We approximate by discrete Г-convergence a functional proposed by Mumford-Shah for a variational approach to image segmentation. Such a functional is first relaxed with a sequence of nonconvex functionals, which in turn, are dis-cretized by piecewise linear finite elements. Under a suitable relation between the relaxation parameter ∊and the meshsize h, the convergence of the discrete functionals and the compactness of any sequence of discrete minimizers are proved. The proof relies on the techniques of Г-convergence and on the properties of the Lagrange interpolation and Clement operators.
No takes yet. Share an insight, caveat, or question.
Bellettini et al. (1994) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: