Please use this identifier to cite or link to this item:
|Title:||Acquisition and Processing of Retinal Images for Blood Vessel Detection|
|Keywords:||Bioscience & Engineering|
|Abstract:||The quantification of blood vessel features, such as length, width, tortuosity and branching pattern, among others, can provide new insights to diagnose and stage pathologies which affect the morphological and functional characteristics of blood vessels. However, when the vascular network is complex, or the number of images is large, manual measurements can become tiresome or even impossible. A feasible solution is the use of automated analysis, which is nowadays commonly accepted by the medical community. Retinal images are influenced by all the factors that affect the body vasculature in general. Moreover the eye is the very unique region of the human body where the vascular condition can be directly observed in vivo. Along with the fovea and optic disc, the vascular tree constitutes one of the main features of an ocular fundus image and several of its properties are noticeably affected by worldwide major diseases such as diabetes, hypertension, and arteriosclerosis. Other eye diseases, such as choroidal neovascularization and retinal artery occlusion, also induce changes in the retinal vasculature. For the reasons here stated, the segmentation of retinal images can be a valuable aid for the detection and follow-up of several pathological states, as a mean of detecting and characterizing over time any of the changes in the blood vessels. The extraction of blood vessels from retinal images is an important and challenging task in medical analysis and diagnosis. In this thesis we presented methods of automated blood vessels segmentation from retinal fundus images. Such automated tool should prove useful to eye care specialists for purposes of patient screening, treatment evaluation, and clinical study. We studied two different approaches here Template matching methods and Karnel or windowing method. In template matching method we use many temples such as Sobel, Prewitt1, Prewitt2, Robinson, Frei & Chen and Kirsch Templates. In second proposed method of automated retinal blood vessel detection, consists of five steps: a) acquiring retinal fundus images from fundus camera, b) pre-processing of the images c) convolution of retinal images with the two-dimensional matched filter and second order Gaussian filter d) Thresholding, e) Length filtering followed by adaptive median filtering. Gray-level profile of the cross section of a blood vessel is approximated by a Gaussian shaped curve. With the pre-processing technique unwanted noises are removed and brightness of the images are adjusted. Blood vessels usually have poor local contrast, two-dimensional matched filter and second order gaussian filter karnels is designed to convolve with the original image in order to enhance the blood vessels. Since vessels may appear in any orientation, a set of 2-D segment profiles in equiangular rotations is used. Twelve different karnels have been constructed to span all possible blood vessel orientations. Blood vessels are segmented by thresholding technique. Many thresholding schemes are studied to do this job efficiently. Adaptive median filtering followed Length filtering are applied then to remove misclassified pixels.|
|Appears in Collections:||Bioscience & Engineering - Master's Degree Theses|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.