Low complexity depth map intra prediction mode decision algorithm for 3D-HEVC

The presented 3D video coding extension of High Efficiency Video Coding (3D-HEVC) was developed for multiview video plus depth (MVD) format. MVD is depth-enhanced 3D video format that uses a few views to synthesize more immediate virtual views with depth map. In the original 3D-HEVC encoder, as a no...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Optik (Stuttgart) Ročník 127; číslo 16; s. 6291 - 6302
Hlavní autori: Huang, Xinpeng, Zhang, Qiuwen, Wang, Xiao, Wu, Qinggang, Huang, Kunqiang, Gan, Yong
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier GmbH 01.08.2016
Predmet:
ISSN:0030-4026, 1618-1336
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:The presented 3D video coding extension of High Efficiency Video Coding (3D-HEVC) was developed for multiview video plus depth (MVD) format. MVD is depth-enhanced 3D video format that uses a few views to synthesize more immediate virtual views with depth map. In the original 3D-HEVC encoder, as a novel tool Depth Modeling Modes (DMM) is added into intra prediction mode candidate list, which aims at improving as much coding efficiency as possible for 3D-HEVC. However, this promotion leads to extremely complicated computation simultaneously which prevents 3D-HEVC encoder from real-time application. In this paper, we propose a low complexity depth map intra prediction mode decision algorithm, including fast angular modes decision algorithm and early skipping DMM decision algorithm. The basic idea of the proposed algorithm is to predict the current depth block intra prediction mode based on depth map texture feature and early skip DMM decision. Experimental results demonstrate that the proposed low complexity algorithm can achieve 38% coding time reduction with just 1.00% bitrate augment on average under “all-intra” case compared to the original 3D-HEVC encoder.
Bibliografia:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0030-4026
1618-1336
DOI:10.1016/j.ijleo.2016.04.099