Browsing by Author "Assefa, Dawit (PhD.)"

Now showing 1 - 6 of 6
  • No Thumbnail Available
    Item
    An Automated Segmentation at Retinal Images for Use in Diabetic Retinopathy Studies
    (Addis Ababa University, 2014-10) Moges, Daniel; Assefa, Dawit (PhD.)
    Automated computer aided detection of retinal lesions associated with Diabetic Retinopathy (DR) offers many potential benefits. In a screening setting, it allows the examination of large number of images in less time and more objectively than traditional observer driven techniques. In a clinical setting, it can be an important diagnostic aid by reducing the workload of trained graders and other costs. However, the segmentation of major pathological structures and their subsequent follow–ups are not easy because of various artifacts such as presence of anatomical structures with highly correlated pixels with that of lesion, illumination variability, noise and movement of the eye during multiple visits by the patient. This study presents a novel mathematical scheme for analysis of color retinal images acquired through digital fundus cameras from patients treated for DR. The proposed scheme uses a holistic representation of the color images in the three (trinion) space and applies trinion based Fourier transforms to extract useful imaging features for the purpose of classification and segmentation of retinal images. A suitable color space transformation and a way of extracting robust higher order features are included in the method. The scheme has been applied in analyzing images acquired from standard retinal image databases. Results have showed that the algorithm achieved 86.06% sensitivity, 96.06% specificity, and 92.65% accuracy for pixel base segmentation of Hard Exudates (HEs) the most prevalent lesions that appears in the earliest stages of DR, while it achieved 96.67% sensitivity, 100% specificity and 97.3% accuracy for image base classification of abnormalities due to DR.
  • No Thumbnail Available
    Item
    Design and Development of Advanced External Fixator for Treatment of Femoral Bone Fracture
    (Addis Ababa University, 2017-02) Misganaw, Talaksew; Assefa, Dawit (PhD.)
    The incidence of accidents causing major injuries globally is increasing every year. Previous studies indicated that more than 90% of the incidences happen in the developing world. In case when the scope of the accident is high, that often creates fractures particularly on strong and heavy bones (femoral and tibia in the lower extremities). In this regard different methods are available to treat fractured bones including tractions, internal and external fixators and their variants. Traction is one of the oldest treatment methods currently used in many clinics as a temporary treatment of long bone fractures through suspension of weights using ropes over pulleys and attached to the patient bed. Such a procedure does not allow patients movement and could result in many known complications: depression, bed sores, pneumonia and urinary infections to mention a few. The internal fixators like intramedullary nails are the gold standard implants for fixation of long bone fractures. However high cost, lack of adequate professionals and instruments in low resource settings, location of the accident, need for multiple surgeries and cultural and/or physiological patient/family related issues for operation significantly hinder patients from getting the service. External fixation management applies aligning/realigning fractures using pins, wires, clamps, and bars or rings. It has advantages of simplicity, ease of adjustability, and increased access for fracture related wound care and wound monitoring. However, the solid bars often used to make fractured fragments rigid has no flexibility, do not employ a force on the muscles tensioned by the fixators and have no micro movement setup. This creates immediate functionality problem after recovery and reduce fast healing process for the fractured bone because loading has considerable effect on tissue repair and remodeling. A thorough investigation of literatures on fixators (especially external fixators) carried out in the current study showed that applying continuous load/force on fractured limbs increases the functionality of the muscles during the bone healing process thereby reducing the rehabilitation time after recovery. In this regard, based on solidwork, an advanced external fixator has been proposed in the current study that makes use of a spring load mechanism to exert a continuous load/force. Detailed 2d and 3d design and simulation of the mechanism has been presented. The proposed external fixator also comes with a traction functionality.
  • No Thumbnail Available
    Item
    Holistic Multi-Parametric Magnetic Resonance Image Processing for Identification of Neoplasm and Normal Brain Tissues
    (Addis Ababa University, 2016-12) Chernet, Teferi; Assefa, Dawit (PhD.)
    Abreakthrough advance in Magnetic Resonance Imaging (MRI) into standard clinical practices has enabled greater accuracy in delineating tumor volumes. The technique allows autonomous detection through noninvasive imaging of malignant brain tumors like gliomas and similar other body tumors for the requirement of efficient and effective procedures like biopsy and radiation therapy as well as for quantification of patients’ responses to treatment. The role of MR imaging in regard of characterization schemes of neoplasms has been tremendous. There are many approaches suggested in the literature for this purpose, one of which is color processing concept. Most color images have three components and each pixel/voxel on such color images can be represented as a three-component vector. Analogously, an MRI voxel, for example, can be seen as a vector with as many channels as available MR image parameters: T1, T2, PD, ADC, rcbv, Ktrans, T2*, etc. And there can be various methods of representing such images as colors. Hence using these methodologies and by combining multiple MRI parameters as we need, we can maximize the information we want to display. The research study exploits advanced mathematical and image analysis assets for use in analysis of MP-MR medical images making use of algorithmic procedures. Accuracy and potentials of the method is compared against gold standards (available ground truth). The data used is composed of standard co-registered multiple MR image parameters taken from patients treated for the highest grade and most aggressive of the gliomas. This approach comes with a great potential to assist accurate and effective detection, visualization, and analysis of useful medical image information for physicians and/or researchers as well. Even though MR images are the subjects of interest in this study, in principle the method developed could be applied to other medical images with multiple components acquired through different modalities. Keywords: Quaternion Fourier Transform, Trinion Fourier Transform, Image Processing, Principal Component Analysis.
