The world of mobile content delivery moves fast--very fast. Not that long ago, we were discussing multimedia messaging; now, here we are discussing the delivery of 3D media that supports multiple viewpoints. Broadcasting and distributing 3D video that can be rendered from many different viewpoints is not a trivial undertaking, hence this project in which the authors build on top of the multiview video coding extensions of the H.264/MPEG-4 standard.

The starting premise is a video encoding method that contains multiple streams of independent depth and texture data. Each additional stream enhances the quality of the depth or texture information available for the rendering node. Typically, streams for more than two viewpoints are provided so that the end display device can render custom viewpoints. The authors focus on two problems. The first is how to choose how many (and which) of the depth and texture streams to send to a mobile device for the best possible rendering quality of the unknown user viewpoint. The second is how to schedule the data into radio frames for transfer between the base station and the mobile device.

Assuming a positive linear relationship between the bit rate of depth or texture stream and the synthesized viewpoint quality, the authors decompose the first problem into the nondeterministic polynomial time (NP)-complete multiple-choice knapsack problem. They then apply an existing linear time partitioning algorithm for solving the linear programming relaxed multiple-choice knapsack problem.

To solve the radio frame allocation problem, they use double buffering at the mobile node and schedule burst transfers to the device that will run out of buffered media the soonest. This is not an ideal solution given that both problems should be optimized jointly rather than independently for the best user experience. I would also question the energy savings claimed by the authors, as even a naive round-robin implementation could be shown to have similar energy-saving gains.

This paper will be of most interest to practitioners and researchers in the telecommunications industry working in the area of 3D content delivery via mobile networks.