ABSTRACT
This chapter focuses on image segmentation and points out the possible relevance of both models in other fields of image processing. It shows some numerical tests where an X-ray image is segmented, and compares several active contour models on the same image. The level set (LS) method, proposed in the late 80s, has proven a very successful technique for the analysis of front propagation problems and has permitted the treatment of many different physical and conceptual models within the same theoretical framework. The major asset of LS methods is their capability to handle the onset of singularities and topological changes in the front. This review is focused on Finite Difference and Semi-Lagrangian numerical techniques for LS models and will introduce the most basic concepts of LS methods. A regularizing curvature term could be used to smooth out the active contour in case of noisy images.
