A Review and Comparison on Video Stabilization Alorithms

Abstract

In recent years mobile phones and hand-held video cameras are gaining increasing popularity. They allow people to easily film videos but also bring unwanted camera shakes and jitters that affect the video quality. Although mechanical devices, optical devices, and electronic devices can help remove unwanted shakes, those methods are usually expensive and impractical for mobile phones and hand-held cameras. Whereas digital video stabilization techniques only require raw footage maybe plus the gyro data. In this paper, we focus on and compare the effects of different approaches to motion estimation and motion compensation, the two crucial parts of video stabilization algorithms that have a large impact on the quality of video stabilization. Based on that, we conclude that the future of video stabilization lies within using a gyroscope to get accurate camera motion and using neural networks to achieve amazing video quality.