In many video-coding standards, the motion vector field is one of the most
important data in the compressed bit-stream, and its loss can lead to sever
e degradation in the decoded picture quality. We propose the multiple descr
iption motion coding (MDMC) algorithm to enhance the robustness of the moti
on vector field against transmission errors. In MDMC, the motion vector fie
ld is encoded into two descriptions, which are transmitted over distinct ch
annels to the decoder. The decoder is designed to provide an acceptable qua
lity prediction image, even if one of the descriptions is lost during the t
ransmission. Moreover, the decoder can reconstruct a higher quality predict
ion image, when both the descriptions are received without error. A complet
e multiple description video coder, based on the MDMC, is implemented by mo
difying the syntax of the H.263 standard, and tested intensively in a reali
stic error-prone environment. It is shown that the proposed algorithm provi
des much better objective and subjective performances than the H.263 coder
in the error-prone environment.