Abstract: One or more template predictions are generated for a template of a current block based on one or more prediction modes. The template includes neighboring samples of the current block. Each of the one or more template predictions is generated based on a respective one of the one or more prediction modes. One or more filters are derived for the current block. Each of the one or more filters is derived based on (i) filter index information or (ii) a respective one of the one or more template predictions and a template reconstruction of the template. One or more predictions of the current block are determined. Each of the one or more predictions is determined based on a respective one of the one or more prediction modes. A final prediction of the current block is determined by applying the one or more filters to the one or more predictions.

Abstract: An example method of video coding includes receiving a video bitstream comprising a plurality of blocks and a plurality of quantized transform coefficients, and determining a quantized level for a quantized transform coefficient of the plurality of quantized transform coefficients. The method also includes selecting a quantizer from a set of three or more quantizers based on the quantized level, and deriving, using the selected quantizer, a dequantized transform coefficient for a first quantized transform coefficient of the plurality of quantized transform coefficients. The method further includes reconstructing a first block using the dequantized transform value.

Abstract: Aspects of the disclosure includes methods and apparatuses for video decoding and encoding and a method of processing visual media data. The method for video decoding includes receiving coded information in a bitstream indicating that a current block is predicted based on a combination of a plurality of intra prediction modes. The method includes determining a plurality of intra predictions of the current block based on the respective intra prediction modes, determining a fused prediction of the current block based on a weighted summation of the plurality of intra predictions where the weighted summation is according to respective weights associated with the plurality of intra predictions, and reconstructing the current block based on the fused prediction. Each weight is based on one of a plurality of weighting functions that depends on a sample location (x, y) and the intra prediction mode of the intra prediction associated with the respective weight.

Abstract: Some aspects of the disclosure provide a method of video decoding. In an example, a coded video bitstream is received. The coded video bitstream includes coded information of a current block in a current picture. When the current block is predicted in a learning-based prediction mode, a first transform kernel for the current block is determined, the first transform kernel is different from a second transform kernel that is applied when a non-learning based prediction is used on the current block. An inverse transform with the first transform kernel is applied on transform coefficients decoded from the coded video bitstream to obtain a residual block. The current block is reconstructed based on a combination of the residual block and a prediction block of the current block according to the learning-based prediction mode.

Abstract: Some aspects of the disclosure provide a method of video decoding. In an example, a coded video bitstream is received. The coded video bitstream includes coded information of a current block in a current picture. The current block is coded by a non-conventional intra prediction mode that is in a pre-defined set of non-conventional intra prediction modes. A virtual intra prediction mode is derived according to available information associated with the non-conventional intra prediction mode. A transform kernel is selected from at least two transform kernel sets according to the virtual intra prediction mode. The current block is reconstructed based on the transform kernel.

Abstract: A method for video decoding includes for each combination of (i) a pair of vectors in one or more pairs of vectors and (ii) a GPM split mode in a list of GPM split modes, determining a temporary reconstructed block of a current block based on the respective combination and determining a sum of gradients along one of a top boundary and a left boundary of the current block based on (i) boundary samples in the temporary reconstructed block along the one of the top boundary and the left boundary and (ii) respective reconstructed neighboring samples of the current block. A GPM split mode is determined from a plurality of GPM split modes including the list of GPM split modes based at least on the sums of gradients associated with the respective combinations. The current block is reconstructed according to the GPM and the GPM split mode.

Abstract: In a method of video decoding, a coded video bitstream that includes coded information of a current block in a current picture is received. In the method, an initial collocated block position in a collocated picture is determined based on a position of the current block in the current picture. A refined collocated block position of a refined collocated block in the collocated picture is determined based on a search and the initial collocated block position. A reference block in a current reference picture of the current block is determined based on motion information of the refined collocated block. The motion information indicates a collocated reference block in a collocated reference picture of the refined collocated block. In the method, the current block is reconstructed based on the reference block in the current reference picture of the current block.

