High-Quality Computational Ghost Imaging with a Conditional GAN

In this study, we demonstrated a framework for improving the image quality of computational ghost imaging (CGI) that used conditional generative adversarial network (cGAN). With set low-quality images from CGI system and their corresponding ground-truth counterparts, cGAN was trained could generate high-quality new images. The results showed compared with traditional method based on compressed sensing, greatly improved when sampling ratio low.