Nonlocal Mumford-Shah Model for Image Segmentation
Wenqi Lu · 2015
Mumford-Shah (MS) model is a popular model among image segmentation area.A multiphase nonlocal Mumford-Shah (NLMS) model which combines nonlocal operators and original MS model is proposed in this paper to segment multiphase images.In order to segment different patterns simultaneously, multiple region partition strategy which uses n-1 level set functions to segment n regions is adopted.Furthermore, corresponding split Bregman algorithm is designed to improve computational efficiency.Finally, numerical experiments are shown to validate the efficiency and effectiveness of the proposed model.