China 
Africa 
Explore chapters and articles related to this topic
Studies on Registration and Fusion of Retinal Images
Published in Rick S. Blum, Zheng Liu, Multi-Sensor Image Fusion and Its Applications, 2018
France Laliberté, Langis Gagnon
In Ref. 9, Symlet wavelets of lengths 4 to 20 are tested and the best results are obtained with the largest length. Using the consistency check technique for the high pass coefficient combination, the authors fused three pairs of images. The first fused image, which is a case of diabetic retinopathy, emphasizes the spatial relationship between the exudates, which are only visible in the color image, and the small focal areas of hypofluorescence visible on the fluorescein angiogram, which are indicative of capillary closure or early ischemia. This could allow the determination of the position of the exudates with respect to the fovea, which is relevant in deciding if the patient should undergo laser treatment. For the second case (cystoid macular edema), some subtle changes in the retinal pigment epithelium can be seen in the color image. The fluorescein angiogram shows swelling in the intraretinal tissue, which appears as a petaloid. The fused image allows the positioning of the swelling in the green band. For the third case (branch retinal vein occlusion), the color image shows a vein that appears thicker and tortuous. Hemorrhages are also visible near the macula and in the upper half of the retina. On the fluorescein angiogram, the abnormal blood return produces leakages of dye, which are scattered through the retina, and swollen pockets of tissues above the macula. The fused image emphasizes the relationship between the leakage sources and the hemorrhages.
Protein, Cellular and Soft Tissue Interactions with Polyurethanes
Published in Nina M. K. Lamba, Kimberly A. Woodhouse, Stuart L. Cooper, Polyurethanes in Biomedical Applications, 2017
Nina M. K. Lamba, Kimberly A. Woodhouse, Stuart L. Cooper
Ex vivo studies are relevant to the study of blood-biomaterial interactions. Ex vivo procedures involve the extracorporeal contact of blood with the test material.9,13–16 Blood usually flows directly onto or through the test piece. Blood contact may be single pass, where the blood is not returned to the donor, or recirculation, where the material is contained within a closed loop and the blood is returned to the donor. The presence of flow allows the study of thrombus formation and embolization, and the degree of occlusion of the test circuit. Although a single pass human ex vivo procedure has been reported,9,16 animal models are predominantly used, which introduces differences between the hematological profiles of the respective model and human subjects.17,18 Subhuman primate models are preferred for such studies, as this reduces the hematological and evolutionary differences. However, this is not always practical, on the grounds of the relative scarcity of such models and high cost. Canine and rabbit models have therefore been employed more widely, in such studies. Canine models are considered sensitive models for thrombus formation as the adhesion of canine platelets to biomaterials is significantly higher than human platelets. Therefore, studies in a canine model are considered to present a worst case scenario. Most ex vivo studies have examined protein adsorption, platelet deposition, or thrombus formation. Figure 1 shows a schematic of a canine ex vivo experiment.15
Imaging the Living Eye
Published in Margarida M. Barroso, Xavier Intes, In Vivo, 2020
Brian T. Soetikno, Lisa Beckmann, Hao F. Zhang
Branch retinal vein occlusion (BRVO) is an ischemic disease of the retina that results from decreased blood flow in one or more major retinal veins. Figures 14.9D–F show OCTA images of a BRVO model created by imaging-guided photocoagulation, taken with a NIR OCT system (Soetikno et al., 2017). Figure 14.9D shows an en face OCTA image of the vasculature before occlusion. The major arteries and veins are labeled as an a and v, respectively. The vein branch, indicated by the white circle, was occluded. The OCTA image after occlusion (Figure 14.9E) shows capillary nonperfusion in the area (white-dashed area) surrounding the occlusion site. This area increased on Day 1 as detected by the OCTA image shown in Figure 14.9F.
