Deep Learning-Based Automated MRI Image Segmentation

منو

" Deep Learning-Based Automated MRI Image Segmentation "
Zhao, Tianyun Huang, Chuan

Document Type	:	Latin Dissertation
Language of Document	:	English
Record Number	:	1110233
Doc. No	:	TLpq2478808272
Main Entry	:	Huang, Chuan
	:	Zhao, Tianyun
Title & Author	:	Deep Learning-Based Automated MRI Image Segmentation\ Zhao, TianyunHuang, Chuan
College	:	State University of New York at Stony Brook
Date	:	2020
student score	:	2020
Degree	:	M.S.
Page No	:	54
Abstract	:	This thesis's work focuses on the clinical applications of U-Net on MRI image segmentation, with various dimensions and depths based on the clinical question. Chapter 1 is a brief introduction to magnetic resonance imaging. Chapter 2 introduces convolutional neural network and U-Net. Chapter 3 focuses on 2D U-Net based lumbar segmentation used for a radiomic pipeline to predict osteoporosis; reliability was assessed by comparing to manual segmentation. Chapter 4 includes a 3D U-Net developed to help with automated breast density measurement. The segmentation reliability was assessed by studying test-retest reliability of the breast density. Chapter 5 contains a 2.5D U-Net used to segment fat pads for future radiomic study attempting to predict osteoarthritis progression.
Subject	:	Artificial intelligence
	:	Biomedical engineering
	:	Medical imaging