Recurrent neural network reveals transparent objects through scattering media
© 2021 Optical Society of America. Scattering generally worsens the condition of inverse problems, with the severity severity depending on the statistics of the refractive index gradient and contrast. Removing scattering artifacts from images has attracted much work in the literature, including rece...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
The Optical Society
2021
|
| Online Access: | https://hdl.handle.net/1721.1/138464 |
| _version_ | 1811072058663108608 |
|---|---|
| author | Kang, Iksung Pang, Subeen Zhang, Qihang Fang, Nicholas Barbastathis, George |
| author_facet | Kang, Iksung Pang, Subeen Zhang, Qihang Fang, Nicholas Barbastathis, George |
| author_sort | Kang, Iksung |
| collection | MIT |
| description | © 2021 Optical Society of America. Scattering generally worsens the condition of inverse problems, with the severity severity depending on the statistics of the refractive index gradient and contrast. Removing scattering artifacts from images has attracted much work in the literature, including recently the use of static neural networks. S. Li et al. [Optica 5(7), 803 (2018)] trained a convolutional neural network to reveal amplitude objects hidden by a specific diffuser; whereas Y. Li et al. [Optica 5(10), 1181 (2018)] were able to deal with arbitrary diffusers, as long as certain statistical criteria were met. Here, we propose a novel dynamical machine learning approach for the case of imaging phase objects through arbitrary diffusers. The motivation is to strengthen the correlation among the patterns during the training and to reveal phase objects through scattering media. We utilize the on-axis rotation of a diffuser to impart dynamics and utilize multiple speckle measurements from different angles to form a sequence of images for training. Recurrent neural networks (RNN) embedded with the dynamics filter out useful information and discard the redundancies, thus quantitative phase information in presence of strong scattering. In other words, the RNN effectively averages out the effect of the dynamic random scattering media and learns more about the static pattern. The dynamical approach reveals transparent images behind the scattering media out of speckle correlation among adjacent measurements in a sequence. This method is also applicable to other imaging applications that involve any other spatiotemporal dynamics. |
| first_indexed | 2024-09-23T09:00:12Z |
| format | Article |
| id | mit-1721.1/138464 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T09:00:12Z |
| publishDate | 2021 |
| publisher | The Optical Society |
| record_format | dspace |
| spelling | mit-1721.1/1384642021-12-14T03:34:46Z Recurrent neural network reveals transparent objects through scattering media Kang, Iksung Pang, Subeen Zhang, Qihang Fang, Nicholas Barbastathis, George © 2021 Optical Society of America. Scattering generally worsens the condition of inverse problems, with the severity severity depending on the statistics of the refractive index gradient and contrast. Removing scattering artifacts from images has attracted much work in the literature, including recently the use of static neural networks. S. Li et al. [Optica 5(7), 803 (2018)] trained a convolutional neural network to reveal amplitude objects hidden by a specific diffuser; whereas Y. Li et al. [Optica 5(10), 1181 (2018)] were able to deal with arbitrary diffusers, as long as certain statistical criteria were met. Here, we propose a novel dynamical machine learning approach for the case of imaging phase objects through arbitrary diffusers. The motivation is to strengthen the correlation among the patterns during the training and to reveal phase objects through scattering media. We utilize the on-axis rotation of a diffuser to impart dynamics and utilize multiple speckle measurements from different angles to form a sequence of images for training. Recurrent neural networks (RNN) embedded with the dynamics filter out useful information and discard the redundancies, thus quantitative phase information in presence of strong scattering. In other words, the RNN effectively averages out the effect of the dynamic random scattering media and learns more about the static pattern. The dynamical approach reveals transparent images behind the scattering media out of speckle correlation among adjacent measurements in a sequence. This method is also applicable to other imaging applications that involve any other spatiotemporal dynamics. 2021-12-13T19:29:14Z 2021-12-13T19:29:14Z 2021 2021-12-13T19:23:58Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/138464 Kang, Iksung, Pang, Subeen, Zhang, Qihang, Fang, Nicholas and Barbastathis, George. 2021. "Recurrent neural network reveals transparent objects through scattering media." Optics Express, 29 (4). en 10.1364/OE.412890 Optics Express Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf The Optical Society OSA Publishing |
| spellingShingle | Kang, Iksung Pang, Subeen Zhang, Qihang Fang, Nicholas Barbastathis, George Recurrent neural network reveals transparent objects through scattering media |
| title | Recurrent neural network reveals transparent objects through scattering media |
| title_full | Recurrent neural network reveals transparent objects through scattering media |
| title_fullStr | Recurrent neural network reveals transparent objects through scattering media |
| title_full_unstemmed | Recurrent neural network reveals transparent objects through scattering media |
| title_short | Recurrent neural network reveals transparent objects through scattering media |
| title_sort | recurrent neural network reveals transparent objects through scattering media |
| url | https://hdl.handle.net/1721.1/138464 |
| work_keys_str_mv | AT kangiksung recurrentneuralnetworkrevealstransparentobjectsthroughscatteringmedia AT pangsubeen recurrentneuralnetworkrevealstransparentobjectsthroughscatteringmedia AT zhangqihang recurrentneuralnetworkrevealstransparentobjectsthroughscatteringmedia AT fangnicholas recurrentneuralnetworkrevealstransparentobjectsthroughscatteringmedia AT barbastathisgeorge recurrentneuralnetworkrevealstransparentobjectsthroughscatteringmedia |