US 12,439,029 B2
Video encoding and decoding using deep learning based inter prediction
Je Won Kang, Seoul (KR); Seung Wook Park, Yongin-si (KR); and Wha Pyeong Lim, Hwaseong-si (KR)
Assigned to HYUNDAI MOTOR COMPANY, Seoul (KR); KIA CORPORATION, Seoul (KR); and EWHA UNIVERSITY—INDUSTRY COLLABORATION FOUNDATION, Seoul (KR)
Appl. No. 18/019,437
Filed by HYUNDAI MOTOR COMPANY, Seoul (KR); KIA CORPORATION, Seoul (KR); and EWHA UNIVERSITY—INDUSTRY COLLABORATION FOUNDATION, Seoul (KR)
PCT Filed Aug. 6, 2021, PCT No. PCT/KR2021/010425
§ 371(c)(1), (2) Date Feb. 2, 2023,
PCT Pub. No. WO2022/031115, PCT Pub. Date Feb. 10, 2022.
Claims priority of application No. 10-2020-0098414 (KR), filed on Aug. 6, 2020; and application No. 10-2021-0103569 (KR), filed on Aug. 6, 2021.
Prior Publication US 2023/0291926 A1, Sep. 14, 2023
Int. Cl. H04N 19/105 (2014.01); H04N 19/109 (2014.01); H04N 19/13 (2014.01); H04N 19/132 (2014.01); H04N 19/139 (2014.01); H04N 19/176 (2014.01); H04N 19/42 (2014.01); H04N 19/52 (2014.01)
CPC H04N 19/105 (2014.11) [H04N 19/109 (2014.11); H04N 19/13 (2014.11); H04N 19/132 (2014.11); H04N 19/139 (2014.11); H04N 19/176 (2014.11); H04N 19/42 (2014.11); H04N 19/52 (2014.11)] 7 Claims
OG exemplary drawing
 
1. A video decoding apparatus comprising:
an entropy decoder configured to decode a motion vector for a current block and residual values from a bitstream;
an inter predictor configured to generate first predicted samples for the current block using reference samples in a reference picture and the motion vector;
a module configured to include a neural network and generate second predicted samples based on all or some of the motion vector, reference samples of at least one reference picture, and the first predicted samples; and
an adder configured to add the residual values to the first predicted samples or the second predicted samples to generate a restoration block for the current block,
wherein the motion vector is acquired by applying an advanced motion vector prediction (AMVP) mode or a merge mode, and wherein whether or not to use the second predicted samples obtained by the module is determined based on information decoded from the bitstream, and
wherein, when an inter prediction mode of the current block is an AMVP mode or a merge mode, and neighboring blocks of the current block are predicted using the neural network, priorities of the motion vectors corresponding to the neighboring blocks in a motion vector candidate list are set to be high.