血管造影图像的量化分析和应用研究
|
摘要
造影图像在临床疾病诊断和基于图像引导的计算机辅助手术治疗等领域中有很多重要的作用。针对血管造影图像的分析和应用是现代医学数字图像研究的一个重要方向。随着人类生活水平的提高和预期寿命的延长,心脑血管疾病(cardiovascular disease,CVD)成为人类的头号死因,流行病学与临床研究表明人体外围动脉阻塞增加了心脑血管病变和死亡的危险,对心脑血管疾病的早期定量诊断和风险评估对延长人类预期寿命,提高人类的生活质量,起着非常关键的作用。X射线血管造影对于诊断和治疗心血管的各种疾病具有重要的价值。本文对血管造影图像的量化分析和主要应用进行了研究。
X射线血管造影(X-Ray Angiogram,XRA)是通过造影剂显示血管的一种方法。它是一种介入式血管疾病诊断与治疗工具,其成像的基本原理是利用注入血管的造影剂来表现血管特性,并根据造影产生的差异来对血管进行分析。目前比较普遍的应用集中在脑部和心脏冠状动脉,特别是在心脏部位,其主要临床应用方向包括针对循环系统及各器官血管性病变以及肿瘤性病变供血血管的诊断,数字减影(DSA)引导下的介入手术治疗和基于三维重建图像引导的计算机辅助手术治疗等等。它的意义在于可以在给疾病定性的同时,还可以给以后外科治疗提供详细的血管解剖图像。更有意义的是对于相当一部分疾病,在通过血管造影明确诊断的同时进行介入性治疗。在这些应用的基础上,随着现代医学和计算机信息学的发展,血管造影图像的临床应用也在不断扩展和深化,基于血管造影图像精确量化分析基础上的计算机辅助诊断(CAD)和基于量化特征为主要内容的基于内容的医学血管造影图像检索(Content Based ImageRetrieval,CBIR)目前已经成为医学图像分析和应用研究的热点之一。
从造影图像中准确分割出血管树、对血管重要参数(如血管中心线及其分支的位置、长度、逐点宽度、曲率等)进行精确的量化描述,是进行可视化三维血管结构重建及心脑血管诊断和治疗的重要前提,也是进行辅助诊断和医学血管造影图像内容检索的重要保障。鉴于血管具有复杂的形态结构,以及在成像环境中存在的噪声、造影剂衰减和分布不均匀等复杂情况,它的造影图像中经常存在血管目标模糊和局部形态特异性等各种问题,造成无法对目标对象进行准确的分割和量化测量。当前的许多经典算法对此都未能具备较高的自动化性和鲁棒性。本文针对这些问题,在随机、模糊、优化等理论框架下对血管造影图像的结构特点进行了颇有成效的实验研究,提出了基于跟踪的血管提取模型,提出了血管中心线和宽度测量新方法。以此实现了一种新的血管自动分割与测量技术,较好地完成了对血管的准确分割和精确的量化描述。在此基础上从临床应用的需要出发,结合前面的研究成果,针对XRA图像的三维血管可视化重建和基于内容的医学图像检索两种主要的实际应用进行了初步探索,建立了两个具有实际价值的应用模型。主要工作包括:
(一)提出了一种基于相似性测度和似然模型的血管造影图像分割新方法。
从造影图像中准确地分割出血管树是后续三维可视化和图像检索等研究的重要基础,该工作涉及到血管分割和血管局部和全局结构的识别。经典的基于跟踪的方法主要思路是利用一个局部算子或函数作用于血管上某一点,通过构造不同的跟踪算法进行跟踪,如利用边缘连接信息引导跟踪、利用中轴线连续性跟踪、利用搜索方向和收缩范围等图像特征跟踪等。这些跟踪方法主要的不足在于由于缺少正则化局部条件和局部测度,导致血管跟踪时很容易受环境噪声和弯曲程度等的影响。Shrijver通过构造了圆周轮廓函数来解决这一问题,但是其血管相似性函数并不具备正则化的高斯分布形式,且无法满足血管转弯过大的情况,在分支点判断方面,仅通过圆周测度观察分支点的局部极值,并没有考虑到分支点处更为复杂的血管形态:如交叉、卷曲、折叠等造成的伪分支情况,因此准确性也是有限的。为提高血管段跟踪的鲁棒性,Park提出了基于血管边缘轮廓的极大似然概率函数,该函数迫使跟踪器采样是选择双边缘中心线上的点,然而Park并未给出血管节点判别情况,如血管分支。针对经典的基于跟踪的一类方法的以上不足,通过比较现有各种血管分割方法,提出了集特征提取、知识引导、相似性测度、概率跟踪于一体的新的血管结构识别算法,给出了新的血管特征提取算法。提出了一种鲁棒的最大似然血管跟踪模型,该模型建立在多种局部相似性测度和智能知识引导的基础上,能够精确估计出血管轴线并正确判断分支点位置。
(二)提出了一种测量血管中心线和宽度的新方法。
血管参数(中心线和直径)测量面临着如下挑战:(1)血管网络从外围动脉顶部到末端、从大动脉到胫骨和肋骨动脉,呈现复杂的空间形态;(2)由于造影剂的分布不均、衰减、曝光不均等复杂情况导致造影图像血管目标和背景反差极小;(3)血管直径由几十象素单位变化到4～5个象素单位,很难准确清晰地提取局部和全局血管结构;(4)部分容积效应也严重影响着估计的精度。
传统的基于CSP的方法经常由于血管轮廓不易定位、血管目标较弱、边缘提取很粗糙而失败。James Lowell等针对此提出一种高斯和高斯微分的二维轮廓模型,用来优化迭代逼近血管目标,其血管宽度的计算精度达到子象素级。Lowell的不足正像他承认的一样:不能很好地作用于血管节点(分支和重叠点)。因此,如果能够有效增强血管目标并提取出目标方向特征,则CSP方法具有很大的挖掘潜力。
鉴于上述情况,本文将提出一种鲁棒的基于截面梯度轮廓(CSP)的血管参数测量新方法,解决经典方法存在的缺陷。该方法的主要特点是:首先,利用Gabor方向滤波器增强造影图像中的血管目标、压制图像背景噪声,并获取图像中血管目标的方向场;其次,提取血管骨架,通过细化处理后,获得分布在血管区域内部的初始轴线;最后,利用边缘梯度算子和血管方向场计算并获取垂直于血管方向的血管截面轮廓,从而测量血管参数。该算法采用了统计均值和亮度控制的搜索技术,可以精确提取血管重要参数的量化描述,解决了经典方法不提供血管宽度,以及无法测量节点(交叉、分支)处血管中心线等缺陷。由于该算法还能够估计血管节点处的血管参数,因此不失为一种血管参数测量的整体解决方案。
(三)提出了一种三维血管重建的思路
目前国内针对血管可视化方面的研究主要体现在基于模型匹配的血管提取、基于分叉模型的三维重建、基于目标引导的三维重建、基于参数化弹性模型的重建、样条运动估计等等方面。可以了解到目前国内针对血管XRA图像的3D重建的研究成果主要还是停留在XRA重建和仿真阶段,与临床实践还存在一定的距离。从国外现状来看,基于X射线造影图像的血管树的三维重建技术正成为研究的热点,国外的研究主要针对临床采集的真实数据,并面向解决实际问题,其发展趋势由早期的参数化、模型匹配、先验知识引导、局部统计优化等等方法发展为近年来以统计优化建模、人工智能、模式识别等为主的方法,并更加注重采用参数法和统计法相结合的思路。
本文从DSA造影设备的成像原理出发,详细分析了两幅单平面造影图像的血管三维重建所涉及的一些关键性技术,包括DSA系统设备的标定、血管骨架的分割、特征点的识别和匹配、血管空间点的重建和血管段的拟合、血管的截面的描述与血管树的三维显示技术等等。本文还对DSA血管三维重建技术的一些难点进行了详细的分析和讨论。本文从应用的角度出发,在血管分析与量化描述的研究成果基础上,提出三维血管重建的主要思路,在该思路下将血管跟踪与多平面造影图像重建结合起来,从而可以解决全自动真三维重建的难题。在此基础上探索建立了一套有临床价值的应用模型,可以很好地为医学临床诊断和治疗提供一些可滋借鉴的方法和工具。
(四)提出了一种医学图像内容检索的应用模型
基于内容的图像检索(Content Based Image Retrieval,CBIR)是目前图像分析的一个广泛关注的热点课题,在各个应用领域都有不同的应用和研究。但是在医学领域的研究还处于起步阶段,这一方面是由于医学图像具有其自身独特的特点,如图像之间的相似度较高、空间和灰度分辨率高、信息量大等;另一方面也是由于医学图像的成像原理、解剖部位特征和临床应用需要的复杂性。当前针对医学图像内容检索的主要研究集中在特征选取算法,降低维数提高有效性以及建立判定准则等方面。
图像特征的提取与表达是基于内容的图像检索技术的基础,本文分析了通用图像检索的原理和各种算法,结合医学图像的共性特征,在通用图像检索模型基础上建立了一个多特征综合医学图像检索模型。采用分形维数FD、颜色矩、循环Moran自相关、灰度共生矩阵GLCM、Gabor纹理、空间灰度差共生矩阵SGLD等几组颜色和纹理特征组合作为特征选择和特征提取的依据,有效地降低了维数,提高了检索速度和检索准确度。该模型可以有效地集成到PACS系统中,为医学图像的管理检索和建立在PACS系统平台上的更多图像内容方面的应用奠定基础。
造影图像量化分析的结果作为血管造影类图像内容检索所需要的特征具有典型性。精确的量化结果,包括血管中心线、宽度和直径、末梢节点、分段血管长度、半径和方向、分支和交叉的位置等各种量化参数,不但可以作为建立针对血管造影这一类医学图像的更加有效的检索模型和检索应用的特征选择,以此为基础建立更高层次上的基于语义的检索模型,是更有价值和意义的研究课题,本文对此也进行了探讨。
本文针对血管造影图像的量化分析和应用研究做了有意义的尝试,通过大量的实验验证了所提出方案和算法的有效性,并从应用角度出发,尝试建立满足实际临床需要的应用模型,达到理论研究与实际应用相结合,医工结合的目的,为最终针对临床的医学图像分析处理和检索系统打下了研究基础。
X-Ray angiogram play a very important role in the fields of clinic-diagnosing and the computer aided and image guided surgery. Analysis and application for XRA is an important research direction in the modern digital medical imaging research. With the improvement of living condition and prolonged life expectancy, cardiovascular diseases (CVD) have been the number one cause of death in modem society. The early quantitative diagnosis and accurate evaluation of CVD are critical to improving quality of life and prolong life expectancy. XRA has great value for the diagnosis of cardiovascular diseases and catheter intervened surgeries. In this thesis, we propose several quantitative descriptions and applications for XRA image. The main contribution of this thesis is list as following.
