Information Technology Reference
In-Depth Information
displays [1, 2], a new era of 3D video has started. On one hand, new methods are be-
ing proposed to generate 3D video content in different forms, and on the other hand
compression techniques for 3D video data are being developed [3, 4]. In the enter-
tainment sector, 3D video is being used in filming, 3D video animation, and games. In
the security sector, surveillance companies and authorities are using 3D video cameras
to capture and analyze 3D video of sensitive locations such as airports, railway sta-
tions, public gatherings, office buildings, and car parks. High definition 3D images are
also used in medical research facilities to aid analysis.
Recently, 3D home cinema for entertainment has become a reality. Service
providers and high-tech industry leaders are making alliances to provide consum-
ers with high quality, yet affordable in-home 3D entertainment [5]. A large num-
ber of TV users are already enjoying stereoscopic 3D content. The growth in sales
volume of 3D LCDs, and 3D enabled DLP HDTVs is already attracting the inter-
est and attention of many of the key players in video communications. A typical
home entertainment scenario will have 3D high definition viewing devices, broad-
band connection for delivering multimedia content form the server to the user's
premises, and Wireless Local Area Network (WLAN) for transmission of video
and other data to various terminals inside the home.
Video data transmission over wireless networks has always been a challenging
problem due to the highly variable nature of wireless networks. Packets may be
dropped during transmission or may only reach the destination after significant de-
lay. The effects of errors propagate into succeeding frames in almost every form
of video coding process. Various schemes exist in the literature that mitigate the
effects of errors during transmission of 2D video data [6, 7]. However, the trans-
mission of compressed 3D video data still needs to be analyzed for various wire-
less conditions. In this work, 3D video transmission over wireless networks is
considered in a home entertainment scenario.
Video data passes through a number of network layers during its distribution. Re-
cent research findings, [8], have demonstrated that sub optimal perceptual quality is
achieved if encoding algorithms at higher level layers do not consider the techniques
applied at the lower level layers, e.g. scheduling, routing, etc. However, adapting the
coding and transmission strategies jointly across the layers helps in maintaining a con-
sistently high quality at the receiving end. Various cross layer design approaches have
been suggested for optimization of 2D video transmission. A weighted fair scheduling
algorithm based on adaptive rate control is presented in [9]. An adaptive cross layer
video multicast streaming algorithm for multi-rate WLANs has been proposed in [10],
which optimizes the data rate at the physical layer for every multicast receiver accord-
ing to its perceived channel conditions. However, these schemes have not considered
any of the content attributes for optimization.
Our main objective is to design a cross layer optimisation approach in order to
achieve efficient 3D video transmission over wireless networks. It is assumed that in
most videos, objects closer to the camera are more important for perceptual quality
compared to those that lie farther away in the video scene. Two reasons can be given
for this assumption. First, objects closer to the camera may be rendered so that they
appear in front of the display, and will thus draw most of the viewers attention. Sec-
ondly, objects closer to the camera undergo more warping during the 3D stereoscopic
