Differentiable Imaging: A New Tool for Computational Optical Imaging

Ni Chen; Liangcai Cao; Ting-Chung Poon; Byoungho Lee; Edmund Y. Lam

doi:10.1002/apxr.202200118

Differentiable Imaging: A New Tool for Computational Optical Imaging

Ni Chen, Liangcai Cao, Ting-Chung Poon, Byoungho Lee, Edmund Y. Lam

Research output: Contribution to journal › Article › peer-review

Abstract

The field of computational imaging has made significant advancements in recent years, yet it still faces limitations due to the restrictions imposed by traditional computational techniques. Differentiable programming offers a solution by combining the strengths of classical optimization and deep learning, enabling the creation of interpretable model-based neural networks. Through the integration of physics into the modeling process, differentiable imaging, which employs differentiable programming in computational imaging, has the potential to overcome challenges posed by sparse, incomplete, and noisy data. As a result, it has the potential to play a key role in advancing the field of computational imaging and its various applications.

Original language	English (US)
Journal	Advanced Physics Research
DOIs	https://doi.org/10.1002/apxr.202200118
State	Published - Mar 23 2023

Bibliographical note

KAUST Repository Item: Exported on 2023-03-28
Acknowledgements: The authors express gratitude to Dr. Congli Wang (UC Berkeley) for his valuable input on an early draft of this paper, Dr. Xin Li (Beijing Institute of Technology) for his insightful feedback and discussions, and Prof. Sigurdur Thoroddsen (King Abdullah University of Science and Technology) for his English editing assistance.

Access to Document

10.1002/apxr.202200118

https://repository.kaust.edu.sa/bitstream/10754/686576/6/Advanced%20Physics%20Research%20-%202023%20-%20Chen%20-%20Differentiable%20Imaging%20%20A%20New%20Tool%20for%20Computational%20Optical%20Imaging.pdf

Cite this

@article{84537d60cc0947e4812b04a18024a7d5,

title = "Differentiable Imaging: A New Tool for Computational Optical Imaging",

abstract = "The field of computational imaging has made significant advancements in recent years, yet it still faces limitations due to the restrictions imposed by traditional computational techniques. Differentiable programming offers a solution by combining the strengths of classical optimization and deep learning, enabling the creation of interpretable model-based neural networks. Through the integration of physics into the modeling process, differentiable imaging, which employs differentiable programming in computational imaging, has the potential to overcome challenges posed by sparse, incomplete, and noisy data. As a result, it has the potential to play a key role in advancing the field of computational imaging and its various applications.",

author = "Ni Chen and Liangcai Cao and Ting-Chung Poon and Byoungho Lee and Lam, {Edmund Y.}",

note = "KAUST Repository Item: Exported on 2023-03-28 Acknowledgements: The authors express gratitude to Dr. Congli Wang (UC Berkeley) for his valuable input on an early draft of this paper, Dr. Xin Li (Beijing Institute of Technology) for his insightful feedback and discussions, and Prof. Sigurdur Thoroddsen (King Abdullah University of Science and Technology) for his English editing assistance.",

year = "2023",

month = mar,

day = "23",

doi = "10.1002/apxr.202200118",

language = "English (US)",

journal = "Advanced Physics Research",

issn = "2751-1200",

publisher = "Wiley Open Access",

}

TY - JOUR

T1 - Differentiable Imaging: A New Tool for Computational Optical Imaging

AU - Chen, Ni

AU - Cao, Liangcai

AU - Poon, Ting-Chung

AU - Lee, Byoungho

AU - Lam, Edmund Y.

N1 - KAUST Repository Item: Exported on 2023-03-28 Acknowledgements: The authors express gratitude to Dr. Congli Wang (UC Berkeley) for his valuable input on an early draft of this paper, Dr. Xin Li (Beijing Institute of Technology) for his insightful feedback and discussions, and Prof. Sigurdur Thoroddsen (King Abdullah University of Science and Technology) for his English editing assistance.

PY - 2023/3/23

Y1 - 2023/3/23

N2 - The field of computational imaging has made significant advancements in recent years, yet it still faces limitations due to the restrictions imposed by traditional computational techniques. Differentiable programming offers a solution by combining the strengths of classical optimization and deep learning, enabling the creation of interpretable model-based neural networks. Through the integration of physics into the modeling process, differentiable imaging, which employs differentiable programming in computational imaging, has the potential to overcome challenges posed by sparse, incomplete, and noisy data. As a result, it has the potential to play a key role in advancing the field of computational imaging and its various applications.

AB - The field of computational imaging has made significant advancements in recent years, yet it still faces limitations due to the restrictions imposed by traditional computational techniques. Differentiable programming offers a solution by combining the strengths of classical optimization and deep learning, enabling the creation of interpretable model-based neural networks. Through the integration of physics into the modeling process, differentiable imaging, which employs differentiable programming in computational imaging, has the potential to overcome challenges posed by sparse, incomplete, and noisy data. As a result, it has the potential to play a key role in advancing the field of computational imaging and its various applications.

UR - http://hdl.handle.net/10754/686576

UR - https://onlinelibrary.wiley.com/doi/10.1002/apxr.202200118

U2 - 10.1002/apxr.202200118

DO - 10.1002/apxr.202200118

M3 - Article

SN - 2751-1200

JO - Advanced Physics Research

JF - Advanced Physics Research

ER -

Differentiable Imaging: A New Tool for Computational Optical Imaging

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this