded based on the selected MV candidate.

Abstract: Aspects of the disclosure provide methods, apparatuses, and non-transitory computer-readable storage mediums for video encoding/decoding. An apparatus includes processing circuitry that partitions a current block of a current picture based on a geometric partitioning mode (GPM). The current block is partitioned into two partitions in the GPM mode. Each of the partitions is associated with a respective predictor. A weighting index for a sample of the current block is determined based on a position of the sample. A weighting factor is calculated based on the weighting index of the sample according to an equation that converts the weighting index to the weighting factor. The sample is encoded based on the weighting factor and the predictor corresponding to the sample.

Abstract: An example method of video coding includes deriving, for a sub-block of a current block in the current picture, a first displacement vector based on a motion vector of a spatial neighboring block to the left of the current block in the current picture, the first displacement vector identifying a sub-block of a co-located reference picture. The method also includes, when the sub-block identified by the first displacement vector overlaps with two or more sub-blocks corresponding to a motion field grid of the co-located reference picture, retrieving two or more sub-block motion vectors used for the two or more sub-blocks. The method further includes deriving a final motion vector predictor for the sub-block of the current block based on a weighted average corresponding to the two or more sub-block motion vectors.

Abstract: Aspects of the disclosure includes methods and apparatuses for video decoding and video encoding and a method of processing visual media data. A current block in a first coding tree unit (CTU) is predicted using samples in a current picture. The first CTU and a second CTU are processed in parallel. The method for video decoding includes determining one of (i) a reference block of the current block in an allowed search range that is constrained for parallel processing of the first CTU and the second CTU in the current picture and (ii) an intra prediction mode of the current block based on a template adjacent to the current block that is constrained for parallel processing of the first CTU and the second CTU, and reconstructing the current block using the one of (i) the reference block and (ii) the intra prediction mode.

Abstract: Aspects of the disclosure include methods and apparatuses for video coding. One of the apparatuses includes processing circuitry that receives a current block in a bitstream. The current block is predicted with intra prediction fusion comprising multiple candidate intra prediction modes. The processing circuitry derives a dominant intra prediction mode from the multiple candidate intra prediction modes. The processing circuitry applies a transform from a transform set associated with the dominant intra prediction mode for the current block and decodes the current block according to the transform.

Abstract: An example method of video coding includes receiving a video bitstream for a plurality of blocks. The method also includes parsing, from the video bitstream, an indicator indicating that a current block of the plurality of blocks is to be decoded using an extrapolation filter intra prediction (EIP) mode. The method includes identifying transform information for the current block according to an EIP feature, the EIP feature including directionality information associated with the EIP mode and/or an index of the EIP mode; and the method includes generating a transform block for the current block using the transform information.

Abstract: Aspects of the disclosure includes methods and apparatuses for video decoding and video encoding and a method of processing visual media data. Coded information in a bitstream is received. The coded information indicates whether filtering is to be applied to at least one of a current template of a current block in a current picture and a reference template of a reference block. The current block is predicted based on the reference block. When the coded information indicates that the filtering is to be applied, a plurality of samples in the at least one of the current template and the reference template is filtered. A linear model between the current template and the reference template is determined based on the filtered plurality of samples. The current block is reconstructed based on the linear model and the reference block.

Abstract: A method of video decoding includes receiving a coded video bitstream comprising coded information of a current block in a current picture. The coded information is indicative of a bi-directional prediction with a refinement. The refinement determines an offset based on a plurality of calculations, and each calculation includes reference samples of a first reference picture and a second reference picture for the current picture. The method further includes determining that a first calculation in the plurality of calculations uses a first reference sample that is an out of boundary (OOB) prediction sample. An OOB prediction sample of a reference picture is out of a frame boundary of the reference picture. The method also includes performing the refinement that determines the offset based on the plurality of calculations with at least the first reference sample being excluded, and reconstructing the current block according to the bi-directional prediction with the offset.