X-ray angiogram (XRA) is a type of technique to show blood vessel by contrast agent. It's an interventional diagnosis and therapy facility for blood vessel disease, the basic principle of XRA imaging is to manifest blood vessel features by contrast agent injected into the blood vessel and to analyse the blood vessel through the difference generated by the angiography. At present, the common application of XRA focused on cerebral vascular and coronary artery, especially on cardiovascular, the main clinic applications including diagnosis of vascular disease in circulatory system and organs and blood supply vessel of tumors, digital subtraction angiography (DSA) guiding interventional surgical therapy, and 3D image guiding computer aided surgery, etc. The meaning is that it can qualitative describing the disease and provide detailed vascular anatomic images to surgery for later therapy. The most significant thing is that for quite a part of diseases, it can carry out simultaneous interventional therapy when definite diagnosis conclusion is made. On these basis, the clinic applications of XRA is keep extending and deepening with the rapid progress of modern medical science and computer informatics, nowadays, computer aided diagnosis (CAD) based on precise quantitative analysis of XRA and content based medical image retrieval (CBIR) based on quantitative characters of XRA have become hotspots of medical image analysis and application research.
Accuracy segmentation of blood vessel tree from XRA image and precise quantitative description of blood vessel parameters (such as centerlines, position of branches, length, width and curvature of blood vessel, etc) is the very important premise of visualized three dimensional vessel structure reconstruction and diagnosis and treatment of cardiac and cerebral vascular disease, it's also the important guarantee of assisted diagnosis and blood vessel angiogram image content retrieval. It's difficult to make precise segmentation and measurement to the object under the condition of fuzzy object or local shape difference and such problems due to the blood vessels in the angiograms take on complicated modalities, as well as complicated conditions during imaging environment such as noise, contrast agent decay and distributioninhomogeneous. The existing classic methods can't be provided with universality and robustion. In this paper, the integrated results and conclusions have been attained with the features of X-ray angiogram under the theory frameworks of stochastic, fuzzy and optimum, a tracing based blood vessel extraction model is established. A new approach to centerlines and widths measurement of blood vessels is proposed. A new full automated segmentation and measurement methods are established, which make a satisfied result of quantitative descriptions of vessel modality. Based upon these analyses, combined with clinic application needs, we made a discussion on two of XRA major practical applications: three dimension visualization reconstruction and content based medical image retrieval, two profitable application models are given for practical purpose. The main contributions are detailed as follows:
(1) A robust Bayesian tracking model is proposed in this thesis.
Accuracy segmentation of blood vessel tree from XRA image is the basis of subsequent research on three dimensional visualization and image content retrieval. Classic tracking based methods construct different tracking algorithms for tracking by means of using a local operator or a local function to a certain point in the blood vessel, the tracking algorithms including edge link information guiding tracking, centerline continuity tracking, image characteristics such as search direction and contraction range tracking, etc. The main deficiency of these methods is that the tracking process is easy to be affected by environment noise and flexural degree due to lack of regularized local conditions and local measurement. Shrijver try to solve this problem by constructing circle contour function, but the blood vessel comparability function also doesn't have regularized Gaussian distribution form, also it can not satisfied the condition that the vessel turning is too excessive, on respects of branch point judgment, it observes local extreme only by circle measurement, without considering more complicated blood vessel morphology in the branch point, eg., pseudo branches caused by crossover, curling, folding, thus the accuracy of this method is also limited. Park proposed a maximum likelihood probability function based on blood vessel edge contour, its forces the tracker choose the points in the centerline of dual edge for sampling, however, Park doesn't give discrimination conditions of blood vessel nodes such as blood vessel branches.
According to the deficiency of the classical tracking based methods, by comparing with existing vessel segmentation methods, a new structure identification algorithm as well as vessel features extraction method, with the integrative of feature extraction, intelligent guidance. Comparability measurement and probability tracing are presented to solve the problems of the classic tracking-based methods such as losing the object in tracking the vessel tree. A robust Bayesian tracking model is presented to precise estimate the vessel axis and infers the vessel structure, which is based on multi-comparability measurement and intelligent guidance.
(2) A new cross-section profile (CSP) based approach for the vessel parameters (centerlines and widths) measurement is proposed
Blood vessel parameters ( centerlines and widths) measurement faced the following challenge: (1) vessel network present complex spatial morphology from peripheral artery top to end, from large artery to tibia and rib artery; (2) The complicated conditions due to uneven distribution, decay and inhomogeneous exposure of contrast agent lead to minimum contrast between vessel target and background in the XRA image; (3) It's difficult to extract local and full vessel structure accurately and clearly due to the vessel diameter variation from several tenth pixels to 4～5 pixels; (4) Partial volume effect also seriously affect the precision of estimation.
Due to hard location problem of vessel profile, weak vessel target and rough edge extraction, classic CSP based approaches are easier to fail. James Lowel,et al. proposed a Gaussian and Gaussian differential based two dimensional profile model used to optimal iterative approximation the blood vessel target, its calculated vessel width precision reached the degree of sub-pixel level. The deficiency of Lowell's approach, just as he admitted, is that it can't well act on vessel nodes (branches and overlapped points).So CSP approach still has much potentiality if we can enhance the vessel target and extract its orientation characteristics effectively.
According to the above description, a new robust cross-section profile (CSP) based approach is proposed for the vessel parameters (centerlines and widths) measurement. The main features is: firstly, use Gabor directional filter to enhance the vessel target and suppressing background noise and also obtain the direction field of target in the XRA image; secondly, extract vessel framework ,. refining the framework then get the initial axis lines distributed in the vessel area; finally, use edge gradient operator and vessel directional field to calculate and acquire the cross section profile vertical to the vessel direction so as to measure the vessel parameters. The proposed method utilizes a statistic-average and luminance-controlled searching techniques to precise extract the vascular description of the important vessel parameters. The CSP based algorithm is better than the classic one of the center of gravity: Two serious disadvantages in the classic algorithms have been solved, i.e. the measurement of vessel widths and the complexion of blood vessel nodes were not considered in their algorithm models.
(3) Present a new idea of 3D vessel reconstruction
At present most of the researches in vessel visualization in China are mainly on several fields: model matching based vessel extraction, fraction model based three dimensional reconstructions, target guiding based three dimensional reconstruction, parametric elastic model based reconstruction, spline motion estimation, etc. Now three dimensional XRA vessel reconstruction research in China mostly stay on XRA reconstruction and simulation. On the side of foreign situation, vessel tree three dimensional reconstructions based on X-ray angiogram image is becoming the research hotspot. Foreign research now aim at real date acquired from clinic and practical issues, their development trend has changed from early methods such as parameterized, model matching, prior knowledge guiding, local statistical optimal and so on to recent developed statistical optimal modeling, artificial intelligence, model recognition methods, and pay more attention to the thought of using parametric methods combined with statistical methods.
In this thesis, starting from the imaging principle of DSA angiography equipment, we detailed analyzed the critical techniques involving in three dimensional reconstruction of two single plane XRA images, including DSA equipment calibration, vessel framework segmentation, characteristic points recognition and matching, vessel spatial points reconstruction and vessel section fitting, vessel cross section description and vessel tree three dimension display technique and so on. We also made a detailed analysis and discussion on the difficulties in DSA three dimensional reconstructions. The idea of 3D reconstruction is developed from application aspect, under the basis of vessel analysis and quantitative description, the idea is provided through associating the multi planar vessel-tracing with the 3D reconstruction. On this condition, we make a farther step into a clinic application modal which we assuredly present some useful approaches and tools to the diagnosing and curing.
(4) A content based medical image retrieval application model is presented.
Content based image retrieval (CBIR) is a present hotpot image analysis subject, it has large amount of applications and researches in various fields. But it's still in starting stage in medical field. This is due to the particular characteristics of medical images, such as the high similarity between images, high spatial and grayscale resolution, and large amount of information; on the other hand, this is also because of the complexity in imaging principle, anatomic part characteristics and clinic application needs of medical image. Now researches on CBIR mainly concentrated on the aspects of feature selection algorithm, decrease dimensions to increase efficiency and establish judgment criterion.
Characteristics extraction and expression is the basis of CBIR, in this thesis we analyzed general image retrieval principles and various algorithms, combining with generality of medical image, a multi-characteristics integrated medical image retrieval model is established based upon common image retrieval model. By selecting the combination of several groups color and texture characteristics, including Fractal Dimensions(FD), Color Moments(CM), Circle Moran Auto-correlation(CMAC), Gray Level Co-occurrence Matrix(GLCM), Gabor texture, Grey-Level Difference(SGLD), for characteristics selection and extraction, effectively decreased dimensions Spatial and increased retrieval speed and accuracy. The model can be integrated to PACS for the purpose of medical image administration and retrieval and further more application upon PACS platform.
Quantitative analysis result of angiography image has typicality as characteristics of CBIR used in XRA. Precise quantitative description, including centerlines, widths and diameters, ending nodes, segment vessel lengths, radius and directions, position of branches and crosses , etc, can not only be used as selected characteristics for special retrieval model of XRA medical image, but also can be used to establish higher level semantic based retrieval model, this is the more valuable and meaningful research subject.
Research of quantitative analysis and application in X-ray angiogram image are discussed in this thesis. A lot of experiments are also illustrated to prove the validity of the models and their corresponding approaches mentioned in the paper. Application models are established for clinic and practical needs.
