Abstract: The automatic detection of objects that are abandoned or removed in a video scene is an interesting area of computer vision, with key applications in video surveillance. There are various methods to detect abandoned or removed objects. Tracking based approaches for detecting abandoned or removed object often become unreliable in complex surveillance videos due to occlusion, lighting changes and other factors. Therefore some authors present other methods like mixture of three Gaussian or using multiple cues like intensity, motion and shape to detect accurate static region. All of these methods are efficient to run in real time applications.

Keywords: Image Processing, abandoned or removed object detection, and visual surveillance.