Optimized Watermarking for Light Field Rendering based Free-View TV

In Free-View Television the viewers select freely
the viewing position and angle of the transmitted multiview
video. It is apparent that copyright and copy protection
problems exist, since a video of this arbitrarily selected
view can be recorded and then misused. In this context, the
watermark in FTV should not only be resistant to common
video processing and multi-view video processing
operations (as in 2D case), but it should also be extracted
from a generated video of an arbitrary view. Based on this
remark, this paper focuses on this problem by evaluating the
functionality and the efficiency of the watermarks and their
corresponding Mathematical Distributions, in terms of
“robustness” and “successful detection” from new
constructed views of FTV, using Light Field Rendering
(LFR) techniques. We studied the values which characterize
the watermark’s performance and the parameters introduced
by the watermark’s insertion-extraction scheme. Therefore,
we ended up to the best five Mathematical Distributions,
and we concluded that the watermark’s robustness in FTV
case does not depend only on the FTV image’s
characteristics, but it also relies on the characteristics of the
Mathematical Distribution that is used as watermark
generator.