Adaptively post-encoding multiple description video coding

In video coding and transmission, the network conditions are crucial, i.e., bandwidth, delay and other factors. However, how to effectively take these factors into account is still a challenge, as heterogeneous networks are dynamic, and therefore it is difficult to predict their changes. This paper first reports a post-encoding of Scalable Multiple Description Coding (SMDC), which is towards self-adaptive video delivery under different conditions of the network. The proposed scheme contains three major steps: (1) spatiotemporal wavelet transformation of input video sequence; (2) context-based adaptive binary arithmetic coding; and (3) rate allocation under the conditions of the network and analysis of the principle relationship of the rate–distortion slope ratio with packet-loss probability in network links. The performance of the SMDC is compared with that of scalable video coding and scalable H.264, and SMDC is demonstrated to effectively optimize the video delivery to adapt to the dynamics of heterogeneous networks. Part of this work has been published on Data Compression Conference 2009 as a short abstract version [37].