All ETDs from UAB

Advisory Committee Chair

Murat M Tanik

Advisory Committee Members

Karthikeyan Lingasubramanian

Jerry A Higgs

Buren E Wells

Document Type

Dissertation

Date of Award

2019

Degree Name by School

Doctor of Philosophy (PhD) School of Engineering

Abstract

Deep Learning (DL) has rapidly become a methodology of choice for analyzing medical images and increasingly attracts researcher’s attention in the medical research community. The brain is one of the most important organs in the human body. Within the context of human brain disease, research and care, accurately detecting, evaluating and segmenting human brain abnormalities play an important role in brain disease diagnosis, prognosis, and treatment planning. A significant challenge in developing good brain abnormalities segmentation methods is the high variation of brain abnormalities such as differences in shape, size, location, appearance, and regularity. Deep Learning approach ad-dresses this challenge very well by allowing operating flexibility on variable datasets to accurately segment brain abnormalities. In this dissertation, we present a study of developing Deep Learning approaches to perform human brain abnormalities segmentation tasks. We begin with a brief introduction of the background of Deep Learning. Then, we proposed and developed Deep Learning approaches to segment human brain tumor, tumor substructures, and brain White Matter Hyper-intensities (WMH) regions automatically using Magnetic Resonance Imaging (MRI) scans. We have demonstrated the proposed models have obtained promising results both qualitatively and quantitatively. Another important aspect is that we describe a Deep Learning Framework for Medical Image Segmentation (DLFMIS) that provides a road map for efficiently designing and developing Deep Learning approaches to perform medical image segmentation tasks. With clearly defined roles and assigned responsibilities, DLFMIS framework enables researchers to deliver their Deep Learning applications fast and meet their expectations.

Included in

Engineering Commons

Share

COinS