MEDICAL IMAGES SEGMENTATION USING ACTIVE CONTOURS DRIVEN BY GLOBAL AND LOCAL IMAGE FITTING ENERGY

摘要

In this paper, we propose a global and local Chan-Vese model which utilizes both global image information and local image information for image segmentation. We define an energy functional with a global term, which incorporates global image information to improve the robustness of the proposed method, and a local term which is dominant near the object boundaries. The regularization term is added to the energy functional to avoid the time-consuming re-initialization. The comparisons with the C-V model, LBF model and LGIF model show that our model can segment images with intensity inhomogeneity in less iteration steps and take less time.