基于C-V模型的医学图像血管钙化分割算法
|
摘要
为了准确分割出医学图像中血管的钙化点,设计并实现了一种基于C-V模型的水平集图像分割方法。首先进行去噪和对比度增强预处理,接下来分割出图像中感兴趣的血管和钙化点区域,然后利用C-V模型水平集分割方法分割血管壁上的钙化点目标,最后采用形态学方法消除分割结果中孤立的噪声和孔洞。针对大量的临床血管钙化图像进行了算法的测试,实验结果表明:能有效分割出血管中的钙化灶,准确检测出血管中钙化的位置、大小、形态等。将C-V模型分割方法与OTSU阈值分割、登山法分割方法进行比较,结果表明C-V模型分割方法对于钙化点的分割更准确,边缘更平滑,更清晰,方便对钙化点进行进一步的测量和诊断。
In order to accurately segment the calcification of the vascular in medical images,a level set segmentation method based on C-V model is designed. Firstly,the image preprocessing is carried out,which including removing the image noise and enhancing the image contrast. Next,the interested regions including the vascular and calcification are segmented. And then,the level set segmentation method based on the C-V model is adopt to segment the calcified points which adhere to the wall of the vascular. Finally,the morphological operation is used to eliminate the noise and hole. A large number of clinical vascular calcification images were tested,and the experimental results clearly indicated that the method in this paper can effectively segments the calcification,and detected the location,size,shape,etc.. At the same time,the C-V model was compared with OTSU threshold segmentation and hill climbing method.,and the contrast results showed that the C-V model is more accurate,the edge is smoother and more clear,and is convenient for further measurement and diagnosis of the calcification.
In order to accurately segment the calcification of the vascular in medical images,a level set segmentation method based on C-V model is designed. Firstly,the image preprocessing is carried out,which including removing the image noise and enhancing the image contrast. Next,the interested regions including the vascular and calcification are segmented. And then,the level set segmentation method based on the C-V model is adopt to segment the calcified points which adhere to the wall of the vascular. Finally,the morphological operation is used to eliminate the noise and hole. A large number of clinical vascular calcification images were tested,and the experimental results clearly indicated that the method in this paper can effectively segments the calcification,and detected the location,size,shape,etc.. At the same time,the C-V model was compared with OTSU threshold segmentation and hill climbing method.,and the contrast results showed that the C-V model is more accurate,the edge is smoother and more clear,and is convenient for further measurement and diagnosis of the calcification.
引文
[1]唐朝枢,齐永芬.关注血管钙化发病新机制的研究[J].中国医学前沿杂志(电子版),2010,2(3):5-8.
[2]齐永芬.关注血管钙化的基础和临床研究[J].中国动脉硬化杂志,2015,23(5):433-436.
[3]田飞,杨丰,刘国庆.一种适用于血管图像分割的活动轮廓模型[J].生物医学工程学杂志,2010,27(5):968-973.
[4]邢栋,杨丰.结合硬斑块特征的心血管内超声图像中-外膜边缘检测[J].中国生物医学工程学报,2012,31(1):25-31.
[5]国凯.CT影像中主动脉血管钙化的分割与重建[D].青岛:青岛大学,2013.
[6]张石,董建威,佘黎煌.医学图像分割算法的评价方法[J].中国图象图形学报,2009,14(9):1872-1880.
[7]江贵平,秦文健,周寿军,等.医学图像分割及其发展现状[J].计算机学报,2015,38(6):1222-1237.
[8]王志东,汪友生,董路,等.血管内超声图像初始轮廓提取算法的研究[J].计算机测量与控制,2014,22(5):1490-1492.
[9]计玉芳,医学图像分割算法的研究与应用[D].无锡:江南大学,2009.
[10]Jing HU,Xiaofeng MA,Cheng WX,et al.Scene classification based details preserving histogram equalization[J].Journal of Computer Applications,2014,34(7):2001-2004.
[11]方明,李洪娜,雷立宏,等.低照度视频图像增强算法综述[J].长春理工大学学报:自然科学版,2016,39(3):56-64.
[12]Shen M,Weng J,Zhang Q,et al.An improved C-V level set model[C].Electronics,Computer and Applications,IEEE,2014:894-897.
[13]Acho SN,Rae WI.Dependence of shape-based descriptors and mass segmentation areas on initial contour placement using the chan-vese method on digital mammograms[J].Computational and Mathematical Method in Medicine,2015:349874.
[14]职占江.基于Chan-Vese模型的图像分割算法[D].郑州:河南大学,2008.
[15]刘军伟.基于水平集的图像分割方法研究及其在医学图像中的应用[D].合肥:中国科技大学,2009.
[16]高杨.基于超声图像的颈动脉内中膜厚度的测定[D].沈阳:东北大学,2012.
[17]危自福,毕笃彦,马时平.一种改进的快速C-V水平集红外图像分割[J].数据采集与处理,2010,25(3):347-352.
[2]齐永芬.关注血管钙化的基础和临床研究[J].中国动脉硬化杂志,2015,23(5):433-436.
[3]田飞,杨丰,刘国庆.一种适用于血管图像分割的活动轮廓模型[J].生物医学工程学杂志,2010,27(5):968-973.
[4]邢栋,杨丰.结合硬斑块特征的心血管内超声图像中-外膜边缘检测[J].中国生物医学工程学报,2012,31(1):25-31.
[5]国凯.CT影像中主动脉血管钙化的分割与重建[D].青岛:青岛大学,2013.
[6]张石,董建威,佘黎煌.医学图像分割算法的评价方法[J].中国图象图形学报,2009,14(9):1872-1880.
[7]江贵平,秦文健,周寿军,等.医学图像分割及其发展现状[J].计算机学报,2015,38(6):1222-1237.
[8]王志东,汪友生,董路,等.血管内超声图像初始轮廓提取算法的研究[J].计算机测量与控制,2014,22(5):1490-1492.
[9]计玉芳,医学图像分割算法的研究与应用[D].无锡:江南大学,2009.
[10]Jing HU,Xiaofeng MA,Cheng WX,et al.Scene classification based details preserving histogram equalization[J].Journal of Computer Applications,2014,34(7):2001-2004.
[11]方明,李洪娜,雷立宏,等.低照度视频图像增强算法综述[J].长春理工大学学报:自然科学版,2016,39(3):56-64.
[12]Shen M,Weng J,Zhang Q,et al.An improved C-V level set model[C].Electronics,Computer and Applications,IEEE,2014:894-897.
[13]Acho SN,Rae WI.Dependence of shape-based descriptors and mass segmentation areas on initial contour placement using the chan-vese method on digital mammograms[J].Computational and Mathematical Method in Medicine,2015:349874.
[14]职占江.基于Chan-Vese模型的图像分割算法[D].郑州:河南大学,2008.
[15]刘军伟.基于水平集的图像分割方法研究及其在医学图像中的应用[D].合肥:中国科技大学,2009.
[16]高杨.基于超声图像的颈动脉内中膜厚度的测定[D].沈阳:东北大学,2012.
[17]危自福,毕笃彦,马时平.一种改进的快速C-V水平集红外图像分割[J].数据采集与处理,2010,25(3):347-352.