Geodesic Active Contour Model Algorithm Research And Application

Geodesic Active Contour Model (GACM) was originally proposed by Caselles as a novel approach for image segmentation and object exaction. The principle of GACM is the same with one of the traditional model of active contour, that is, the model searched and exacted the object contour by cooperation of the curve inner control force as well as the image force. However, the kinematics in the former doesn't include any parameters that have no relation with the curve's geometric structure, so GACM can automatically process the topological structure transformation when exacting the object contours.GACM is based on the theory of curve evolution and Level Set. And Level Set, proposed by Sethian to study curves'evolution at the curvature-related velocity, is used to depict the evolution course of curves or surfaces. Level Set is a strong robust algorithm about curves'evolution by taking the evolving curves as level set in the higher dimensions, utilizing the computing approach based on hyperbolic conversation law, and applying the similarity between of curves evolution and Hamiton-Jacobi equations. Nonetheless, the energy function of the GACM and its solving algorithm has many restrictions, such as instability of numerical computation, probability of divergence of control points. This paper discusses the original model at length, analyzes its pro and con, and finally, proposes a chain-code-based level set algorithm under the application background. This method connects the discrete points processed in the level set method to form a smooth curve, thus transforms the discrete problem to continuous one. Constraint entry is added to the curve according to the actual demand, and satisfactory result is obtained in experiments. The results show that the algorithm is effective in the actual applications.