Adaptive temporal error concealment method based on the MB behavior estimation in the video

摘要

In this paper, an adaptive error concealment method is proposed to recover the motion vectors of degraded macroblocks (MBs) in a frame, based on analyzing the behavior of the MBs in the frames. In this method, the behavior of the degraded MB is first estimated considering the obtained information from motion vectors of neighboring MBs in the current frame and collocated MBs in the previous frames. Then, an appropriate algorithm is selected for recovering the motion vector of degraded MB using the adaptive algorithm. Results of simulation performed on various video sequences show that in comparison with the outer boundary matching algorithm (OBMA), the proposed method can reach the maximum frame performance and average frame performance of 2.4dB and 0.2dB per video sequence, respectively.