A fast active contour algorithm for object tracking in complex background

Publisher Summary 2D object tracking is a hit research topic in dynamic scene analysis. There are different methods of 2D object tracking: image region based tracking algorithms, feature point based tracking algorithms, and line segment based tracking algorithms. In general, these methods require an explicit definition of a dynamic model of the moving objects. Many objects cannot be described by simple geometric shapes but need to be represented with complex contours. Active contour is a powerful tool to object tracking. However, the existing models are only applicable to track on simple image. Based on the idea of the original greedy algorithm, a fast greedy tracking algorithm is presented to face the problem of tracking on complex real image. Algorithms on tracking complex shape object are also demonstrated on complex background.