US 12,335,514 B2
Methods and apparatus of video coding using subblock-based temporal motion vector prediction
Yi-Wen Chen, San Diego, CA (US); and Xianglin Wang, San Diego, CA (US)
Assigned to BEIJING DAJIA INTERNET INFORMATION TECHNOLOGY CO., LTD., Beijing (CN)
Filed by BEIJING DAJIA INTERNET INFORMATION TECHNOLOGY CO., LTD., Beijing (CN)
Filed on Dec. 1, 2023, as Appl. No. 18/527,112.
Application 18/527,112 is a continuation of application No. 17/707,807, filed on Mar. 29, 2022, granted, now 11,936,901.
Application 17/707,807 is a continuation of application No. 17/187,587, filed on Feb. 26, 2021, granted, now 11,368,713, issued on Jun. 21, 2022.
Application 17/187,587 is a continuation of application No. PCT/US2019/048864, filed on Aug. 29, 2019.
Claims priority of provisional application 62/724,506, filed on Aug. 29, 2018.
Prior Publication US 2024/0098301 A1, Mar. 21, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 19/52 (2014.01); H04N 19/70 (2014.01)
CPC H04N 19/52 (2014.11) [H04N 19/70 (2014.11)] 15 Claims
OG exemplary drawing
 
1. A method of video decoding, the method comprising:
selecting, according to syntax elements signalled in a video bitstream, a reference picture of a plurality of reference pictures of a current picture as a collocated picture of the current picture, wherein the plurality of reference pictures includes a first list of reference pictures and a second list of reference pictures different from the first list of reference pictures, wherein the syntax elements include a first syntax element and a second syntax element, wherein the first syntax element indicates whether the collocated picture is selected from the first list of reference pictures or the second list of reference pictures, and wherein the second syntax element indicates which reference picture in a list of reference pictures indicated by the first syntax element is the collocated picture;
determining a temporal vector between the collocated picture and the current picture by:
checking reference pictures associated with a spatially neighbouring block of a current coding unit (CU) in the current picture according to a fixed order until one reference picture associated with the spatially neighbouring block is same as the collocated picture, and
choosing a motion vector of the spatially neighbouring block pointing to the reference picture as the temporal vector for the current CU;
splitting the current CU into a plurality of sub-CUs, each sub-CU corresponding to a respective subblock of the current picture;
obtaining a temporal motion vector predictor for each sub-CU of the current CU based on:
the temporal vector between the collocated picture and the current picture, and
motion information of a block in the collocated picture that corresponds to the respective subblock of the current picture; and
decoding the current CU according to temporal motion vector predictors of the plurality of sub-CUs of the current CU,
wherein the obtaining a temporal motion vector predictor for each sub-CU of the current CU further comprises:
identifying a block in the collocated picture corresponding to a subblock of a sub-CU at or close to a center of the current CU in the current picture based on the temporal vector between the collocated picture and the current picture;
determining motion information of the identified block in the collocated picture; and
deriving a motion vector from the determined motion information of the identified block as a default temporal motion vector predictor of a sub-CU of the current CU whose corresponding block in the collocated picture does not have motion information, wherein the corresponding block in the collocated picture is at a same relative location as the subblock of the sub-CU in the current picture based on the temporal vector between the collocated picture and the current picture.