A DASH-Based 3D Multi-view Video Rate Control System

In this paper, we propose a dynamic adaptive rate control system and its associated rate-distortion model for multi-view 3D video transmission, which will improve the user's quality of experience in the face of varying network bandwidth. Our rate control system has been built on top of two state-of-the-art key technologies: High Efficiency Video coding (HEVC), and MPEG's Dynamic Adaptive Streaming over HTTP (DASH). We show how to prepare the content at the server side and present a policy for the client to choose content from the server based on our distortion model for views reconstruction. The proposed system is tested under different network conditions. We also provide a user-based test for subjective evaluation of the rendered views, to decide on the number of views and quality of each view to be encoded for a specific network condition. The results of our simulation and the subjective test show that our proposed rate control system for 3D multi-view video allows for the transmission of different bitstreams at a higher quality, compared to non-dynamic adaptive rate control, given network bandwidth fluctuations.