Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Lane Department of Computer Science and Electrical Engineering

Committee Chair

Donald A Adjeroh

Committee Co-Chair

Susan K Lemieux


Magnetic Resonance Imaging (MRI) is one of the widely used medical technologies for diagnosis of various degenerative diseases like Alzheimer's disease. In degenerative diseases the brain atrophies resulting in enlarged ventricles. The segmentation of the ventricle and the total ventricle volume measurement can be important in the analysis of the neurological disease severity and in studies of animal models of disease used to advance treatments for people. This thesis is focused on using Gabor wavelets for ventricular measurements in a rabbit model of Alzheimer's disease.;Standard intensity-based tools like Statistical Parametric Mapping (SPM) used for human brain segmentation requires a priori knowledge of the data. As there is no available rabbit brain atlas incorporated into the algorithm, SPM failed for rabbit brain segmentation. Hence, texture-based segmentation was used to solve the problem. Gabor wavelet decomposition of the textures in an image provides a powerful mathematical tool for feature extraction. The features contain important frequency information that can be estimated from biological parameters to choose the wavelet set. Our goal is to develop an automated method for extracting ventricular structures from whole-head rabbit MRI.;Rabbit brain segmentation is a complex process in low contrast MRI images. Hence the Gabor wavelets were first studied on the high resolution digital histological rabbit brain images. With the success of ventricle segmentation in the histology images, the Gabor parameters were modified for the rabbit brain MR images. Initially, manually guided segmentation results were obtained using Gabor wavelets. This work continued further to automate the process of brain region of interest (ROI) selection to automate the ventricle extraction process.