0.8% Nyquist computational ghost imaging via non-experimental deep learning

We present a framework for computational ghost imaging based on deep learningand customized pink noise speckle patterns. The deep neural network in thiswork, which can learn the sensing model and enhance image reconstructionquality, is trained merely by simulation. To demonstrate the sub-Nyquist levelin our work, the conventional computational ghost imaging results,reconstructed imaging results using white noise and pink noise via deeplearning are compared under multiple sampling rates at different noiseconditions. We show that the proposed scheme can provide high-quality imageswith a sampling rate of 0.8% even when the object is outside the trainingdataset, and it is robust to noisy environments. This method is excellent forvarious applications, particularly those that require a low sampling rate, fastreconstruction efficiency, or experience strong noise interference.