In characterization of metal nanoparticle doped spherical composites, the two-dimensional nature of transmission electron microscopy (TEM) images leads to ambiguities about the true location of the nanoparticles. Walking-in of simulated projections in comparison with actual TEM images leads to quantitative results such as location-dependent particle sizes and particle number density. This method takes advantage of the strength of fuzzy neural network computations via the human hunter-gatherer's visual system's evolved superiority while still allowing quantitative results by use of exact numerical simulations.
