AN IMPROVED LOW-COMPLEXITY MULTIPLE DESCRIPTION CODING FOR PEER-TO-PEER VIDEO STREAMING

Multiple description scalable coding based on T+2D wavelet decomposition structure is highly flexible for peer-to-peer (P2P) video streaming. Finding the optimal truncation point of each code block (CB) within each description is an NP-hard problem. To implement an efficient low-complexity solution, we propose a simple clustering algorithm for partitioning the CBs into a limited number of clusters, such that one can find the optimal cluster-level redundancy-rate assignment matrix using a low-complexity full search. This approach improves the decoding quality compared to the co-echelon frameworks [1] [6] in which a non-optimal rate assignment matrix is used. In addition, the proposed clustering approach may be analytically represented by closed-form relations for low-complexity computation of optimal encoding parameters. The simulation results demonstrate that the adaptive proposed framework outperforms the approaches in [1] by (0.25 similar to 1dB), the scheme of [6] by (1.3 similar to 1.6dB), and the non-adaptive multiple description coding by (2.3 similar to 4.3dB). Furthermore, the proposed clustering approach requires %52-%88 less computations compared to the framework in [1].