Deep-learning denoising computational ghost imaging

•A deep learning denoising computational ghost imaging method is proposed.•A deep neural network is developed for ghost imaging image reconstruction.•The object image is restored directly from the one-dimensional bucket signals.•The proposed scheme have wide potential applications.•The proposed scheme can obtain clear images with sub-Nyquist sampling ratios. We propose a deep learning denoising computational ghost imaging (CGI) method to obtain a clear object image with a sub-Nyquist sampling ratio. We develop an end-to-end deep neural network (DDANet) for CGI image reconstruction. DDANet uses a one-dimensional (1-D) bucket signals (BSs) and multiple tunable noise-level maps as input, and outputs a clear image. We train DDANet with simulated BSs and ground-truth pairs, and then retrieve the object image directly from an experimental obtained 1-D BSs. The effectiveness of the proposed method is experimentally investigated. The proposed method has practical applications in image denoising and enhancement of the CGI and single-pixel computational imaging.