Phase imaging with an untrained neural network
Name
s41377-020-0302-3.pdf
Description
Published version
Size
1.41 MB
Format
Adobe PDF
Checksum (MD5)
b736668ee175c4270dc7117b92afb3a9
Author(s)
Wang, Fei
Bian, Yaoming
Wang, Haichao
Lyu, Meng
Pedrini, Giancarlo
Osten, Wolfgang
Barbastathis, George
Situ, Guohai
Date Issued
May 2020
Journal
Light: Science and Applications
Publisher
Springer Science and Business Media LLC
Citation
Wang, Fei et al., "Phase imaging with an untrained neural network." Light: Science and Applications 9 (May 2020): no. 77 doi. 10.1038/s41377-020-0302-3 ©2020 Author(s)
Version
Final published version
Abstract
Most of the neural networks proposed so far for computational imaging (CI) in optics employ a supervised training strategy, and thus need a large training set to optimize their weights and biases. Setting aside the requirements of environmental and system stability during many hours of data acquisition, in many practical applications, it is unlikely to be possible to obtain sufficient numbers of ground-truth images for training. Here, we propose to overcome this limitation by incorporating into a conventional deep neural network a complete physical model that represents the process of image formation. The most significant advantage of the resulting physics-enhanced deep neural network (PhysenNet) is that it can be used without training beforehand, thus eliminating the need for tens of thousands of labeled data. We take single-beam phase imaging as an example for demonstration. We experimentally show that one needs only to feed PhysenNet a single diffraction pattern of a phase object, and it can automatically optimize the network and eventually produce the object phase through the interplay between the neural network and the physical model. This opens up a new paradigm of neural network design, in which the concept of incorporating a physical model into a neural network can be generalized to solve many other CI problems. ©2020, The Author(s).
MIT Department
Massachusetts Institute of Technology. Department of Mechanical Engineering
Terms of Use
Creative Commons Attribution 4.0 International license
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1038/s41377-020-0302-3
