US 12,034,921 B2
Apparatus and method for applying artificial neural network to image encoding or decoding
Tae Young Na, Seoul (KR); Sun Young Lee, Seoul (KR); Jae Seob Shin, Seoul (KR); Se Hoon Son, Seoul (KR); Hyo Song Kim, Seoul (KR); and Jeong Yeon Lim, Seoul (KR)
Assigned to SK TELECOM CO., LTD., Seoul (KR)
Filed by SK TELECOM CO., LTD., Seoul (KR)
Filed on Jan. 14, 2022, as Appl. No. 17/576,000.
Application 17/576,000 is a continuation of application No. 17/064,304, filed on Oct. 6, 2020, granted, now 11,265,540.
Application 17/064,304 is a continuation of application No. PCT/KR2019/002654, filed on Mar. 7, 2019.
Claims priority of application No. 10-2018-0021896 (KR), filed on Feb. 23, 2018; application No. 10-2018-0022254 (KR), filed on Feb. 23, 2018; application No. 10-2018-0040588 (KR), filed on Apr. 6, 2018; application No. 10-2018-0072499 (KR), filed on Jun. 25, 2018; application No. 10-2018-0072506 (KR), filed on Jun. 25, 2018; application No. 10-2018-0081123 (KR), filed on Jul. 12, 2018; and application No. 10-2018-0099166 (KR), filed on Aug. 24, 2018.
Prior Publication US 2022/0141462 A1, May 5, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 19/00 (2014.01); G06N 3/00 (2023.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01); G06N 5/04 (2023.01); H04N 19/117 (2014.01); H04N 19/124 (2014.01); H04N 19/176 (2014.01); H04N 19/82 (2014.01); H04N 19/86 (2014.01)
CPC H04N 19/117 (2014.11) [G06N 3/04 (2013.01); G06N 3/08 (2013.01); G06N 5/04 (2013.01); H04N 19/124 (2014.11); H04N 19/176 (2014.11); H04N 19/82 (2014.11); H04N 19/86 (2014.11)] 9 Claims
OG exemplary drawing
 
1. An intra-prediction based video decoding method, comprising:
decoding, from a bitstream, transform coefficients corresponding to a current block to be decoded in a current picture;
configuring an input data using at least one reference area that is included in the current picture and has been decoded prior to the current block, wherein the input data is configured using average pixel values which are generated by performing an averaging operation on values of pixels in the reference area;
filtering the input data using a filter coefficient set, which is in a two or three dimensional matrix form to thereby set the filtered input data as predicted pixels for the current block;
generating residual signals for the current block by inversely transforming the transform coefficients; and
reconstructing the current block by adding the predicted pixels and the residual signals corresponding thereto,
wherein the method further comprises decoding, from the bitstream, first information for selecting the filter coefficient set to apply to the current block among a plurality of filter coefficient sets each of which are in the two or three dimensional matrix form,
wherein the filtering is performed using the filter coefficient set determined based on the second information.