Phase Retrieval Method Based on Deep Learning with Single Image Training in Holographic Data Storage

In this paper, a non-interferometric phase retrieval method based on deep learning is proposed. The relationship between the diffraction intensity and the encoded data page is established through the convolutional neural network (CNN). After training, the phase information can be detected directly from a single diffraction image. Moreover, by designing the encoded data page, only one pair of intensity-phase images are needed to finish the train of the neural network. The proposed method solves the problem that supervised end-to-end neural networks rely on a large amount of training data and cannot be used in practical applications due to the lack of sufficient numbers of training images from the experiment.