Moving Object Tracking Based On B-Snake

Moving object detecting and tracking are the main topics of computer vision and image coding, they concern many subjects such as image process, pattern recognition, auto control, application mathematic, AI and so on and have many applications in robotic navigate, intelligent monitor system, medical image analysis, video image compression and transmission and so on.Object tracking based on Snake model(Active Contour model) is the hot point in computer vision in recent years. Because Snake model can blend the prior knowledge of the problem to be solved and the image feature, it has a high practical value. The methods of tracking based on Snake model can be classified to two classes: method based on energy minimization and method based on physical kinetics function. This paper focus on these two methods.Firstly knowledge of mathematics and physics required by Snake model is introduced in this paper, following by optical flow, then is the Snake models used in this paper such as B-Snake model, GVF Snake model, Velocity Snake model. The main contribution of this paper is: for Velocity model, the contour is presented by B-spline curve and a new B-Snake model is induced by combined with Lagrangian kinetics function;for method based on energy minimization, the initial contour of the object of very frames is presented by B-spline curve and predicted by optical flow and converged to the boundary of object by GVF. Experimental results show that the proposed algorithm is better than the B-snake model in tracking.