ABSTRACT
This chapter provides a mathematical and conceptual treatment of image reconstruction from projections, a method commonly used in electron paramagnetic resonance imaging. Projection reconstruction has been used in several fields. The ultimate goal of image reconstruction is the recovery of some function f that vanishes outside a region bounded in an n-dimensional space. Projection reconstruction is the recovery of a function from its projections. Filtered backprojection reconstruction is a direct implementation of Radon's inversion formulas. With the exception of dead-time corrections, Fourier image reconstruction of the data proceeds in a manner exactly analogous to that used in nuclear magnetic resonance image reconstruction. One additional category is included since it directs the reader to a general class of reconstruction algorithms that readily accommodate projection data. The choice of an interpolator involves a tradeoff between accuracy and resolution in the image and the computational time required to perform the interpolation.
