Characterization Goal

Plaque characterization MRI is a relatively new field. Thus, the goals have been to (1) to develop a validation technique such that MR images taken from atherosclerotic plaques in vivo can be correlated to histological examination of plaque specimens, (2) explore hardware and software approaches such that both vessel luminal and outer wall boundaries can be accurately identified on the images and detailed information of plaque compositions visualized, and (3) determine the accuracy and...

Theory and Algorithm of Ellipsoidal Filtering

Consider an image I and its Taylor expansion in the neighborhood of a point x0 as I xo Sxo, a I xq, op Sxor Vo J SxpT Ho J Sxo ---- 2.1 This expansion approximates the structure of the image up to the second order. V0 and H0,a are the gradient vector and Hessian matrix of the image computed at x0 at scale a. To calculate the differential operators I, the scale-space framework is adapted. In this framework, the differentiation is defined as a convolution with Derivative of Gaussian DoG . This is...

Flow Enhanced Gradient Echo Sequence

Calcifications Removed From Sinuses

Flow-enhanced techniques refer to the gradient echo-based imaging sequences that are typically used to acquire MR angiograms. The enhancement bright blood is due to the time-of-flight TOF effects of blood flow. Techniques such as gradient recalled echo GRE , spoiled gradient recalled echo SPGR , TOF, and true FISP are all variations of such techniques. With flow enhancement, the lumen appears hyperintense relative to the adjacent vessel wall. Compared to spin echo SE sequences, bright blood...

Ground Truth from Conflicting Observers

To determine the best way to combine different observer's segmentation results, one must first consider what constitutes a correct and incorrect vessel segmentation. Essentially, there are three possible ways to define correct vessel segmentation for a particular pixel. A conservative method is to declare pixels to be part of a vessel or part of the background when all observers' segmentations agree. If such total agreement exists, then that pixel would be marked as vessel or background in the...

Quantifying the Vulnerable Plaque

Figure 8.3 depicts histological specimens from three different carotid plaques. The first example has a thin fibrous cap separating a lipid-rich core from the lumen. The second exhibits calcified nodules extruding into the lumen. The third has large, densely packed neovasculature in the shoulder region of the plaque. Although these plaques show considerable compositional variability, all have been characterized as vulnerable because of these features. Assuming that the features of vulnerable...

Figure 41

Axial images from a knee MRA Left A The bright blood two-dimensional TOF, acquired in the axial plane. Right B The black blood T2-weighted three-dimensional FSE, acquired in the sagittal plane, reformatted here in the axial plane. The imaging time for three-dimensional FSE was comparable with two-dimensional TOF but covered a larger volume. In A , the arrows point to vessels but they also have some black neighborhood areas around them. In B , the black-blood image shows perfect vessel capture....

IVUS for Abdominal Aorta Aneurysm Analysis

Abdominal Aneurysm Ivus

Intravascular ultrasound images are proving their importance for coronary vessels. However, one should not think that cardiology is the only useful application of these images. Given their excellent properties intraoperative images as well as providing morphological information about tissue structure and composition , IVUS images are generating increasing clinical interest in diagnosis and therapy of abdominal aortic aneurysms. It has been reported that determination of fixation sites and...

Key Words in Chapter 5 Segmentation of Retinal Fundus Vasculature in Nonmydriatic Camera Images Using Wavelets

Wavelet transform mathematical morphology proliferative retinopathy complexity preprocessing postprocessing Resolution gray-level singularity Mexican hat derivative of Gaussian edge Key Words in Chapter 6 Automated Model-Based Segmentation, Tracing, and Analysis of Retinal Vasculature from Digital Fundus Images

Info

The major disadvantages of the system included 1. The model was not capable of understanding the forking or bifurcation behavior of the vessels. Thus, it was very difficult to design all the possible three-dimensional configurations, at least the detection of these deviations from the model and the re-initialization of the tracking procedure along these branches. 2. The window must be adaptive in size and only one vessel should be observed during the tracking procedure. 3. The fundamental...

References

Sutherland. Intracoronary ultrasound current state of the art, Br. Heart J., 73 2 , 1995. 2. A. Colombo, P. Hall, et al. Intravascular stenting without anticoagulation accomplished with intravascular ultrasound guidance, Circulation, 91 2 , 1676-1688, 1995. 3. X. Zhang, C. R. McKay, and M. Sonka. Tissue characterization in intravascular ultrasound, IEEE Trans. Med. Imag., 17 6 , 889-899, 1998. 4. J. K. Morton. The Cardiac Catheterization Handbook, Mosby-Year...

N J V N

Where k is the iteration count, pk and pk 1 denote the current and new locations of the trace, a is a step size, sk e 0, 1, 2 N 1 is an integer index specifying one of N discretized angular directions usually N 16 , and k is a lateral displacement vector that centers the new point pk 1 on the vessel. In Figure 6.15, this is illustrated for a pair of intersecting vessels. The left and right directional kernels at 0 and 45 are also illustrated for the lower branch in Figure 6.15. The angle at...

Vascular Image Segmentation

One of the most common adaptations of global thresholding is connected components 24 . Connected components refers to a class of methods that augment intensity thresholding with the requirement that the points belonging to an object of interest must be continuous that is, they must be linked together by other object points having appropriate intensities in the volume. The behavior of this method is intuitive. It is illustrated in Figure 11.15 and in Figure 11.16. Variations on this method...