DocumentCode
57037
Title
Edge-Based Intramode Selection for Depth-Map Coding in 3D-HEVC
Author
Chun-Su Park
Author_Institution
Dept. of Digital Contents, Sejong Univ., Seoul, South Korea
Volume
24
Issue
1
fYear
2015
fDate
Jan. 2015
Firstpage
155
Lastpage
162
Abstract
The 3D video extension of High Efficiency Video Coding (3D-HEVC) is the state-of-the-art video coding standard for the compression of the multiview video plus depth format. In the 3D-HEVC design, new depth-modeling modes (DMMs) are utilized together with the existing intraprediction modes for depth intracoding. The DMMs can provide more accurate prediction signals and thereby achieve better compression efficiency. However, testing the DMMs in the intramode decision process causes a drastic increase in the computational complexity. In this paper, we propose a fast mode decision algorithm for depth intracoding. The proposed algorithm first performs a simple edge classification in the Hadamard transform domain. Then, based on the edge classification results, the proposed algorithm selectively omits unnecessary DMMs in the mode decision process. Experimental results demonstrate that the proposed algorithm speeds up the mode decision process by up to 37.65% with negligible loss of coding efficiency.
Keywords
Hadamard transforms; computational complexity; data compression; decision theory; edge detection; image classification; video coding; 3D video extension; 3D-HEVC design; DMM; Hadamard transform domain; compression efficiency; computational complexity; depth intracoding; depth-modeling modes; edge classification; edge-based intramode selection; fast mode decision algorithm; high efficiency video coding; intramode decision process; intraprediction modes; multiview video-plus-depth format compression; prediction signals; Algorithm design and analysis; Encoding; Image coding; Partitioning algorithms; Prediction algorithms; Software algorithms; Transforms; 3D-HEVC; Hadamard transform; depth maps; depth-modeling modes; intra coding; mode decision;
fLanguage
English
Journal_Title
Image Processing, IEEE Transactions on
Publisher
ieee
ISSN
1057-7149
Type
jour
DOI
10.1109/TIP.2014.2375653
Filename
6966781
Link To Document