Abstract
Recent years have seen the rapid growth of new approaches to optical imaging, with an emphasis on extracting three-dimensional (3D) information from what is normally a two-dimensional (2D) image capture. Perhaps most importantly, the rise of computational imaging, defined as the synergistic design of optical systems in conjunction with image reconstruction algorithms, enables both new physical layouts of optical components and new algorithms to be implemented. This paper concerns the convergence of two advances: the development of transparent photodetectors with high responsivity, and the rapid expansion of the capabilities of machine learning including the development of powerful neural networks. In particular, we demonstrate that the use of transparent photodetector arrays stacked vertically along the optical axis of an imaging system, called a focal stack, together with a feedforward neural network, provides a powerful new approach to real-time 3D optical imaging including object tracking. The focal stack imaging system is realized through the development of graphene transparent photodetector arrays. As a proof-of concept, 3D tracking of point-like objects was successfully demonstrated with multilayer feedforward neural networks, which was then extended for tracking of multi-point objects in position. Our computer model further demonstrates how this optical system can track extended objects in 3D, highlighting the promise of combining nanophotonic devices, new optical system designs, and machine learning for new frontiers in 3D imaging.