Understanding ghost imaging from a machine learning perspective
Journal of the Optical Society of America A · 01.01.2025 · 42 (1) · 89
DOI: 10.1364/josaa.537888 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the Optical Society of America A, online erschienen am 24.12.2024. Verfasst von Wenjie Liu und 7 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 32 Literaturangaben.
Abstract (Originalsprache)
In this paper, we analyze the mechanism of computational ghost imaging and its mathematical similarity to the linear regression process in machine learning. We point out that the imaging process in computational ghost imaging essentially involves solving a linear regression problem, where the bucket detector plays the role of a perceptron with a linear activation function. We validated these conclusions through simulations and experiments, and several algorithms from machine learning were applied for imaging and were compared with traditional ghost imaging algorithms (including Hadamard speckle imaging and compressed sensing). We believe that this research can help discover new algorithms to improve the imaging quality and noise resistance of computational ghost imaging, while also providing an approach for implementing neural network computation in the physical world.
Zitieren
APA
Liu, W., Zhou, Y., Liu, J., Zheng, H., Chen, H., He, Y., Li, F., & Xu, Z. (2025). Understanding ghost imaging from a machine learning perspective. Journal of the Optical Society of America A, 42(1), 89. https://doi.org/10.1364/josaa.537888
Vancouver
Liu W, Zhou Y, Liu J, Zheng H, Chen H, He Y, et al. Understanding ghost imaging from a machine learning perspective. Journal of the Optical Society of America A. 2025;42(1):89. doi:10.1364/josaa.537888
BibTeX
@article{Liu2025,
author = {Liu, Wenjie and Zhou, Yu and Liu, Jianbin and Zheng, Huaibin and Chen, Hui and He, Yuchen and Li, Fuli and Xu, Zhuo},
title = {Understanding ghost imaging from a machine learning perspective},
journal = {Journal of the Optical Society of America A},
year = {2025},
volume = {42},
number = {1},
pages = {89},
doi = {10.1364/josaa.537888}
}
EUVIMED