期刊
COMPUTER COMMUNICATIONS
卷 85, 期 -, 页码 89-97出版社
ELSEVIER
DOI: 10.1016/j.comcom.2016.04.001
关键词
HTTP adaptive streaming; Priority-based; Multi-view; Bandwidth utilization
资金
- National Science Foundation of China [61472317, 61428206, 61221063, 91118005, 91218301]
- MOE Innovation Research [IRT13035]
- National Key Technologies R&D Program of China [2013BAK09B01]
- Co-ordinator Innovation Project for the Key Lab of Shaanxi Province [2013SZS05-Z01]
- China Scholarship Council
Adaptive streaming over Hypertext Transport Protocol (HTTP) has been widely used for the transmission of video content. Most of the existing studies about HTTP adaptive streaming (HAS) concentrate on improving resource efficiency, fairness among users, and quality of service for single-source videos. However, there is a growing number of live streaming applications, such as video-casts of courses, where service providers want to transmit video streams of different scenes simultaneously over the bandwidth constrained network. Since concurrent streams further deteriorate the insufficient bandwidth situation, a systematic solution is highly needed to utilize the resources efficiently and guarantee the quality of multi-view video services. In this paper, we propose a priority-based adaptive scheme for multi-view live streaming based on HTTP streaming. Firstly, we adopt an integrated bandwidth prediction approach to calculate the available bandwidth, solving the low bandwidth utilization problem existing in the benchmark method. Secondly, we design a unified segment request strategy to restrict the interval between each request, and guarantee the synchronization of different live video streams. Finally, an adaptive scheduling algorithm is proposed with respect to priorities, to dynamically adjust the quality level of multiple video streams and improve the quality of multi-view video service. Simulations show that the proposed scheme makes considerable improvements compared to the benchmarks. (C) 2016 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据