A Multiple Description Distributed Compressed Video Sensing for Robust Transmission

This paper proposes a multiple description distributed compressed video sensing (MD-DCVS) for the robust transmission of low-power video devices. In this scheme, the encoder firstly divides the video sequence into two sub-sequences (two descriptions). Then, both descriptions are compressed by the distributed compressed video sensing (DCVS) which behaves low-complexity encoding property. At the decoder side, two kinds of decoders, central decoder and side decoder are explored respectively for the case when two descriptions or only one description is received, which guarantees that the recovered video is acceptable even when one description is lost. The experimental results show the superiority of our MD-DCVS especially when the measurement rate is low.