Please use this identifier to cite or link to this item:
Title: Segmentation and Classification of Multiple Sclerosis Tissues in Two of FLAIR Series Using Feed Forward Neural Network
Authors: Hanem Mahmoud Abd Elrahman Ellethy 
Supervisor: Dr. Mohammed Alroussan
Keywords: Sclerosis Tissues , FLAIR Series , Neural Network
Issue Date: 2016
Publisher:  Kuwait university - college of graduate studies
Abstract: iv ABSTRACT Multiple Sclerosis (MS) is an inflammatory and nervous disorder disease. Young adults are the most attacked by MS and get disability effects. The Magnetic Resonance Imaging (MRI) images of brain are the most effective way for both MS detection and follow up. Identification and segmentation of MRI images are done usually by neurologists' visual inspection. An automatic segmentation and classification system with minimum human intervention can be a helpful tool in this field. The objective of this thesis is to implement an intelligent system that automatically segments and classifies MS tissues (lesions). Image preprocessing techniques are used to prepare MRI images for the segmentation step. Global threshold and mathematical operation are used in segmentation phase to set the tissues to the feature extraction step and dataset building. An effective and simple segmentation technique is used to simplify the preprocessing steps and reduce the number of unwanted tissues thus reducing the overall processed dataset. It does not need the removal of Skull and Cerebrospinal Fluid (CSF) or separate gray and white matter as previous studies required. A feed forward neural network is used to train and test the proposed system. Our proposed system is built based on Fluid Attenuated Inversion Recovery (FLAIR) series only. Moreover, SAG (sagittal) FLAIR series system is proposed for the first time in MS segmentation and classification field. Significant recognition rate is achieved for the proposed systems which reached up to 98.5%. A relatively high Dice Coefficient (DC) value is carried out when testing new images; namely 0.71±0.18.
Appears in Programs:0612 Computer Engineering

Files in This Item:
File Description SizeFormat 
Hanem_Ellethy_ thesis.pdf1,8 MBAdobe PDFView/Open    Request a copy
Show full item record

Page view(s)

Last Week
Last month
checked on Sep 27, 2020


checked on Sep 27, 2020

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.