The ITER Radial Neutron Camera (RNC) is a neutron diagnostic intended for the measurement of the neutron emissivity radial profile and the estimate of the total fusion power. This paper presents a proof of-principle method based on neural networks to estimate the neutron emissivity profile in different ITER scenarios and for different RNC architectures. The design, optimization and training of the implemented neural network is presented together with a decision algorithm to select, among the multiple trained neural networks, which one provides the inverted neutron emissivity profile closest to the input one. Examples are given for a selection of ITER scenarios and RNC architectures. The results from this study indicate that neural networks for the neutron emissivity recognition in ITER can achieve an accuracy and precision within the spatial and temporal requirements set by ITER for such a diagnostic.