用户体验质量驱动的立体视频可靠传输理论与技术研究

3D视频可以带给用户身临其境的视觉体验，是多媒体信息技术的发展方向。3D视频应用的普及必须有快速高效的传播手段支撑，通过移动互联网获取信息具有方便、灵活、内容丰富等特点，是未来3D多媒体信息传播的主要渠道。针对移动互联网带宽波动范围大、信道误码/丢包不可避免等特点，研究与3D视频特性相匹配的可靠传输机制是亟待解决的关键问题。本项目首先研究网络3D视频服务所需的QoE预测模型,用视频客观参数预测感知质量；在此基础上研究以体验质量最大化为目标的3D视频自适应传输机制,针对网络视频业务，充分利用人眼的感知特性，设计符合MPEG-DASH协议的3D视频自适应传输方法；针对无线视频传输，研究结合应用层、物理层前向纠错的跨层3D视频流不等差错保护机制，使用户在延迟、带宽受限的环境下，获得最佳的感知质量。其研究成果有助于提高视频服务质量和效率，对推动基于移动互联网的3D视频服务发展有重要意义

3D media can bring the user with immersive audio-visual experiences, it is the main streams of multimedia information technology. The popularity of 3D video applications must have the support of fast and efficient distribution methods. Mobile Internet which has the advantages of convenient, flexible, rich in content and etc., undoubtedly, will be the main channels of the upcoming 3D multimedia information transportation. Mobile Internet provides a best of effort service, suffer from bandwidth variation and packets-loss. Designing reliable transmission mechanism cater to 3D video is a challenge to both industry and academic. Building QoE prediction model for internet based 3D media service is the first research topic of this project. Base on the QoE prediction model, 3D video adaptive transmission mechanism in mobile internet will be explored. This project will take the latest advances in video coding ,control theory and human visual system characteristics as a integral parts，develope a 3D video adaptive transmission scheme conforming to MPEG-DASH standards. Aiming at the applications based on UDP protocal, unequal error protection scheme jointly considering the application layer FEC,physical layer FEC and the inherent hierarchical attributes of compressed video stream is developed. It provides the users with the best perception quality under bandwidth and delay constrains. The outcome of this projects will improve the quality of video service and utilization of network bandwidth，make significant sense for the popularity of 3D media application.