X射线血管造影(X-Ray Angiogram,XRA)是通过造影剂显示血管的一种方法。它是一种介入式血管疾病诊断与治疗工具,其成像的基本原理是利用注入血管的造影剂来表现血管特性,并根据造影产生的差异来对血管进行分析。目前比较普遍的应用集中在脑部和心脏冠状动脉,特别是在心脏部位,其主要临床应用方向包括针对循环系统及各器官血管性病变以及肿瘤性病变供血血管的诊断,数字减影(DSA)引导下的介入手术治疗和基于三维重建图像引导的计算机辅助手术治疗等等。它的意义在于可以在给疾病定性的同时,还可以给以后外科治疗提供详细的血管解剖图像。更有意义的是对于相当一部分疾病,在通过血管造影明确诊断的同时进行介入性治疗。在这些应用的基础上,随着现代医学和计算机信息学的发展,血管造影图像的临床应用也在不断扩展和深化,基于血管造影图像精确量化分析基础上的计算机辅助诊断(CAD)和基于量化特征为主要内容的基于内容的医学血管造影图像检索(Content Based ImageRetrieval,CBIR)目前已经成为医学图像分析和应用研究的热点之一。
从造影图像中准确分割出血管树、对血管重要参数(如血管中心线及其分支的位置、长度、逐点宽度、曲率等)进行精确的量化描述,是进行可视化三维血管结构重建及心脑血管诊断和治疗的重要前提,也是进行辅助诊断和医学血管造影图像内容检索的重要保障。鉴于血管具有复杂的形态结构,以及在成像环境中存在的噪声、造影剂衰减和分布不均匀等复杂情况,它的造影图像中经常存在血管目标模糊和局部形态特异性等各种问题,造成无法对目标对象进行准确的分割和量化测量。当前的许多经典算法对此都未能具备较高的自动化性和鲁棒性。本文针对这些问题,在随机、模糊、优化等理论框架下对血管造影图像的结构特点进行了颇有成效的实验研究,提出了基于跟踪的血管提取模型,提出了血管中心线和宽度测量新方法。以此实现了一种新的血管自动分割与测量技术,较好地完成了对血管的准确分割和精确的量化描述。在此基础上从临床应用的需要出发,结合前面的研究成果,针对XRA图像的三维血管可视化重建和基于内容的医学图像检索两种主要的实际应用进行了初步探索,建立了两个具有实际价值的应用模型。主要工作包括:
(一)提出了一种基于相似性测度和似然模型的血管造影图像分割新方法。
从造影图像中准确地分割出血管树是后续三维可视化和图像检索等研究的重要基础,该工作涉及到血管分割和血管局部和全局结构的识别。经典的基于跟踪的方法主要思路是利用一个局部算子或函数作用于血管上某一点,通过构造不同的跟踪算法进行跟踪,如利用边缘连接信息引导跟踪、利用中轴线连续性跟踪、利用搜索方向和收缩范围等图像特征跟踪等。这些跟踪方法主要的不足在于由于缺少正则化局部条件和局部测度,导致血管跟踪时很容易受环境噪声和弯曲程度等的影响。Shrijver通过构造了圆周轮廓函数来解决这一问题,但是其血管相似性函数并不具备正则化的高斯分布形式,且无法满足血管转弯过大的情况,在分支点判断方面,仅通过圆周测度观察分支点的局部极值,并没有考虑到分支点处更为复杂的血管形态:如交叉、卷曲、折叠等造成的伪分支情况,因此准确性也是有限的。为提高血管段跟踪的鲁棒性,Park提出了基于血管边缘轮廓的极大似然概率函数,该函数迫使跟踪器采样是选择双边缘中心线上的点,然而Park并未给出血管节点判别情况,如血管分支。针对经典的基于跟踪的一类方法的以上不足,通过比较现有各种血管分割方法,提出了集特征提取、知识引导、相似性测度、概率跟踪于一体的新的血管结构识别算法,给出了新的血管特征提取算法。提出了一种鲁棒的最大似然血管跟踪模型,该模型建立在多种局部相似性测度和智能知识引导的基础上,能够精确估计出血管轴线并正确判断分支点位置。
(二)提出了一种测量血管中心线和宽度的新方法。
血管参数(中心线和直径)测量面临着如下挑战:(1)血管网络从外围动脉顶部到末端、从大动脉到胫骨和肋骨动脉,呈现复杂的空间形态;(2)由于造影剂的分布不均、衰减、曝光不均等复杂情况导致造影图像血管目标和背景反差极小;(3)血管直径由几十象素单位变化到4～5个象素单位,很难准确清晰地提取局部和全局血管结构;(4)部分容积效应也严重影响着估计的精度。
传统的基于CSP的方法经常由于血管轮廓不易定位、血管目标较弱、边缘提取很粗糙而失败。James Lowell等针对此提出一种高斯和高斯微分的二维轮廓模型,用来优化迭代逼近血管目标,其血管宽度的计算精度达到子象素级。Lowell的不足正像他承认的一样:不能很好地作用于血管节点(分支和重叠点)。因此,如果能够有效增强血管目标并提取出目标方向特征,则CSP方法具有很大的挖掘潜力。
鉴于上述情况,本文将提出一种鲁棒的基于截面梯度轮廓(CSP)的血管参数测量新方法,解决经典方法存在的缺陷。该方法的主要特点是:首先,利用Gabor方向滤波器增强造影图像中的血管目标、压制图像背景噪声,并获取图像中血管目标的方向场;其次,提取血管骨架,通过细化处理后,获得分布在血管区域内部的初始轴线;最后,利用边缘梯度算子和血管方向场计算并获取垂直于血管方向的血管截面轮廓,从而测量血管参数。该算法采用了统计均值和亮度控制的搜索技术,可以精确提取血管重要参数的量化描述,解决了经典方法不提供血管宽度,以及无法测量节点(交叉、分支)处血管中心线等缺陷。由于该算法还能够估计血管节点处的血管参数,因此不失为一种血管参数测量的整体解决方案。
(三)提出了一种三维血管重建的思路
目前国内针对血管可视化方面的研究主要体现在基于模型匹配的血管提取、基于分叉模型的三维重建、基于目标引导的三维重建、基于参数化弹性模型的重建、样条运动估计等等方面。可以了解到目前国内针对血管XRA图像的3D重建的研究成果主要还是停留在XRA重建和仿真阶段,与临床实践还存在一定的距离。从国外现状来看,基于X射线造影图像的血管树的三维重建技术正成为研究的热点,国外的研究主要针对临床采集的真实数据,并面向解决实际问题,其发展趋势由早期的参数化、模型匹配、先验知识引导、局部统计优化等等方法发展为近年来以统计优化建模、人工智能、模式识别等为主的方法,并更加注重采用参数法和统计法相结合的思路。
本文从DSA造影设备的成像原理出发,详细分析了两幅单平面造影图像的血管三维重建所涉及的一些关键性技术,包括DSA系统设备的标定、血管骨架的分割、特征点的识别和匹配、血管空间点的重建和血管段的拟合、血管的截面的描述与血管树的三维显示技术等等。本文还对DSA血管三维重建技术的一些难点进行了详细的分析和讨论。本文从应用的角度出发,在血管分析与量化描述的研究成果基础上,提出三维血管重建的主要思路,在该思路下将血管跟踪与多平面造影图像重建结合起来,从而可以解决全自动真三维重建的难题。在此基础上探索建立了一套有临床价值的应用模型,可以很好地为医学临床诊断和治疗提供一些可滋借鉴的方法和工具。
(四)提出了一种医学图像内容检索的应用模型
基于内容的图像检索(Content Based Image Retrieval,CBIR)是目前图像分析的一个广泛关注的热点课题,在各个应用领域都有不同的应用和研究。但是在医学领域的研究还处于起步阶段,这一方面是由于医学图像具有其自身独特的特点,如图像之间的相似度较高、空间和灰度分辨率高、信息量大等;另一方面也是由于医学图像的成像原理、解剖部位特征和临床应用需要的复杂性。当前针对医学图像内容检索的主要研究集中在特征选取算法,降低维数提高有效性以及建立判定准则等方面。
图像特征的提取与表达是基于内容的图像检索技术的基础,本文分析了通用图像检索的原理和各种算法,结合医学图像的共性特征,在通用图像检索模型基础上建立了一个多特征综合医学图像检索模型。采用分形维数FD、颜色矩、循环Moran自相关、灰度共生矩阵GLCM、Gabor纹理、空间灰度差共生矩阵SGLD等几组颜色和纹理特征组合作为特征选择和特征提取的依据,有效地降低了维数,提高了检索速度和检索准确度。该模型可以有效地集成到PACS系统中,为医学图像的管理检索和建立在PACS系统平台上的更多图像内容方面的应用奠定基础。
造影图像量化分析的结果作为血管造影类图像内容检索所需要的特征具有典型性。精确的量化结果,包括血管中心线、宽度和直径、末梢节点、分段血管长度、半径和方向、分支和交叉的位置等各种量化参数,不但可以作为建立针对血管造影这一类医学图像的更加有效的检索模型和检索应用的特征选择,以此为基础建立更高层次上的基于语义的检索模型,是更有价值和意义的研究课题,本文对此也进行了探讨。
本文针对血管造影图像的量化分析和应用研究做了有意义的尝试,通过大量的实验验证了所提出方案和算法的有效性,并从应用角度出发,尝试建立满足实际临床需要的应用模型,达到理论研究与实际应用相结合,医工结合的目的,为最终针对临床的医学图像分析处理和检索系统打下了研究基础。
X-Ray angiogram play a very important role in the fields of clinic-diagnosing and the computer aided and image guided surgery. Analysis and application for XRA is an important research direction in the modern digital medical imaging research. With the improvement of living condition and prolonged life expectancy, cardiovascular diseases (CVD) have been the number one cause of death in modem society. The early quantitative diagnosis and accurate evaluation of CVD are critical to improving quality of life and prolong life expectancy. XRA has great value for the diagnosis of cardiovascular diseases and catheter intervened surgeries. In this thesis, we propose several quantitative descriptions and applications for XRA image. The main contribution of this thesis is list as following.
X-ray angiogram (XRA) is a type of technique to show blood vessel by contrast agent. It's an interventional diagnosis and therapy facility for blood vessel disease, the basic principle of XRA imaging is to manifest blood vessel features by contrast agent injected into the blood vessel and to analyse the blood vessel through the difference generated by the angiography. At present, the common application of XRA focused on cerebral vascular and coronary artery, especially on cardiovascular, the main clinic applications including diagnosis of vascular disease in circulatory system and organs and blood supply vessel of tumors, digital subtraction angiography (DSA) guiding interventional surgical therapy, and 3D image guiding computer aided surgery, etc. The meaning is that it can qualitative describing the disease and provide detailed vascular anatomic images to surgery for later therapy. The most significant thing is that for quite a part of diseases, it can carry out simultaneous interventional therapy when definite diagnosis conclusion is made. On these basis, the clinic applications of XRA is keep extending and deepening with the rapid progress of modern medical science and computer informatics, nowadays, computer aided diagnosis (CAD) based on precise quantitative analysis of XRA and content based medical image retrieval (CBIR) based on quantitative characters of XRA have become hotspots of medical image analysis and application research.
Accuracy segmentation of blood vessel tree from XRA image and precise quantitative description of blood vessel parameters (such as centerlines, position of branches, length, width and curvature of blood vessel, etc) is the very important premise of visualized three dimensional vessel structure reconstruction and diagnosis and treatment of cardiac and cerebral vascular disease, it's also the important guarantee of assisted diagnosis and blood vessel angiogram image content retrieval. It's difficult to make precise segmentation and measurement to the object under the condition of fuzzy object or local shape difference and such problems due to the blood vessels in the angiograms take on complicated modalities, as well as complicated conditions during imaging environment such as noise, contrast agent decay and distributioninhomogeneous. The existing classic methods can't be provided with universality and robustion. In this paper, the integrated results and conclusions have been attained with the features of X-ray angiogram under the theory frameworks of stochastic, fuzzy and optimum, a tracing based blood vessel extraction model is established. A new approach to centerlines and widths measurement of blood vessels is proposed. A new full automated segmentation and measurement methods are established, which make a satisfied result of quantitative descriptions of vessel modality. Based upon these analyses, combined with clinic application needs, we made a discussion on two of XRA major practical applications: three dimension visualization reconstruction and content based medical image retrieval, two profitable application models are given for practical purpose. The main contributions are detailed as follows:
(1) A robust Bayesian tracking model is proposed in this thesis.
Accuracy segmentation of blood vessel tree from XRA image is the basis of subsequent research on three dimensional visualization and image content retrieval. Classic tracking based methods construct different tracking algorithms for tracking by means of using a local operator or a local function to a certain point in the blood vessel, the tracking algorithms including edge link information guiding tracking, centerline continuity tracking, image characteristics such as search direction and contraction range tracking, etc. The main deficiency of these methods is that the tracking process is easy to be affected by environment noise and flexural degree due to lack of regularized local conditions and local measurement. Shrijver try to solve this problem by constructing circle contour function, but the blood vessel comparability function also doesn't have regularized Gaussian distribution form, also it can not satisfied the condition that the vessel turning is too excessive, on respects of branch point judgment, it observes local extreme only by circle measurement, without considering more complicated blood vessel morphology in the branch point, eg., pseudo branches caused by crossover, curling, folding, thus the accuracy of this method is also limited. Park proposed a maximum likelihood probability function based on blood vessel edge contour, its forces the tracker choose the points in the centerline of dual edge for sampling, however, Park doesn't give discrimination conditions of blood vessel nodes such as blood vessel branches.
