Active Contour Model Based on Genetic Algorithm
LIU Zhi-jian (Department of Automatic Control, National University of Defence Technology, Changsha 410073)
Active Contour Model introduced by Kass et al is a energy-minimizing curve in essential. It is a new method of image object extraction based on top-down mechanism, which makes use of high level information to improve the speed and veracity of object extraction. It has been used more and more widely in applications of image analysis and computer vision. The original algorithm of active contour model involves four steps: setting up a variational integral on the continuous, deriving a pair of Euler equations, discretizing them, and solving the discrete equations. This algorithm suffers a number of problems. In this paper, we will firstly discuss the original algorithm and some improved algorithms of active contour model, then propose a algorithm based on the genetic algorithm and present the experiment result. The result proves that genetic algorithm settles the problem of original model that run into the local least value end enhance the success ratio of the object extraction.