We realize a numerical investigation of error rates in rough particle sizing using interferometric out-of-focus imaging. We show that an automatic image processing based on a 2D-Fourier transform of the patterns enables size estimations with accuracy better than algorithms based on 2D-autocorrelation of the patterns. When ILIDS enables size determination of droplets whose out-of-focus images exhibit 2 or 3 bright fringes, we show that the minimum size measurable in the case irregular rough particles is around two or three times higher than for spherical droplets. © 2017 Elsevier GmbH