According to the deficiency of the classical tracking based methods, by comparing with existing vessel segmentation methods, a new structure identification algorithm as well as vessel features extraction method, with the integrative of feature extraction, intelligent guidance. Comparability measurement and probability tracing are presented to solve the problems of the classic tracking-based methods such as losing the object in tracking the vessel tree. A robust Bayesian tracking model is presented to precise estimate the vessel axis and infers the vessel structure, which is based on multi-comparability measurement and intelligent guidance.
(2) A new cross-section profile (CSP) based approach for the vessel parameters (centerlines and widths) measurement is proposed
Blood vessel parameters ( centerlines and widths) measurement faced the following challenge: (1) vessel network present complex spatial morphology from peripheral artery top to end, from large artery to tibia and rib artery; (2) The complicated conditions due to uneven distribution, decay and inhomogeneous exposure of contrast agent lead to minimum contrast between vessel target and background in the XRA image; (3) It's difficult to extract local and full vessel structure accurately and clearly due to the vessel diameter variation from several tenth pixels to 4～5 pixels; (4) Partial volume effect also seriously affect the precision of estimation.
Due to hard location problem of vessel profile, weak vessel target and rough edge extraction, classic CSP based approaches are easier to fail. James Lowel,et al. proposed a Gaussian and Gaussian differential based two dimensional profile model used to optimal iterative approximation the blood vessel target, its calculated vessel width precision reached the degree of sub-pixel level. The deficiency of Lowell's approach, just as he admitted, is that it can't well act on vessel nodes (branches and overlapped points).So CSP approach still has much potentiality if we can enhance the vessel target and extract its orientation characteristics effectively.
According to the above description, a new robust cross-section profile (CSP) based approach is proposed for the vessel parameters (centerlines and widths) measurement. The main features is: firstly, use Gabor directional filter to enhance the vessel target and suppressing background noise and also obtain the direction field of target in the XRA image; secondly, extract vessel framework ,. refining the framework then get the initial axis lines distributed in the vessel area; finally, use edge gradient operator and vessel directional field to calculate and acquire the cross section profile vertical to the vessel direction so as to measure the vessel parameters. The proposed method utilizes a statistic-average and luminance-controlled searching techniques to precise extract the vascular description of the important vessel parameters. The CSP based algorithm is better than the classic one of the center of gravity: Two serious disadvantages in the classic algorithms have been solved, i.e. the measurement of vessel widths and the complexion of blood vessel nodes were not considered in their algorithm models.
(3) Present a new idea of 3D vessel reconstruction
At present most of the researches in vessel visualization in China are mainly on several fields: model matching based vessel extraction, fraction model based three dimensional reconstructions, target guiding based three dimensional reconstruction, parametric elastic model based reconstruction, spline motion estimation, etc. Now three dimensional XRA vessel reconstruction research in China mostly stay on XRA reconstruction and simulation. On the side of foreign situation, vessel tree three dimensional reconstructions based on X-ray angiogram image is becoming the research hotspot. Foreign research now aim at real date acquired from clinic and practical issues, their development trend has changed from early methods such as parameterized, model matching, prior knowledge guiding, local statistical optimal and so on to recent developed statistical optimal modeling, artificial intelligence, model recognition methods, and pay more attention to the thought of using parametric methods combined with statistical methods.
In this thesis, starting from the imaging principle of DSA angiography equipment, we detailed analyzed the critical techniques involving in three dimensional reconstruction of two single plane XRA images, including DSA equipment calibration, vessel framework segmentation, characteristic points recognition and matching, vessel spatial points reconstruction and vessel section fitting, vessel cross section description and vessel tree three dimension display technique and so on. We also made a detailed analysis and discussion on the difficulties in DSA three dimensional reconstructions. The idea of 3D reconstruction is developed from application aspect, under the basis of vessel analysis and quantitative description, the idea is provided through associating the multi planar vessel-tracing with the 3D reconstruction. On this condition, we make a farther step into a clinic application modal which we assuredly present some useful approaches and tools to the diagnosing and curing.
(4) A content based medical image retrieval application model is presented.
Content based image retrieval (CBIR) is a present hotpot image analysis subject, it has large amount of applications and researches in various fields. But it's still in starting stage in medical field. This is due to the particular characteristics of medical images, such as the high similarity between images, high spatial and grayscale resolution, and large amount of information; on the other hand, this is also because of the complexity in imaging principle, anatomic part characteristics and clinic application needs of medical image. Now researches on CBIR mainly concentrated on the aspects of feature selection algorithm, decrease dimensions to increase efficiency and establish judgment criterion.
Characteristics extraction and expression is the basis of CBIR, in this thesis we analyzed general image retrieval principles and various algorithms, combining with generality of medical image, a multi-characteristics integrated medical image retrieval model is established based upon common image retrieval model. By selecting the combination of several groups color and texture characteristics, including Fractal Dimensions(FD), Color Moments(CM), Circle Moran Auto-correlation(CMAC), Gray Level Co-occurrence Matrix(GLCM), Gabor texture, Grey-Level Difference(SGLD), for characteristics selection and extraction, effectively decreased dimensions Spatial and increased retrieval speed and accuracy. The model can be integrated to PACS for the purpose of medical image administration and retrieval and further more application upon PACS platform.
Quantitative analysis result of angiography image has typicality as characteristics of CBIR used in XRA. Precise quantitative description, including centerlines, widths and diameters, ending nodes, segment vessel lengths, radius and directions, position of branches and crosses , etc, can not only be used as selected characteristics for special retrieval model of XRA medical image, but also can be used to establish higher level semantic based retrieval model, this is the more valuable and meaningful research subject.
Research of quantitative analysis and application in X-ray angiogram image are discussed in this thesis. A lot of experiments are also illustrated to prove the validity of the models and their corresponding approaches mentioned in the paper. Application models are established for clinic and practical needs.
引文
[1]A.Murat Tekalp.Digital Video Processing.Prentice-Hall,Inc.1996.
[2]Stan Z.Li.Markov random field modeling in image analysis[M].Springer-Verlag Tokyo,2001.
[3]X.Han,C.Xu,D.Tosun,and J.L.Prince,"Cortical surface reconstruct-tion using a topology preserving geometric deformable model," in Proc.5th IEEE Workshop MMBIA 2001,pages 213--220,Kauai,HI,Dec 2001.
[4]Cemil Kirbas,Francis K.H.Quek:A review of vessel extraction techniques and algorithms.ACM Computing Surveys(CSUR),Vol.36,Issue 2,pp:81-121,2004.
[5]Jiri Hladuvka,Andreas K(o|¨)nig,Meister Eduard Gr(o|¨)ller.Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation.Proceedings of the 9th International Conference in Central Europe on Computer Graphics(WSCG),Visualization and Computer Vision 2001,vol.1,pp.124-129,February 2001.
[6]郁道银,黄家祥,谢洪波,徐智,孙正.冠状动脉树三维重建理论模型的研究.工程图学学报,2003,第4期.
[7]R.Pajak.Use of two-dimensional matched filters for estimating a length of blood vessels newly created in angiogenesis process.OPTO-ELECTRONICS REVIEW 11(3),237.241(2003)
[8]A.E Frangi,W.J.Niessen,K.L.Vincken,M.A.Viergever.Multiscale vessel enhancement filtering.In Medical Image Computing and Computer-Assisted Intervention,MICCAI'98,1998.
[9]L.Torresani,D.B.Yang,E.J.Alexander,and C.Bregler.Tracking and Modeling Non-Rigid Objects withRa nk Constraints.In CVPR,2001.
[10]Cemil Kirbas and Francis K.H.Quek.Vessel Extraction Techniques and Algorithms:A Survey.Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering(BIBE),2003.
[11]Marc Schrijver and Comelis H.Slump.Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections.Proceedings of ProRISC 2002November 28-29.Veldhoven,the Netherlands,2002.
[12]Park,S.,Lee,J.,Koo,J.,Kwon,O.,and Hong,S.1997.Adaptive tracking algorithm based on direction field using ml estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol. 2.671—675,1997.
[13] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[14]Subhasis Chaudhuri, Shankar Chatterjee, Norman Katz. Mark Nelson, and Michael Goldbaum. Detection of Blood Vessels in Retinal Images Using Two-Dimensional Matched Filters. IEEE Transactions on Medical Imaging, Vol.8, No. 3, September 1989.
[15]Simona E. Grigorescu, Nicolai Petkov, and Peter Kruizinga. Comparison of Texture Features Based on Gabor Filters. IEEE Transactions on Image Processing,Vol. 11, no. 10, pp:1160-1167, Oct. 2002
[16]Cosmin Grigorescu, Student Member, IEEE, Nicolai Petkov, and Michel A.Westenberg. Contour Detection Based on Nonclassical Receptive Field Inhibition.IEEE Transactions On Image Processing, Vol. 12, No. 7, July 2003.
[17] Calibration of x-ray projections in 3D reconstruction Wang, Ping; Mou, Xuanqin;Qin, Zhongyuan; Cai, Yuanlong Author Affiliation: Image Processing Center,Xi'an Jiaotong University Source: Proceedings of SPIE - The International Society for Optical Engineering v 4553 Oct 22-24 2001 2001 Sponsored by:SPIE The International Society for Optical Engineer- ing pp: 44-49 0277-786X In English.
[18] John Puentes, Member, IEEE, Christian Roux,* Senior Member, IEEE, Mireille Garreau and Jean Louis Coatrieux, Fellow, IEEE. Dynamic Feature Extraction of Coronary Artery Motion Using DSA Image Sequences. IEEE Transactions on Medical Imaging, Vol. 17, No. 6, December 1998.
[19]A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[20]S.Y. Chen, John D. Carroll, "Kinematic and Deformation Analysis of 4-D Coronary Arterial Trees Reconstructed fromCine Angiograms". IEEE Trans. Med.Imaging 22(6): 710-721, 2003
[21]S.Y. Chen, John D. Carroll, John C, "Messenger: Quantitative analysis of reconstructed 3-D coronary arterial tree and intracoronary devices". IEEE Trans. Med. Imaging 21(7): 724-740, 2002
[22]C.Canero, P. Radena, R.Toledo, J.J. Villanueva, and J.Mauri. "3D curve reconstruction by biplane snakes". In Proc. IEEE of ICPR, volume 4, pages 563-566,2000.
[23]Henning Muller, Nicolas Michoux, David Bandon, Antoine Geissbuhler, A review of content-based image retrieval systems in medical applications——clinical benefits and future directions, International Journal of Medical Informatics(2004) 73,1-23
[24]L.H.Y. Tang, R. Hanka, H.H.S. Ip, A review of intelligent, content-based indexing and browsing of medical images, Health Informatics J. 5 (1999) 40—49.
[1] Michael H. F. Wilkinson, Tsjipke Wijbenga, Gijs de Vries, and Michel A. Blood vessel segmentation using moving-window robust automatic threshold selection.Proceedings of 2003 International Conference on Image Processing (ICIP),Volume: 2, pp: II-1093-6 vol.3,14-17 Sept. 2003.
