流量付费因素影响下用户视频满意度评估研究

With the popularization of Internet, and the proposed policies of heightening the network speed and reducing telecoms charges, the consumption of mobile video content will become a trend. However, the payment of mobile data can have an immediate impact on the customer’s satisfaction of the video service. How to provide video services with lower-cost and higher-quality for end-users based on accurate evaluation and forecast for video satisfaction has already became a crucial problem demanding a prompt solution. To this end, we focus on the research of assessment for customer satisfaction of mobile video services, and plan to reveal the relationship between the customer satisfaction and the cost of mobile data consumption, as well as the principle and regularity of interaction between customer satisfaction and joint selection of video encoding parameters and transmission modes. Then, the personalized customer satisfaction model is developed. The ultimate goal is to propose a joint optimization of video coding and transmission scheme driven by customer satisfaction with consideration of the related cost. The innovation points includes: 1) The cost of mobile data and users’ personalized features are introduced into video quality assessment system with the user’s satisfaction at the core. 2) To reveal the intrinsic laws between mobile data payment and customer satisfaction, the method of blending the empirical analysis and mathematical modeling is adopted. 3) Solutions to satisfaction-driven video coding and transmission based on various application scenarios will be proposed.

随着互联网的广泛应用和我国“提网速、降网费”政策的落地，使用数据流量观看移动视频将成为一种趋势，而流量付费对用户视频满意度将会产生直接影响。如何在准确的视频满意度评估和预测基础上，为用户提供高质量低花费的视频服务，已经成为亟待解决的重要问题。本项目以视频满意度评估为研究对象，探索流量付费与用户视频满意度之间的定性定量关系，揭示视频编码参数与传输模式的联合选取影响用户视频满意度的机理和内在规律性，建立流量付费影响下的个性化视频满意度评估模型；其目标是提出一套全新的满意度驱动下的视频编码-传输联合优化方案。创新点：1）以“用户满意度”为核心，将流量付费和用户个性化特征两个因素纳入评估体系；2）采用“实证分析”和“数学建模”相结合的方法，揭示流量付费影响视频满意度的机理和内在规律；3）基于不同应用场景，提出满意度驱动的视频编码-传输方案。

随着移动互联网的广泛应用和我国“提网速、降网费”政策的落地，使用数据流量观看移动视频已经成为了一种趋势。但是，流量付费对用户视频满意度将会产生直接影响。因此，如何在准确的视频满意度评估和预测的基础上，为用户提供高质量、低花费的视频服务，是一个亟待解决的重要问题。本项目以图像/视频满意度评估为研究对象，探究了流量付费、语义场景对用户视频/图像满意度的影响。并且从DCT域和深度特征感知域等方面，揭示了视频编码参数与传输模式的联合选取影响用户视频满意度的机理和内在规律性，建立了特定语义场景和流量付费影响下的个性化视频满意度评估模型。在此基础上，通过探究端到端视频质量与编码参数与视频内容的内在联系，提出了一套全新的针对实时移动网络视频的抗误码-低复杂度编码传输方案。通过主观视频质量测试实验，验证了提出了流量付费影响下以及特定语义场景下的视频满意度模型能够准确预测和描述用户的主观满意度体验。提出的基于DCT域和深度特征感知域的图像质量评价模型在对公开数据库的测试集的质量评价准确率结果中能够超过现有模型。提出的针对实时移动网络视频的编码传输方案能够有效提升移动视频在有损网络传输下的抗误码能力，同时复杂度的大幅度降低满足了实时视频服务对低时延的需求，提升了用户对移动视频服务的满意度。