Abstract
We propose to use autoencoder and a two-stage training strategy for image reconstruction in multimode fiber imaging system. Compared with the conventional method of U-net, this approach achieves higher reconstruction accuracy and faster convergence speed.
© 2020 The Author(s)
PDF ArticleMore Like This
Xinlei Zhang, Zhenming Yu, Ziyi Meng, Kaiqiang Ding, Zhenyu Ju, and Kun Xu
T4A.4 Asia Communications and Photonics Conference (ACP) 2019
Baurzhan Muminov and Luat T. Vuong
FW7A.3 Frontiers in Optics (FiO) 2020
Le Yang, Yangyang Xiang, Jiali Yang, Junhui Li, Mingying Lan, and Li Gao
FD5.3 Frontiers in Optics (FiO) 2023