Video stabilization seeks to create stable versions of casually shot video, which can not only improve people’s visual comfort but also be a preprocessing of some other procedures, such as object tracking, object detection, video compression and so on, resulting in increas- ed precision and robustness. The project focuses on improving the efficiency of video stabilization algorithms as well as developing bet- ter deshaking assessment tools to identify shaking and evaluate the exsiting algorithms.
We have proposed a video stabilization algorithm based on L1-L2 optimization, which can remove unwanted camera movements as well as keep the original video information to the greatest extent, and a video stabilization algorithm based on attitude sensors, which can stabilize shaky videos using gyroscope sensors. We also designed a shaking video synthesis method for performance evaluation. We have applied our stabilization algorithm into cloud-based stabilization, in which the stabilization process is on the server, and users only need to upload their shaky videos and download the stabilized ones after a while.
Demo for video stabilization
Software available for cloud based stabilization