Approaches to oscillating-gradient fast scanning in two-dimensional nuclear magnetic resonance imaging
Various improvements to the theory and practice of oscillating-gradient NMR imaging are presented. In reconstruction theory, a relationship is established between the two major types of reconstruction, these being the one- and two-dimensional Fourier-transform methods. Using the analysis that results, a study is made of exact reconstruction methods, where exact is defined to mean as good as slow-imaging methods or to be perfect reconstruction when there is no noise or experimental errors in the data. Such exactness is tested by using discrete sets of data in computer-simulation files, making from them simulated NMR signals, then reconstructing these signals into images that have no Gibbs error due to the discreteness of the original data and are correct copies of the original data to within the accuracy of the computer's arithmetic capability. A first exact method was developed using the single large one-dimensional Fourier transform. A second exact method, due to Haacke, with faster speed and lower signal-to-noise ratio was studied and improved. A third exact reconstruction method that used the two-dimensional Fourier transform is given for comparison.
- Research Organization:
- California Univ., Irvine, CA (USA)
- OSTI ID:
- 7066898
- Country of Publication:
- United States
- Language:
- English
Similar Records
A comparison of the noise characteristics of projection reconstruction and two-dimensional fourier transformations in NMR imaging
Two-dimensional maximum entropy reconstruction of radio brightness