Slice based video rate control for network adaptation

The video traffic over Internet is increasing and new applications with higher frame rates and resolution are being deployed. Such applications are likely to put heavy demands on network resources. Network congestion could cause some packets to be dropped which can have detrimental effects on video traffic. Also, the video data rate needs to be reduced in case of congestion to enable the other network traffic to continue. If the compressed video data allows certain video units to be dropped then the rate adaptation becomes possible. The retransmission of missing data in case of a packet loss is generally not feasible due to strict delay requirements of video. This could however be addressed by providing forward error correction (FEC). A low cost rate control scheme is proposed to enable rate adaptation in case of congestion. It is shown that a range of different rate points can be obtained by reducing the temporal and quality information which could optimally match the available bandwidth. The results show that the scheme provides the optimum PSNR for a given set of bandwidth constraints.