Graphics Reference
In-Depth Information
3.3 3DVideo Coding
Scenery captured from different angles/viewpoints simultaneously results a large amount
of raw video data. For instance, stereoscopic video, which is one of the simplest forms
of 3D video, requires twice the size of storage capacity and double the bandwidth com-
paredtotherequirementsof2Dvideo.Thus,3Dvideocodingiscrucialifimmersivevideo
applicationsaretobeavailableforthemassconsumermarketinthenearfuture.TheMPEG
Ad-hoc group which worked on exploration of 3DAV (3D Audio-Visual) application scen-
arios and technologies, identified several MPEG codingtools which are directly applicable
for different representations of 3D video and also identified the areas where further stand-
ardization is required
[5].
The coding approaches for 3D video may be diverse depending
on the representation of 3D video. For example, available principles of classical video cod-
ing can be utilized to compress pixel-type data including stereo video, multi-view video,
and associated depth or disparity maps. However, efficient coding approaches and stand-
ardization for some 3D representations such as multi-view video are yet to be discovered.
Thecompressionalgorithmsfor3Dmeshmodelshavealsoreachedahighlevelofmaturity.
Even though there are efficient compression algorithms for static geometry models, com-
pression of dynamic 3D geometry is still an active field of research. A survey of coding
algorithmsfordifferent3Dapplicationscenariosispresentedin
[44].
Thecodingofimage-
based3Drepresentations(e.g.stereoscopicvideo)is addressed in this topic.Henceforth,the
encodingofimage based 3D representations is discussed in this chapter.
