Fractional Calculus for improving Edge-Based Active Contour Models.

Abstract

In this paper, an attention is given to edge-stop function (ESF) in active contour models which are based on a gradient calculus. Usually this kind of algorithms uses the gradient of a smoothed image by a Gaussian kernel. In this work, a fractional calculus is introduced to the edge-stop function formulation. The regular gradient in the ESF formulation has been substituted by a fractional one. The Grunwald-Letnikov definition has been used. The proposed method has been tested on MRI database. Obtained results are good enough compared to existing methods in literature.