ABSTRACT
In Section 2 we briefly review the total variation restoration model and some numerical methods for the solution of the Euler-Lagrange equations. We also discuss modular solvers for the solution of these equations when constraints are present. We will next present two types of preconditioning techniques for the iterative solution of the linearized Euler-Lagrange equations, one based on fast transforms and the other on multigrid techniques. In Section 3, we extend the total variation restoration technique to vector images (including color), and discuss techniques to lessen the tendency of TV to over-sharpening smooth images. Finally, in Sections 4 and 5 we present some techniques for image segmentation and active contours based on the Mumford-Shah model.
