IP Library › Granted Patent US 12,641,219
Granted Patent B2
US 12,641,219 · App. 18/762,319 · Granted May 26, 2026

Video encoding/decoding method and apparatus for constructing merge candidate list by generating pairwise merge candidates

Inventors: Jin Heo (Yongin-si, KR); Seung Wook Park (Yongin-si, KR)
Assignees: HYUNDAI MOTOR COMPANY; KIA CORPORATION
H04N19/105H04N19/139H04N19/176
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,641,219
App. No.
18/762,319
Granted
May 26, 2026
Kind
B2
Abstract

A method and an apparatus for encoding/decoding a video are provided. A method for decoding a video includes generating one or more spatial merging candidates, generating a temporal merging candidate. The method also includes generating one or more history-based motion vector prediction candidates. The method also includes generating a first merging candidate list based on the one or more spatial merging candidates, the temporal merging candidates, and the one or more history-based motion vector prediction candidates. The method also includes generating one or more pairwise average merging candidates when the first merging candidate list is not fulfilled with merging candidates. The method also includes generating a second merging candidate list based on the one or more pairwise average merging candidates and the first merging candidate list. The method also includes generating a prediction block of the current block based on the second merging candidate list.

Claims (65)

1 . A method of decoding a video, the method comprising:

generating one or more spatial merging candidates based on neighboring blocks that are spatially adjacent to a current block;

generating a temporal merging candidate based on neighboring blocks that are temporally adjacent to the current block;

generating one or more history-based motion vector prediction candidates based on motion information of a block encoded before the current block;

generating a first merging candidate list based on the one or more spatial merging candidates, the temporal merging candidate, and the one or more history-based motion vector prediction candidates;

generating one or more pairwise average merging candidates when the first merging candidate list is not fulfilled with merging candidates;

generating a second merging candidate list based on the one or more pairwise average merging candidates and the first merging candidate list; and

generating a prediction block of the current block based on the second merging candidate list,

wherein generating the one or more pairwise average merging candidates comprises:

calculating cost values by performing a template matching on one or more merging candidates in the first merging candidate list;

reordering, based on the cost values, the one or more merging candidates in the first merging candidate list;

selecting two merging candidates from the reordered one or more merging candidates; and

generating the one or more pairwise average merging candidates by averaging or weight summing the two merging candidates,

wherein the two merging candidates are randomly selected.

2 . The method of claim 1 , wherein generating the one or more pairwise average merging candidates comprises:

calculating the cost values by performing the template matching on the one or more merging candidates in the first merging candidate list;

reordering, based on the cost values, the one or more merging candidates in the first merging candidate list; and

generating a single pairwise average merging candidate by using a first merging candidate and a second merging candidate among the reordered one or more merging candidates.

3 . The method of claim 2 , wherein generating the single pairwise average merging candidate comprises:

generating the single pairwise average merging candidate by averaging or weight summing the first merging candidate and the second merging candidate.

4 . The method of claim 1 , wherein generating the second merging candidate list comprises:

filling up the first merging candidate list based on the one or more pairwise average merging candidates;

generating a zero-motion vector merging candidate when the first merging candidate list is not fulfilled; and

generating the second merging candidate list by adding the zero-motion vector merging candidate to the first merging candidate list.

5 . A method of encoding a video, the method comprising:

generating one or more spatial merging candidates based on neighboring blocks that are spatially adjacent to a current block;

generating a temporal merging candidate based on neighboring blocks that are temporally adjacent to the current block;

generating one or more history-based motion vector prediction candidates based on motion information of a block encoded before the current block;

generating a first merging candidate list based on the one or more spatial merging candidates, the temporal merging candidate, and the one or more history-based motion vector prediction candidates;

generating one or more pairwise average merging candidates when the first merging candidate list is not fulfilled with merging candidates;

generating a second merging candidate list based on the one or more pairwise average merging candidates and the first merging candidate list; and

generating a prediction block of the current block based on the second merging candidate list,

wherein generating the one or more pairwise average merging candidates comprises:

calculating cost values by performing a template matching on one or more merging candidates in the first merging candidate list;

reordering, based on the cost values, the one or more merging candidates in the first merging candidate list;

selecting two merging candidates from the reordered one or more merging candidates; and

generating the one or more pairwise average merging candidates by averaging or weight summing the two merging candidates,

wherein the two merging candidates are randomly selected.

