Soft video multicasting over wireless networks based on BM3D and approximate message passing

Recently, a soft video multicasting scheme called distributed compressed sensing-based multicast (DCS-cast) has achieved better video quality when multiple receivers have heterogeneous channel characteristics. However, its performance is still unsatisfactory when applied to a noisy channel. In this paper, we put forward a scheme called BM3D-AMP-cast, which combines the block-matching with 3-D filtering (BM3D) denoising algorithm and the approximate message passing (AMP) based compressed sensing recovery algorithm, with the goal of improving the performance of DCS-cast in noisy environments. Experimental results show that through comparing to SoftCast and DCS-cast schemes, the proposed BM3D-AMP-cast is more effective in unicast and multicast scenarios on the condition that high packet loss rate. Moreover, visual quality is better.performance