A variational level set method image segmentation model with application to intensity inhomogene magnetic resonance imaging

• DOI: 10.4103/digm.digm_44_17
• Publication Date: 2018-05-18T06:01:27Z
• Authors: Chun Li; Jinhe Su; Longlong Yu; Le Wang; Luo Ze
• Keywords: Regularization; Level set method; Smoothing

ABSTRACT Background and Objectives: In this article, we propose an image segmentation model based on Chan-Vese (CV) for segmentation. By taking into account the local features of image, new proposed can successfully segment images with intensity nonuniformity. Materials Methods: We quantitatively compare our method other two state-of-the-art algorithms, namely, CV binary fitting (LBF) in segmenting synthetic MR ground truth from BrainWeb; data be available at: https://www.mni/mcgill.ca/brainweb/. For missing weak boundaries, to deal inhomogeneity, LBF model, introduced convex total variation regularization term, explicit smoothing level set function ø. The evolution equation will solved through calculus variations. Results: experimental processing, use some real magnetic resonance imaging brain as images, validate stabilization algorithm. results comprehensive sincerity show outstanding reference availability. Conclusions: a information introduce functional is keep smooth. Finally, various low-contrast showing which powerful type including that would difficult gradient-based methods. addition, advantages are better than model. Our effectively image.