A novel active contour model for medical image segmentation

作者：Fu, Zeng Liang; Ye, Ming; Su, Yong Lin; Lin, Yan Ping; Wang, Cheng Tao^*

来源：Journal of Shanghai Jiaotong University (Science), 2010, 15(5): 549-555.

DOI：10.1007/s12204-010-1047-6

• Accurate computations
• Active contour model
• Active contours
• Edge indicators
• Gradient informations
• Intensity inhomogeneity
• Region-based
• Segmentation methods

A novel segmentation method for medical image with intensity inhomogeneity is introduced. In the proposed active contour model, both region and gradient information are taken into consideration. The former, i.e., region-based fitting energy, draws upon the region information and guarantees the accurate extraction of inhomogeneous image's local information. The latter, i.e., an edge indicator, weights the length penalizing term to consider the gradient constrain. Moreover, signed distance penalizing term is also added to ensure accurate computation and avoid the time-consuming re-initialization of evolving level set function. Experiments for synthetic and real images demonstrate the feasibility and superiority of the proposed model.

• 出版日期2010
• 单位上海交通大学