The challenge of models which account for optical SEDs is to identify an intrinsic distribution of stars and dust which is consistent with the observed integrated SEDs and surface brightness distributions. To assess whether a particular choice for the geometry of stars and dust is consistent with observations, calculations of the transfer of radiation through the galaxy must be done. Thus, models must incorporate a radiative transfer code. Different radiation transfer codes have been discussed in the review talk of Kylafis & Xilouris .
The model of Xilouris et al.  derived the relative distribution of stars and dust by fitting optical/NIR images of galaxies, assuming the simplest, yet realistic, distributions of stars and dust that could be used to describe such systems (i.e., exponential disks plus a de Vaucouleurs bulge; see Fig. 1, left panel). This work was done for edge-on galaxies, which allows not only the scalelength but also the scaleheights of dust and stars to be extracted. We should also mention that, since opacities were independently derived at each wavelength, the extinction law was extracted for each modelled galaxy, giving some constraints on the dust model. Apart from the detailed knowledge of the distributions of stars and dust in individual galaxies, some general trends became evident (Xilouris et al. ), namely: all the galaxies modeled were able to reproduce Milky Way extinction laws; the optically emitting stars have a scaleheight which is about twice that of the dust; the dust scalelength is about 1.4 times larger than that of the stars and the dust is more radially extended than the stars. This inference about the relative radial extent of stars and dust can in principle be tested through FIR observations of the dust emission from the regions beyond the stellar disk. This has been done in the case on NGC 891 (Popescu & Tuffs ; see also Fig. 2 from the review of Tuffs & Popescu ).
Figure 1. Left panel: Images of NGC 891 in K, J, I, V, B bands (top to bottom) taken from Xilouris et al. . The left hand half of each galaxy image is the model image and the right half is the real galaxy image (folded). Right panel: Comparison of R- and H-band Monte Carlo model images of UGC 7321 (second and bottom rows) with real data at these wave bands (top and third row), taken from Matthews & Wood .
Synthetic images have also been compared with observed images by Matthews & Wood  (see Fig. 1, right panel), but in this work the geometry was fixed and only the face-on optical depth was varied.