点击次数：

论文类型：会议论文

发表时间：2016-09-25

收录刊物：EI、CPCI-S、Scopus

卷号：2016-August

页面范围：4092-4096

关键字：local trajectories; outlier remove; mesh transformation; video stabilization

摘要：Video stabilization improves video quality by removing undesirable jitter. This paper proposes a new approach for video stabilization. We directly smooth the local trajectory matrix, obtained by complete trajectories, and every frame is smoothed in a separate local trajectory matrix. which is suitable for modeling the nonlinear video stream, and more effective than conventional methods based on feature trajectories. In addition, the coarse to fine outlier detection technology is applied to guarantee that the Miler feature points belong to background scenes. Finally, to make the image composition more fluent and authentic, we exploit the robust mesh transformation to warp the input frame based on the displacement of feature points. Experiments show that our method can generate comparable results with regard to some other state-of-the-art video stabilization methods.