[2] Jiri Hladuvka, Andreas Konig, Meister Eduard Groller. Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation. In conference proceedings of the 9th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision 2001, WSCG'2001, University of West Bohemia, Pilsen,Czech Republic, Vaclav Skala (ed.), February 2001, vol. 1, pp. 124-129
[3] A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer- Assisted Intervention, MICCAI'98, 1998.
[4] Cemil Kirbas, Francis K. H. Quek: A review of vessel extraction techniques and algorithms. ACM Computing Surveys (CSUR), Vol. 36, Issue 2, pp:81-121,2004.
[5] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering (BIBE), 2003.
[6] Marc Schrijver and Cornelis H. Slump. Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[1] Marc Schrijver and Comelis H. Slump. Automatic segmentation of the coronary artery tree in angiographic projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[2] Ali Can, Hong Shen, James N.Turner, Howard L.Tanenbaum, and Badrinath Roysam, IEEE Member. Rapid automated tracking and feature extraction from retinal fundus images using direct exploratory algorithms. IEEE Tran. On Information Technology in Biomedicine, Vol.3, No.1. March 1999.
[3] Park, S., Lee, J., Koo, J., Kwon, O., and Hong, S. Adaptive tracking algorithm based on direction field using ML estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol.2. 671—675,1997.
[4] H. Shen, C. Stewart, B. Roysam. Optimal scheduling of tracing computations for real-time vascular landmark extraction from retinal fundus images. IEEE Transactions on Information Technology in Biomedicine, 5(1):77—91, Mar 2001.
[5] Tsai, Chia-Ling, Stewart, Charles V, Tanenbaum, H. L., and Roysam, B.,"Model-Based Method for Improving The Accuracy and Repeatability of Estimating Vascular Bifurcations and Crossovers From Retinal Fundus Images,"IEEE Transactions on IT for Biomedicine, vol. 8, No. 2, June 2004.
[6] Li Wang and A. Bhalerao. Model based detection of branching structures, In Proc of Medical Image Understanding and Analysis (MIUA), Portsmouth, UK. July,2002.
[7] A. Bhalerao, R. Wilson. Estimating Local and Global Structure using a Gaussian Intensity Model, In Proceedings of Medical Image Understanding and Analysis (MIUA), July 2001.
[8] Li Wang, A. Bhalerao and R. Wilson. Robust Modelling of Local Image Structures and its Application to Medical Imagery. In Proc. Int. Conf. Pattern Recognition, ICPR'04. Cambridge, 2004.
[9] Jiri Hladuvka, Andreas Konig, Meister Eduard Groller. Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation. In conference proceedings of the 9th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision 2001, WSCG'2001, University of West Bohemia, Pilsen,Czech Republic, Vaclav Skala (ed.), February 2001, vol. 1, pp. 124-129
[10]A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[11]R. Pajak. Use of two-dimensional matched filters for estimating a length of blood vessels newly created in angiogenesis process. OPTO-ELECTRONICS REVIEW 11(3), 237.241 (2003).
[1] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering (BIBE), 2003.
[2] Cemil Kirbas, Francis K. H. Quek: A review of vessel extraction techniques and algorithms. ACM Computing Surveys (CSUR), Vol. 36, Issue 2, pp:81-121,2004.
[3] Marc Schrijver and Cornelis H. Slump. Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[4] Park, S., Lee, J., Koo, J., Kwon, O., and Hong, S. 1997. Adaptive tracking algorithm based on direction field using ml estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol. 2.671—675,1997.
[5] Ali Can, Hong Shen, James N.Turner, Howard L.Tanenbaum, and Badrinath Roysam, IEEE Member. Rapid automated tracking and feature extraction from retinal fundus images using direct exploratory algorithms. IEEE Tran. On Information Technology in Biomedicine, Vol.3, No.1. March 1999.
[6] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[7] A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[8] L. Lecornu, C. Toumoulin, C. Roux, and L. Senhadji, Using svd and analyzing wavelets for the analysis and the segmentation of digital subtraction angiography images, Proceedings Computer in Cardiology, 1992, pp. 655-658.
[9] Sato, Y., Nakajima, S., Shiraga, N., Atsumi, H., Yoshida, S., Koller, T., Gerig, G, and Kikinis,R.:'3D multi-scale filter line filter for segmentation and visualization of curvilinear structures in medical images',Med.Image Anal.,1998,2,(2),pp.143-168
[10]O.Chutatape,Senior Member,IEEE,Liu Zheng,and S.M.Krishnan,Senior Member,IEEE.Retinal blood vessel detection and tracing by matched gaussian and kalman filters.In Proc.of the 20th Annual International Conference of the IEEE Engineering in Medicine and Biology Society,(EMBS'98),volume 20,pages 3144-3149,1998.
[11]Sun,Y.,"Automated identification of vessel contours in coronary arteriograms by an adaptive tracking algorithm," IEEE Trans.Med.Imag.,Vol.8,pp.78-88,1989.
[12]Yannis A.Tolias,Student,IEEE,and Stavros M.Panas,Member,IEEE.A Fuzzy Vessel Tracing Algorithm for Retinal Images Based on Fuzzy Clustering.IEEE Trans.on Med.Imag.,Vol.17,No.2,263-273.April 1998.
[13]E.Grisan,M.Foracchia,and A.Rugged,"Normalization of luminance and contrast in fundus images," Abstract of 3rd CAFIA workshop,European Journal of Ophthalmology,vol.13,no.2,pp.228-229,March 2003.
[14]Marc Lalonde,Langis Gagnon and Marie-Carole Boucher,"Non-recursive paired tracking for vessel extraction from retinal images",In Proc.of the Conference Vision Interface 2000,pages 61-68,2000.
[15]Khalid A.Al-Kofahi,Ali Can,Member,IEEE,et al.Median-Based Robust Algorithms for Tracing Neurons from Noisy Confocal Microscope Images.IEEE Transactions on Information Technology in Biomedicine,Vol.7,No.4,pp:302-317,December 2003.
[16]M.Hart and L.Holley,"A method of automated coronary artery tracking in unsubtracted angiograms," in Proc.Computers Cardiology'93,1993,pp.93-96.
[17]Park,S.,Lee,J.,Koo,J.,Kwon,O.,and Hong,S.1997.Adaptive tracking algorithm based on direction field using ml estimation in angiogram.In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications.Vol.2.671-675,1997.
[18]Shen,H.,"Optical Instrumentation and Real-Time Image Processing Algorithms for Simultaneous ICG and Red-Free Video Angiography of the Retina," M.S.Thesis,Rensselaer Polytechnic Institute,1996.
[1] James Lowell, Andrew Hunter, David Steel, Ansu Basu, Robert Ryder, and R.Lee Kennedy. Measurement of Retinal Vessel Widths From Fundus Images Based on 2-D Modeling. IEEE Transactions on Medical Imaging, Vol. 23, No. 10,pp: 1196-1204, October 2004.
[2] A. La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Proceedings of Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, pp: 115-120,2004.
[3] C. Lorenz, I.-C. Carsen, T. Buzug, C. Fassnacht, and J. Weese. Multi-scale line segmentation with automatic estimation of width, contrast and tangential direction in 2d and 3d medical images. In CVRMed-MRCAS'97, volume 1205,pages 213{222, Grenoble, France, Mar. 1997.
[4] T. Lindeberg. Edge detection and ridge detection with automatic scale selection.In CVPR, page 465, San Francisco, June 1996.
[5] A. Frangi, W. Niessen, K. Vincken, and M. Viergever. Multiscale vessel enhancement
[6] _ltering. In W.Weils, A. Colchester, and S. Delp, editors, MICCAI'98, volume 1496 of lncs, pages 130{137, Oct. 1998.
[7] J. Canny. A computational approach to edge detection. PAMI-8:679{698,Nov.1986.
[8] L. M. Lorigo, O. Faugeras, W. Grimson, R. Keriven, R. Kikinis, and C.-F. Westin.Co-dimension 2 geodesic active contours for mra segmentation. In Inter. Conf.On Information Processing in Medical Imaging (IPMI), volume 1613 of Lecture Notes in Computer Science, pages 126{139. Springer-Verlag, 1999.
[9] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on Biolnformatics and BioEngineering (BIBE), 2003.
[10] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[11]C. Xu and J.L. Prince, "Snakes, shapes, and gradient vector flow", IEEE Trans.on Image Proces., vol. 7, pp. 359-369,1998.
[12]Maclaughlin. A.: Randomized Hough Transform: Improved ellipse detection with comparison. In Pattern Recognition Letters 19 (1998), pp. 299-305. 3,4
[13] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels.Joint EUROGRAPHICS IEEE TCVG Symposium on Visualization,2004.
[1]吴恩惠,医学影像学(第四版),北京:人民卫生出版社,2001.1-9,172-176
[2]R.Toledo."Cardiac Workstation and Dynamic Model to Assist in Coronary Tree Analysis".PhD thesis,University of Barcelona,Spain,July 2001.
[3]A.C.M.Dumay,J.H.C.Reiber,and J.J.Gerbrands."Determination of optimal angiographic viewing angles:Basic principles and evaluation study".IEEE Trans.Med.Imag.,13(1),March 1994.
[4]Windyga P,Garreau M,LeBreton H,"2D and 3D knowledge combination for the reconstruction of coronary arteries:first results on real data",Page(s):620-621vol.1,Year:3-6 Nov 1994
[5]AB merle,G Finet,J Lienard,"3D reconstruction of the deformable coronary tree skeleton from two X-Ray angiographic views",IEEE Computers in Cardiology,1998,25:7572760.
[6]S.J.Chen,J.D.Carroll,"3-D reconstruction of coronary arterial tree to optimize angiographic visualization," IEEE Trans.Medical Imaging,vol.19,pp.318-336,Apr.2000.
[7]S.Y.Chen,John D.Carroll,John C,"Messenger:Quantitative analysis of reconstructed 3-D coronary arterial tree and intracoronary devices".IEEE Trans.Med.Imaging 21(7):724-740,2002
[8]S.Y.Chen,John D.Carroll,"Kinematic and Deformation Analysis of 4-D Coronary Arterial Trees Reconstructed fromCine Anglograms".IEEE Trans.Med.Imaging 22(6):710-721,2003
[9]C.Canero,P.Radena,R.Toledo,J.J.Villanueva,and J.Mauri."3D curve reconstruction by biplane snakes".In Prec.IEEE of ICPR,volume 4,pages 563-566,2000.
[10]O.Faugeras."Three-Dimensional Computer Vision:A Geometric Viewpoint".The MIT Press,1993.
[11]E.Coste,D.Gibon,and J.Rousseau."Assessment of image intensifier and distortion for DSA localization studies".The British Journal of Radiology,0:70-73,1997
[12]L.Launay et al."Quantitative evaluation of an algorithm for correcting geometrical distorsions in DSA images: application to stereotaxy". In SPIE,volume 2434, pages 520-529, San Diego, California, February 1995.
[13]H. Wollschlager et al. "Improvement of quantitative angiography by exact calculation of radiological magnification factors". In Computers in Cardiology,pages 483-486,1985.
[14]D.GW onnasch and GP.M. prause, "Geometric image correction and iso-center calibration at oblique biplane angiographic view". In Proc. IEEE Computers in Cardiology, pages 647-650,1992.
[15]L.Launay et al. "Quantitative evaluation of an algorithm for correcting geometrical distorsions in DSA images: applications to stereotaxy". In SPIE,volue 2434, pages 520-529, San Diego, California, February 1995.
