Not yet reviewed by Pith; the record is open.

Observational data and theoretical calculations show that significant small-scale substructures are present in dark molecular clouds. These inhomogeneities can provide precious hints on the physical conditions inside the clouds, but can also severely bias extinction measurements. We present NICEST, a novel method to account and correct for inhomogeneities in molecular cloud extinction studies. The method, tested against numerical simulations, removes almost completely the biases introduced by sub-pixel structures and by the contamination of foreground stars. We applied NICEST to 2MASS data of the Pipe molecular complex. The map thereby obtained shows significantly higher (up to 0.41 mag in A_K) extinction peaks than the standard NICER (Lombardi et al. 2001) map. This first application confirms that substructures in nearby molecular clouds, if not accounted for, can significantly bias extinction measurements in regions with A_K > 1 mag; the effect, moreover, is expected to increase in more distant molecular cloud, because of the poorer physical resolution achievable.