Graphics Reference
In-Depth Information
As the base layer of this configuration is compatible with H.264/AVC decoders, users with
a H.264/ AVC decoder will be able to decode the colour image sequence, whereas users
with a SVC decoder and 3D rendering hardware will be able to decode the depth map
sequence and experience the benefits of stereoscopic video. Therefore, the backward com-
patibility nature of this scalable coding configuration can be employed to enhance or scale
existing 2D video applications into stereoscopic video applications. Furthermore,thiscon-
figuration canbe utilized toexploit asymmetric codingof the colourand depth map image
sequencesincescalableH.264/AVCsupportsarangeoftemporal,spatialandqualityscalable
layers. For instance, during encoding, the resolution of colour and depth image sequences
can be independently changed based on their inherent characteristics, without affecting
the perceptual qualities of 3D video. The proposed asymmetric coding methods with this
scalablecodingconfigurationarepresentedin
[131].
Furthermore, inter-layer redundancies
canbeexploitedbasedonthecorrelationbetweenthecolouranddepthimages.Theresults
presented here make use of the adaptive inter-layer prediction option in the JSVM (Joint
Scalable Video Model) reference software codec, which selects the coding mode for each
MB (Macro-Block) using Rate-Distortion (R-D) optimization. Moreover, this coding con-
figuration complies withthe ISO/IEC 23002-3(MPEG-C part 3)standard,which specifies
the inter-operability among applications based on colour plus depth/disparity sequences
[28].
The output of this coding configuration is a single SVC bit-stream which can be par-
tially decoded. Therefore, the transmission and synchronization of this content over com-
munication channel is not difficult compared to sending colour and depth streams separ-
ately. Furthermore, the supportive technologies (e.g. RTP format for SVC) are emerging to
support the delivery of SVC data streams over communication channels
[43].
Subsection
3.5.1 discusses the coding efficiency of the proposed SVC configuration compared to
MPEG-4 MAC and H.264/AVC based stereoscopic video coding approaches.
Stereoscopic Video Coding with MPEG-4 Multiple Auxiliary Components (MAC):
The MAC (Multiple Auxiliary Components) is added to Version 2 of the MPEG-4 Visual
part
[57]
in order to describe the transparency of video objects. The MAC is defined for
a video object plane (VOP) on a pixel-by-pixel basis and contains data related to video
objects such as disparity, depth, and additional texture. As MPEG-4 MAC allows the en-
coding of auxiliary components (e.g. depth, disparity, shape) in addition to the Y, U and
V components present in 2D video, this coding approach can be utilized to compress col-
ThecodingofmonoscopicvideoplusdepthmapwithMPEG-4MACisillustratedinFigure