[16]E. Kerrien et al. "Machine precision assessment for 3D/2D digital substracted angiography images registration". In SPIE, Medical Imaging, Image processing,volume 3338, San Diego, California, February 1998.
[17]Cemil Kirbas, Francis Quek, "A review of vessel extraction techniques and algorithms". ACM Computing Surveys(CSUR) (June 2004),Volume: 36,Issue: 2
[18]A. Sarwal and A.P. Dhawan, "3-d reconstruction of coronary arteries", in IEEE Conf. Eng. in Medicine and Bio., vol. 1, pp. 504-505,1994.
[19]D. Guo and P. Richardson, "Automatic vessel extraction from angiogram images", IEEE Computers in Cardiology, vol. 25, pp. 441-444, 1998.
[20]M. Kass, A. Witkin, and D. Terzoopoulos, "Snakes: Active contour models", Int.J. of Comp. Vision, vol. 1, pp. 321-331,1988.
[21]V. Caselles, F. Catte, T. Coll, and F. Dibos, "A geometric model for active contours in image processing", NM, vol. 66, pp. 1-32,1993.
[22]S.Aylward, S. Pizer, E. Bullitt, and D. Eberl, "Intensity ridge and widths for tabular object segmentation and registration", in Wksp on Math. Methods in Biomed. Image Analysis, pp. 131-138,1996.
[23]S. Lu and S. Eiho, "Automatic detection of the coronary arterial contours with sub-branches from an x-ray angiogram", in IEEE Computers in Cardiology, pp.575-578,1993.
[24]C. Smets, G.Verbeeck, P. Suetens, and A. Oosterlinck, "A knowledge-based system for the delineation of blood vessels on subtraction angiograms", Pattern Rec. Let., vol. 8, pp. 113-121, September 1988.
[25] S.A. Stansfiled, "Angy: A rule-based expert system for automatic segmentation of coronory vessels from digital subtracted angiograms",PAMI,vol.8,pp.188-199,March 1986.
[26]B.D.Ripley,"Pattern Recognition and Neural Networks",Cambridge Un.Press,1996.
[27]Richard Hartley,Andrew Zisserman,"Multiple View Geometry in Computer Vision(Second edition)",Cambridge University Press,March 2004.
[28]XIA Shtm-ren,LV Wei2xue,"Progress in 3D reconstruction of coronary artery from biplane angiograms',Foreign Medical Sciences(Biomedical Engineering Fascicle),1993,16(5):2492254.(in Chinese)
[29]C.Xu and J.L.Prince,"Snakes,shapes,and gradient vector flow",IEEE Trans.on Image Proces.,vol.7,pp.359-369,1998.
[30]Donald Hearn,M.Pauline Baker."Computer Graphics with OpenGL(Third Edition)",Publishing House of Electronics Industry,Beijing,2004.
[31]Albert K.W.Law,Hui Zhu,and Francis H.Y."Chan,3D Reconstruction of Coronary Artery using Biplane Angiography",Proceeding of the 25th Annual International Conference of the IEEE EMBS Cancun,Mexico,September 17-21,2003.
[32]http://www.ge-china.com/product/jrj.htm
[1]Henning M(u|¨)ller,Nicolas Michoux,David Bandon,Antoine Geissbuhler,A review of content-based image retrieval systems in medical applications-—clinical benefits and future directions,International Journal of Medical Informatics(2004) 73,1-23
[2]Arnold W.M.Smeulders,Senior Member,IEEE,Marcel Worring,Simone Santini,Member,Amarnath Gupta,Member,IEEE,and Ramesh Jain,Fellow,IEEE,Content-Based Image Retrieval at the End of the Early Years,IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,VOL.22,NO.12,DECEMBER 2000
[3]L.H.Y.Tang,R.Hanka,H.H.S.Ip,A review of intelligent,content-based indexing and browsing of medical images,Health Informatics J.5(1999) 40-49.
[4]A.Mojsilovis,J.Gomes,Semantic based image categorization,browsing and retrieval in medical image databases,in:IEEE International Conference on Image Processing (ICIP' 2000), Rochester, NY, USA, 2000.
[5] H. J. Zhang and D. Zhong, "A Scheme for visual feature-based image indexing,"Proc. of SPIE conf. on Storage and Retrieval for Image and Video Databases III,pp. 36-46, San Jose, Feb. 1995.
[6] 66. D. R. Xu. Research on the imagery generation in Design. Ph.D dissertation,Zhejiang University, Hangzhou, 1995.
[7] R. M. Haralick, S. R. Sternberg, and X. Zhuang, "Image analysis using mathematical morphology," IEEE Trans. Pattern Anal. Machine Intell., vol.PAMI-9,pp. 532-550, Apr. 1987.
[8] 68. P. Brodatz, "Textures: A photographic album for artists & designers," Dover,NY, 1966.
[9] Y. Rui, T.S. Huang, S.-F. Chang, Image retrieval: past, present and future, in: M.Liao (Ed.), Proceedings of theInternational Symposium on Multimedia Information Processing, Taipei, Taiwan, 1997.
[10]J.P. Eakins, M.E. Graham, Content-based image retrieval, Tech. Rep. JTAP-039,JISC Technology Application Program, Newcastle upon Tyne, 2000.
[11]O. Chapelle, V. Vapnik, O. Bousquet, and S. Mukherjee. Choosing kernel parameters for support vector machines. Machine Learning, 2000.
[12]B. Moghaddam, A. Pentlan. Beyond linear eigenspaces: Bayesian matching for face recognition. Face Recognition: from Theory to Application. Springer-Verlag 1998. 230-243
[13]A.W.M. Smeulders, M. Worring, S. Santini, A. Gupta, R.Jain, Content-based image retrieval at the end of theearly years, IEEE Trans. Pattern Anal. Machine Intel.22 (12) (2000) 1349—1380.
[14]H. Schneiderman, T. Kanade. A statistical method for 3D object detection applied to faces and cars. IEEE Conf on Computer Vision and Pattern Recognition,Hilton Head Island, South Carolina, USA: IEEE Computer Society, 2000
[15]L.H.Y. Tang, R. Hanka, H.H.S. Ip, A review of intelligent content-based indexing and browsing of medical images,Health Informatics J. 5 (1999) 40—49.
[16]R. E. Schapire, Y. Singer. Improved boosting algorithms using confidence-rated predictions. Machine Learning, 1999, 37(3). 297-336
[17]T.M. Lehmann, M.O. Guld, C. Thies, B. Fischer, M. Keysers, D. Kohnen, H.Schubert, B.B. Wein, Content-based image retrieval in medical applications for picture archiving and communication systems, in: Proceedings of the SPIE Conference on Medical Imaging,vol.5033,San Diego,CA,USA,2003.
[18]M.R.Ogiela,R.Tadeusiewicz,Semantic-oriented syntactic algorithms for content recognition and understanding of images in medical databases,in:Proceedings of the second International Conference on Multimedia and Exposition(ICME' 2001),IEEE Computer Society,IEEE
[1]Albert K.W.Law,Hui Zhu,and Francis H.Y."Chan,3D Reconstruction of Coronary Artery using Biplane Angiography",Proceeding of the 25th Annual International Conference of the IEEE EMBS Cancun,Mexico,September 17-21,2003.
[2]http://www.ge-china.com/product/jrj.htm
[3]Li Wang and A.Bhalerao.Model based detection of branching structures,In Proc of Medical Image Understanding and Analysis(MIUA),Portsmouth,UK.July,2002.
[4]Li Wang,A.Bhalerao and R.Wilson.Robust Modelling of Local Image Structures and its Application to Medical Imagery.In Proc.Int.Conf.Pattern Recognition,ICPR'04.Cambridge,2004.
[5]Wong,S.T.C.Content-Based Access of Image and Video Libraries,1998.Proceedings.IEEE Workshop on Volume,Issue,21 Jun 1998 Page(s):114-117
[6]A.Marchiod,C.Brodley,L.Broderick,J.OF,C.Pavlopoulou,A.C.Kak and A.Aisen,"CBIR for Medical Images-An Evaluation Trial",in Proceedings IEEE Workshop on Content-Based Access of Image and Video Databases,Hawaii,2001.
[2]Stan Z.Li.Markov random field modeling in image analysis[M].Springer-Verlag Tokyo,2001.
[3]X.Han,C.Xu,D.Tosun,and J.L.Prince,"Cortical surface reconstruct-tion using a topology preserving geometric deformable model," in Proc.5th IEEE Workshop MMBIA 2001,pages 213--220,Kauai,HI,Dec 2001.
[4]Cemil Kirbas,Francis K.H.Quek:A review of vessel extraction techniques and algorithms.ACM Computing Surveys(CSUR),Vol.36,Issue 2,pp:81-121,2004.
[5]Jiri Hladuvka,Andreas K(o|¨)nig,Meister Eduard Gr(o|¨)ller.Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation.Proceedings of the 9th International Conference in Central Europe on Computer Graphics(WSCG),Visualization and Computer Vision 2001,vol.1,pp.124-129,February 2001.
[6]郁道银,黄家祥,谢洪波,徐智,孙正.冠状动脉树三维重建理论模型的研究.工程图学学报,2003,第4期.
[7]R.Pajak.Use of two-dimensional matched filters for estimating a length of blood vessels newly created in angiogenesis process.OPTO-ELECTRONICS REVIEW 11(3),237.241(2003)
[8]A.E Frangi,W.J.Niessen,K.L.Vincken,M.A.Viergever.Multiscale vessel enhancement filtering.In Medical Image Computing and Computer-Assisted Intervention,MICCAI'98,1998.
[9]L.Torresani,D.B.Yang,E.J.Alexander,and C.Bregler.Tracking and Modeling Non-Rigid Objects withRa nk Constraints.In CVPR,2001.
[10]Cemil Kirbas and Francis K.H.Quek.Vessel Extraction Techniques and Algorithms:A Survey.Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering(BIBE),2003.
[11]Marc Schrijver and Comelis H.Slump.Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections.Proceedings of ProRISC 2002November 28-29.Veldhoven,the Netherlands,2002.
[12]Park,S.,Lee,J.,Koo,J.,Kwon,O.,and Hong,S.1997.Adaptive tracking algorithm based on direction field using ml estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol. 2.671—675,1997.
[13] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[14]Subhasis Chaudhuri, Shankar Chatterjee, Norman Katz. Mark Nelson, and Michael Goldbaum. Detection of Blood Vessels in Retinal Images Using Two-Dimensional Matched Filters. IEEE Transactions on Medical Imaging, Vol.8, No. 3, September 1989.
[15]Simona E. Grigorescu, Nicolai Petkov, and Peter Kruizinga. Comparison of Texture Features Based on Gabor Filters. IEEE Transactions on Image Processing,Vol. 11, no. 10, pp:1160-1167, Oct. 2002
[16]Cosmin Grigorescu, Student Member, IEEE, Nicolai Petkov, and Michel A.Westenberg. Contour Detection Based on Nonclassical Receptive Field Inhibition.IEEE Transactions On Image Processing, Vol. 12, No. 7, July 2003.
