基于水平集方法的显微细胞图像分割

基于水平集方法的显微细胞图像分割

罗宏文¹, 马驷良¹, 徐中宇¹,²

1. 吉林大学数学研究所, 长春 130012； 2. 长春工业大学计算机科学与工程学院, 长春 130012

收稿日期:2006-02-14 修回日期:1900-01-01 出版日期:2006-09-26 发布日期:2006-09-26
通讯作者: 马驷良

Segmentation of Microscopic Cell Image Based on Level Set Methods

LUO Hongwen¹, MA Siliang¹, XU Zhongyu¹,²

1. Institute of Mathematics, Jilin University, Changchun 130012, China; 2. College of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China

Received:2006-02-14 Revised:1900-01-01 Online:2006-09-26 Published:2006-09-26
Contact: MA Siliang

摘要/Abstract

摘要： 利用变分法推导了图像边缘检测Snake模型对应的Euler方程, 证明了水平集方法3种离散格式(向前差分、向后差分、中心差分)的不稳定性, 采用迎风差分格式构造水平集算法, 对显微细胞图像进行边缘检测, 实验结果表明, 该算法是稳定的和收敛的.

关键词: 边缘检测与分割, 水平集, 迎风差分格式, 显微细胞图像

Abstract: Two methods of edge detection and segmentation based on partial differential equation (PDE) are introduced, they are Snake Model and Level Set Methods. We worked out Euler equation corresponding to Snake using variational method and proved nonstability of the three discrete schemes which are forward differential, backward differential and center differential. We detectd the edge of microscopic cell image by constructing algorithms with upwind differential scheme. The results suggest it is stable and convergent.

Key words: edge detection and segmentation, level set, upwind differential scheme, microscopic cell image

中图分类号:

TP391.41

罗宏文, 马驷良, 徐中宇,. 基于水平集方法的显微细胞图像分割[J]. J4, 2006, 44(05): 760-763.

LUO Hongwen, MA Siliang, XU Zhongyu,. Segmentation of Microscopic Cell Image Based on Level Set Methods[J]. J4, 2006, 44(05): 760-763.

[1]	徐文杰, 王昕. 融合KFCM与改进DRLSE模型的甲状腺结节图像分割[J]. 吉林大学学报(理学版), 2016, 54(05): 1123-1130.
[2]	罗岑弘. 基于梯度向量流的活动轮廓模型[J]. 吉林大学学报(理学版), 2016, 54(05): 1103-1108.
[3]	黄龙光, 储理才. 锥拟凸向量优化问题解的稳定性及广义适定性[J]. 吉林大学学报(理学版), 2016, 54(03): 461-469.
[4]	柴振华, 罗宏文, 苗闯, 马驷良. 一种保持拓扑结构的隐式活动轮廓图像分割方法[J]. J4, 2009, 47(05): 981-986.
[5]	李松涛, 张旭利, 杨荣, 黄明游. JohnsonSegalman模型非定常解的估计与数值计算[J]. J4, 2008, 46(03): 389-392.