Motion vector recovery based colour information

In this paper, motion vector (MV) recovery algorithm using the colour information of the surrounding macroblocks (MBs) on the left, top-left, top, top-right, right, bottom-right, bottom and bottom-left is proposed. The technique is based on the assumption that a group of neighbouring MBs in a small region that belongs to the same object likely to have similar colour information. Hence, the lost MVs of the corrupted MBs can be estimated using a set of corresponded MVs of the surrounding MBs that have similar colour information. The proposed algorithm has been tested on several test video sequences. The experimental results are compared with conventional error concealment methods and higher performance is achieved in both objective peak signal-to-noise ratio (PSNR) measurements and subjective visual quality.