Multiresolution encoding with frames in MRI /
Magnetic Resonance Imaging (MRI) has become a major imaging modality in medicine and has a wide application to other disciplines. The conventional Fourier imaging technique, however, is very slow. Many new imaging techniques have been proposed to overcome this drawback. Observing that these new enco...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1999.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=731686541&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | Magnetic Resonance Imaging (MRI) has become a major imaging modality in medicine and has a wide application to other disciplines. The conventional Fourier imaging technique, however, is very slow. Many new imaging techniques have been proposed to overcome this drawback. Observing that these new encoding techniques uses bases that are in the L₂([]) space, based on the frame theory, we treat all of them from a united perspective and categorize them as "frame imaging'' . Though frame imaging has many potential advantages, there exist some difficulties in its implementation, among which radiofrequency (RF) pulse design and image reconstruction are most prominent. We propose a near-resonance expansion of the solution to the Bloch equations in the presence of an RF pulse. The first-order approximation explicitly demonstrates the nonlinear nature of the Bloch equations and precisely relates the excitation profiles with the RF pulse when the flip angle is less than []/2. The image reconstruction problem is solved with the frame theory. After measuring the excitation profiles as discrete sequences, which constitute a frame, the dual frame is constructed and used for image reconstruction. Analyses of the SNR and resolution are carried out with the frame algorithm. Results show that frame imaging leads to a lower SNR than Fourier imaging but provides the capability of multiresolution. After modifying the pulse sequence of single-slice frame imaging, we propose a numerical approach to basis selection. SNR and imaging speed are the criteria used to compare different bases. With our pulse design method and image reconstruction algorithm, the imaging process is simulated and analyzed for three orders of spline bees. Simulation shows that frame imaging can be more than ten times faster than Fourier imaging. For comparison, a dynamic imaging algorithm is also designed and implemented for normal frame imaging that does not have modification to the pulse sequence. The dynamic imaging algorithm speeds up imaging. Dynamic imaging, however, is quite slower than the modified version of frame imaging. |
|---|---|
| Item Description: | Vita. "Major Subject: Electrical Engineering". |
| Physical Description: | xii, 101 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilm Inc. |
| Bibliography: | Includes bibliographical references (leaves 92-96). |