Retinal vessel extraction by matched filter with first-order derivative of Gaussian

Accurate extraction of retinal blood vessels is an important task in computer aided diagnosis of retinopathy. The matched filter (MF) is a simple yet effective method for vessel extraction. However, a MF will respond not only to vessels but also to non-vessel edges. This will lead to frequent false vessel detection. In this paper we propose a novel extension of the MF approach, namely the MF-FDOG, to detect retinal blood vessels. The proposed MF-FDOG is composed of the original MF, which is a zero-mean Gaussian function, and the first-order derivative of Gaussian (FDOG). The vessels are detected by thresholding the retinal image's response to the MF, while the threshold is adjusted by the image's response to the FDOG. The proposed MF-FDOG method is very simple; however, it reduces significantly the false detections produced by the original MF and detects many fine vessels that are missed by the MF. It achieves competitive vessel detection results as compared with those state-of-the-art schemes but with much lower complexity. In addition, it performs well at extracting vessels from pathological retinal images.     

基于数学形态学的视网膜血管提取算法

目的 影像中血管的分割与特征提取,对疾病的早期诊断具有重要意义。针对很多视网膜血管提取算法分割精度不高的问题,提出了运用数学形态学中的高帽变换的方法对其进行检测。方法 首先,选取结构元素为“圆盘形”的形态学对图像进行高帽变换,经过高帽变换后的图像平滑了图像的背景,同时增强了血管在图像中的对比度。其次,对变换后的图像利用Otsu＇s自动分割法对图像进行阈值分割得到血管的二值图像。再次,根据血管在图像中的结构信息和几何信息,利用基于连通域度量的方法,设置连通域的“面积”和“长宽比”两个阈值,去除虚假目标。最后,为保持血管的连续性,对图像进行一次膨胀运算,可将断裂的血管连接起来,减小了实验的误差。结果 通过上述步骤实现了对血管的提取。结论 结果表明,本文算法能有效提取视网膜眼底图像的血管网络,有较强的分割精度。     

Image feature analysis and computer-aided diagnosis in digital radiography. 2. Computerized determination of vessel sizes in digital subtraction angiography

We developed an iterative deconvolution technique to determine the size of a "blurred" vessel in a digital subtraction angiographic (DSA) image by taking into account the unsharpness of the DSA system. Initially, a region of interest over a small segment of the contrast-filled vessel was selected in a DSA image, and the center line of the opacified vessel was determined by polynomial curve fitting of the locations of the peak pixel values along the vessel image. The blurred image profile was then obtained from pixel values across the vessel in a direction perpendicular to the center line. This measured profile was compared iteratively with a calculated profile for various size vessels, which was obtained from a cylindrical vessel model and from the line spread function, until the root-mean-square difference between the two profiles was minimized. The size of a cylindrical vessel yielding the matched profile was considered the best estimate of the unknown vessel size. Studies with a blood vessel phantom indicated that vessels larger than 0.5 mm could be measured with an accuracy and precision of approximately 0.1 mm, which is about 1/3 of the pixel size used in our DSA system. Details of our approach and some clinical vessel images with and without simulated stenotic lesions are presented.     

基于中心线提取的视网膜血管分割

精确估计眼底图像中的血管中心线和血管宽度是眼底血管疾病量化和可视化诊断的先决条件。在深入研究眼底图像和视网膜血管特征的基础上,提出一种基于中心线提取的视网膜血管分割算法,该算法以离散高斯核函数的偏微分为模板卷积图像,利用血管的微分几何特征定位血管中心线,并测量血管宽度。本方法的血管分割精度达到亚像素级,解决了传统方法无法直接计算血管宽度的问题,且在速度方面有较大提高。应用本算法对各类眼底图像进行了血管分割的实验,结果表明,该算法可以快速、准确地进行眼底血管分割和中心线提取。     

基于Gabor滤波器的糖尿病视网膜新生血管检测

目的:采用Gabor滤波器实现眼底图像中新生血管检测,帮助医生准确确定糖尿病视网膜病变的分期。 方法:对眼底图像进行预处理,并使用不同尺度参数和方向参数Gabor滤波器作用于预处理图像,并在尺度参数确定的情况下取各方向输出结果的最大值作为最后Gabor滤波器的输出。 结果:对比分析不同尺度参数的Gabor滤波器的结果,发现小尺度参数的Gabor滤波器在新生血管部分具有较强的输出。 结论:本研究提出的Gabor滤波器可以很好地区分眼底图像中正常血管与新生血管结构。     

