QoE-driven and Network-assisted Naked-eye 3D Adaptive Video Streaming over Wireless Network

摘要

Fluctuations in wireless channel conditions have a negative impact on naked-eye three-dimensional (3D) video streaming. While offering free stereo perception, it poses a challenge to wireless networks for the large volumes of video traffic. To obtain better viewing experiences, a trade-off between stereo perception improvement and freeze avoidance is necessary. Though adaptive streaming seems to work well, traditional bandwidth estimation methods are at risk of overestimation which may lead to stall or underestimation which may cause resource waste. To improve the quality of experience (QoE), a novel scheme named QoE-driven and Network-assisted Naked-eye 3D Adaptive Video Streaming over Wireless Network (QNANA) is proposed to transmit naked-eye 3D video streaming over wireless networks adaptively with network assistance. In view of the constraint of bandwidth capacity, the scheme can select the video of the highest QoE under the current bandwidth conditions, so as to achieve the highest QoE overall. Besides, the clients can follow the fluctuating channel conditions timely and accurately. Results show that the scheme can realize the lowest prediction error and the best resource utility with up to 54% reduction in total stall durations compared to traditional bandwidth estimation methods.