convergence of FBP as the discretization step size goes to zero,
which turns out to be the best possible bound by previous results.
Sean Holman
University of Washington
Date: May 1, 2006
Rieder and Faridani consider the FBP for semidiscrete data from the
Radon transform in two dimensions. Semidiscrete in this case means
that only the first variable, the radial one, is discretized. By
considering certain specialized convolution kernels, they are able to
prove that the FBP algorithm, using these types of kernels, can be
rewritten as the exact Radon inversion formula with two interpolation
operators added in. Then, using techniques from approximation theory
about interpolation operators they are able to prove a bound for the
convergence of FBP as the discretization step size goes to zero,
which turns out to be the best possible bound by previous results.