Reusability report : Unpaired deep-learning approaches for holographic image
Zhang, Yuhe LU ; Ritschel, Tobias and Villanueva Perez, Pablo LU
(2024)
In Nature Machine Intelligence
6(3).
p.284-290
- Abstract
- Deep-learning methods using unpaired datasets hold great potential for image reconstruction, especially in biomedical imaging where obtaining paired datasets is often difficult due to practical concerns. A recent study by Lee et al. (Nature Machine Intelligence 2023) has introduced a parameterized physical model (referred to as FMGAN) using the unpaired approach for adaptive holographic imaging, which replaces the forward generator network with a physical model parameterized on the propagation distance of the probing light. FMGAN has demonstrated its capability to reconstruct the complex phase and amplitude of objects, as well as the propagation distance, even in scenarios where the object-to-sensor distance exceeds the range of the... (More)
- Deep-learning methods using unpaired datasets hold great potential for image reconstruction, especially in biomedical imaging where obtaining paired datasets is often difficult due to practical concerns. A recent study by Lee et al. (Nature Machine Intelligence 2023) has introduced a parameterized physical model (referred to as FMGAN) using the unpaired approach for adaptive holographic imaging, which replaces the forward generator network with a physical model parameterized on the propagation distance of the probing light. FMGAN has demonstrated its capability to reconstruct the complex phase and amplitude of objects, as well as the propagation distance, even in scenarios where the object-to-sensor distance exceeds the range of the training data. We performed additional experiments to comprehensively assess FMGAN’s capabilities and limitations. As in the original paper, we compared FMGAN to two state-of-the-art unpaired methods, CycleGAN and PhaseGAN, and evaluated their robustness and adaptability under diverse conditions. Our findings highlight FMGAN’s reproducibility and generalizability when dealing with both in-distribution and out-of-distribution data, corroborating the results reported by the original authors. We also extended FMGAN with explicit forward models describing the response of specific optical systems, which improved performance when dealing with non-perfect systems. However, we observed that FMGAN encounters difficulties when explicit forward models are unavailable. In such scenarios, PhaseGAN outperformed FMGAN.
(Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/3144bef6-0304-4a57-a1fd-9eb291c75aa7
- author
- Zhang, Yuhe
LU
; Ritschel, Tobias
and Villanueva Perez, Pablo
LU
- organization
- publishing date
- 2024-02-15
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Nature Machine Intelligence
- volume
- 6
- issue
- 3
- pages
- 7 pages
- publisher
- Springer Nature
- external identifiers
-
- scopus:85185146734
- ISSN
- 2522-5839
- DOI
- 10.1038/s42256-024-00798-7
- language
- English
- LU publication?
- yes
- id
- 3144bef6-0304-4a57-a1fd-9eb291c75aa7
- date added to LUP
- 2024-05-20 15:12:11
- date last changed
- 2024-05-21 08:11:44
@article{3144bef6-0304-4a57-a1fd-9eb291c75aa7,
abstract = {{Deep-learning methods using unpaired datasets hold great potential for image reconstruction, especially in biomedical imaging where obtaining paired datasets is often difficult due to practical concerns. A recent study by Lee et al. (Nature Machine Intelligence 2023) has introduced a parameterized physical model (referred to as FMGAN) using the unpaired approach for adaptive holographic imaging, which replaces the forward generator network with a physical model parameterized on the propagation distance of the probing light. FMGAN has demonstrated its capability to reconstruct the complex phase and amplitude of objects, as well as the propagation distance, even in scenarios where the object-to-sensor distance exceeds the range of the training data. We performed additional experiments to comprehensively assess FMGAN’s capabilities and limitations. As in the original paper, we compared FMGAN to two state-of-the-art unpaired methods, CycleGAN and PhaseGAN, and evaluated their robustness and adaptability under diverse conditions. Our findings highlight FMGAN’s reproducibility and generalizability when dealing with both in-distribution and out-of-distribution data, corroborating the results reported by the original authors. We also extended FMGAN with explicit forward models describing the response of specific optical systems, which improved performance when dealing with non-perfect systems. However, we observed that FMGAN encounters difficulties when explicit forward models are unavailable. In such scenarios, PhaseGAN outperformed FMGAN.<br/><br/>}},
author = {{Zhang, Yuhe and Ritschel, Tobias and Villanueva Perez, Pablo}},
issn = {{2522-5839}},
language = {{eng}},
month = {{02}},
number = {{3}},
pages = {{284--290}},
publisher = {{Springer Nature}},
series = {{Nature Machine Intelligence}},
title = {{Reusability report : Unpaired deep-learning approaches for holographic image}},
url = {{http://dx.doi.org/10.1038/s42256-024-00798-7}},
doi = {{10.1038/s42256-024-00798-7}},
volume = {{6}},
year = {{2024}},
}