Lee, Wen-Chi Christina
(2004)
Intramodality and Intermodality Registration of the Liver.
Master's Thesis, University of Pittsburgh.
(Unpublished)
Abstract
Radiological imaging of the liver is an important medical problem. The ever increasing amount of data acquired when imaging the liver makes integration of information desirable and crucial in building up a comprehensive diagnostic picture of the patient. The foundation of all such image integration is image registration.</br></br>Image registration is the process of aligning images so that corresponding features can easily be related, including: (1) landmark-driven methods, (2) surface-based methods, and (3) voxel similarity-based methods. A challenge with registering the liver is that the liver moves within the abdomen with respiration. Therefore any effective alignment of the liver must first separate the liver from the remainder of the image. With this as a constraint, the goal of this research effort was to determine the feasibility and efficacy of surface-based and voxel similarity-based schemes in registering abdominal CT and MR images with and without contrast.</br></br>A multi-scale surface fitting technique was implemented based on the Head and Hat algorithm. Equivalent surfaces from the in vivo images were extracted manually. The hand segmentation approach was validated by ensuring the volume of the liver of each image from the same patient was consistently within +/- 7% of one another. The registration transformation was determined by iteratively transforming the hat with respect to the head surface, until the closest fit of the hat onto the head was found. </br></br>In addition, registration of in vivo CT and MR images was performed using a multi-resolution mutual information scheme distributed with the ITK Insight software package (National Library of Medicine, Bethesda, MD). As an independent measure of registration accuracy, the mean displacement of automatically selected point landmarks was evaluated. For the multi-resolution mutual information approach, mean misregistrations were in the range of 7.7-8.4mm for CT-CT intramodality registration, 8.2mm for MR-MR intramodality registration, and 14.0-18.9mm for CT-MR intermodality registration. For the Head and Hat surface registration scheme, mean misregistrations were in the range of 9.6-11.1mm for CT-CT intramodality registration, 9.2-12.4mm for MR-MR intramodality registration, and 15.2-19.0mm for MR-CT intermodality registration.
Share
Citation/Export: |
|
Social Networking: |
|
Details
Item Type: |
University of Pittsburgh ETD
|
Status: |
Unpublished |
Creators/Authors: |
Creators | Email | Pitt Username | ORCID  |
---|
Lee, Wen-Chi Christina | WCL3@pitt.edu | WCL3 | |
|
ETD Committee: |
|
Date: |
2 February 2004 |
Date Type: |
Completion |
Defense Date: |
7 November 2003 |
Approval Date: |
2 February 2004 |
Submission Date: |
2 December 2003 |
Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
Institution: |
University of Pittsburgh |
Schools and Programs: |
Swanson School of Engineering > Bioengineering |
Degree: |
MSBeng - Master of Science in Bioengineering |
Thesis Type: |
Master's Thesis |
Refereed: |
Yes |
Uncontrolled Keywords: |
liver registration; mutual information; surface fitting |
Other ID: |
http://etd.library.pitt.edu/ETD/available/etd-12022003-131104/, etd-12022003-131104 |
Date Deposited: |
10 Nov 2011 20:07 |
Last Modified: |
15 Nov 2016 13:52 |
URI: |
http://d-scholarship.pitt.edu/id/eprint/9940 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |