北京生物医学工程

基于快速推进法的血管内超声图像序列的三维分割

Three Dimensional Segmentation of Intravascular Ultrasound Image Sequence Based on Fast Marching

作者：杨宇

单位：解放军理工大学理学院（南京 211101）

分类号：

出版年·卷·期（页码）：2011·30·6（582-587）

摘要：

目的血管内超声（intravascular ultrasound，IVUS）是近年来临床诊断血管病变的一项新技术。对IVUS图像进行分割，提取出图像中血管壁的内膜和中-外膜轮廓，是IVUS图像序列定量分析和三维重建的重要步骤。方法本文提出一种基于快速推进法的IVUS图像序列三维并行分割方法。在完成对原始图像的滤波去噪、抑制导管伪影等预处理后，获取IVUS序列纵向视图，并从中提取出血管壁的内膜和中-外膜轮廓，然后将该轮廓映射到每帧IVUS横向视图中，得到各IVUS帧中血管壁的内膜和中-外膜轮廓，最后采用快速推进法对初始轮廓进行演化变形，最终提取出目标轮廓。结果对临床IVUS图像数据进行实验，与逐帧处理的串行分割方法相比，本文方法明显提高了处理效率，克服了传统串行处理方法运算效率低的缺点。结论该方法对血管疾病的临床诊治具有重要意义。

Objective Intravascular ultrasound (IVUS) is used in clinical diagnosis of vascular disease as a new technology in recent years. The detection of vascular wall from IVUS image to get the intima contour and the media-adventitia contour of vessel wall is an important step in quantitative analysis and three-dimensional reconstruction of IVUS image sequence. Methods We proposed a three dimensional parallel segmentation of IVUS image sequence based on fast marching method. After the pretreatments of original images, such as filtering, smoothing and suppression of ring-down artifacts, IVUS longitudinal cut was obtained and its detected vessel wall contours were mapped into each cross-sectional slice to obtain initial vessel wall contours in each IVUS frame. Finally, with the fast marching algorithm, initial contours deformed until stopping at the actual contours. Results The data of the experiment on clinical IVUS image showed that this method greatly improved the efficiency and overcame the shortcomings of the traditional serial processing methods, compared with the serial segmentation on frame processing. Conclusions This proposed method is of importance in clinical diagnosis and interventional treatment of coronary artery diseases.

参考文献：

［1］胡春红.感兴趣血管段最佳视角和血管内超声与冠脉造影融合研究［D］.天津，天津大学，2006.
［2］董海艳，王惠南,李虹.基于血管内超声图像序列的自动三维边缘检测［J］.南京航空航天大学学报，2007,39(4)：514-520.
［3］范幸义.计算机图形学［M］.重庆：重庆大学出版社,2008:110-120.
［4］Osher S, Paragios N. Geometric level set method in imaging vision and graphics［M］. New York:Springer Verlag,2003:25-30.
［5］杨振林，李传富，周康源，等.一种基于水平集的前列腺超声图像自动分割算法［J］.北京生物医学工程，2009,28(1):17-21.
［6］Brox T, Weickert J. Level set based image segmentation with multiple regions. Patten Recognition. Lecture Notes in Computer Seience,2004,3175:415-423.
［7］Ranchin F, Dibos F. Variational level set methods: from continuous to discrete setting, applications in video segmentation and tracking［C］. Genova:Proceedings of International Conference on Image Processing，2005:10-15.
［8］Tsai YH, Osher S. Total variation and level set based methods in image science［J］. Aetna Nmueriea,2005:1-61.
［9］Paragios N，Deriche R. Geodesic active regions and level set methods for motion estimation and tracking［J］. Computer Vision and Image Undesrtnading，2005,3:259-282.
［10］Cecil T, Marthaler D. A variational approach to path planning in three dimensions using level set methods［R］. ICES Report,2005,10:5-11.
［11］Lefohn AE, Kniss JM, Hnasen CD, et al. A streaming narrow-band algorithm: interactive computation and visualization of level sets［J］. IEEE Transactions Visualization and Computer Grpahies,2004,10(4):422-433.
［12］Osher S, Sehtian J. Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations［J］. Journal of Computational Physics,1988,79:12-49.
［13］Sehtian J. A fast marching level set method of monotonically advancing fronts［J］. Applied Mathematics,1996,93(4):1591-1595.
［14］Cardinal MR, Meunier J, Soulez G, et al. Intravascular ultrasound image segmentation: a three-dimensional fast-marching method based on gray level distributions［J］. IEEE transactions on medical imaging, 2006,25(5):590-601.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录