A new decoder side video stabilization using particle filter

Machine vision systems which are being extensively used for intelligent transportation applications, such as traffic monitoring and automatic navigation, suffer from image instability caused by environmental unstable conditions. On the other hand, increase in use of home video cameras and need to remove unwanted camera movements which caused by cameraman shaking hands, video stabilization algorithms are being considered. Motion estimation process is the main time consumer phase in video stabilization algorithms. In this paper, by extracting motion vectors from H.264 compressed video, motion estimation process is removed. Moreover, eliminating iterative outlier removal preprocessing and adaptive selection of motion vectors will increase the speed of algorithm. The proposed method was simulated and it was be demonstrated that the extracted motion parameters from motion vectors using particle filter can produce appropriate results for video stabilization problem.