Document Type


Date of Degree

Spring 2016

Degree Name

MS (Master of Science)

Degree In

Biomedical Engineering

First Advisor

Vincent Magnotta


Quantifying tissue volumes in pediatric brains from magnetic resonance (MR) images can provide insight into etiology and onset of neurological disease. Unbiased volumetric analysis can be applied to large population studies when automated image processing is possible. Standard segmentation strategies using adult atlases fail to account for varying tissue contrasts and types associated with the rapid growth and maturational changes seen in early neurodevelopment. The goal of this project was to develop an automated pipeline and two age-specific atlases capable of providing accurate tissue classification despite these challenges.

The automated pipeline consisted of a stepwise initial atlas-to-subject registration, expectation maximization (EM) atlas based segmentation, and a post-processing level set segmentation for improved white/gray matter separation. This level set segmentation is a 3D and multiphase adaptation of a 2D method intended for use on images with the types of intensity Inhomogeneities found in MR images.

The initial tissue maps required to determine spatial priors for the one-year-old atlas were created by manually cleaning the results of an adult atlas and the automated pipeline. Additional tissue maps were incrementally added until the spatial priors were sufficiently representative. The neonate atlas was similarly created, starting with the one-year-old atlas.

Public Abstract

Automated image processing can provide unbiased data in large scale population studies to contribute insight into the progression and onset of neurologic disease. Due to the rapid growth and development of the human brain in the first few years of life, automated medical image processing of magnetic resonance (MR) images from infant subjects to investigate neurological conditions is especially challenging. The goal of this work was to create methods and tools suitable to overcome these challenges. An atlas is a collection of prior knowledge that is used in tissue classification.

Because of the fast growth of the brain, a collection of knowledge that is representative of a limited age group can be more precise and useful. For this reason, an atlas was created for the use of two age groups: neonate (images taken shortly after birth) and one-year-old subjects (images taken at 10-18 months of age). To determine the likelihood of a tissue based on its relative spatial positioning and expected image intensity, a set of subjects from each age group were segmented and manually corrected.

MR images of these age groups also suffer from excessive noise and motion artifacts. This can complicate the differentiation of white and gray matter in the cortex. To address this issue, a level set segmentation method, which can be particularly robust to the complex topologies found in the cortex, was adapted to correct tissue classification in an additional post-processing step.


publicabstract, atlas-based segmentation, automated pipeline, Image Processing, level set segmentation, Magnetic Resonance images, tissue classification


viii, 32




Copyright 2016 Andrew Metzger