Usage
  • 48 views
  • 75 downloads

Computer-Aided Detection of Epileptogenic Lesions Based on Brain MRI Image Analysis

  • Author / Creator
    Wang, Huiquan
  • The invention of magnetic resonance imaging (MRI) has advanced the diagnosis of diseases dramatically as it is capable of producing high-quality images of soft tissues non-invasively and safely. However, the high volume of MRI images also means heavy workloads for physicians, which results in slow diagnosis and even misdiagnosis. As a consequence, computer-aided detection (CAD) systems, which can process the input images by computer and generate diagnosis results for radiologists to improve the diagnostic efficiency and accuracy, are in high demand. Some CAD systems for brain disorders, such as brain tumors, Alzheimer’s disease, and multiple sclerosis, have already been proposed and actively researched. But there is no CAD system for screening epileptogenic lesions even though there is a large population of epilepsy patients worldwide. In this thesis, several automated detection techniques that can aid build a CAD system for epileptogenic lesions are developed.The contributions of this thesis are threefold. First, an automated detection technique for cavernous malformation based on MRI image analysis is proposed. The technique is a coarse to fine framework. After obtaining the candidate cavernous malformation regions using skull stripping and template matching, image features including texture, the histogram of oriented gradients, and local binary patterns of each candidate are extracted and then classified using support vector machines to exclude false detections. Second, an automated detection technique for mesial temporal sclerosis (MTS) is proposed. After the segmentation and 3D reconstruction of the hippocampus, the hippocampal volume, shape, and cerebrospinal fluid features are calculated. Two support vector machines are then used to detect and lateralize MTS. Third, an automated detection technique for focal cortical dysplasia is proposed using MRI images and deep learning. After bias field correction, intensity normalization, and registration with a brain atlas, cortical patches are extracted and fed to a deep convolutional neural network with five convolutional layers, a max pooling layer, and two fully-connected layers. Image patches with focal cortical dysplasia are detected after classification and postprocessing.The proposed techniques are evaluated thoroughly using both a publicly available MRI dataset and images obtained from the University of Alberta Hospital. Experimental results indicate that the proposed techniques could provide superior performance compared with existing methods in the literature, thus showing potential in assisting neuroradiologists in the detection of epileptogenic lesions using brain MRI analysis.

  • Subjects / Keywords
  • Graduation date
    Spring 2019
  • Type of Item
    Thesis
  • Degree
    Doctor of Philosophy
  • DOI
    https://doi.org/10.7939/r3-mrt6-t535
  • License
    Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.