Abstract

Methods for diffraction tomography, which are both numerically feasible and mathematically rigorous, have required that the scattering object only interact weakly with the incident field¹,². The approximations used include the first-order Born and the Rytov methods, which are rarely appropriate in practice, thus limiting their usefulness. More general methods or "exact" inversion procedures have proved extremely difficult, if not impossible, to implement. Recent developments have been made which do extend the domain of validity of the Born and Rytov approximations to some extent² these are based on the distorted- wave Born or Rytov approximations. These methods assume that the strongly scattering component of the object is known, and that an unknown perturbation to this satisfies the Born or Rytov approximation. Thus, some a priori information about the scatterer must be acquired and which represents a (strongly scattering) background, against which small fluctuations in permittivity are imaged.

© 1998 Optical Society of America