Applications of Video Distortion Estimation Algorithms for Efficient Video Streaming

This paper applies three low complexity methods, previously proposed by the authors, to estimate the distortion deriving from packet losses in error-prone video communication systems. In a first application scenario, the distortion estimated by the encoder is encapsulated in each video packet and used as side information by a modular router, in which the architecture is properly modified for taking into account the distortion impact produced by the loss of each packet. Computational complexity considerations, compared with traditional Rate Distortion Optimization (RaDiO) techniques, are also discussed, showing that the proposed distortion estimation algorithms may be easily used and integrated in a wireless router with limited CPU resources. In the paper a multi-hop scenario is also considered as possible application environment, in which the prioritization and selective discarding policy are based on the estimated distortion and on the loss pattern deriving form previous hop transmissions.