A NEW ACTIVE CONTOUR TRACKING ALGORITHM

Active contour models, known as "snakes", introduced by M. Kass et al. in 1987, havebeen used more and more widely in applications of image analysis and computer vision. However,contour finding and tracking algorithms available still exist some problems, for example, thealgorithm can not guarantee the convergence and some of the controlling parameters can not determinedadaptively. In this paper, a new tracking algorithm for active contours,new-snake, is proposed. Itretains the improvements of the available contour tracking methods and overcomes the most of theirdisadvantages. It adds the following new features: first, the algorithm converges quickly; second,the number of controlling points of snakes changes adaptively to guarantee the stable tracking;third,the hard constrains is introduced into the energy formula directly; fourth,the color constrainscan be added easily; finally, the influence by the changing of parameters is weakened. Experimentsresults are given to demonstrate the robustness and practicality of the approach.