Abstract: A video bitstream including coded information of a current block in a current picture of a video is received. The coded information indicates that a subblock-level cross-component model (CCM) is applied to the current block. A luma prediction block and a chroma prediction block for the current block are determined based on a prediction mode. The luma prediction block includes a plurality of luma prediction subblocks and the chroma prediction block includes a plurality of chroma prediction subblocks. The subblock-level CCM is determined based on a first luma prediction subblock of the plurality of luma prediction subblocks and a first chroma prediction subblock of the plurality of chroma prediction subblocks that corresponds to the first luma prediction subblock. A first chroma subblock of the current block is reconstructed based on the first chroma prediction subblock, the subblock-level CCM, and the first luma prediction subblock.

Abstract: A method for video decoding includes receiving a bitstream including coded information indicating that a current chroma block is filtered with cross-component filtering. One or more cross-component filtering models are determined based on first luma samples in a current luma block that is collocated with a current chroma block and based on first chroma samples of the current chroma block. A second luma sample is obtained by filtering one of the first luma samples with a luma in-loop filter. A cross-component filtering model of the one or more cross-component filtering models is then determined based on the second luma sample. The cross-component filtering model is applied to the second luma sample to obtain a cross-component filtered value. A second chroma sample is obtained based on the cross-component filtered value and one of the first chroma samples that is collocated with the one of the first luma samples.

Abstract: Coded information indicates that a bi-directional subblock-based spatial motion vector prediction (SbSMVP) is applied to a current block. A first subblock motion vector field is determined based on a plurality of top neighboring subblocks positioned at a top side of the current block. A second subblock motion vector field is determined based on a plurality of side neighboring subblocks positioned at a left side of the current block. According to a spatial direction, a motion vector for a first subblock of the current block is determined based on at least one of a first motion vector determined from the first subblock motion vector field and a second motion vector determined from the second subblock motion vector field. The first subblock of the current block is reconstructed based on the determined motion vector for the first subblock.

Abstract: In a method, a first affine-translational merge candidate is determined from a candidate list for a prediction of a current block in a current picture when the first affine-translational merge candidate for the prediction of the current block is allowed. The first affine-translational merge candidate provides affine motion information associated with a first reference picture in a first reference list and translational motion information associated with a second reference picture in a second reference list. A first prediction for a sample in the current block is generated according to the affine motion information associated with the first reference picture. A second prediction for the sample in the current block is generated according to the translational motion information associated with the second reference picture. The sample in the current block is encoded in a bitstream according to a combination of the first prediction and the second prediction.

Abstract: An example method of video coding includes receiving a video bitstream for a plurality of blocks. When an extrapolation filter-based intra prediction (EIP) mode is active for a current block of the plurality of blocks, the method also includes determining, when a boundary condition is met for the current block, one or more EIP mode parameters from a first set of EIP parameters. When the boundary condition is not met for the current block, the method also includes determining the one or more EIP mode parameters from a second set of EIP parameters, where the second set of EIP parameters includes one or more parameters not included in the first set of EIP parameters. The method further includes reconstructing the current block using the one or more EIP mode parameters.

Abstract: In a method of decoding performed in a decoder, a syntax element indicating a prediction mode associated with a current block in a current picture is received, the prediction mode being one of a uni-prediction mode and a bi-prediction mode. When the bi-prediction mode is applied to a first subblock of a plurality of subblocks of the current block and one of (i) a boundary of a first reference subblock is beyond a boundary of a first reference picture or (ii) a boundary of a second reference subblock is beyond a boundary of a second reference picture, the first subblock of the current block is reconstructed based on the uni-prediction mode such that the first subblock is reconstructed based on one of the first reference picture and the second reference picture.

Abstract: Aspects of the disclosure provide a method and an apparatus including processing circuitry that decodes a current block in a current picture with a subblock-based temporal motion vector prediction (SbTMVP) mode. A first collocated block in a first collocated picture is determined based on a first displacement vector candidate of the current block corresponding to a first SbTMVP candidate. The processing circuitry determines first motion information of a current template of the current block based on one or more pieces of motion information of the first collocated block or a neighboring block of the first collocated block. The processing circuitry determines one of a first reference template and a first subblock reference template in a first reference picture based on the first motion information and determines a first template matching cost based on the current template and the one of the first reference template and the first subblock reference template.

