LAUSR

In this paper, we present a method for reconstructing video from event data alone captured by an event camera. We employ point-spread-function engineering in the imaging optics to compress spatiotemporal information into each event. Dynamic scenes are reconstructed solely from the sparse, signed-polarity event data using a compressive sensing algorithm. We numerically validate the concept and characterize the video-imaging performance of the compressive event camera. This method enhances event-to-video reconstruction while preserving the advantages of event cameras-high speed, high dynamic range, and high energy efficiency.