Image Superresolution in Single-Pixel Imaging with Generative Adversarial Networks

Imaging of physical objects using single-pixel cameras is an actively developing area at the intersection of optics and computational mathematics. Image restoration algorithms used in single-pixel cameras usually provide low resolution due to practical limitations on realistic computing resources. In this paper, we demonstrate an increase in the resolution of images obtained in single-pixel imaging using a generative adversarial neural network and discuss its application using the example of chest X-ray images from the MedMNIST dataset.