视网膜血管高血压交叉压迫征的量化研究

研究视网膜血管高血压交叉压迫征的量化方法,采用局部血管图像增强的匹配滤波进行预处理、数学形态学的闭合运算进行图像平滑、线化无级扩大法确定血管边界,实施管径可逆量化自动寻优算法,为获取相关疾病诊断的数据提供一种辅助手段。     

带有预处理的计划CT和在线CT的形变配准方法

CT扫描中,水溶性碘造影的存在使得计划CT和在线CT图像中血管内的HU值出现非常大的偏差,从而导致计划CT和在线CT图像错配。针对该问题,本研究提出了一种基于预处理的计划CT和在线CT形变配准方法。首先,根据CT图像组织和结构的信息,利用阈值分割方法分割出血管,并将所有分割中最大的联通区域作为初始分割的强化血管;其次,利用分割得到的强化血管区域外扩5 mm,作为外扩的强化血管,并将血管用固定的HU值进行填充;最后,对完成填充后的图像利用Demons算法进行形变配准。实验结果显示本文提出的带有预处理的形变配准方法,可以较好地解决水溶性碘造影剂引起的CT错配问题。     

Decision support system for the detection and grading of hard exudates from color fundus photographs

Diabetic retinopathy is a major cause of blindness, and its earliest signs include damage to the blood vessels and the formation of lesions in the retina. Automated detection and grading of hard exudates from the color fundus image is a critical step in the automated screening system for diabetic retinopathy. We propose novel methods for the detection and grading of hard exudates and the main retinal structures. For exudate detection, a novel approach based on coarse-to-fine strategy and a new image-splitting method are proposed with overall sensitivity of 93.2% and positive predictive value of 83.7% at the pixel level. The average sensitivity of the blood vessel detection is 85%, and the success rate of fovea localization is 100%. For exudate grading, a polar fovea coordinate system is adopted in accordance with medical criteria. Because of its competitive performance and ability to deal efficiently with images of variable quality, the proposed technique offers promising and efficient performance as part of an automated screening system for diabetic retinopathy.     

Detection of needle puncture to blood vessel using puncture force measurement

Although blood sampling is frequently performed, a system to take blood samples automatically has not yet been developed. In the paper, as a first step towards automatic blood sampling, an examination of an automatic method for puncturing blood vessels is described. The force waveforms produced by puncturing rabbit ear veins were measured. A characteristic peak, possibly associated with the needle piercing the vessel wall, was observed in each waveform of ten successful cases of 14 trials. An algorithm that allowed the detection of this peak was developed, and parameters of a filter to improve the accuracy of the algorithm were determined. Using this algorithm, automatic needle punctures were performed in a rabbit ear vein and then were simulated using the data derived from manual needle puncture on four other rabbits. The results gave 30 fully successful cases of 33 trials of needle puncture and showed that measurement of the puncture force may be important for automatic needle puncture of blood vessels.     

视网膜血管形态识别方法的研究

本文利用人体生物特征的视网膜血管作为个人身份鉴别的特征 ,在建立高层次保密和安全应用的理论依据的研究中 ,提出了一种形态识别的相关网点法 ,包括方法的构思和依据 ,图像特征交点的提取 ,网点质心和最远网点的水平方位对准 ,图像相关网点匹配的判定 ,并给出了形态识别的结果     

Astrocyte invasion and vasculogenesis in the developing ferret retina

We studied the time course of astrocyte invasion and blood vessel formation in the developing ferret retina using glial fibrillary acidic protein (GFAP)-immunohistochemistry for astrocytes and isolectin B4 histochemistry for blood vessels. As in other mammals, strongly GFAP positive astrocytes invade the ferret retina from the optic nerve. At birth, strongly GFAP positive astrocytes have reached about 22% of the distance between optic disc and outer retinal edge whereas weakly GFAP positive processes already extend to the edge of the retina. At postnatal days P30–P37 about 82% of the distance between optic disc and outer retinal edge and in the adult 88% of this distance is covered with strongly labelled astrocytes. Superficial blood vessels form from the optic disc. They reach up to about 24% of the retinal radius at birth and grow radially across the retina during further development. At P30–P37, the whole retina is covered with superficial blood vessels. The deep vascular layer forms later (around P30) through sprouting from superficial vessels. The radial pattern of astrocyte and vessel growth from the optic disc is not affected by the formation of the area centralis and visual streak.     

