SHAPE MODELING WITH FRONT PROPAGATION - A LEVEL SET APPROACH

Shape modeling is an important constituent of computer vision as well as computer graphics research. Shape models aid the tasks of object re presentation and recognition. This paper presents a new approach to sh ape modeling which retains some of the attractive features of existing methods and overcomes some of their limitations. Our techniques can b e applied to model arbitrarily complex shapes, which include shapes wi th significant protrusions, and to situations where no a priori assump tion about the object's topology is made. A single instance of our mod el, when presented with an image having more than one object of intere st, has the ability to split freely to represent each object. This met hod is based on the ideas developed by Osher and Sethian to model prop agating solid/liquid interfaces with curvature dependent speeds. The i nterface (front) is a closed, nonintersecting, hypersurface flowing al ong its gradient field with constant speed or a speed that depends on the curvature, It is moved by solving a ''Hamilton-Jacobi'' type equat ion written for a function in which the interface is a particular leve l set. A speed term synthesized from the image is used to stop the int erface in the vicinity-of object boundaries. The resulting equation of motion is solved by employing entropy-satisfying upwind finite differ ence schemes. We present a variety of ways of computing evolving front , including narrow bands, reinitializations, and different stopping cr iteria. The efficacy of the scheme is demonstrated with numerical expe riments on some synthesized images and some low contrast medical image s.