Towards QoE-Driven Optimization of Multi-Dimensional Content Streaming

Abstract

Whereas adaptive video streaming for 2D video is well established and frequently used in streaming services, adaptation for emerging higher-dimensional content, such as point clouds, is still a research issue. Moreover, how to optimize resource usage in streaming services that support multiple content types of different dimensions and levels of interactivity has so far not been sufficiently studied. Learning-based approaches aim to optimize the streaming experience according to user needs. They predict quality metrics and try to find system parameters maximizing them given the current network conditions. With this paper, we show how to approach content and network adaption driven by Quality of Experience (QoE) for multi-dimensional content. We describe components required to create a system adapting multiple streams of different content types simultaneously, identify research gaps and propose potential next steps.