Block-based temporal error concealment for video packet using motion vector extrapolation

When transmitted over error prone networks, the compressed video can suffer severe degradation. One approach to combat the error effect is the concealment technique. A novel temporal concealment method is proposed. The technique first extrapolates motion vectors of macroblock (MB) to the corrupted frame. Then, the overlapped areas between the damaged block and the motion extrapolation MBs are estimated, and the best estimation motion vector is obtained. Finally, the corrupted block is concealed with the block in the previous frame. This method overcomes the disadvantages of incorrect MB displacement, and decreases the computational complexity compared with the methods based on pixels. Simulation results, with one entire frame lost or interlace group of blocks (GOBs) lost and an H.263+ codec, show satisfactory concealment performance of the proposed technique and the suitability for real-time video packet applications.