Plato-NeRF is a computer vision system that uses lidar and machine learning to reconstruct 3D scenes, including hidden objects, from a single camera view by exploiting shadows. This technique has potential applications in autonomous vehicles, AR/VR headsets, and warehouse robots.