Classical ghost imaging (GI) is a computational imaging technique developed in the visible-light optics community. It is very similar to single-pixel cameras but relies on patterned illumination rather than patterned detection. I will describe how my collaborators and I have translated this technology to weakly interacting probes such as x rays and neutrons. We have developed transmission ghost imaging and ghost tomography, as well as emission GI through x-ray fluorescence. Recently we have also experimentally demonstrated GI in reverse or “ghost projection” that may be the most practical method to produce images with these weakly interacting probes. GI has potential for super-resolution and, importantly for ionising radiation, dose reduction. I will present a proof-of-principle demonstration of super-resolution GI using neutrons and outline the paths for GI to realise dose reduction. A significant aspect of this is that GI fidelity is subject to the patterns employed; I will show experimentally the advantages and disadvantages of various types of aperiodic patterns.