US 12,294,730 B2
Pruning methods for neural network based video coding
Xiaozhong Xu, State College, PA (US); Wei Jiang, San Jose, CA (US); Shan Liu, San Jose, CA (US); and Wei Wang, Palo Alto, CA (US)
Assigned to TENCENT AMERICA LLC, Palo Alto, CA (US)
Filed by TENCENT AMERICA LLC, Palo Alto, CA (US)
Filed on Jun. 20, 2023, as Appl. No. 18/337,986.
Application 18/337,986 is a continuation of application No. 17/362,184, filed on Jun. 29, 2021, granted, now 11,765,376.
Claims priority of provisional application 63/089,481, filed on Oct. 8, 2020.
Prior Publication US 2023/0336762 A1, Oct. 19, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 7/12 (2006.01); G06N 3/082 (2023.01); H04N 19/172 (2014.01); H04N 19/174 (2014.01); H04N 19/176 (2014.01); H04N 19/463 (2014.01); H04N 19/80 (2014.01)
CPC H04N 19/463 (2014.11) [G06N 3/082 (2013.01); H04N 19/172 (2014.11); H04N 19/174 (2014.11); H04N 19/176 (2014.11); H04N 19/80 (2014.11)] 12 Claims
OG exemplary drawing
 
1. A pruning apparatus for neural network based video coding of a current block of a picture of a video sequence, the pruning apparatus comprising:
at least one memory configured to store computer program code; and
at least one processor configured to access the at least one memory and operate according to the computer program code, the computer program code comprising:
first setting code configured to cause the at least one processor to set a sequence level flag indicating whether pruning is to be performed in a coded bitstream;
second setting code configured to cause the at least one processor to, based on the sequence level flag indicating that pruning is to be performed in the coded bitstream, set, for each picture in the coded bitstream, a picture level flag that indicates that one or more neural network parameters associated with a respective picture are to be pruned;
first transmitting code configured to cause the at least one processor to transmit, to a decoder, a total number of groups of parameters; and
second transmitting code configured to cause the at least one processor to transmit, to the decoder, the sequence level flag and the respective picture level flag,
wherein, based on the respective picture level flag, the respective pictures are processed using non-pruned parameters.