提出了一种基于压缩感知理论的新型无线视频传输方案残差分布式压缩感知组播(RDCScast).在编码端,我们提出运用残差压缩感知技术来表示非参考帧,然后把参考帧的原始观测值和非参考帧的残差观测值一起打包进行传输,最主要的是我们提出了对每一个包里面的残差观测值和原始观测值进行最佳能量分配方案。在解码端,不同信道特性的接收者能够获得相应数量的包,最后运用运动估计重建视频帧。实验结果表明对于无线视频组播,RDCS-cast方法比现有的Softcast和DCS-cast方法更有效且保持了低复杂度的编码。
This paper proposed a novel wireless video transmission scheme named residual distributed compressed- sensing multicast (RDCS-cast) based on compressed sensing theory. At the encoder, we first proposed to use re- sidual compressed sensing technique to measure the non-reference frames and the residual measurements are pack- aged together with the original measurements of reference frame. More importantly, we derive an optimal power al- location scheme for the residual and original measurements within each packet. At the decoder, the receivers with different channel characteristics obtain a certain number of packets and then reconstruct video frames by exploiting motion estimation. Experimental results show that RDCS-cast is more effective than the-state-of-art Soflcast and DCS-cast for wireless videocast and preserves a low-complexity encoding.