[17] Calibration of x-ray projections in 3D reconstruction Wang, Ping; Mou, Xuanqin;Qin, Zhongyuan; Cai, Yuanlong Author Affiliation: Image Processing Center,Xi'an Jiaotong University Source: Proceedings of SPIE - The International Society for Optical Engineering v 4553 Oct 22-24 2001 2001 Sponsored by:SPIE The International Society for Optical Engineer- ing pp: 44-49 0277-786X In English.
[18] John Puentes, Member, IEEE, Christian Roux,* Senior Member, IEEE, Mireille Garreau and Jean Louis Coatrieux, Fellow, IEEE. Dynamic Feature Extraction of Coronary Artery Motion Using DSA Image Sequences. IEEE Transactions on Medical Imaging, Vol. 17, No. 6, December 1998.
[19]A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[20]S.Y. Chen, John D. Carroll, "Kinematic and Deformation Analysis of 4-D Coronary Arterial Trees Reconstructed fromCine Angiograms". IEEE Trans. Med.Imaging 22(6): 710-721, 2003
[21]S.Y. Chen, John D. Carroll, John C, "Messenger: Quantitative analysis of reconstructed 3-D coronary arterial tree and intracoronary devices". IEEE Trans. Med. Imaging 21(7): 724-740, 2002
[22]C.Canero, P. Radena, R.Toledo, J.J. Villanueva, and J.Mauri. "3D curve reconstruction by biplane snakes". In Proc. IEEE of ICPR, volume 4, pages 563-566,2000.
[23]Henning Muller, Nicolas Michoux, David Bandon, Antoine Geissbuhler, A review of content-based image retrieval systems in medical applications——clinical benefits and future directions, International Journal of Medical Informatics(2004) 73,1-23
[24]L.H.Y. Tang, R. Hanka, H.H.S. Ip, A review of intelligent, content-based indexing and browsing of medical images, Health Informatics J. 5 (1999) 40—49.
[1] Michael H. F. Wilkinson, Tsjipke Wijbenga, Gijs de Vries, and Michel A. Blood vessel segmentation using moving-window robust automatic threshold selection.Proceedings of 2003 International Conference on Image Processing (ICIP),Volume: 2, pp: II-1093-6 vol.3,14-17 Sept. 2003.
[2] Jiri Hladuvka, Andreas Konig, Meister Eduard Groller. Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation. In conference proceedings of the 9th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision 2001, WSCG'2001, University of West Bohemia, Pilsen,Czech Republic, Vaclav Skala (ed.), February 2001, vol. 1, pp. 124-129
[3] A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer- Assisted Intervention, MICCAI'98, 1998.
[4] Cemil Kirbas, Francis K. H. Quek: A review of vessel extraction techniques and algorithms. ACM Computing Surveys (CSUR), Vol. 36, Issue 2, pp:81-121,2004.
[5] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering (BIBE), 2003.
[6] Marc Schrijver and Cornelis H. Slump. Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[1] Marc Schrijver and Comelis H. Slump. Automatic segmentation of the coronary artery tree in angiographic projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[2] Ali Can, Hong Shen, James N.Turner, Howard L.Tanenbaum, and Badrinath Roysam, IEEE Member. Rapid automated tracking and feature extraction from retinal fundus images using direct exploratory algorithms. IEEE Tran. On Information Technology in Biomedicine, Vol.3, No.1. March 1999.
[3] Park, S., Lee, J., Koo, J., Kwon, O., and Hong, S. Adaptive tracking algorithm based on direction field using ML estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol.2. 671—675,1997.
[4] H. Shen, C. Stewart, B. Roysam. Optimal scheduling of tracing computations for real-time vascular landmark extraction from retinal fundus images. IEEE Transactions on Information Technology in Biomedicine, 5(1):77—91, Mar 2001.
[5] Tsai, Chia-Ling, Stewart, Charles V, Tanenbaum, H. L., and Roysam, B.,"Model-Based Method for Improving The Accuracy and Repeatability of Estimating Vascular Bifurcations and Crossovers From Retinal Fundus Images,"IEEE Transactions on IT for Biomedicine, vol. 8, No. 2, June 2004.
[6] Li Wang and A. Bhalerao. Model based detection of branching structures, In Proc of Medical Image Understanding and Analysis (MIUA), Portsmouth, UK. July,2002.
[7] A. Bhalerao, R. Wilson. Estimating Local and Global Structure using a Gaussian Intensity Model, In Proceedings of Medical Image Understanding and Analysis (MIUA), July 2001.
[8] Li Wang, A. Bhalerao and R. Wilson. Robust Modelling of Local Image Structures and its Application to Medical Imagery. In Proc. Int. Conf. Pattern Recognition, ICPR'04. Cambridge, 2004.
[9] Jiri Hladuvka, Andreas Konig, Meister Eduard Groller. Exploiting Eigenvalues of the Hessian Matrix for Volume Decimation. In conference proceedings of the 9th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision 2001, WSCG'2001, University of West Bohemia, Pilsen,Czech Republic, Vaclav Skala (ed.), February 2001, vol. 1, pp. 124-129
[10]A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[11]R. Pajak. Use of two-dimensional matched filters for estimating a length of blood vessels newly created in angiogenesis process. OPTO-ELECTRONICS REVIEW 11(3), 237.241 (2003).
[1] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on BioInformatics and BioEngineering (BIBE), 2003.
[2] Cemil Kirbas, Francis K. H. Quek: A review of vessel extraction techniques and algorithms. ACM Computing Surveys (CSUR), Vol. 36, Issue 2, pp:81-121,2004.
[3] Marc Schrijver and Cornelis H. Slump. Automatic Segmentation of the Coronary Artery Tree in Angiographic Projections. Proceedings of ProRISC 2002 November 28-29. Veldhoven, the Netherlands, 2002.
[4] Park, S., Lee, J., Koo, J., Kwon, O., and Hong, S. 1997. Adaptive tracking algorithm based on direction field using ml estimation in angiogram. In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications. Vol. 2.671—675,1997.
[5] Ali Can, Hong Shen, James N.Turner, Howard L.Tanenbaum, and Badrinath Roysam, IEEE Member. Rapid automated tracking and feature extraction from retinal fundus images using direct exploratory algorithms. IEEE Tran. On Information Technology in Biomedicine, Vol.3, No.1. March 1999.
[6] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[7] A.F. Frangi, W.J. Niessen, K.L. Vincken, M.A. Viergever. Multiscale vessel enhancement filtering. In Medical Image Computing and Computer-Assisted Intervention, MICCAI'98,1998.
[8] L. Lecornu, C. Toumoulin, C. Roux, and L. Senhadji, Using svd and analyzing wavelets for the analysis and the segmentation of digital subtraction angiography images, Proceedings Computer in Cardiology, 1992, pp. 655-658.
[9] Sato, Y., Nakajima, S., Shiraga, N., Atsumi, H., Yoshida, S., Koller, T., Gerig, G, and Kikinis,R.:'3D multi-scale filter line filter for segmentation and visualization of curvilinear structures in medical images',Med.Image Anal.,1998,2,(2),pp.143-168
[10]O.Chutatape,Senior Member,IEEE,Liu Zheng,and S.M.Krishnan,Senior Member,IEEE.Retinal blood vessel detection and tracing by matched gaussian and kalman filters.In Proc.of the 20th Annual International Conference of the IEEE Engineering in Medicine and Biology Society,(EMBS'98),volume 20,pages 3144-3149,1998.
[11]Sun,Y.,"Automated identification of vessel contours in coronary arteriograms by an adaptive tracking algorithm," IEEE Trans.Med.Imag.,Vol.8,pp.78-88,1989.
[12]Yannis A.Tolias,Student,IEEE,and Stavros M.Panas,Member,IEEE.A Fuzzy Vessel Tracing Algorithm for Retinal Images Based on Fuzzy Clustering.IEEE Trans.on Med.Imag.,Vol.17,No.2,263-273.April 1998.
[13]E.Grisan,M.Foracchia,and A.Rugged,"Normalization of luminance and contrast in fundus images," Abstract of 3rd CAFIA workshop,European Journal of Ophthalmology,vol.13,no.2,pp.228-229,March 2003.
[14]Marc Lalonde,Langis Gagnon and Marie-Carole Boucher,"Non-recursive paired tracking for vessel extraction from retinal images",In Proc.of the Conference Vision Interface 2000,pages 61-68,2000.
[15]Khalid A.Al-Kofahi,Ali Can,Member,IEEE,et al.Median-Based Robust Algorithms for Tracing Neurons from Noisy Confocal Microscope Images.IEEE Transactions on Information Technology in Biomedicine,Vol.7,No.4,pp:302-317,December 2003.
[16]M.Hart and L.Holley,"A method of automated coronary artery tracking in unsubtracted angiograms," in Proc.Computers Cardiology'93,1993,pp.93-96.
[17]Park,S.,Lee,J.,Koo,J.,Kwon,O.,and Hong,S.1997.Adaptive tracking algorithm based on direction field using ml estimation in angiogram.In IEEE Conference on Speech and Image Technologies for Computing and Telecommunications.Vol.2.671-675,1997.
[18]Shen,H.,"Optical Instrumentation and Real-Time Image Processing Algorithms for Simultaneous ICG and Red-Free Video Angiography of the Retina," M.S.Thesis,Rensselaer Polytechnic Institute,1996.
[1] James Lowell, Andrew Hunter, David Steel, Ansu Basu, Robert Ryder, and R.Lee Kennedy. Measurement of Retinal Vessel Widths From Fundus Images Based on 2-D Modeling. IEEE Transactions on Medical Imaging, Vol. 23, No. 10,pp: 1196-1204, October 2004.
[2] A. La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Proceedings of Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, pp: 115-120,2004.
[3] C. Lorenz, I.-C. Carsen, T. Buzug, C. Fassnacht, and J. Weese. Multi-scale line segmentation with automatic estimation of width, contrast and tangential direction in 2d and 3d medical images. In CVRMed-MRCAS'97, volume 1205,pages 213{222, Grenoble, France, Mar. 1997.
[4] T. Lindeberg. Edge detection and ridge detection with automatic scale selection.In CVPR, page 465, San Francisco, June 1996.
[5] A. Frangi, W. Niessen, K. Vincken, and M. Viergever. Multiscale vessel enhancement
[6] _ltering. In W.Weils, A. Colchester, and S. Delp, editors, MICCAI'98, volume 1496 of lncs, pages 130{137, Oct. 1998.
[7] J. Canny. A computational approach to edge detection. PAMI-8:679{698,Nov.1986.
[8] L. M. Lorigo, O. Faugeras, W. Grimson, R. Keriven, R. Kikinis, and C.-F. Westin.Co-dimension 2 geodesic active contours for mra segmentation. In Inter. Conf.On Information Processing in Medical Imaging (IPMI), volume 1613 of Lecture Notes in Computer Science, pages 126{139. Springer-Verlag, 1999.
[9] Cemil Kirbas and Francis K.H. Quek. Vessel Extraction Techniques and Algorithms: A Survey. Proceedings of the Third IEEE Symposium on Biolnformatics and BioEngineering (BIBE), 2003.
[10] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels. Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization, 2004.
[11]C. Xu and J.L. Prince, "Snakes, shapes, and gradient vector flow", IEEE Trans.on Image Proces., vol. 7, pp. 359-369,1998.
