A novel active contour model for medical image segmentation |
| |
Authors: | Zeng-liang Fu Ming Ye Yong-lin Su Yan-ping Lin Cheng-tao Wang |
| |
Institution: | Institute of Biomedical Manufacturing and Life Quality Engineering, Shanghai Jiaotong University, Shanghai 200240, China |
| |
Abstract: | 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. |
| |
Keywords: | |
本文献已被 CNKI 维普 SpringerLink 等数据库收录! |
|