Global motion estimation from randomly selected motion vector groups and GM/LM based applications

Fast global motion estimation has been paid much attention in video compression and analysis. In this paper, a global motion estimation method is proposed by randomly selected motion vector groups in the compression domain directly. It is carried out by refining the centroid of the global motion parameters corresponding to the motion vector groups. Simulation results on different global motions show its effectiveness and robustness against noise and motion vector loss. Finally, two applications, namely the text occluded region recovery and the error concealment, are presented using the global motion/local motion information. Experimental results show the effectiveness of the proposed method. This work is supported in part by China National Natural Science Foundation (CNSF) under Project No.60572045, the Ministry of Education of China Ph. D. Program Foundation under Project No.20050698033, and by a Cooperation Project (2005.7- 2007.7) with Microsoft Research Asia.