Proceedings of SPIE - the International Society for Optical Engineering
This paper presents a multiple description scalable video coding scheme based on overcomplete motion compensated temporal filtering, named MD-OMCTF, for robust video transmission over wireless and packet loss networks. The intrinsic nature of the structure of OMCTF and embedded coding with modified SPIHT algorithm enable us to provide fully scalable properties for the proposed scheme. We show that multiple description coding is very effective in combating with channel failures in both Internet and wireless video. The integration of MD with OMCTF allows us to achieve both loss resilience and complete scalability. In order to further improve error-resilience to channel bit error for this scheme and reduce error propagation in error-prone network, we apply error resilient entropy coding (EREC) to the multiple bitstreams to gain additional error resilience. With EREC, multiple bitstreams are reorganized into fixed-length slots so that synchronization of the beginning of each bitstream can be automatically obtained at the receiver. The integration of scalable coding and EREC with MDC enables the coded video bitstream to be adaptive to the varying channel condition and to be resilient to both transmission losses and bit errors. We also develop corresponding error concealment scheme to recover the lost or erroneous information during video transmission. Experimental results show that the proposed scheme is able to achieve robust video transmission over both wireless and packet loss networks.
Zheng, H., Yu, L., & Chen, C. W. (2005). Robust video transmission based on multiple description scalable coding with EREC. Paper presented at the Proceedings of SPIE - the International Society for Optical Engineering, 5960(4) 1891-1902. doi:10.1117/12.633197