Semi-Implicit Additive Operator Splitting Scheme for Image Segmentation Using the Chan-Vese Model

Abstract

Active contour models are designed to evolve an initial curve, called level set, to extract the desired object(s) in an image. Most approaches are based on semi-implicit schemes which are stable for all time steps. Various models are used for the global segmentation such as Chan-Vese (CV) model. The CV model has the global segmentation property to segment all objects in an image. The problem with this model is the high time computing. In order to reduce it, our contribution in this work is the association of a semi-implicit Additive Operator Splitting (AOS) technique with the CV model in biphase and multiphase cases. The basic idea behind AOS schemes is to decompose a multi-dimensional problem into one-dimensional ones that can be solved very efficiently. In this paper, we present the new association in biphase and multiphase cases with simulations showing the efficiency of the proposed method.

Keywords

• Image segmentation
• Active contours
• Chan Vese
• AOS scheme
• Level set.