An adaptive QoE-based multipath video streaming algorithm for Scalable Video Coding (SVC)

In the domain of video streaming, Scalable Video Coding (SVC) comes as a solution that adapts to network bandwidth fluctuation and to terminals heterogeneity. The multilayer-coding criterion of SVC has increased its correlation with the streaming over multiple paths. However, an important challenge that we still need to face is how to maximize the overall video quality in a way that satisfies the end-user, without greedily consuming the network resources. To this purpose, in this paper we aim at providing a powerful system that is based on the Quality of Experience (QoE) evaluations, to adjust the streaming of SVC video flows over multiple paths. Our method dynamically selects the best overlay paths using the available bandwidth estimations. Keeping/updating the selected paths is then automatically done based on the feedback of the quality as it is perceived by the end-user. For evaluating the QoE at the recipient, we use an SVC-compatible module of the hybrid Pseudo-Subjective Quality assessment (PSQA) tool. Results show how our method can guarantee the deliverance of the best possible video quality, while keeping the balance between end-user requirements and network bandwidth wise consumption.