US 12,483,723 B2
Method and apparatus for encoding or decoding video data with sub-pixel motion vector refinement
Guillaume Laroche, Saint Aubin d'Aubigne (FR); Patrice Onno, Rennes (FR); Jonathan Taquet, Talensac (FR); and Christophe Gisquet, Acigne (FR)
Assigned to Canon Kabushiki Kaisha, Tokyo (JP)
Filed by CANON KABUSHIKI KAISHA, Tokyo (JP)
Filed on Nov. 20, 2023, as Appl. No. 18/514,259.
Application 18/514,259 is a continuation of application No. 17/833,338, filed on Jun. 6, 2022, granted, now 11,849,142.
Application 17/833,338 is a continuation of application No. 16/627,996, granted, now 11,394,997, issued on Jul. 19, 2022, previously published as PCT/EP2018/067197, filed on Jun. 27, 2018.
Claims priority of application No. 1710747 (GB), filed on Jul. 4, 2017.
Prior Publication US 2024/0089498 A1, Mar. 14, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 19/523 (2014.01); H04N 19/176 (2014.01); H04N 19/56 (2014.01)
CPC H04N 19/523 (2014.11) [H04N 19/56 (2014.11); H04N 19/176 (2014.11)] 5 Claims
OG exemplary drawing
 
1. A method for decoding video data comprising frames, each frame being split into blocks, the method for decoding comprising:
determining first and second motion vectors for a block to be decoded in a frame to be decoded, the first motion vector specifying a first area in a first frame different from the frame to be decoded, and the second motion vector specifying a second area in a second frame different from the frame to be decoded;
determining, by using at least sample values of the first and second areas, whether to perform refinement for the first and second motion vectors in the method for decoding, in a case where a mode in which the refinement for the first and second motion vectors is capable of being performed is used in the method for decoding;
wherein in a case where it is determined that the refinement is performed, performing the refinement for the first and second motion vectors; and
wherein in a case where the refinement is performed, decoding the block to be decoded using an inter prediction using the refined first and second motion vectors, wherein inverse transformation is used for decoding the block of pixels;
wherein a plurality of patterns for refining is capable of being used in the mode in the method for decoding;
wherein the plurality of patterns includes a horizontal pattern and a vertical pattern,
wherein, when using the horizontal pattern, a plurality of horizontal sub-pixel positions on the right of a position to be refined and a plurality of horizontal sub-pixel positions on the left of the position to be refined are candidates of a horizontal position of the refined motion vector, and
wherein, when using the vertical pattern, a plurality of vertical sub-pixel positions above the position to be refined and a plurality of vertical sub-pixel positions below the position to be refined are candidates of a vertical position of the refined motion vector.