[12]Maclaughlin. A.: Randomized Hough Transform: Improved ellipse detection with comparison. In Pattern Recognition Letters 19 (1998), pp. 299-305. 3,4
[13] Alexandra La Cruz. Accuracy Evaluation of Different Centerline Approximations of Blood Vessels.Joint EUROGRAPHICS IEEE TCVG Symposium on Visualization,2004.
[1]吴恩惠,医学影像学(第四版),北京:人民卫生出版社,2001.1-9,172-176
[2]R.Toledo."Cardiac Workstation and Dynamic Model to Assist in Coronary Tree Analysis".PhD thesis,University of Barcelona,Spain,July 2001.
[3]A.C.M.Dumay,J.H.C.Reiber,and J.J.Gerbrands."Determination of optimal angiographic viewing angles:Basic principles and evaluation study".IEEE Trans.Med.Imag.,13(1),March 1994.
[4]Windyga P,Garreau M,LeBreton H,"2D and 3D knowledge combination for the reconstruction of coronary arteries:first results on real data",Page(s):620-621vol.1,Year:3-6 Nov 1994
[5]AB merle,G Finet,J Lienard,"3D reconstruction of the deformable coronary tree skeleton from two X-Ray angiographic views",IEEE Computers in Cardiology,1998,25:7572760.
[6]S.J.Chen,J.D.Carroll,"3-D reconstruction of coronary arterial tree to optimize angiographic visualization," IEEE Trans.Medical Imaging,vol.19,pp.318-336,Apr.2000.
[7]S.Y.Chen,John D.Carroll,John C,"Messenger:Quantitative analysis of reconstructed 3-D coronary arterial tree and intracoronary devices".IEEE Trans.Med.Imaging 21(7):724-740,2002
[8]S.Y.Chen,John D.Carroll,"Kinematic and Deformation Analysis of 4-D Coronary Arterial Trees Reconstructed fromCine Anglograms".IEEE Trans.Med.Imaging 22(6):710-721,2003
[9]C.Canero,P.Radena,R.Toledo,J.J.Villanueva,and J.Mauri."3D curve reconstruction by biplane snakes".In Prec.IEEE of ICPR,volume 4,pages 563-566,2000.
[10]O.Faugeras."Three-Dimensional Computer Vision:A Geometric Viewpoint".The MIT Press,1993.
[11]E.Coste,D.Gibon,and J.Rousseau."Assessment of image intensifier and distortion for DSA localization studies".The British Journal of Radiology,0:70-73,1997
[12]L.Launay et al."Quantitative evaluation of an algorithm for correcting geometrical distorsions in DSA images: application to stereotaxy". In SPIE,volume 2434, pages 520-529, San Diego, California, February 1995.
[13]H. Wollschlager et al. "Improvement of quantitative angiography by exact calculation of radiological magnification factors". In Computers in Cardiology,pages 483-486,1985.
[14]D.GW onnasch and GP.M. prause, "Geometric image correction and iso-center calibration at oblique biplane angiographic view". In Proc. IEEE Computers in Cardiology, pages 647-650,1992.
[15]L.Launay et al. "Quantitative evaluation of an algorithm for correcting geometrical distorsions in DSA images: applications to stereotaxy". In SPIE,volue 2434, pages 520-529, San Diego, California, February 1995.
[16]E. Kerrien et al. "Machine precision assessment for 3D/2D digital substracted angiography images registration". In SPIE, Medical Imaging, Image processing,volume 3338, San Diego, California, February 1998.
[17]Cemil Kirbas, Francis Quek, "A review of vessel extraction techniques and algorithms". ACM Computing Surveys(CSUR) (June 2004),Volume: 36,Issue: 2
[18]A. Sarwal and A.P. Dhawan, "3-d reconstruction of coronary arteries", in IEEE Conf. Eng. in Medicine and Bio., vol. 1, pp. 504-505,1994.
[19]D. Guo and P. Richardson, "Automatic vessel extraction from angiogram images", IEEE Computers in Cardiology, vol. 25, pp. 441-444, 1998.
[20]M. Kass, A. Witkin, and D. Terzoopoulos, "Snakes: Active contour models", Int.J. of Comp. Vision, vol. 1, pp. 321-331,1988.
[21]V. Caselles, F. Catte, T. Coll, and F. Dibos, "A geometric model for active contours in image processing", NM, vol. 66, pp. 1-32,1993.
[22]S.Aylward, S. Pizer, E. Bullitt, and D. Eberl, "Intensity ridge and widths for tabular object segmentation and registration", in Wksp on Math. Methods in Biomed. Image Analysis, pp. 131-138,1996.
[23]S. Lu and S. Eiho, "Automatic detection of the coronary arterial contours with sub-branches from an x-ray angiogram", in IEEE Computers in Cardiology, pp.575-578,1993.
[24]C. Smets, G.Verbeeck, P. Suetens, and A. Oosterlinck, "A knowledge-based system for the delineation of blood vessels on subtraction angiograms", Pattern Rec. Let., vol. 8, pp. 113-121, September 1988.
[25] S.A. Stansfiled, "Angy: A rule-based expert system for automatic segmentation of coronory vessels from digital subtracted angiograms",PAMI,vol.8,pp.188-199,March 1986.
[26]B.D.Ripley,"Pattern Recognition and Neural Networks",Cambridge Un.Press,1996.
[27]Richard Hartley,Andrew Zisserman,"Multiple View Geometry in Computer Vision(Second edition)",Cambridge University Press,March 2004.
[28]XIA Shtm-ren,LV Wei2xue,"Progress in 3D reconstruction of coronary artery from biplane angiograms',Foreign Medical Sciences(Biomedical Engineering Fascicle),1993,16(5):2492254.(in Chinese)
[29]C.Xu and J.L.Prince,"Snakes,shapes,and gradient vector flow",IEEE Trans.on Image Proces.,vol.7,pp.359-369,1998.
[30]Donald Hearn,M.Pauline Baker."Computer Graphics with OpenGL(Third Edition)",Publishing House of Electronics Industry,Beijing,2004.
[31]Albert K.W.Law,Hui Zhu,and Francis H.Y."Chan,3D Reconstruction of Coronary Artery using Biplane Angiography",Proceeding of the 25th Annual International Conference of the IEEE EMBS Cancun,Mexico,September 17-21,2003.
[32]http://www.ge-china.com/product/jrj.htm
[1]Henning M(u|¨)ller,Nicolas Michoux,David Bandon,Antoine Geissbuhler,A review of content-based image retrieval systems in medical applications-—clinical benefits and future directions,International Journal of Medical Informatics(2004) 73,1-23
[2]Arnold W.M.Smeulders,Senior Member,IEEE,Marcel Worring,Simone Santini,Member,Amarnath Gupta,Member,IEEE,and Ramesh Jain,Fellow,IEEE,Content-Based Image Retrieval at the End of the Early Years,IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,VOL.22,NO.12,DECEMBER 2000
[3]L.H.Y.Tang,R.Hanka,H.H.S.Ip,A review of intelligent,content-based indexing and browsing of medical images,Health Informatics J.5(1999) 40-49.
[4]A.Mojsilovis,J.Gomes,Semantic based image categorization,browsing and retrieval in medical image databases,in:IEEE International Conference on Image Processing (ICIP' 2000), Rochester, NY, USA, 2000.
[5] H. J. Zhang and D. Zhong, "A Scheme for visual feature-based image indexing,"Proc. of SPIE conf. on Storage and Retrieval for Image and Video Databases III,pp. 36-46, San Jose, Feb. 1995.
[6] 66. D. R. Xu. Research on the imagery generation in Design. Ph.D dissertation,Zhejiang University, Hangzhou, 1995.
[7] R. M. Haralick, S. R. Sternberg, and X. Zhuang, "Image analysis using mathematical morphology," IEEE Trans. Pattern Anal. Machine Intell., vol.PAMI-9,pp. 532-550, Apr. 1987.
[8] 68. P. Brodatz, "Textures: A photographic album for artists & designers," Dover,NY, 1966.
[9] Y. Rui, T.S. Huang, S.-F. Chang, Image retrieval: past, present and future, in: M.Liao (Ed.), Proceedings of theInternational Symposium on Multimedia Information Processing, Taipei, Taiwan, 1997.
[10]J.P. Eakins, M.E. Graham, Content-based image retrieval, Tech. Rep. JTAP-039,JISC Technology Application Program, Newcastle upon Tyne, 2000.
[11]O. Chapelle, V. Vapnik, O. Bousquet, and S. Mukherjee. Choosing kernel parameters for support vector machines. Machine Learning, 2000.
[12]B. Moghaddam, A. Pentlan. Beyond linear eigenspaces: Bayesian matching for face recognition. Face Recognition: from Theory to Application. Springer-Verlag 1998. 230-243
[13]A.W.M. Smeulders, M. Worring, S. Santini, A. Gupta, R.Jain, Content-based image retrieval at the end of theearly years, IEEE Trans. Pattern Anal. Machine Intel.22 (12) (2000) 1349—1380.
[14]H. Schneiderman, T. Kanade. A statistical method for 3D object detection applied to faces and cars. IEEE Conf on Computer Vision and Pattern Recognition,Hilton Head Island, South Carolina, USA: IEEE Computer Society, 2000
[15]L.H.Y. Tang, R. Hanka, H.H.S. Ip, A review of intelligent content-based indexing and browsing of medical images,Health Informatics J. 5 (1999) 40—49.
[16]R. E. Schapire, Y. Singer. Improved boosting algorithms using confidence-rated predictions. Machine Learning, 1999, 37(3). 297-336
[17]T.M. Lehmann, M.O. Guld, C. Thies, B. Fischer, M. Keysers, D. Kohnen, H.Schubert, B.B. Wein, Content-based image retrieval in medical applications for picture archiving and communication systems, in: Proceedings of the SPIE Conference on Medical Imaging,vol.5033,San Diego,CA,USA,2003.
[18]M.R.Ogiela,R.Tadeusiewicz,Semantic-oriented syntactic algorithms for content recognition and understanding of images in medical databases,in:Proceedings of the second International Conference on Multimedia and Exposition(ICME' 2001),IEEE Computer Society,IEEE
[1]Albert K.W.Law,Hui Zhu,and Francis H.Y."Chan,3D Reconstruction of Coronary Artery using Biplane Angiography",Proceeding of the 25th Annual International Conference of the IEEE EMBS Cancun,Mexico,September 17-21,2003.
[2]http://www.ge-china.com/product/jrj.htm
[3]Li Wang and A.Bhalerao.Model based detection of branching structures,In Proc of Medical Image Understanding and Analysis(MIUA),Portsmouth,UK.July,2002.
[4]Li Wang,A.Bhalerao and R.Wilson.Robust Modelling of Local Image Structures and its Application to Medical Imagery.In Proc.Int.Conf.Pattern Recognition,ICPR'04.Cambridge,2004.
[5]Wong,S.T.C.Content-Based Access of Image and Video Libraries,1998.Proceedings.IEEE Workshop on Volume,Issue,21 Jun 1998 Page(s):114-117
[6]A.Marchiod,C.Brodley,L.Broderick,J.OF,C.Pavlopoulou,A.C.Kak and A.Aisen,"CBIR for Medical Images-An Evaluation Trial",in Proceedings IEEE Workshop on Content-Based Access of Image and Video Databases,Hawaii,2001.