US 12,206,894 B2
Apparatus, method and storage medium for image encoding/decoding to perform intra prediction using artificial neutral network
Dong-Hyun Kim, Daejeon (KR); Ji-Hoon Do, Daejeon (KR); Youn-Hee Kim, Daejeon (KR); Se-Yoon Jeong, Daejeon (KR); Hyoung-Jin Kwon, Daejeon (KR); Jong-Ho Kim, Daejeon (KR); Joo-Young Lee, Daejeon (KR); Jin-Soo Choi, Daejeon (KR); and Tae-Jin Lee, Daejeon (KR)
Assigned to ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE, Daejeon (KR)
Filed by Electronics and Telecommunications Research Institute, Daejeon (KR)
Filed on Sep. 20, 2021, as Appl. No. 17/479,811.
Claims priority of application No. 10-2020-0123292 (KR), filed on Sep. 23, 2020; and application No. 10-2021-0124655 (KR), filed on Sep. 17, 2021.
Prior Publication US 2022/0094977 A1, Mar. 24, 2022
Int. Cl. H04N 19/593 (2014.01); G06N 3/04 (2023.01); H04N 19/11 (2014.01); H04N 19/136 (2014.01); H04N 19/176 (2014.01); H04N 19/184 (2014.01); H04N 19/436 (2014.01)
CPC H04N 19/593 (2014.11) [G06N 3/04 (2013.01); H04N 19/11 (2014.11); H04N 19/136 (2014.11); H04N 19/176 (2014.11); H04N 19/184 (2014.11); H04N 19/436 (2014.11)] 16 Claims
OG exemplary drawing
 
1. A decoding method, comprising:
deriving an intra-prediction mode for a target block; and
performing intra-prediction for the target block that uses the derived intra-prediction mode,
wherein a Most Probable Mode (MPM) list comprising a plurality of candidates of the MPM list is generated,
wherein at least one of the plurality of the candidates in the MPM list is configured using an artificial neural network,
wherein the artificial neural network selects at least one of the plurality of the candidates of the MPM list from a plurality of directional intra-prediction modes and at least one non-directional intra-prediction mode,
wherein the artificial neural network outputs information indicating which intra-prediction mode is selected to be comprised in the MPM list as the at least one of the candidates of the MPM list,
wherein the artificial neural network determines whether each of at least one non-directional intra-prediction mode is comprised in the MPM list as one of the plurality of the candidates,
wherein the at least one non-directional intra-prediction mode comprises at least one of a Planar mode and a DC mode,
wherein the intra-prediction mode for the target block is derived using the MPM list for the target block,
wherein the MPM list is configured to comprise the DC mode in a case that the artificial neural network determines that the DC mode is comprised in the MPM list, and
wherein the MPM list is configured not to comprise the DC mode in a case that the artificial neural network determines that the DC mode is not comprised in the MPM list.