Robust methods for background extraction in video

Abstract

In this paper a framework is presented to automaticallyextract a sequence of images of the background of a scene from a shakyfilm. That is, the input video sequence may have local and global motionbut the output video must contain exclusively the static backgroundscene. Applying robust procedures to this end is one of the main goalsof this work, since the aim is to get a procedure not only resistant tolow scale noise but to occasional high scale noise. The median is used asan estimate of the background, the median absolute deviation (MAD) isused to establish a threshold to locate foreground and M-estimation forregression is used to stabilize the video sequence.