Object Tracking based on Snake and Sequential Monte Carlo Method

Snake has found a number of applications in recent years in computer vision. A snake is an elastic curve, which dynamically adjusts its initial position to the object shape. Snake is sensitive to parameters values and initialization, and moreover, it is a popular method for object contour localization while not suitable for state estimation in time series. This paper presents a method to extend snake for object contour tracking. Sequential Monte Carlo method is used to improve the performance of initialization. Experiments demonstrate that our method leads to considerable improvement over Sequential Monte Carlo method and snake