Abstract: An apparatus includes processing circuitry that is configured to determine whether at least one sample of a plurality of samples in a first template of a current block is in an intra coded block that is a spatial neighboring block of the current block, the first template including one or more samples of (1) top neighboring blocks and (2) left neighboring blocks of the current block. The processing circuitry is configured to, when (i) the at least one sample is included in the intra coded block and (ii) a template matching mode is applied to the current block, exclude the at least one sample from the first template to generate a second template for the current block. The processing circuitry is configured to determine a motion vector of the current block based on the second template. The processing circuitry is configured to reconstruct the current block based on the MV.

Abstract: A method includes receiving a bitstream that includes coded information of a sequence of pictures. The coded information indicates an inter prediction of a current block in a current picture, and an adjustment factor to a linear formula that is used in a formula-based inter prediction of the inter prediction. The formula-based inter prediction generates a prediction sample of the current block based on a linear formula with one or more reconstructed samples in a reference block being input to the linear formula, the linear formula includes one or more parameters derived based on a current template of the current block and a reference template of the reference block. The method also includes applying the adjustment factor on the linear formula to generate an adjusted linear formula, and determining at least a reconstructed sample of the current block according to the adjusted linear formula.

Abstract: Aspects of the disclosure provide a method and an apparatus including processing circuitry that determines, based on a syntax element in a coded video bitstream, that a current block including a plurality of subblocks is coded in a subblock-based temporal motion vector prediction (SbTMVP) mode. Motion vector offset (MVO) information indicating an MVO is received. The MVO indicates a motion offset of a displacement vector (DV) used to adjust a location of a collocated block in a collocated reference picture. An updated DV of the current block is determined based on the DV and the MVO. SbTMVP information of a respective subblock in the plurality of subblocks is derived based on motion information of a corresponding subblock in the collocated block indicated by the updated DV. The plurality of subblocks in the SbTMVP mode is reconstructed based on the SbTMVP information of the subblock in the plurality of subblocks.

Abstract: An apparatus includes processing circuitry that is configured to receive a bitstream including coded information of a current block. The processing circuitry is configured to partition the current block into a first partition and a second partition based on a geometric partition mode (GPM). The processing circuitry is configured to determine a first prediction mode for the first partition based on a first number of direct neighboring reconstructed samples of the current block being larger than a threshold value. The processing circuitry is configured to, when the first number of direct neighboring reconstructed samples of the current block is less than the threshold value, infer that the first prediction mode is inter prediction mode, a flag indicating whether the current block is intra coded or inter coded not being signaled in the bitstream, and reconstruct the first partition of the current block based on the first prediction mode.

Abstract: An example method of video decoding includes receiving a video bitstream that includes a current block. The method also includes obtaining a set of subblock-based motion vector prediction (SbTMVP) candidates for the current block and parsing a syntax element from the video bitstream to identify a merge motion vector difference (MMVD) for the current block. The method further includes generating a vector offset value for the current block based on the MMVD and generating a set of motion vectors (MVs) for the set of SbTMVP candidates using the vector offset value. When a MV in the set of MVs meets one or more criteria, including a corresponding SbTMVP candidate into a candidate list for the current block. When the MV in the set of MVs does not meet the one or more criteria, not including the corresponding SbTMVP candidate into the candidate list for the current block.

Abstract: A method includes receiving a bitstream that comprises coded information of a current block, the coded information of the current block indicates a state transition path of a state machine, the state transition path of the state machine includes at least a first state transition associated with a first quantization shifting offset of one or more first transform coefficients in transform coefficients of the current block. The method also includes determining the first quantization shifting offset associated with the one or more first transform coefficients according to the first state transition; reconstructing the one or more first transform coefficients based on the first quantization shifting offset; calculating residuals in a spatial domain of the current block based on at least the one or more first transform coefficients; and reconstructing the current block according to the residuals in the spatial domain.