Multiple description video coding using motion-compensated algorithm

This paper proposes a multiple description (MD) video coder, which uses motion-compensated predictions. We suggested a new algorithm of MD named polyphase downsampling that describes intra and inter frames. Results show that by using the coder scheme and the error environment like those suggested by A. Rebman, et. al. [IEEE Trans. CSVT; vol. 12, No.3, p.193-204, March 2002] our MD algorithm improves the performance of the system both in coding efficiency and error-resilience. Moreover our MD coding approach allows the introduction of a multi level scalability. This last topic is the direction follows in the future research.