A model based method for retinal blood vessel detection

Retinal blood vessels are important structures in ophthalmological images. Many detection methods are available, but the results are not always satisfactory. In this paper, we present a novel model based method for blood vessel detection in retinal images. It is based on a Laplace and thresholding segmentation step, followed by a classification step to improve performance. The last step assures incorporation of the inner part of large vessels with specular reflection. The method gives a sensitivity of 92% with a specificity of 91%. The method can be optimized for the specific properties of the blood vessels in the image and it allows for detection of vessels that appear to be split due to specular reflection.     

Modelling of pulsatile blood flow in arterial trees of retinal vasculature

The paper presents a numerical investigation of the pulsatile blood flow in the detailed arterial vasculatures of a mouse retina using the mathematical model based on frequency domain incorporating an appropriate outlet boundary impedance at the end of the terminal vessels of the arterial trees. The viscosity in the vessels was evaluated considering the Fahraeus-Lindqvist effect, the plasma skimming effect and in vivo viscosity effect in the microcirculation. Comparative studies of the pulsatile circulation were carried out for cases of rigid vessels, constant viscosity, zero and non-zero outlet boundary impedances. In addition, the dependence of the oscillating input impedance at the inlet of the arterial trees on angular frequencies of the oscillation and vessel elasticises was also studied. The study shows that the pressure wave continues in the pre-capillary vessels throughout the retina. In elastic vessels, the amplitude of oscillatory velocity and wall shear stress in larger vessels and in vessels at the periphery region of the retina is amplified. The pulsatile blood flow is significantly influenced by the outlet boundary (or load) impedance which simulates the effect of the capillary and venous vasculatures. The oscillating input impedance at the inlet of the arterial trees is also found to be dependent on the angular frequency and the Young modulus of the vessel segment. Insights into the potential variations of the dynamic responses of the system under retinal pathological condition of arteriosclerosis may be inferred from the findings of the present study.     

眼底数码图像血管分割预处理关键技术研究

目的本文以眼底图像为研究对象,针对眼底图像血管分割的预处理关键技术进行了深入的研究和探讨。方法首先通过RGB通道选择,显示出较为清晰的血管图像,并通过自适应直方图均衡化增强图像对比度。其次,利用灰度倒置和灰度增强,使细小血管更加清晰,并采用空域滤波对血管轮廓边缘进行增强。最后进行图像填充,去除背景中的噪声干扰。结果通过MATLAB仿真实验,获得了血管脉络清晰的眼底数码图像。结论本研究为血管的自动分割以及血管成分的定性和定量分析奠定了基础。     

Retinal Vessel Detection and Measurement for Computer-aided Medical Diagnosis

Since blood vessel detection and characteristic measurement for ocular retinal images is a fundamental problem in computer-aided medical diagnosis, automated algorithms/systems for vessel detection and measurement are always demanded. To support computer-aided diagnosis, an integrated approach/solution for vessel detection and diameter measurement is presented and validated. In the proposed approach, a Dempster–Shafer (D–S)-based edge detector is developed to obtain initial vessel edge information and an accurate vascular map for a retinal image. Then, the appropriate path and the centerline of a vessel of interest are identified automatically through graph search. Once the vessel path has been identified, the diameter of the vessel will be measured accordingly by the algorithm in real time. To achieve more accurate edge detection and diameter measurement, mixed Gaussian-matched filters are designed to refine the initial detection and measures. Other important medical indices of retinal vessels can also be calculated accordingly based on detection and measurement results. The efficiency of the proposed algorithm was validated by the retinal images obtained from different public databases. Experimental results show that the vessel detection rate of the algorithm is 100 % for large vessels and 89.9 % for small vessels, and the error rate on vessel diameter measurement is less than 5 %, which are all well within the acceptable range of deviation among the human graders.     

Digital subtraction peripheral angiography using image stacking: initial clinical results.

Using clinically acquired x-ray angiography image sequences, we compared three algorithms for creating a single diagnostic quality image that combined input images containing flowing contrast agent. These image-stacking algorithms were: maximum opacity with the minimum gray-scale value across time recorded at each spatial location, (REC) recursive temporal filtering followed by a maximum opacity operation, and (AMF) an approximate matched filter consisting of a convolution with a kernel approximating the matched filter followed by a maximum opacity operation. Eighteen clinical exams of the peripheral arteries of the legs were evaluated. AMF gave 2.7 times greater contrast to noise ratio than the single best subtraction image and 1.3 times improvement over REC, the second best stacking algorithm. This is consistent with previous simulations showing that AMF performs nearly equal to the optimal result from matched filtering without the well-known limitations. For example, unlike matched filtering, AMF filter coefficients were obtained automatically using an image-processing algorithm. AMF effectively brought out small collateral arteries, otherwise difficult to see, without degrading artery sharpness or stenosis grading. Comparing results using reduced and full contrast agent volumes demonstrated that contrast agent load could be reduced to one-third of the conventional amount with AMF processing. By simulating reduced x-ray exposures on clinical exams, we determined that x-ray exposure could be reduced by 80% with AMF processing. We conclude that AMF is a promising, potential technique for reducing contrast agent load and for improving vessel visibility, both very important characteristics for vascular imaging.     

