US 12,225,201 B2
Image or video coding based on temporal motion information in units of subblocks
Jie Zhao, Seoul (KR); and Seunghwan Kim, Seoul (KR)
Assigned to LG ELECTRONICS INC., Seoul (KR)
Appl. No. 17/617,279
Filed by LG ELECTRONICS INC., Seoul (KR)
PCT Filed Jun. 11, 2020, PCT No. PCT/KR2020/007563
§ 371(c)(1), (2) Date Dec. 7, 2021,
PCT Pub. No. WO2020/251270, PCT Pub. Date Dec. 17, 2020.
Claims priority of provisional application 62/860,234, filed on Jun. 11, 2019.
Prior Publication US 2022/0232219 A1, Jul. 21, 2022
Int. Cl. H04N 19/137 (2014.01); H04N 19/105 (2014.01); H04N 19/132 (2014.01); H04N 19/176 (2014.01)
CPC H04N 19/137 (2014.11) [H04N 19/105 (2014.11); H04N 19/132 (2014.11); H04N 19/176 (2014.11)] 9 Claims
OG exemplary drawing
 
1. An image decoding method performed by a decoding apparatus, the method comprising:
deriving reference subblocks on a collocated reference picture for subblocks of a current block;
deriving motion information for the subblocks of the current block based on motion information for the reference subblocks;
generating prediction samples based on the motion information for the subblocks of the current block; and
generating reconstructed samples based on the prediction samples,
wherein for a reference subblock in which a motion vector is not available among the reference subblocks, base motion information is derived, and
wherein the base motion information is used as the motion information for a subblock of the current block corresponding to the reference subblock,
wherein the base motion information is derived based on one of the reference subblocks in a reference block located corresponding to the current block on the collocated reference picture,
wherein the base motion information is derived based on a first reference subblock having an initially available motion vector, by checking whether a motion vector is available according to a determined order with respect to the reference subblocks in the reference block, and
wherein, based on the base motion information being not derived, a subblock-based temporal motion vector prediction (sbTMVP) is specified as available.