期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2017
卷号：5
期号：3
页码：3865
DOI：10.15680/IJIRCCE.2017.0503030
出版社：S&S Publications
摘要：In image processing motion detection is a very important problem and using this concept a novelsaliency-based video object extraction (VOE) framework is presented for theft detection. Using motion and visualsaliency information extracted from the input video sequence, the background and foreground will be separated. Aconditional Random Field (CRF) is applied for effective integration of the saliency induced features. A video camera isused to detect the motion and capture the image and send the captured image to the owner via email. Rather than usingsignals to show the light the infuse gas like chloroform is used to make the suspect unconscious. Then the framescontinue to be saved and compared with the process iteration taking place once again.