Application of Morphological Bit Planes in Retinal Blood Vessel Extraction

The appearance of the retinal blood vessels is an important diagnostic indicator of various clinical disorders of the eye and the body. Retinal blood vessels have been shown to provide evidence in terms of change in diameter, branching angles, or tortuosity, as a result of ophthalmic disease. This paper reports the development for an automated method for segmentation of blood vessels in retinal images. A unique combination of methods for retinal blood vessel skeleton detection and multidirectional morphological bit plane slicing is presented to extract the blood vessels from the color retinal images. The skeleton of main vessels is extracted by the application of directional differential operators and then evaluation of combination of derivative signs and average derivative values. Mathematical morphology has been materialized as a proficient technique for quantifying the retinal vasculature in ocular fundus images. A multidirectional top-hat operator with rotating structuring elements is used to emphasize the vessels in a particular direction, and information is extracted using bit plane slicing. An iterative region growing method is applied to integrate the main skeleton and the images resulting from bit plane slicing of vessel direction-dependent morphological filters. The approach is tested on two publicly available databases DRIVE and STARE. Average accuracy achieved by the proposed method is 0.9423 for both the databases with significant values of sensitivity and specificity also; the algorithm outperforms the second human observer in terms of precision of segmented vessel tree.     

Method for modelling cerebral blood vessels and their bifurcations using circular,homogeneous, generalised cylinders

A method for automatic modelling of blood vessels and their bifurcations from 3D scans of the brain is presented. The method is three-step procedure. First, a skeleton of the cerebral blood vessels is developed, and then the surfaces of the blood vessels are located using an active contour approach. The active contour approach uses circular homogeneous generalised cylinders (CHGCs) to model the thin, elongated blood vessels. Finally, a novel method for modelling the surfaces of the bifurcations in a vessel tree is presented. The method was tested on simulated data: a computed tomography angiography (CTA) and four magnetic resonance angiography (MRA) volumes. Furthermore, the method was tested on ten magnetic resonance images (MRIs) to demonstrate its robustness. The test on the simulated data indicated that the approach for the surface modelling of vessels had a mean radius error of less than 0.1 mm and a mean localisation error of 0.1 mm. Surface models evaluated by an expert in vascular neurosurgery were found to have a smooth appearance and generally agreed with the image data. The test on the MRI scans indicated that the method performed well in noisy environments.     

Reconstruction and representation of caudal vasculature of zebrafish embryo from confocal scanning laser fluorescence microscopic images

Three-dimensional (3D) reconstruction from a series of sections is an important technique in medical imaging, particularly for visualization of blood vessels from angiography. Here, we present a framework for automatic segmentation and registration of different kind of blood vessels from 2-day-old zebrafish embryos. Series of optical sections were acquired from confocal microscopy with the blood vessels labeled by fluorescent microbeads (0.02 microm) injected into blood stream of 2-day-old zebrafish embryos. Blood vessels were extracted and their morphological parameters, including length and diameter, were calculated. At the same time, individual blood vessels were registered automatically. Vasculature was represented by attributed vessel represent graph (AVRG), which contained morphological data and connectivity of every blood vessel. Using AVRG to represent a vasculature made the comparison between vasculatures of different embryos more easy. Visualization, as well as quantification, of reconstructed 3D model of AVRG was presented in an interactive interface. The framework was implemented by Visual C++ as Windows-based program.     

基于相位一致性模型的眼底微动脉瘤检测方法

眼底微动脉瘤是糖尿病视网膜病变最早期的症状,准确检测眼底图像中的微动脉瘤对糖尿病视网膜病变的筛查具有重要意义。提出一种基于相位一致性模型的微动脉瘤检测方法。首先采用相位一致性模型获取微动脉瘤候选者,然后通过构建灰度剖面图去除图像中血管片段等无关信息,从而筛选出真正的微动脉瘤。通过对ROC网站提供的50幅眼底图像进行实验,在图像水平上实现了灵敏度94%、特异性100%、准确率96%的检测效果。结果表明,该方法对图像的亮度、对比度不敏感,能够高效自动地检测出彩色眼底图像中的微动脉瘤。