In nuclear safeguards, one of the most important tasks is the regular inspection of Spent Nuclear Fuel (SNF) assemblies discharged from nuclear power reactors, in order to verify that nuclear material is not removed illicitly. At Chalmers University of Technology, a novel non-intrusive methodology for accurate detection and characterization of the content of a SNF assembly is under investigation. This methodology is based on a new detector that can measure inside the assemblies the spatial distribution of the neutrons emitted by the spent nuclear fuel, and the use of Artificial Neural Network (ANN) models to process such measurements and identify whether nuclear material is missing or not.