排序方式: 共有1条查询结果,搜索用时 2 毫秒
1
1.
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. 相似文献
1