# US 11,451,820 B2 · DMVR-based inter-prediction method and device > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,451,820 B2 - Document ID: US11451820B2 - Status: Granted - Kind: B2 - Application number: 17/470,899 - Filed: 2021-09-09 - Published: 2022-03-03 - Granted: 2022-09-20 - Earliest claimed: 2019-03-13 - Assignees: LG Electronics Inc. - Applicants: LG Electronics Inc. - Inventors: Naeri Park; Junghak Nam; Hyeongmoon Jang - Web page: https://patentgrep.com/patent/US11451820B2 ## CPC - H04N 19/52: by predictive encoding - H04N 19/103: Selection of coding mode or of prediction mode - H04N 19/105: Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction - H04N 19/132: Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking - H04N 19/176: the region being a block, e.g. a macroblock - H04N 19/186: the unit being a colour or a chrominance component - H04N 19/577: Motion compensation with bidirectional frame interpolation, i.e. using B-pictures - H04N 19/70: characterised by syntax aspects related to video coding, e.g. related to compression standards - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION ## Abstract An image decoding method includes: acquiring, from a bitstream, luma weight L0 flag information indicating whether there is an L0 prediction-related weight factor and luma weight L1 flag information indicating whether there is an L1 prediction-related weight factor; determining to apply decoder-side motion vector refinement (DMVR) to an L0 motion vector and L1 motion vector for a current block, when the luma weight L0 flag information and the luma weight L1 flag information are both zero; when it has been determined to apply DMVR, deriving a refined L0 motion vector and a refined L1 motion vector by applying the DMVR to the current block; deriving prediction samples for the current block on the basis of L0 prediction using the refined L0 motion vector and L1 prediction using the refined L1 motion vector; and generating reconstruction samples for the current block on the basis of the prediction samples. ## Claims ### Claim 1 An image decoding method performed by a decoding apparatus, the image decoding method comprising: obtaining luma weight L0 flag information related to indicating whether a weight factor for L0 prediction is present and luma weight L1 flag information related to indicating whether a weight factor for L1 prediction is present from a bitstream; determining to apply decoder-side motion vector refinement (DMVR) to an L0 motion vector and an L1 motion vector of a current block, based on a case where a value of the luma weight L0 flag information and a value of the luma weight L1 flag information are both equal to 0; deriving a refined L0 motion vector and a refined L1 motion vector by applying the DMVR to the current block; deriving prediction samples for the current block based on the L0 prediction using the refined L0 motion vector and the L1 prediction using the refined L1 motion vector; and generating reconstructed samples for the current block based on the prediction samples, wherein in the deriving of the prediction samples for the current block, based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0, the prediction samples are derived by applying bi-directional optical flow (BDOF) to the current block. ### Claim 2 (depends on 1) The image decoding method of claim 1, wherein comprising deriving DMVR flag information related to indicating whether to apply the DMVR, wherein the DMVR flag information related to applying the DMVR is derived as 1 or true, wherein the DMVR flag information related to not applying the DMVR is derived as 0 or false, and wherein a value of the DMVR flag information is derived as 1 or true based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0. ### Claim 3 (depends on 2) The image decoding method of claim 2, wherein comprising obtaining bi-prediction weight index information of the current block, and wherein the value of the DMVR flag information is derived as 1 or true based on a value of the bi-prediction weight index information being equal to 0. ### Claim 4 (depends on 3) The image decoding method of claim 3, wherein a ½ weight is applied to each of the L0 prediction and the L1 prediction based on the value of the bi-prediction weight index information being equal to 0. ### Claim 5 (depends on 1) The image decoding method of claim 1, wherein comprising deriving BDOF flag information related to indicating whether to apply the BDOF, wherein the BDOF flag information related to applying the BDOF is derived as 1 or true, wherein the BDOF flag information related to not applying the BDOF is derived as 0 or false, and wherein a value of the BDOF flag information is derived as 1 or true based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0. ### Claim 6 (depends on 5) The image decoding method of claim 5, wherein comprising obtaining bi-prediction weight index information of the current block, and wherein the value of the BDOF flag information is derived as 1 or true based on a value of the bi-prediction weight index information being equal to 0. ### Claim 7 (depends on 1) The image decoding method of claim 1, wherein the L0 motion vector and the L1 motion vector are derived based on a motion vector of a neighboring block indicated by merge index in a merge candidate list derived based on neighboring blocks of the current block, based on a case where a merge mode is applied to the current block, and wherein the refined L0 motion vector and the refined L1 motion vector are derived based on a sample position corresponding to a minimum sum of absolute differences (SAD) by calculating the minimum SAD based on reference samples in an L0 reference picture derived based on the L0 motion vector and reference samples in an L1 reference picture derived based on the L1 motion vector. ### Claim 8 (depends on 1) The image decoding method of claim 1, wherein in the deriving of the prediction samples for the current block, based on a case where the BDOF is applied to the current block, L0 prediction samples are derived based on