Time-efficient object recognition in quantum ghost imaging

Ghost imaging utilises the optical correlations between two beams to form an image of an object. One beam interacts with the object, and the other possesses spatial information. It is not possible to reconstruct the image from either beam separately, but only via the correlations between them [1] . The origin of such correlations does not constitute a restriction, thus it has been demonstrated in the quantum and classical regimes [2] . Due to the low light levels and the possibility of non-degenerate photon correlations, quantum ghost imaging has great potential in areas such as biological imaging. This particular use has been hindered by the extensive image reconstruction times. Proposed solutions include the use of compressive sensing, novel reconstruction algorithms, and deep learning techniques. Here, we followed a three fold strategy to reduce the time required to reconstruct the image: optimizing the experimental parameters, performing image processing and introducing a recognition algorithm.