When transmitting compressed video over a data network, one has to deal wit
h how channel errors affect the decoding process. This is particularly a pr
oblem with data loss or erasures. In this paper we describe techniques to a
ddress this problem in the contest of Asynchronous Transfer Mode (ATM) netw
orks. Our techniques can be extended to other types of data networks such a
s wireless networks. In ATM networks channel errors or congestion cause dat
a to be dropped, which results in the loss of entire macroblocks when MPEG
video is transmitted. In order to reconstruct the missing data, the locatio
n of these macroblocks must be known. We describe a technique for packing A
TM cells with compressed data, whereby the location of missing macroblocks
in the encoded video stream can be found. This technique also permits the p
roper decoding of correctly received macroblocks, and thus prevents the los
s of ATM cells from affecting the decoding process. The packing strategy ca
n also be used for wireless or other types of data networks. We also descri
be spatial and temporal techniques for the recovery of lost microblocks. In
particular, we develop several optimal estimation techniques for the recon
struction of missing macroblocks that contain both spatial and temporal inf
ormation using a Markov random field model. We further describe a sub-optim
al estimation technique that can be implemented in real time.