A QoE-driven optimization strategy for dynamic adaptive streaming over HTTP

In this paper, we first propose a Quality of Experience (QoE) evaluation model for dynamic adaptive streaming over HTTP (DASH) services. The proposed model predicts the perceived quality of user based on segment media quality, playback continuity and perceptual quality fluctuations caused by bitrate switching. Large quantities of subjective mean-opinion-score (MOS) tests demonstrate that our QoE evaluation model can evaluate users' perception on DASH services quality accurately. Based on the model, we further propose a dynamic selection mechanism of adaptive algorithms. In the case where the network bandwidth fluctuated, the mechanism determine next segment through selecting the optimal adaptive algorithm. At the end of the paper, we conduct experiment in various different environment of bandwidth so as to verify the performance of the mechanism. The mechanism is proved to have very good effect on improving QoE in different network environment.