Video quality assessment and QoE-driven adjustment scheme in wireless networks

With the further development of video, the assessment and optimization of video Quality of Experience (QoE) have become an important issue. Meanwhile, QoE is the main influential factor determining the success of video applications. The paper proposes a novel QoE real-time assessment and QoE-driven Send Bitrate (SBR) adjustment scheme. Firstly, in order to assess the user experience, the NS2 simulation platform is used to generate video streams and the Evalvid tool is adopted to evaluate QoE of these streams, which takes the video attributes and network parameters into account to get the Peak Signal to Noise Ratio (PSNR) value that is mapped to Mean Opinion Score (MOS) to represent QoE. Then we obtain an assessment model through neural networks. Secondly, based on the obtained model, the server provides the optimal video according to the network condition, namely adjusting the video SBR by the PSNR index and predicted bandwidth. The proposed scheme performs well in real-time video QoE assessment and makes the video server adjust video SBR, which matchs the network conditions and maximizes video QoE.