摘要
针对当前的图谱分割方法,在一定能高程度上缩短了计算时间,但是对于要求较高的应用软件来说,在精确度和平滑度上不能得到理想的分割结果。提出了一种最优模板选择和水平集的图谱分割算法。方法一方面使用正规化的交互信息获得"最优"模板。另一方面在图谱分割中提出基于水平集图谱分割方法,在基于光强度的配准算法的类型中引入附加约束条件,对约束条件可提高轮廓的平滑度,同时使引入的先验信息(如亮度分布或被分割物体可采纳的形状)更加局部化。实验结果表明,方法比传统的图谱分割方法有更高配准的精确度和平滑度。
Current atlas - based segmentation methods can decrease the computation time. However this is not yet sufficient to guarantee the desired quality of segmentation for the most demanding applications. A method of optimum template selection and level set atlas - based segmentation is proposed. On the one hand, the method automatically chooses the ＇ best＇ template based on normalized mutual information, and on the other hand, additional constraints are included in these types of intensity - based registration algorithms. These constraints should improve the smoothness of the contours while introducing more local a priori information such as the intensity distribution or the admissible shapes of objects to be segmented. The example is presented to show the superior performance of the proposed thresholding algorithm compared to that of the existing algorithms of the atlas - based segnentation.
出处
《计算机仿真》
CSCD
北大核心
2009年第3期213-216,共4页
Computer Simulation
关键词
最优模板选择
水平集
图谱
Optimum template selection
Level set
Atla
