欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
脑核磁共振（Magnetic Resonance简称MR）图像中存在灰度不均匀现象使得传统方法很难得到理想的分割与偏移场矫正结果.针对这一问题,本文首先提出Legendre基函数拟合偏移场下的参数化互信息度量,建立脑MR图像的分割与偏移场矫正的变分模型.最后,给出了基于分裂Bregman迭代方法的快速分割与偏移场矫正算法.实验结果表明本文方法可以得到较准确的分割和偏移场矫正结果,而且具有较快的收敛速度.
英文摘要：
Due to the intensity inhomogeneous in brain MR image,it is difficult for the traditional methods to obtain accurate segmentation results.In this paper,by using the bias fitting field with Legendre basis,a new parameterize mutual information metric is firstly proposed,and a unified variational model is proposed for the optimizing of segmentation and bias correction.Finally,we present a fast algorithm based on split Bregman iteration methods.Comparative results demonstrate that our method can obtain more accurate segmentation and bias correction results with a faster convergence rate.
同期刊论文项目
同项目期刊论文
期刊信息
位置：