6 . The method of claim 5 , wherein generating the one or more pairwise average merging candidates comprises:

calculating the cost values by performing the template matching on the one or more merging candidates in the first merging candidate list;

reordering, based on the cost values, the one or more merging candidates in the first merging candidate list; and

generating a single pairwise average merging candidate by using a first merging candidate and a second merging candidate among the reordered one or more merging candidates.

7 . The method of claim 6 , wherein generating the single pairwise average merging candidate comprises:

generating the single pairwise average merging candidate by averaging or weight summing the first merging candidate and the second merging candidate.

8 . The method of claim 5 , wherein generating the second merging candidate list comprises:

filling up the first merging candidate list based on the one or more pairwise average merging candidates;

generating a zero-motion vector merging candidate when the first merging candidate list is not fulfilled; and

generating the second merging candidate list by adding the zero-motion vector merging candidate to the first merging candidate list.

9 . A method for transmitting a bitstream associated with video data, the method comprising:

generating the bitstream by encoding the video data; and

transmitting the bitstream to a video decoding apparatus,

wherein generating the bitstream comprises:

generating one or more spatial merging candidates based on neighboring blocks that are spatially adjacent to a current block;

generating a temporal merging candidate based on neighboring blocks that are temporally adjacent to the current block;

generating one or more history-based motion vector prediction candidates based on motion information of a block encoded before the current block;

generating a first merging candidate list based on the one or more spatial merging candidates, the temporal merging candidate, and the one or more history-based motion vector prediction candidates;

generating one or more pairwise average merging candidates when the first merging candidate list is not fulfilled with merging candidates;

generating a second merging candidate list based on the one or more pairwise average merging candidates and the first merging candidate list; and

generating a prediction block of the current block based on the second merging candidate list,

wherein generating the one or more pairwise average merging candidates comprises:

calculating cost values by performing a template matching on one or more merging candidates in the first merging candidate list;

reordering, based on the cost values, the one or more merging candidates in the first merging candidate list;

selecting two merging candidates from the reordered one or more merging candidates; and

generating the one or more pairwise average merging candidates by averaging or weight summing the two merging candidates,

wherein the two merging candidates are randomly selected.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 3, 2024
From: HEO, JIN; PARK, SEUNG WOOK
To: HYUNDAI MOTOR COMPANY; KIA CORPORATION
Reel/Frame 067909/0707 →
Priority Claims (2)
KR 10-2022-0000970 · Jan 4, 2022 · national
KR 10-2022-0190778 · Dec 30, 2022 · national
Continuity (2)
Continuation PCTKR2023000131 · Jan 4, 2023
Related Publication 20240357086A1 · Oct 24, 2024
References Cited (26)
US 11202081B2 · Zhang · 2021 [cited by applicant]
US 11509915B2 · Zhang · 2022 [cited by applicant]
US 11523123B2 · Zhang · 2022 [cited by applicant]
US 11558632B2 · Solovyev · 2023 [cited by applicant]
US 11831884B2 · Zhang · 2023 [cited by applicant]
US 11924406B2 · Zhao · 2024 [cited by applicant]
US 11973962B2 · Zhang · 2024 [cited by applicant]
US 20200396465A1 · Zhang · 2020 [cited by applicant]
US 20200404254A1 · Zhao · 2020 [cited by examiner]
US 20200404255A1 · Zhang · 2020 [cited by applicant]
US 20200404260A1 · Zhang · 2020 [cited by applicant]
US 20200413048A1 · Zhang · 2020 [cited by applicant]
US 20210006780A1 · Zhang · 2021 [cited by applicant]
US 20210203958A1 · Zhang · 2021 [cited by applicant]
US 20210281870A1 · Solovyev · 2021 [cited by applicant]
US 20210392341A1 · Zhang · 2021 [cited by applicant]
US 20220078452A1 · Zhang · 2022 [cited by applicant]
US 20220217363A1 · Zhang · 2022 [cited by applicant]
US 20230082043A1 · Hendry · 2023 [cited by applicant]
US 20230115768A1 · Liao · 2023 [cited by examiner]
US 20230125381A1 · Solovyev · 2023 [cited by applicant]
US 20240121410A1 · Zhang · 2024 [cited by applicant]
KR 20200094788A · 2020 [cited by applicant]
WO 2020182147A1 · 2020 [cited by applicant]
WO 2021172881A1 · 2021 [cited by applicant]
International Search Report and Written Opinion cited in corresponding international patent application No. PCT/KR2023/000131; Apr. 24, 2023; 10 pp. [cited by applicant]