Probabilistic model of multiple light scattering based on rigorous computation of the first and the second moments of photon coordinates

We consider a concise method based on recurrent relations that permit rigorous computing of the first and the second moments of the components of the vector locating a randomly walking photon in an infinite homogeneous light-scattering medium. On assumption that the components obey a three-dimensional Gaussian distribution a probability density for the photon locations at the Nth scattering event can readily be written down and the light-intensity distribution in the medium may be calculated. The results from theoretical analyses are compared with the solution of a light-diffusion equation and with results of Monte Carlo simulations and show a better fit with simulated data than the diffusion approximation.