Imaging through glass diffusers using densely connected convolutional networks
Computational imaging through scatter generally is accomplished by first characterizing the scattering medium so that its forward operator is obtained and then imposing additional priors in the form of regularizers on the reconstruction functional to improve the condition of the originally ill-posed...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Oxford University Press
2019
|
| Online Access: | http://hdl.handle.net/1721.1/121037 https://orcid.org/0000-0002-7836-0431 https://orcid.org/0000-0003-4340-0998 https://orcid.org/0000-0002-7225-7580 https://orcid.org/0000-0002-4140-1404 |
| _version_ | 1811094891055284224 |
|---|---|
| author | Li, Shuai Deng, Mo Lee, Justin Wu Sinha, Ayan T Barbastathis, George |
| author2 | Massachusetts Institute of Technology. Institute for Medical Engineering & Science |
| author_facet | Massachusetts Institute of Technology. Institute for Medical Engineering & Science Li, Shuai Deng, Mo Lee, Justin Wu Sinha, Ayan T Barbastathis, George |
| author_sort | Li, Shuai |
| collection | MIT |
| description | Computational imaging through scatter generally is accomplished by first characterizing the scattering medium so that its forward operator is obtained and then imposing additional priors in the form of regularizers on the reconstruction functional to improve the condition of the originally ill-posed inverse problem. In the functional, the forward operator and regularizer must be entered explicitly or parametrically (e.g., scattering matrices and dictionaries, respectively). However, the process of determining these representations is often incomplete, prone to errors, or infeasible. Recently, deep learning architectures have been proposed to instead learn both the forward operator and regularizer through examples. Here, we propose for the first time, to our knowledge, a convolutional neural network architecture called “IDiffNet” for the problem of imaging through diffuse media and demonstrate that IDiffNet has superior generalization capability through extensive tests with well-calibrated diffusers. We also introduce the negative Pearson correlation coefficient (NPCC) loss function for neural net training and show that the NPCC is more appropriate for spatially sparse objects and strong scattering conditions. Our results show that the convolutional architecture is robust to the choice of prior, as demonstrated by the use of multiple training and testing object databases, and capable of achieving higher space–bandwidth product reconstructions than previously reported. |
| first_indexed | 2024-09-23T16:07:08Z |
| format | Article |
| id | mit-1721.1/121037 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T16:07:08Z |
| publishDate | 2019 |
| publisher | Oxford University Press |
| record_format | dspace |
| spelling | mit-1721.1/1210372022-09-29T18:20:16Z Imaging through glass diffusers using densely connected convolutional networks Li, Shuai Deng, Mo Lee, Justin Wu Sinha, Ayan T Barbastathis, George Massachusetts Institute of Technology. Institute for Medical Engineering & Science Massachusetts Institute of Technology. Department of Biology Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology. Department of Mechanical Engineering Massachusetts Institute of Technology. Department of Physics Li, Shuai Deng, Mo Lee, Justin Wu Sinha, Ayan T Barbastathis, George Computational imaging through scatter generally is accomplished by first characterizing the scattering medium so that its forward operator is obtained and then imposing additional priors in the form of regularizers on the reconstruction functional to improve the condition of the originally ill-posed inverse problem. In the functional, the forward operator and regularizer must be entered explicitly or parametrically (e.g., scattering matrices and dictionaries, respectively). However, the process of determining these representations is often incomplete, prone to errors, or infeasible. Recently, deep learning architectures have been proposed to instead learn both the forward operator and regularizer through examples. Here, we propose for the first time, to our knowledge, a convolutional neural network architecture called “IDiffNet” for the problem of imaging through diffuse media and demonstrate that IDiffNet has superior generalization capability through extensive tests with well-calibrated diffusers. We also introduce the negative Pearson correlation coefficient (NPCC) loss function for neural net training and show that the NPCC is more appropriate for spatially sparse objects and strong scattering conditions. Our results show that the convolutional architecture is robust to the choice of prior, as demonstrated by the use of multiple training and testing object databases, and capable of achieving higher space–bandwidth product reconstructions than previously reported. Singapore-MIT Alliance United States. Office of the Director of National Intelligence. Rapid Analysis of Various Emerging Nanoelectronics United States. Department of Energy (DE-FG02-97ER25308) United States. Department of Energy. Computational Science Graduate Fellowship Program 2019-03-19T13:10:21Z 2019-03-19T13:10:21Z 2018-07 2018-06 2019-03-01T12:55:20Z Article http://purl.org/eprint/type/JournalArticle 2334-2536 http://hdl.handle.net/1721.1/121037 Li, Shuai, Mo Deng, Justin Lee, Ayan Sinha, and George Barbastathis. “Imaging through Glass Diffusers Using Densely Connected Convolutional Networks.” Optica 5, no. 7 (July 6, 2018): 803. https://orcid.org/0000-0002-7836-0431 https://orcid.org/0000-0003-4340-0998 https://orcid.org/0000-0002-7225-7580 https://orcid.org/0000-0002-4140-1404 http://dx.doi.org/10.1364/OPTICA.5.000803 Optica 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 Oxford University Press OSA Publishing |
| spellingShingle | Li, Shuai Deng, Mo Lee, Justin Wu Sinha, Ayan T Barbastathis, George Imaging through glass diffusers using densely connected convolutional networks |
| title | Imaging through glass diffusers using densely connected convolutional networks |
| title_full | Imaging through glass diffusers using densely connected convolutional networks |
| title_fullStr | Imaging through glass diffusers using densely connected convolutional networks |
| title_full_unstemmed | Imaging through glass diffusers using densely connected convolutional networks |
| title_short | Imaging through glass diffusers using densely connected convolutional networks |
| title_sort | imaging through glass diffusers using densely connected convolutional networks |
| url | http://hdl.handle.net/1721.1/121037 https://orcid.org/0000-0002-7836-0431 https://orcid.org/0000-0003-4340-0998 https://orcid.org/0000-0002-7225-7580 https://orcid.org/0000-0002-4140-1404 |
| work_keys_str_mv | AT lishuai imagingthroughglassdiffusersusingdenselyconnectedconvolutionalnetworks AT dengmo imagingthroughglassdiffusersusingdenselyconnectedconvolutionalnetworks AT leejustinwu imagingthroughglassdiffusersusingdenselyconnectedconvolutionalnetworks AT sinhaayant imagingthroughglassdiffusersusingdenselyconnectedconvolutionalnetworks AT barbastathisgeorge imagingthroughglassdiffusersusingdenselyconnectedconvolutionalnetworks |