Spatial resolution analysis for time-domain diffuse optical tomography based on a perturbation model

We estimate a limit to spatial resolution in time-domain diffuse optical tomography (DOT) based on a perturbation model by Lyubimov. In the context of structure reconstruction accuracy we consider and compare three approaches to the inverse DOT problem. The first reconstructs diffuse tomograms from straight lines; the second does it from curvilinear average trajectories of photons; and the third uses the total banana-like distributions of photon trajectories. For getting estimates to resolution, we derive analytical expressions for the point spread function and the modulation transfer function, and perform a numerical experiment to reconstruct rectangular scattering objects with circular absorbing inhomogeneities. It is shown that reconstruction with photon trajectory distributions instead of straight lines gives a gain of about order of magnitude in resolution and attains the accuracy of multistep nonlinear DOT algorithms.