Skip to main navigation Skip to search Skip to main content

A new fast algorithm for sample adaptive offset

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

Abstract

Sample Adaptive Offset is a new adopted technology by HEVC in recent years, which improves the visual quality of reconstructed videos significantly. However, there are two problems in current SAO technology. The first is that the statistic phase needs to traverse each pixel to collect relevant information. The other problem is that the complexity of SAO is too high for SAO mode decision stage, which needs to be performed on each CTU. To solve these problems, we proposed a fast SAO algorithm in HEVC encoder. We explore the correlation of SAO type among neighboring CTUs, and then utilize this spatial information to reduce the complexity of SAO. Experimental results demonstrate that our proposed method can achieve about 62%, 80% and 75% SAO encoding time saving on average in AI, RA, and LDB test condition compared with HM16.0 respectively. At the same time, the proposed method just causes negligible compression performance loss.

Original languageEnglish
Title of host publicationAdvances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers
EditorsBing Zeng, Hongliang Li, Qingming Huang, Abdulmotaleb El Saddik, Shuqiang Jiang, Xiaopeng Fan
PublisherSpringer Verlag
Pages388-396
Number of pages9
ISBN (Print)9783319773827
DOIs
StatePublished - 2018
Event18th Pacific-Rim Conference on Multimedia, PCM 2017 - Harbin, China
Duration: 28 Sep 201729 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10736 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference18th Pacific-Rim Conference on Multimedia, PCM 2017
Country/TerritoryChina
CityHarbin
Period28/09/1729/09/17

Keywords

  • HEVC
  • Merge
  • SAO
  • Statistic

Fingerprint

Dive into the research topics of 'A new fast algorithm for sample adaptive offset'. Together they form a unique fingerprint.

Cite this