The present invention discloses a method and apparatus for encoding or decoding a video signal. The method for processing a video signal according to the present invention uses a merging mode in which prediction information on a neighbor unit is used instead of transmitting prediction information on the present unit, so as to improve coding efficiency. In this case, the number of available candidate units for merging among the units in a predetermined position is determined, and information for the merging mode is acquired on the basis of the number of the available candidate units for merging. The unit to be merged is determined using the information for the merging mode, and prediction information on the unit to be merged is acquired. The prediction value for the present unit is acquired using the prediction information on the unit to be merged, and the present unit is restored using the acquired prediction value.

A method for processing a video includes performing a conversion between a current block of visual media data and a corresponding coded representation of the visual media data, wherein the conversion of the current block includes determining whether a use of one or both of a bi-directional optical flow (BIO) technique or a decoder-side motion vector refinement (DMVR) technique to the current block is enabled or disabled, and wherein the determining the use of the BIO technique or the DMVR technique is based on a cost criterion associated with the current block.

A device for decoding video data includes a memory configured to store video data; and one or more processors implemented in circuitry and configured to: determine a deterministic bounding box from which to retrieve reference samples of reference pictures of video data for performing decoder-side motion vector derivation (DMVD) for a current block of the video data; derive a motion vector for the current block according to DMVD using the reference samples within the deterministic bounding box; form a prediction block using the motion vector; and decode the current block using the prediction block.

Video decoder and/or video encoder, configured to determine a set of search area location candidates in a reference picture of a video; match the set of search area location candidates with a current template area adjacent to a current block of a current picture to obtain a best matching search area location candidate; select, out of a search area positioned in the reference picture at the best matching search area location candidate, a set of one or more predictor blocks by matching the current template area against the search area; and predictively decode/encode the current block from/into a data stream based on the set of one or more predictor blocks.

Video decoder and/or video encoder, configured to determine a set of search area location candidates in a reference picture of a video; match the set of search area location candidates with a current template area adjacent to a current block of a current picture to obtain a best matching search area location candidate; select, out of a search area positioned in the reference picture at the best matching search area location candidate, a set of one or more predictor blocks by matching the current template area against the search area; and predictively decode/encode the current block from/into a data stream based on the set of one or more predictor blocks.

Embodiments of video coding systems and methods are described for reducing coding latency introduced by decoder-side motion vector refinement (DMVR). In one example, two non-refined motion vectors are identified for coding of a first block of samples (e.g. a first coding unit) using bi-prediction. One or both of the non-refined motion vectors are used to predict motion information for a second block of samples (e.g. a second coding unit). The two non-refined motion vectors are refined using DMVR, and the refined motion vectors are used to generate a prediction signal of the first block of samples. Such embodiments allow the second block of samples to be coded substantially in parallel with the first block without waiting for completion of DMVR on the first block. In additional embodiments, optical-flow-based techniques are described for motion vector refinement.

Systems and techniques are provided for processing video data. For example, the systems and techniques can include obtaining a current picture of video data and obtaining reference pictures for the current picture from the video data. A merge mode candidate can be determined for the current picture. First and second motion vectors can be identified for the merge mode candidate. A motion vector search strategy can be selected for the merge mode candidate from a plurality of motion vector search strategies. The selected motion vector search strategy can be associated with one or more constraints corresponding to at least one of the first motion vector or the second motion vector. The selected motion vector search strategy can be used to determine refined motion vectors based on the first motion vector, the second motion vector, and the reference pictures. The merge mode candidate can be processed using the refined motion vectors.

Motion information for an internal sub-block of a larger block can be derived for use in encoding or decoding the video block or a coding unit by using the motion information for sub-blocks on the left or top edge of the coding block. The left column of edge sub-blocks and the top row of sub-blocks has motion information, such as motion vectors, derived using such techniques as template matching. The motion vectors of these edge sub-blocks are used in deriving the motion vectors of internal sub-blocks, which leads to better prediction and improved coding efficiency. In another embodiment, other motion information for internal sub-blocks is derived from corresponding information of the edge sub-blocks.

A picture prediction method is provided, which includes: obtaining initial motion information of a current picture block; determining, based on a matching cost criterion, that positions of a pair of reference blocks are a position of a target forward reference block of the current picture block and a position of a target backward reference block of the current picture block, where positions of each pair of reference blocks include a position of a forward reference block and a position of a backward reference block; and for the positions of each pair of reference blocks, a first position offset and a second position offset are in a mirror relationship; and obtaining a predicted value of a pixel value of the current picture block based on a pixel value of the target forward reference block and a pixel value of the target backward reference block.

A video encoding method of estimating a motion vector of a current block in a current frame includes identifying a reference block in a reference frame, a first location of the current block within the current frame corresponding to a second location of the reference block within the reference frame, setting a search range in the reference frame, searching for candidate blocks from the search range of the reference frame, each of the candidate blocks satisfying a corresponding first motion estimation criterion of first motion estimation criteria, searching for a prediction block from among the candidate blocks, the prediction block satisfying a second motion estimation criterion, estimating, as the motion vector of the current block, a motion vector corresponding to a displacement between the reference block and the prediction block, and encoding the current block based on the estimated motion vector.