US 12,081,780 B2
Advanced temporal motion vector prediction
Yu Han, San Diego, CA (US); Kai Zhang, San Diego, CA (US); Li Zhang, San Diego, CA (US); Wei-Jung Chien, San Diego, CA (US); and Marta Karczewicz, San Diego, CA (US)
Assigned to QUALCOMM Incorporated, San Diego, CA (US)
Appl. No. 16/982,013
Filed by Qualcomm Incorporated, San Diego, CA (US)
PCT Filed Mar. 19, 2018, PCT No. PCT/CN2018/079458
§ 371(c)(1), (2) Date Sep. 17, 2020,
PCT Pub. No. WO2019/178721, PCT Pub. Date Sep. 26, 2019.
Prior Publication US 2021/0099724 A1, Apr. 1, 2021
Int. Cl. H04N 19/44 (2014.01); H04N 19/105 (2014.01); H04N 19/137 (2014.01); H04N 19/159 (2014.01); H04N 19/176 (2014.01)
CPC H04N 19/44 (2014.11) [H04N 19/105 (2014.11); H04N 19/137 (2014.11); H04N 19/159 (2014.11); H04N 19/176 (2014.11)] 30 Claims
OG exemplary drawing
 
1. A device for decoding a current block of video data in a current coding picture, the device comprising:
a memory configured to store video data; and
 a processor configured to: receive bits in a bitstream to indicate a mode is an advanced temporal motion vector prediction (ATMVP) mode;
determine which picture is a collocated reference picture in the ATMVP mode;
based on the determination there is a collocated reference picture, perform a decision check, in the ATMVP mode, to (a) determine if the collocated reference picture is the same as (b) a reference picture based on one or more previously decoded spatial neighboring blocks in the current coding picture to (c) determine an initial motion vector between an associated block in the collocated reference picture and the current block in the current coding picture, to determine a location of the associated block of the video data in the collocated reference picture that corresponds to the current block of the video data in the current coding picture based on the initial motion vector determined based on the decision check, where the associated block of the video data of the current bock includes more than one first derived motion vector, wherein each associated block in the collocated reference picture includes more than one motion vector, and wherein the associated block in the collocated reference picture includes M sub-blocks with the more than one first derived motion vector, where M>1;
determine at least one second derived motion vector associated with the current block in the current coding picture, when the initial motion vector points to the collocated reference picture, based on the more than one first derived motion vector associated with the associated block in the collocated reference picture;
determine motion information of the current block of the video data in the current coding picture based on the at least one second derived motion vector includes association of M sub-blocks from the collocated reference picture with M different motion vectors to determine M motion vectors of M sub-blocks of the current block in the ATMVP mode, based on the M sub-blocks from the collocated reference picture to obtain a prediction block; and
generate a reconstructed block based on a combination of the prediction block and a residual block.