In magnetic resonance electrical impedance tomography (MR-EIT), crosssectional images of a conductivity distribution are reconstructed. When current is injected to a conductor, it generates a magnetic field, which can be measured by a magnetic resonance imaging (MRI) scanner. MR-EIT reconstruction algorithms can be grouped into two: current density based reconstruction algorithms (Type-I) and magnetic flux density based reconstruction algorithms
(Type-II). The aim of this study is to implement a series of reconstruction algorithms for MR-EIT, proposed by several research groups, and compare their performance under the same circumstances. Five direct and one iterative Type-I
algorithms, and an iterative Type-II algorithm are investigated. Reconstruction errors and spatial resolution are quantified and compared. Noise levels corresponding to system SNR 60, 30 and 20 are considered. Iterative algorithms
provide the lowest errors for the noise- free case. For the noisy cases, the iterative Type-I algorithm yields a lower error than the Type-II, although it can diverge for SNR lower than 20. Both of them suffer significant blurring effects, especially at SNR 20. Another two algorithms make use of integration in the reconstruction, producing intermediate errors, but with high blurring effects. Equipotential lines are calculated for two reconstruction algorithms. These lines may not be found
accurately when SNR is lower than 20. Another disadvantage is that some pixels may not be covered and, therefore, cannot be reconstructed. Finally, the algorithm involving the solution of a linear system provides the less blurred
images with intermediate error values. It is also very robust against noise.
| Identifer | oai:union.ndltd.org:METU/oai:etd.lib.metu.edu.tr:http://etd.lib.metu.edu.tr/upload/12608250/index.pdf |
| Date | 01 February 2007 |
| Creators | Martin Lorca, Dario |
| Contributors | Eyuboglu, Murat |
| Publisher | METU |
| Source Sets | Middle East Technical Univ. |
| Language | English |
| Detected Language | English |
| Type | M.S. Thesis |
| Format | text/pdf |
| Rights | To liberate the content for public access |
Page generated in 0.0015 seconds