Abstract

Imaging through a scattering medium is of great significance in many areas. Especially, speckle correlation imaging has been valued for its noninvasiveness. In this work, we report a deep learning solution that incorporates the physical model and an additional regularization for high-fidelity speckle correlation imaging. Without large-scale data to train, the physical model and regularization prior provide a correct direction for neural network to precisely reconstruct hidden objects from speckle under different scattering scenarios and noise levels. Experimental results demonstrate that the proposed method presents a significant advance in improving generalization and combating the invasion of noise.

© 2023 Optica Publishing Group

DeepSCI: scalable speckle correlation imaging using physics-enhanced deep learning

Zhiwei Tang, Fei Wang, ZhenFeng Fu, Shanshan Zheng, Ying Jin, and Guohai Situ
Opt. Lett. 48(9) 2285-2288 (2023)

Reconstructing images of two adjacent objects passing through scattering medium via deep learning

Xuetian Lai, Qiongyao Li, Ziyang Chen, Xiaopeng Shao, and Jixiong Pu
Opt. Express 29(26) 43280-43291 (2021)

Plug-and-play algorithm for imaging through scattering media under ambient light interference

Kai Ma, Xia Wang, Si He, and Linhao Li
Opt. Lett. 48(7) 1754-1757 (2023)

Projecting colorful images through scattering media via deep learning

Sitong Huang, Jian Wang, Daixuan Wu, Yin Huang, and Yuecheng Shen
Opt. Express 31(22) 36745-36753 (2023)

Non-invasive imaging through scattering medium and around corners beyond 3D memory effect

YingJie Shi, EnLai Guo, Ming Sun, LianFa Bai, and Jing Han
Opt. Lett. 47(17) 4363-4366 (2022)