CT冠脉造影和血管内超声融合算法的研究及实现
文献类型：学位论文
| 作者 | 易建华 |
| 学位类别 | 工学硕士 |
| 答辩日期 | 2010-05-29 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 中国科学院自动化研究所 |
| 导师 | 何晖光 |
| 关键词 | 冠状动脉 中心线提取 血管内超声 CT冠脉造影 图像融合 coronary artery centerline extraction intravascular ultrasound coronary CTA image fusion |
| 其他题名 | Research and Implement on Fusion Method of Coronary CT Angiography and Intravascular Ultrasound |
| 学位专业 | 模式识别与智能系统 |
| 中文摘要 | 冠心病是目前导致人类死亡的首要原因之一。冠状动脉介入治疗手术是治疗冠心病最主要的手段，而CT冠脉造影和血管内超声是介入治疗手术常用的影像。CT冠脉造影是一种新的损伤性很小的血管成像技术，可以清晰的显示血管全貌，但是不能提供血管内部的信息。血管内超声是近年来诊断血管病变的一种新手段，可以清晰显示管腔形态，对血管的直径及面积进行精确测量，但血管内超声是一系列超声图像的叠加，忽略了超声导管在三维空间中的形状。融合CT冠脉造影和血管内超声，可以结合两种影像各自的优势，给医生诊断冠心病提供更多的有用信息。本文提出了一种CT冠脉造影和血管内超声融合的方法。本文的主要工作及贡献概述如下： 1.冠状动脉中心线提取算法的研究 使用了一种基于Fast Marching(快速行进)算法提取冠状动脉中心线的方法。在该方法中，首先对图像进行下采样，然后利用Fast Marching算法提取下采样图像中初始点之间的最短路径，并使用最短路径在原始图像上计算冠状动脉中心线。使用该方法降低了计算量，并且保证了冠状动脉中心线的准确性。 2.血管内超声内膜和中-外膜分割算法的研究 研究并实现了一种基于模糊聚类和活动轮廓的血管内超声内膜自动分割的方法，该方法首先利用模糊聚类提取纵截面的轮廓线，然后使用样条插值得到横截面的初始轮廓，最后使用活动轮廓方法得到血管内超声的内膜轮廓，该方法最主要的优点是不需要人工描绘初始轮廓。还研究了一种血管内超声内膜、中-外膜半自动分割的方法，该方法结合横截面和纵截面初始化轮廓，减少了人工的交互。 3.CT冠脉造影和血管内超声融合的初步研究 提出了一种基于冠状动脉中心线的CT冠脉造影和血管内超声的融合方法。在该方法中，使用环形最大强度投影算法把CT冠脉造影和血管内超声投影到二维空间中，并使用投影图像进行配准，降低了配准计算的复杂度。 4.CT冠脉造影和血管内超声融合原型系统的开发工作 在上述研究的基础上，基于Medical Imaging Interaction Toolkit(MITK)开发了CT冠脉造影和血管内超声融合的原型系统。该系统具有冠状动脉中心线提取、血管内超声内膜和中-外膜半自动分割及CT冠脉造影和血管内超声融合等功能。该系统能够为医生诊断冠心病提供更多的信息。 |
| 英文摘要 | Coronary heart disease is one of the leading causes of death to humans. Percutaneous coronary intervention is the most primary treatment, with coronary computed tomography angiography (CTA) and intravascular ultrasound (IVUS) commonly used in it. Coronary CTA is a new minimal invasive vascular imaging technique, which can clearly exhibit a three-dimensional impression of the coronary arteries. However, the information inside the vessel cannot be provided by coronary CTA. In recent years, IVUS has been applied clinically to the diagnosis of vascular lesions as a new diagnostic tool. It can show lumen and vessel wall, and can provide an accurate measurement of the diameter and area of vessel. However, it simply performs a straight stacking of adjacent frames, and completely neglects the vessel curvature. If we combine IVUS and coronary CTA together, the advantages of both images can be exploited and more useful information on diagnosing coronary artery disease can be provided. In this paper, a fusion method of coronary CTA and IVUS is proposed. The main works and contributions of this article are as follows: 1. Research on the extraction of the coronary artery centerline A method based on Fast Marching is used to extract coronary artery centerline. In our method, the origin image is first down-sampled; and Fast Marching algorithm is then used to extract the minimal path of the initial points in down-sampled image; finally, the centerline in the original image is calculated using minimal path. This method can reduce computational complexity, and ensure the accuracy of the centerline of coronary artery. 2. Research on the segmentation of luminal and media-adventitia contour in IVUS We investigate and implement a fully automatic algorithm to segment luminal contour in IVUS, which is based on fuzzy clustering and active contour. In our method, firstly, fuzzy clustering is employed to extract the luminal contour in the longitudinal image; secondly, the initial contour is calculated by spline interpolation; thirdly, active contour is used to get the final luminal contour. We also research a semi-automatic algorithm to segment luminal contour and medial-adventitial contour in IVUS. In our algorithm, we combine the longitudinal and transversal image to obtain initial contour, which greatly reduces the user interaction 3. Preliminary research on fusion of coronary CTA and IVUS We propose a fusion method of coronary CTA and IVUS based on coronary art... |
| 语种 | 中文 |
| 公开日期 | 2015-09-08 |
| 其他标识符 | 200728014628012 |
| 源URL | [http://ir.ia.ac.cn/handle/173211/7533]  |
| 专题 | 毕业生_硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 易建华. CT冠脉造影和血管内超声融合算法的研究及实现[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2010. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。