Background Extraction by Fast Independent Component Analysis in Video Surveillance
Rui Feng Ding, Jin Tang, Wang Wei · Advanced materials research · 2014
To solve dynamic background extraction in complicated outdoor surveillance, a method of background extraction based on fast independent component analysis (FastICA) is presented. Since foreground regions and background in an image are considered to be independent, and background images in video show a high correlation coefficient, the method can directly recover the background signal without recover other source signals. In this paper, the principle of FastICA are introduced, and the detailed processes of the method and results are given, which show that the method can realize extracting background image .