Quality of Experience Measurement Tool for SVC Video Coding

The scalable extension of H.264, known as Scalable Video Coding (SVC), is recently finalized and adapted by the Joint Video Team. Scalability is achieved in the temporal, spatial, quality (SNR), or any combination of those domains. One example of using video scalability is in saving bandwidth when the same media content is required to be sent simultaneously at different resolutions to support heterogeneous devices and networks. Meanwhile, Quality of Experience (QoE) is the key criteria for evaluating the video service such as SVC. Unlike QoS metrics (such as bandwidth, delay, jitters, etc.), QoE is more accurate to reflect the user experience as it considers Human Visual System and its complex behavior towards distortions in the displayed video sequence. In order to evaluate QoE, objective assessment tools may not correlate well with the human perceived video quality and at same time, subjective quality assessment methods are costly and time consuming. In this paper, we design an automatic QoE monitoring tool for SVC video coding mechanism. The module is based on PSQA (Pseudo Subjective Quality Assessment tool), which is a hybrid (objective/subjective) assessment tool. PSQA uses RNN (Random Neural Network) to capture the non-linear relation between the video coding as well as network parameters affecting the video quality, and QoE. The results show clearly that our module can accurately estimate QoE for SVC video streams.