the L0 prediction using the refined L0 motion vector and L1 prediction samples are derived based on the L1 prediction using the refined L1 motion vector, and wherein the prediction samples are derived based on a gradient for the L0 prediction samples and a gradient for the L1 prediction samples. ### Claim 9 An image encoding method performed by an encoding apparatus, the image encoding method comprising: determining luma weight L0 flag information related to indicating whether a weight factor for L0 prediction is present and luma weight L1 flag information related to indicating whether a weight factor for L1 prediction is present; determining to apply decoder-side motion vector refinement (DMVR) to an L0 motion vector and an L1 motion vector of a current block, based on a case where a value of the luma weight L0 flag information and a value of the luma weight L1 flag information are both equal to 0; deriving a refined L0 motion vector and a refined L1 motion vector by applying the DMVR to the current block; deriving prediction samples for the current block based on the L0 prediction using the refined L0 motion vector and the L1 prediction using the refined L1 motion vector; deriving residual samples based on the prediction samples; and encoding image information including information on the residual samples, wherein in the deriving of the prediction samples for the current block, based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0, the prediction samples are derived by applying bi-directional optical flow (BDOF) to the current block. ### Claim 10 (depends on 9) The image encoding method of claim 9, wherein comprising deriving DMVR flag information related to indicating whether to apply the DMVR, wherein the DMVR flag information related to applying the DMVR is derived as 1 or true, wherein the DMVR flag information related to not applying the DMVR is derived as 0 or false, and wherein a value of the DMVR flag information is derived as 1 or true based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0. ### Claim 11 (depends on 10) The image encoding method of claim 10, wherein comprising determining bi-prediction weight index information of the current block, and wherein the value of the DMVR flag information is derived as 1 or true based on a value of the bi-prediction weight index information being equal to 0. ### Claim 12 (depends on 11) The image encoding method of claim 11, wherein a ½ weight is applied to each of the L0 prediction and the L1 prediction based on the value of the bi-prediction weight index information being equal to 0. ### Claim 13 (depends on 9) The image encoding method of claim 9, wherein comprising deriving BDOF flag information related to indicating whether to apply the BDOF, wherein the BDOF flag information related to applying the BDOF is derived as 1 or true, wherein the BDOF flag information related to not applying the BDOF is derived as 0 or false, and wherein a value of the BDOF flag information is derived as 1 or true based on the case the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0. ### Claim 14 (depends on 13) The image encoding method of claim 13, wherein comprising determining bi-prediction weight index information of the current block, and wherein the value of the BDOF flag information is derived as 1 or true based on a value of the bi-prediction weight index information being equal to 0. ### Claim 15 A non-transitory computer-readable digital storage medium storing a bitstream generated by a method, the method comprising: determining luma weight L0 flag information related to indicating whether a weight factor for L0 prediction is present and luma weight L1 flag information related to indicating whether a weight factor for L1 prediction is present; determining to apply decoder-side motion vector refinement (DMVR) to an L0 motion vector and an L1 motion vector of a current block, based on a case where a value of the luma weight L0 flag information and a value of the luma weight L1 flag information are both equal to 0; deriving a refined L0 motion vector and a refined L1 motion vector by applying the DMVR to the current block; deriving prediction samples for the current block based on the L0 prediction using the refined L0 motion vector and the L1 prediction using the refined L1 motion vector; deriving residual samples based on the prediction samples; and encoding image information including information on the residual samples to generate the bitstream, wherein in the deriving of the prediction samples for the current block, based on the case where the value of the luma weight L0 flag information and the value of the luma weight L1 flag information are both equal to 0, the prediction samples are derived by applying bi-directional optical flow (BDOF) to the current block. ### Claim 16 (depends on 1) The image decoding method of claim 1, wherein the determining to apply the DMVR to the L0 motion vector and the L1 motion vector of the current block is further based on a case where a first picture order count (POC) difference between a current picture and an L0 reference picture and a second POC difference between the current picture and an L1 reference picture are equal to each other. ### Claim 17 (depends on 9) The image encoding method of claim 9, wherein the determining to apply the DMVR to the L0 motion vector and the L1 motion vector of the current block is further based on a case where a first picture order count (POC) difference between a current picture and an L0 reference picture and a second POC difference between the current picture and an L1 reference picture are equal to each other. ### Claim 18 (depends on 15) The non-transitory computer-readable digital storage medium of claim 15, wherein the determining to apply the DMVR to the L0 motion vector and the L1 motion vector of the current block is further based on a case where a first picture order count (POC) difference between a current picture and an L0 reference picture and a second POC difference between the current picture and an L1 reference picture are equal to each other.