Video stabilization using SIFT-ME features and EM fuzzy clustering

Video stabilization is the process of removing unwanted motion from a video. In this process, the desired motion is maintaining. In this paper we use a digital video stabilization process using SIFT_ME features. Then we use EM algorithm to clustering SIFT features. For separating global motion from local motion, we use fuzzy clustering. In next step, orientation and transition variations between the previous frame and the current frame is calculated. For estimate the desired motion, kalman filtering is used. Finally, undesired motion is compensated. Experimental results show, qualitative analysis and peak signal-to-noise ratio of stabilized video is more than primary video.

Keywords:

SIFT-ME , Motion Estimation , EM-fuzzy clustering , kalman filtering , video stabilization