Three-dimensional optical diffusion imaging in a Bayesian framework

Summary form only given. Spectroscopic specificity and low photon energy are particularly attractive features of optical tissue imaging. This potential has spurred interest in optical diffusion imaging, where the spatially-dependent scatter and absorption are to be reconstructed using measured data with a number of source and detector locations using a diffusion equation forward model. While robust and accurate inversions have been achieved for 2-D test problems, practical optical diffusion geometries require 3-D reconstructions. This need led to the development of efficient multigrid reconstruction algorithms. In this work we demonstrate accurate 3-D reconstruction using a Bayesian prior model with an iteratively estimated data term variance.