Foveation based error resilience for wireless video communication

Based on the observation that foveation analysis can be used to find most sensitive content of human eyes in video, one effective error resilience method is proposed for robust transmission of H.264 inter-coded frames in wireless channel. It firstly exploits the results of automatic foveation selection to find foveated area in inter-coded frames, and then considers the results of pre-error concealment effect analysis to search for error resilient alignment order of MB used in video encoder. Extensive experimental results on different portrait video sequences over random bit error wireless channel demonstrate that this proposed method can achieve better subjective and objective effect than original JM 8.2 H.264 video codec and H.264 its own error resilient tool.