Advances and Challenges of Single-Pixel Imaging Based on Deep Learning

Abstract

Single-pixel imaging technology can capture images at wavelengths outside the reach of conventional focal plane array detectors. However, the limited image quality and lengthy computational times for iterative reconstruction still hinder its practical application. Recently, single-pixel imaging based on deep learning has attracted a lot of attention due to its exceptional reconstruction quality and fast reconstruction speed. In this review, an overview of the current status, and the latest advancements of deep learning technologies in the field of single-pixel imaging are provided. Initially, the fundamental principles of single-pixel imaging and deep learning, followed by a discussion of their integration and associated benefits are presented. Subsequently, a comprehensive review is conducted on the advancements of deep learning in various domains of single-pixel imaging, covering super-resolution single-pixel imaging, single-pixel imaging through scattering media, photon-level single-pixel imaging, optical encryption based on single-pixel imaging, color single-pixel imaging, and image-free sensing. Finally, open challenges and potential solutions are discussed.