基于超像素方法的腹部CT影像多目标器官分割研究
|
摘要
针对医学影像中各个器官间的区域划分不明显,影像噪声较大等问题,本文提出了一种通过构建超像素从腹部医学影像中自动分割多个目标器官的方法。基于超像素的分割方法适应了CT图像中的各种成像条件,并且考虑了多个器官之间的相互联系与制约关系。该方法首先根据像素相关性和位置邻近性对超像素进行聚类,然后再引入器官空间结构分布图,修正多个器官的分割。实验结果表明,该分割方法能有效地完成CT影像中的各个器官分割。
In order to address the typical problems in the medical image, such as unclear boundaries between organs and loud imaging noises, we proposed a method of automatic segmentation to get the target organs' images from abdominal medical images by building super-pixels. The super-pixel based segmental method adapted to various imaging conditions in the CT image and considered the interrelationship and constraints among multiple organs. In this method, we firstly clustered the super-pixels in the light of pixel correlation and location adjacency, then the spatial distribution of organs was used to modify the segmentation process of multiple organs. The experimental results showed that this proposed method could effectively segment the organs in the abdominal CT images compared with some other traditional segmental algorithms.
In order to address the typical problems in the medical image, such as unclear boundaries between organs and loud imaging noises, we proposed a method of automatic segmentation to get the target organs' images from abdominal medical images by building super-pixels. The super-pixel based segmental method adapted to various imaging conditions in the CT image and considered the interrelationship and constraints among multiple organs. In this method, we firstly clustered the super-pixels in the light of pixel correlation and location adjacency, then the spatial distribution of organs was used to modify the segmentation process of multiple organs. The experimental results showed that this proposed method could effectively segment the organs in the abdominal CT images compared with some other traditional segmental algorithms.
引文
[1]李强.医学图像分割进展[J].中国医疗设备,2010,25(5):121-124.
[2]Kamdi S,Krishna RK.Image segmentation and region growing algorithm[J].Inter J Comput Tech Elec Engin,2012,1(2).
[3]徐蔚波,刘颖,章浩伟.基于区域生长的图像分割研究进展[J].北京生物医学工程,2017,36(3):317-322.
[4]彭微.基于区域生长算法的肝脏CT图像分割[J].医疗卫生装备,2014,35(9):49-50.
[5]冯皓.图像区域生长算法的稳定性研究[D].福州:福建师范大学,2012.
[6]Bezdek JC,Hall LO,Clarke LP.Review of MR image segmentation techniques using pattern recognition[J].Med Phys,1993,20(4):1033.
[7]何瀚志,朱红,王竞.基于聚类分析的医学图像分割综述[J].中国科技信息,2017,(15):43-46.
[8]加小红.聚类算法及其在图像分割中的应用研究[D].兰州:兰州交通大学,2017.
[9]Metaxas DN.Physics-based deformable models:applications to computer vision,graphics,and medical imaging[J].Graph Med Imag,1997.
[10]宋熙煜,周利莉,李中国,等.图像分割中的超像素方法研究综述[J].中国图象图形学报,2015,20(5):599-608.
[11]韩文艳.CT能谱成像的基本原理与临床应用优势[J].中国医疗设备,2015,30(12):90-91.
[12]梁启堂,钟向阳.心脏冠状动脉多排CT扫描的优化措施[J].中国医疗设备,2011,26(8):161-162.
[13]郭书杰,吴小欣,黄杰.改进的聚类分析算法及其性能分析[J].计算机时代,2010,(8):4-6.
[14]Bezdek JC,Ehrlich R,Full W.FCM:The fuzzy c-means clustering algorithm[J].Comput Geosci,1984,10(2):191-203.
[15]贺晓建,王福明.基于灰度共生矩阵的纹理分析方法研究[J].山西电子技术,2010,(4):89-90.
[16]张云彬,张永生.基于图像纹理特征的目标快速检索[J].高技术通讯,2004,14(8):11-14.
[17]张树伟.医学图像的高斯混合模型及聚类研究[D].镇江:江苏大学,2010.
[18]胡兵.改进的Otsu算法在多尺度图像分割中的应用研究[D].长沙:中南大学,2012.
[19]Liu X,Linguraru MG,Yao J,et al.Abdominal multi-organ localization on contrast-enhanced CT based on maximum a posteriori probability and minimum volume overlap[A].IEEE International Symposium on Biomedical Imaging:From Nano To Macro[C].IEEE,2011:2083-2086.
[20]Selver MA,Kocao?lu A,Akyar H,et al.Patient oriented neural networks to overcome challenges of abdominal organ segmentation in CT angiography studies[A].International Conference on Electrical and Electronics Engineering[C].New York:IEEE,2009:177-181.
[2]Kamdi S,Krishna RK.Image segmentation and region growing algorithm[J].Inter J Comput Tech Elec Engin,2012,1(2).
[3]徐蔚波,刘颖,章浩伟.基于区域生长的图像分割研究进展[J].北京生物医学工程,2017,36(3):317-322.
[4]彭微.基于区域生长算法的肝脏CT图像分割[J].医疗卫生装备,2014,35(9):49-50.
[5]冯皓.图像区域生长算法的稳定性研究[D].福州:福建师范大学,2012.
[6]Bezdek JC,Hall LO,Clarke LP.Review of MR image segmentation techniques using pattern recognition[J].Med Phys,1993,20(4):1033.
[7]何瀚志,朱红,王竞.基于聚类分析的医学图像分割综述[J].中国科技信息,2017,(15):43-46.
[8]加小红.聚类算法及其在图像分割中的应用研究[D].兰州:兰州交通大学,2017.
[9]Metaxas DN.Physics-based deformable models:applications to computer vision,graphics,and medical imaging[J].Graph Med Imag,1997.
[10]宋熙煜,周利莉,李中国,等.图像分割中的超像素方法研究综述[J].中国图象图形学报,2015,20(5):599-608.
[11]韩文艳.CT能谱成像的基本原理与临床应用优势[J].中国医疗设备,2015,30(12):90-91.
[12]梁启堂,钟向阳.心脏冠状动脉多排CT扫描的优化措施[J].中国医疗设备,2011,26(8):161-162.
[13]郭书杰,吴小欣,黄杰.改进的聚类分析算法及其性能分析[J].计算机时代,2010,(8):4-6.
[14]Bezdek JC,Ehrlich R,Full W.FCM:The fuzzy c-means clustering algorithm[J].Comput Geosci,1984,10(2):191-203.
[15]贺晓建,王福明.基于灰度共生矩阵的纹理分析方法研究[J].山西电子技术,2010,(4):89-90.
[16]张云彬,张永生.基于图像纹理特征的目标快速检索[J].高技术通讯,2004,14(8):11-14.
[17]张树伟.医学图像的高斯混合模型及聚类研究[D].镇江:江苏大学,2010.
[18]胡兵.改进的Otsu算法在多尺度图像分割中的应用研究[D].长沙:中南大学,2012.
[19]Liu X,Linguraru MG,Yao J,et al.Abdominal multi-organ localization on contrast-enhanced CT based on maximum a posteriori probability and minimum volume overlap[A].IEEE International Symposium on Biomedical Imaging:From Nano To Macro[C].IEEE,2011:2083-2086.
[20]Selver MA,Kocao?lu A,Akyar H,et al.Patient oriented neural networks to overcome challenges of abdominal organ segmentation in CT angiography studies[A].International Conference on Electrical and Electronics Engineering[C].New York:IEEE,2009:177-181.