A new data-driven model has been developed to determine 1D/2D particle size distribution (PSD) from measured FBRM chord-length distribution (CLD) data. The structure of the model consists of three steps: first, the measured CLDs are compressed down to a small set of parameters; second, these parameters are correlated with low order moments or a small number of percentiles of the PSD using regression models; third, the PSD low order moments are used as input variables, and a two layer-network sub-model is built to predict the PSD in a form of a histogram. Two key aspects of this modeling strategy are noteworthy, namely the construction of specialized parameterized functions (herein called generating functions) that reduce the number of parameters needed to train the model's two layer network component, as well as its ability to model a 2D size distribution. To demonstrate this paradigm, the model was used to determine the PSD of particles with an elongated morphology. It is shown that even with a limited data set, models could be trained to predict PSD generated by laser diffraction, minor size PSD generated by image analysis, and 2D minor-major PSD measured by image analysis. (C) 2017 Elsevier Ltd. All rights reserved.