Stereoscopic Error Concealment Based on Motion Characteristic Analysis

Traditional video error concealment techniques perform poorly in handling loss MacroBlocks(MBs) with intense or complicated motion. Based on the statistic analysis of stereoscopic video coding prediction mode, this paper classifies the loss MBs into different categories according to the motion and disparity information of their neighboring correctly received or concealed MBs, and uses different methods to conceal different kinds of MBs. Therefor the recovery of loss MBs is more purposeful and efficient. Simulation results show that the proposed method can improve the concealment quality of intense or complicated motion region.