Fast Depth Video Coding Method Using Adaptive Edge Classification
- PDF / 713,524 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 51 Downloads / 257 Views
Fast Depth Video Coding Method Using Adaptive Edge Classification Da-Hyun Yoon · Yo-Sung Ho
Received: 15 February 2011 / Revised: 18 September 2012 / Published online: 9 November 2012 © Springer Science+Business Media New York 2012
Abstract In this paper, we propose a fast-mode decision algorithm for both intra and inter prediction in depth videos. The proposed algorithm can reduce the computational complexity of the depth video coding. According to the depth variation, depth video can be classified into depth-continuity and depth-discontinuity regions. From experiments, we determine a threshold value for classifying these regions. Since the depth-continuity region has an imbalance in the mode distribution, we limit the mode candidates to reduce the complexity of the mode decision process. Experimental results show that our proposed algorithm reduces the encoding time by up to 78 % and 97 % for the intra and inter frames, respectively, compared to JMVC 8.3, with negligible PSNR loss and bitrate increase. Keywords Depth video coding · Macroblock mode decision · Depth compression
1 Introduction Due to advances in three-dimensional (3D) display technologies, the development of 3DTV has realized the human dream of viewing scenes as if in the real world. Via 3DTV, the interactive selection of viewpoint and direction becomes possible within a certain operational range. This is referred to as free viewpoint TV (FTV). FTV has been widely utilized because it transmits and records all spatiotemporal information from the real world [1]. Multiview plus depth (MVD) is another framework used D.-H. Yoon () · Y.-S. Ho Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712, Korea e-mail: [email protected] Y.-S. Ho e-mail: [email protected] url: http://vclab.gist.ac.kr
804
Circuits Syst Signal Process (2013) 32:803–813
to represent 3D scenes. MVD has been used to synthesize intermediate views from captured images and depth maps. In a recent Moving Picture Experts Group (MPEG) meeting, MVD has received increased attention and has been discussed as a nextgeneration FTV format [10]. Since the amount of data and complexity of MVD is proportional to the number of cameras, an efficient encoding method for 3D video scenes must be developed. An efficient multiview video coding (MVC) scheme is needed to transmit and store multiview contents. The MPEG 3D Audio/Video (3DAV) group is working on the standardization of MVC. The joint MVC (JMVC) was released by the MPEG 3DAV group as the reference software. JMVC uses a rate-distortion optimization (RDO) technique [8] to select the optimal coding mode, where the rate-distortion (RD) cost is calculated by J = D + λR
(1)
where J is the RD cost of the current mode, D denotes the distortion between the original and reconstructed macroblock (MB), R stands for the total bits of the MB header, motion vectors, and DCT coefficients, and λ is the Lagrange multiplier. JMVC 8.3 [3] is then used to select the best mode among 14 different macroblock modes: SKIP, Dir
Data Loading...
