Symmetries, Redundant Information, and Noise Control in Tomographic Imaging

Xiaochuan Pan
DOI: https://doi.org/10.4006/1.3028850
2000-01-01
Physics Essays
Abstract:Tomography is a noninvasive imaging approach for revealing the internal structures or processes of an object. In this paper, I compare the tasks of image reconstruction from the data functions measured in two tomographic-imaging modalities: computed tomography (CT) and diffraction tomography (DT). In particular, I investigate the symmetries and redundant information inherent in these data functions and exploit such information for reducing image noise, imaging time, and radiation dose. It is demonstrated that, although CT and DT are based upon entirely different physical principles, the methods for image reconstruction in CT and DT are remarkably similar because their respective data functions possess highly similar symmetries.
What problem does this paper attempt to address?