基于MPEG-DASH的视频自适应共享系统

Adaptive video sharing system based on MPEG-DASH

最新的超高清视频、立体视频等视频因其数据量巨大,无法在现有的网络当中实现自由共享,HEVC及3D-HEVC作为最新的视频及立体视频压缩标准,性能优越,但依然无法实现超高清视频及立体视频的自适应传输,本文基于MPEG-DASH的切片编码方案,将视频切分为不同的时间片段,然后对每个片段采用不同的参数进行编码,生成不同码率的视频码流,以适应不同的网络情形,然后采用基于高斯分布的网络带宽预测法方案,实时预测下一个时刻的网络宽,实现了超高清视频及立体视频在异构网络及网络波动情形下的自由共享。

Due to the huge amount of data,the latest ultra-high definition video,stereo video cannot be freely shared in the existing network.HEVC and 3D-HEVC are the latest video and stereo video compression standards which have superior performance,but still can not achieve adaptive transmission for ultra-high definition video and stereo video.Based on MPEG-DASH segment coding scheme,we develop a system,in which the video is divided into different time segments,and then each segment is encoded with different parameters to generate video streams with different bit rates in order to adapt to different network situations,then bandwidth prediction scheme based on Gaussian distribution is adopted to predict the bandwidth for the next moment in real-time,and at last we realize the free sharing of ultra-high definition video and stereo video in heterogeneous network and network fluctuation situation.