A Generalized Level Set Formulation of the Mumford- Shah Functional for Brain MR Image Segm.pdfVIP

A Generalized Level Set Formulation of the Mumford- Shah Functional for Brain MR Image Segmentation Lishui Cheng1, Jie Yang1, Xian Fan1 and Yuemin Zhu2 1 Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University (SJTU), Shanghai, 200030, P.R.China {lishuicheng,jieyang,Yivanne@sjtu.edu.cn} 2 CREATIS, CNRS UMR 5515 INSERM Unit 630, 69621 Villeurbanne cedex,France Yue-Min.Zhu@creatis.insa-lyon.fr Abstract. Brain MR image segmentation is an important research topic in medical image analysis area. In this paper, we propose an active contour model for brain MR image segmentation, based on a generalized level set formulation of the Mumford-Shah functional. The model embeds explicitly gradient infor- mation into the Mumford-Shah functional, and incorporates in a generic frame- work both regional and gradient information into segmentation process simul- taneously. The proposed method has been evaluated on real brain MR images and the obtained results have shown the desirable segmentation performance. 1. Introduction Segmentation of anatomical structures is of utmost significance both for clinical diag- nosis and visualization purposes. Because of the huge amount of data and the com- plexity of these organs and structures, manual segmentation is extremely tedious and often unfeasible. Computer-based automatic segmentation methods are required to fully exploit medical data. Active contour model, since it was first proposed in [1], has been extensively stud- ied and used in the field of image segmentation. The central idea behind active con- tour model is to evolve a curve or surface based on energy minimization method un- der the influence of image dependent forces, regularity constraints and certain user- specified constraints. Originally, active contour models are parameterized curves or surfaces which iteratively evolve toward the desired locations according to the energy minimization criterion. The