Abstract
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.
© 2022 IEEE
PDF ArticleMore Like This
Jianying Hao, Xiao Lin, Ryushi Fujimura, Yoshito Tanaka, Soki Hirayama, Xiaodi Tan, and Tsutomu Shimura
21a_C205_2 JSAP-Optica Joint Symposia (JSAP) 2022
Jianying Hao, Xiao Lin, Soki Hirayama, Ryushi Fujimura, Yoshito Tanaka, Xiaodi Tan, and Tsutomu Shimura
JW5A.74 Frontiers in Optics (FiO) 2022
Mingjun Xiang, Lingxiao Wang, Yu Sha, Hui Yuan, Kai Zhou, and Hartmut G. Roskos
Tu4A.4 Digital Holography and Three-Dimensional Imaging (DH) 2022