  • No Thumbnail Available
    Item
    Medical Image Segmentation Using Spectral Clustering Based on Hypercomplex Analysis
    (Addis Ababa University, 2016-10) Getaneh, Samuel; Assefa, Dawit (PhD.)
    Images are destined to be segmented, either with our visual system (as we do it every day) or algorithmically using computing machines. However, as critical as medical images, image segmentation has to be consistent which we do not usually have. Therefore, to bring consistency, physicians need some kind of consistent clues about medical images they are analyzing. Particularly, when the images considered are colors with multiple channels/bands, developing such a consistent segmentation tool could take more effort and that often requires a rigorous mathematical computation and algorithm free from human cognition. To create such algorithm for color medical image segmentation, primarily the images have to be considered as collection of pixels that have to be represented properly and holistically with no separation of color components. Secondly, the relationships of pixels have to be defined holistically. Finally, based on these defined relationships, the pixels have to be grouped and segmented. Hence, in order to perform holistic segmentation, quaternion based representation of color pixels and quaternion based spectral clustering technique has been proposed in this thesis. Test results have shown that the proposed scheme can use and also can be used in machine learning applications of image segmentation and pattern recognition. Keywords: Quaternion Rotation, Spectral Clustering, Image Processing, Color Distance Measurement
  • No Thumbnail Available
    Item
    Modeling Simulating and Quantification of Optical Images at Ultrasound Contrast Bubbles
    (Addis Ababa University, 2017-07) Abraham, Habtamu; Assefa, Dawit (PhD.)
    Microbubble based ultrasound contrast agents are being used in clinical settings to Enhance backscattered ultrasound signal from blood pool during perfusion and blood flow Measurements. Often individual bubbles are characterized optically by recording their Vibrations with a fast framing camera and direct quantitative information on their dynamic Behavior can be derived. Nonetheless, when a three-dimensional object, stack of infinitely thin Two-dimensional layers, is imaged through a microscope, the image formed onto the charge Coupled device element consists of contributions from all layers. If a bubble is larger than the Depth of focus, the part of the bubble above the focal plane influences the image formation And therefore the bubble size measured. Thus, this thesis presents a methodology to compute Two-dimensional image formation from three-dimensional objects, hollow spheres and find Under which circumstances the optical image formation leads to a significant deviation in Measurement of the actual size. Two-dimension image formations of the three-dimensional Object was computed by convolving the slices of an artificial object, hollow spheres with the Respective weighted point spread function and summing all convolved slices. Finally, image Processing was applied the optical image formed to quantify the object size and a systematic Error was observed for objects in focus with radius 1:65mm. Also it was concluded that Even though a three-dimensional object is in focus, there is discrepancy of up to 0.66% in size Measurement. In addition, size measurement of an object for the same shift above the focus and Below the focus could differ by up to 3.6%. Moreover, defocusing up to 90% could result up To 64.7 mean percentage error. The results reveal that defocusing above 25% severely deviates Size measurements. This thesis hopes to offer a standard for quantification of optical images of Three-dimensional objects, and in the future actual size of an object could be measured from its Defocused optical images.
  • No Thumbnail Available
    Item
    Robust Edge Detection Applied to Multi-Parametric Magnetic Resonance Images
    (Addis Ababa University, 2016-11) Damenu, Serkalem; Assefa, Dawit (PhD.)
    Various edge detection techniques for color images that have been proposed in the last two decades showed that color images contain 10% additional edge information as compared to their gray scale counterparts. Edge detection is one of the most commonly used operations in medical color image processing. Efficient and accurate edge detection leads to increased performance of subsequent image processing techniques, including image segmentation, object-based image coding, and image retrieval. A color image edge detection algorithm is proposed in this paper that introduces a robust and automated scheme that makes use of higher order statistical features derived from locally computed trinion Fourier transforms. The proposed scheme uses a holistic vectorial representation of the color images in the three (trinion) space and applies trinion based Fourier transforms to extract useful imaging features for the purpose of edge detection of multi parametric magnetic resonance images (MP-MRI). A suitable color space transformation and a way of extracting robust higher order features are included in the method. Performance of the proposed scheme is compared against classical edge detection methods and other vectorial approaches which have been proposed in the literature. Results have shown that none of the classical as well as the other vectorial approaches were able to detect useful edges. Application of the method is shown in edge detection on MP-MR images of brain scans of patients treated for Glioblastoma multiforme (GBM). The algorithm performs well in detecting the tumor edges, that was (qualitatively) in a very good agreement with the ground truth information which is the oncologist‘s contour drawn manually. Key Words: Color Image processing, Edge detection, Trinion, Quaternion, Magnetic Resonance Image.