A method of video decoding performed in a video decoder includes receiving a coded video bitstream including a current block that is divided into a plurality of sub-blocks. The method includes performing sub-block based affine motion compensation on the current block to generate a sub-block prediction for each pixel in each sub-block of the current block. The method further includes determining one or more spatial gradients for each sub-block prediction. The method includes performing, for each sub-block prediction, prediction refinement with an optical flow process using the respective determined one or more spatial gradients and at least one constraint included in the coded video bitstream. The method further includes adding, for each sub-block prediction, an output of the respective prediction refinement to the respective sub-block prediction to generate a final prediction for each pixel in each sub-block of the current block.

The present disclosure provides systems and methods for constructing a merge candidate list used for video processing. One exemplary method includes: inserting a set of spatial merge candidates to a merge candidate list of a coding block, wherein the set of spatial merge candidates are inserted according to an order of: top neighboring block, left neighboring block, top neighboring block, left neighboring block and above-left neighboring block. The method can further include adding to the merge candidate list at least one of: a temporal merge candidate from collocated coding units, a history-based motion vector predictor (HMVP) from a First-In, First-Out (FIFO) table, a pairwise average candidate, or a zero motion vector.

There is provided a method for estimating weighted prediction parameters intended to be used for predicting an image block.

There is provided a method for estimating weighted prediction parameters intended to be used for predicting an image block.

A video decoding apparatus predicts a target block in a current picture to be decoded. The apparatus comprises a predictor configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; and generate a prediction block of the target block using the first and second reference blocks. The predictor includes a first coding tool configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.

A video decoding apparatus predicts a target block in a current picture to be decoded. The apparatus comprises a predictor configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; and generate a prediction block of the target block using the first and second reference blocks. The predictor includes a first coding tool configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.

A method, apparatus, and non-transitory computer-readable medium for adaptive neural image compression by meta-learning using substitute QF settings, which includes generating one or more substitute quality factors via a plurality of iterations using the original quality factors, wherein the substitute quality factors are a modified version of the original quality factors and are associated with a single instance of neural network loop filtering model. The approach may further include determining a neural network based loop filter comprising neural network based loop filter parameters and a plurality of layers, wherein the neural network based loop filter parameters include shared parameters and adaptive parameters, and may further include generating enhanced video data, based on the one or more substitute quality factors and the input video data, using the neural network based loop filter.

Aspects of the disclosure provide methods and an apparatus for video coding. The apparatus includes processing circuitry that decodes coding information of a current block (CB) from a coded video bitstream. The coding information indicates that the CB is coded with a subblock-based affine motion model including affine parameters that are based on multiple control point motion vectors (MVs) for the CB. The processing circuitry determines, based on the coding information, whether to select a subblock characteristic for generating a prediction for a sample in an affine subblock of the CB based on a corresponding subblock MV. In response to selecting the subblock characteristic, the processing circuitry determines the subblock characteristic based on at least one of the affine parameters. The subblock characteristic indicates one of: (i) a subblock size used for generating the prediction for the sample and (ii) an interpolation filter type for the affine subblock.

A method for identifying stationary regions in frames of a video sequence comprises receiving an encoded version of the video sequence, wherein the encoded version of the video sequence includes an intra-coded frame followed by a plurality of inter-coded frames; reading coding-mode information in the inter-coded frames of the encoded version of the video sequence, wherein the coding-mode information is indicative of blocks of pixels in the inter-coded frames being skip-coded; finding, using the read coding-mode information, one or more blocks of pixels that each was skip-coded in a respective plurality of consecutive frames in the encoded version of the video sequence; and designating each found block of pixels as a stationary region in the respective plurality of consecutive frames.

A method for identifying stationary regions in frames of a video sequence comprises receiving an encoded version of the video sequence, wherein the encoded version of the video sequence includes an intra-coded frame followed by a plurality of inter-coded frames; reading coding-mode information in the inter-coded frames of the encoded version of the video sequence, wherein the coding-mode information is indicative of blocks of pixels in the inter-coded frames being skip-coded; finding, using the read coding-mode information, one or more blocks of pixels that each was skip-coded in a respective plurality of consecutive frames in the encoded version of the video sequence; and designating each found block of pixels as a stationary region in the respective plurality of consecutive frames.