An estimated one-third of the global population is now obese or overweight, a figure that has grown approximately 30% in adults since the 1980s. Obesity has been linked to numerous comorbidities and non-communicable diseases such as type 2 diabetes, high blood pressure, stroke, and coronary heart disease, the number one leading cause of death globally. The excess accumulation of adipose tissue, or fat tissue, is the key characteristic of obesity. The traditional whole-fat measure Body Mass Index is the gold standard in fat quantification but does not consistently predict health risks. Rather, recent research has revealed that certain localized regions of adipose tissue have a direct link to obesity-related illnesses and more accurately predict health risks. The quantification of localized adipose tissue is crucial to understanding and assessing the health risks related to obesity. Magnetic resonance imaging (MRI) and ultrasound are two common imaging modalities used for adipose tissue quantification. Manual segmentation of the adipose tissue in either modality can be time-consuming and expensive for a radiologist to perform. In both modalities, automated segmentation algorithms greatly speed up the process of fat quantification.
In this study, first a semi-automated segmentation algorithm is developed to identify three adipose tissue regions from Dixon MRI images and subsequently evaluated against three expert observers. The algorithm was performed on 10 subjects and compared to three expert observers on 635 images in total. Comparison between expert observers and the computer results demonstrate that the computer algorithm performs, on average, at least as good as an expert observer.
The second part of this study contains preliminary work towards the segmentation of adipose tissue around the heart in echocardiograms (ultrasound images). Particularly, an IQ to B-mode conversion algorithm was developed to visualize the raw transducer data from the ultrasound machine. With the generated B-mode images, a semi-automated segmentation algorithm is developed to identify the endocardial boundary in parasternal short-axis echocardiograms and evaluated against an expert observer. A statistically significant correlation was found between the expert observer and computer results, instilling confidence in the accuracy of the computer segmentation algorithm. In the final portion of this study, a software package is created in Python 3.7 that allows loading, segmentation, and visualization of Dixon MRI and echocardiograms in one easy-to-use program.
|Commitee:||LeAnder, Robert, Umbaugh, Scott|
|School:||Southern Illinois University at Edwardsville|
|School Location:||United States -- Illinois|
|Source:||MAI 58/06M(E), Masters Abstracts International|
|Subjects:||Engineering, Electrical engineering|
|Keywords:||Adipose tissue, Dixon MRI, Echocardiography, Segmentation, Subcutaneous, Visceral|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be