Weighted Pairwise MV Candidate Addition with Zero Motion Vector Reduction for Merge Candidate

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In the Versatile Video Coding (VVC), the merge mode lets a decoder to derive a motion vector (MV) autonomously using the motion information of a selected previously inter-coded block. For this purpose, a merge candidate list is maintained by encoder and decoder by filling it up with various candidates. In this process, if the merge candidate list is not sufficiently full, zero motion vector is repetitively added as required, thus, the redundant zero MVs in the merge candidate list may limit diversity of the list, resulting in low coding efficiency. In this paper we propose a new method of adding up additional merge candidates based on pairwise average of merge candidates. The additional MVs replacing the redundant zero MVs are obtained by linearly combining the MVs of the first and second candidate in the merge candidate list. The proposed method attains coding gain of 0.01%, 0.04%, and 0.05% respectively of Y, Cb, and Cr channels against VTM-22.0. The result also shows that the candidate list is diversified more by the proposed method, which consequently increases the coding efficiency.

Original languageEnglish
Title of host publicationInternational Workshop on Advanced Imaging Technology, IWAIT 2024
EditorsMasayuki Nakajima, Phooi Yee Lau, Jae-Gon Kim, Hiroyuki Kubo, Chuan-Yu Chang, Qian Kemao
PublisherSPIE
ISBN (Electronic)9781510679924
DOIs
StatePublished - 2024
Externally publishedYes
Event2024 International Workshop on Advanced Imaging Technology, IWAIT 2024 - Langkawi, Malaysia
Duration: 7 Jan 20248 Jan 2024

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume13164
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

Conference2024 International Workshop on Advanced Imaging Technology, IWAIT 2024
Country/TerritoryMalaysia
CityLangkawi
Period7/01/248/01/24

Keywords

  • inter prediction
  • pairwise average merge candidate
  • regular merge
  • Versatile Video Coding
  • Video Compression
  • zero motion vector

Fingerprint

Dive into the research topics of 'Weighted Pairwise MV Candidate Addition with Zero Motion Vector Reduction for Merge Candidate'. Together they form a unique fingerprint.

Cite this