Review on the current treatment status of vein of Galen malformations and future directions in research and treatment
Published in Expert Review of Medical Devices, 2021
Panagiotis Primikiris, Georgios Hadjigeorgiou, Maria Tsamopoulou, Alessandra Biondi, Christina Iosif
As for the functional outcome, Berenstein et al. [3] reported 66.6% (30/45) of patients who were neurologically and developmentally intact presenting a score of 4 in the scale suggested by Jones et al. [101], 20% (9/45) had a score of 3 with mild impairment, and 8.9% (4/45) had a score of 2 with moderate impairment [3]. The authors reported radiological cure in 82% of patients (37/45), with 13,3% (6/45) of patients being still under treatment and 4,4% (2/45) mortality. Yan et al. [32] reported in their review and meta-analysis of the literature, complete occlusion in 57% (95% CI 48%–65%; I2 = 68,2%) of cases and partial occlusion in 43% (95% CI 34%–51%; I2 = 70.7%) of cases. Out of 667 patients who underwent embolization between 1987 and 2014, the authors reported, based on the 5-point scale according to Jones et al. [101], a good outcome (scores between 4 and 3) in 68% of cases (95% CI 61%–76%; I2 = 77.8%) and a poor outcome (scores of 2, 1, 0) in 31% (95% CI 24%–38%; I2 = 75.6%).
Advances in CSF shunt devices and their assessment for the treatment of hydrocephalus
Published in Expert Review of Medical Devices, 2021
Kamran Aghayev, Sheikh MA Iqbal, Waseem Asghar, Bunyad Shahmurzada, Frank D. Vrionis
As mentioned above tissue ingrowth into catheter holes is the main cause of occlusion. Therefore, the optimal placement should aim the ventricular region away from tissues that can potentially occlude VC. An important step in handling this issue was objectivization of the of catheter position assessment. Most of the proposed classification schemes designate the position into 3–4 categories, such as optimal (grade 1) – floating in the CSF away from the choroid plexus and ventricular wall, suboptimal (grade 2) – catheter in contact with ventricular wall or choroid plexus, highly suboptimal (grade 3) – catheter being partially embedded into brain parenchyma or failed/unacceptable (grade 4) – catheter completely outside of the ventricle [26–28]. It has been shown that shunt survival (i.e. time spent from insertion until revision) is well correlated with the position (grade) of VC placement [26,29–33].
A novel context-aware system to improve driver’s field of view in urban traffic networks
Published in Journal of Intelligent Transportation Systems, 2022
A. Nourbakhshrezaei, M. Jadidi, M. R. Delavar, B. Moshiri
The other part of the process is applying smoothing and blurring that helps to reduce noises and edges also known as high-frequency content reduction. It might sound like a counterintuitive part by reducing edges and details, however, it is useful to find the object that we are interested in (moving objects). In another word, it helps to find the larger structural objects in the image. The miner movements such as leaves movements will be removed from the scene. There are two options that are commonly used for blurring; Simple average, and Gaussian. The result of the simple average filter is less naturally filtered than the result of the Gaussian filter. In a Gaussian filter, a weighted mean is used for blurring and the weights are biased to the pixels that are closer to the central pixel. Moreover, it eliminates noises that are following the Gaussian distribution. The Gaussian filter is used in this paper since the result of this filter is naturally blurred. Equation 4 shows the Gaussian formula. where x and y are the distances to the horizontal and vertical central pixel of the kernel. σ is the standard deviation of the Gaussian kernel. The 9 × 9 kernel was used for this operation. Figure 7d shows the result of applying this filter. The difference between each filtered frame and the RI was calculated using Equation (1). Figure 7e shows the differences between the filtered frame and the RI. After this stage, the threshold was changed to gray levels. The threshold is the simplest method of image segmentation. This is a non-linear operation that converts a gray-scale image into a binary image in which two levels are assigned to pixels that are set below or above the specified threshold value. Figure 7f shows the threshold. After mapping grayscales to the binary image, some parts of the images that are representing the segmentation of the moving objects might have small holes. One of the approaches to fill small holes in objects is called dilation. The dilation operation was used as a morphological transformation to prevent broken objects from being used.The result of the dilation filter is the maximum value of other pixels on the kernel. As an example, a pixel that belongs to class 0 in the input image will be set as class 1 if at least one of the pixels in the neighborhood belongs to class 1. Figure 7g illustrates the threshold before dilation and Figure 7h presents the threshold after the dilation operation. As the system sends the trajectory data of the moving objects to the nearest vehicles, a trajectory line will be mapped on the mobile phone application that shows the path of moving object. The occlusion issue is one of the critical challenges in this research. The dilation filter is used as a morphological operation to decrease occlusion. However, as the main goal is to inform vehicles about the existing moving objects around, and by considering the fact that the occlusion issue occurs when two objects are so close to each other, we conclude that vehicles or drivers can still make a better decision even if the system considers two close moving objects as a single object in some cases. Because knowing that there is a moving object around can apply the same action as when two or more moving objects exist.