US 12,294,734 B2
Spatial neighbor based affine motion derivation
Wei Chen, Beijing (CN); Xiaoyu Xiu, Beijing (CN); Yi-Wen Chen, Beijing (CN); Tsung-Chuan Ma, Beijing (CN); Hong-Jheng Jhu, Beijing (CN); Xianglin Wang, Beijing (CN); and Bing Yu, Beijing (CN)
Assigned to BEIJING DAJIA INTERNET INFORMATION TECHNOLOGY CO., LTD., Beijing (CN)
Filed by BEIJING DAJIA INTERNET INFORMATION TECHNOLOGY CO., LTD., Beijing (CN)
Filed on Sep. 20, 2022, as Appl. No. 17/933,552.
Application 17/933,552 is a continuation of application No. PCT/US2021/023115, filed on Mar. 19, 2021.
Claims priority of provisional application 62/992,704, filed on Mar. 20, 2020.
Prior Publication US 2023/0011286 A1, Jan. 12, 2023
Int. Cl. H04N 19/52 (2014.01); H04N 19/105 (2014.01); H04N 19/139 (2014.01); H04N 19/159 (2014.01); H04N 19/167 (2014.01); H04N 19/176 (2014.01); H04N 19/196 (2014.01); H04N 19/70 (2014.01)
CPC H04N 19/52 (2014.11) [H04N 19/105 (2014.11); H04N 19/139 (2014.11); H04N 19/159 (2014.11); H04N 19/167 (2014.11); H04N 19/176 (2014.11); H04N 19/197 (2014.11); H04N 19/70 (2014.11)] 18 Claims
OG exemplary drawing
 
1. A method of coding video data, comprising:
receiving, from a bitstream of the video data, a first syntax that indicates an affine motion model enabled for a current coding block;
estimating parameters of the affine motion model using gradients of motion vectors of multiple spatial neighboring blocks of the current coding block; and
constructing motion vectors of the affine motion model for the current coding block by using the estimated parameters,
wherein estimating the parameters comprises:
receiving, from the bitstream, a second syntax that indicates the affine motion model is one of a 6-parameter, 4-parameter, and 2-parameter affine motion model; and
in accordance with a determination from the second syntax that the affine motion model is a 6-parameter affine motion model:
calculating approximated values of the parameters a, b, d, e of the 6-parameter affine motion model respectively by:

OG Complex Work Unit Math
wherein (VxiT, VyjT) is motion vector of a selected top spatial neighboring block of the current coding block at a location (xi, yi), (VxiL, VyjL) is motion vector of a selected left spatial neighboring block of the current coding block at a location (xi, yi), NH is number of the selected top spatial neighboring blocks, NV is number of the selected left spatial neighboring blocks, gx is the gradient in a horizontal direction, gy is the gradient in a vertical direction, â, b, d and ê are the approximated values of the parameters a, b, d and e, respectively, a and b are delta zoom and rotation parameters for horizontal direction respectively, d and e are delta zoom and rotation parameters for vertical direction respectively.