Loading…
Memory-less scattering imaging with ultrafast convolutional optical neural networks
The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, w...
Saved in:
Published in: | Science advances 2024-06, Vol.10 (24), p.eadn2205 |
---|---|
Main Authors: | , , , , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c282t-1e621d19d53383ca0b02b17fb3e6cb1e887a65040e7ef626db91b414a3a06d533 |
---|---|
cites | |
container_end_page | |
container_issue | 24 |
container_start_page | eadn2205 |
container_title | Science advances |
container_volume | 10 |
creator | Zhang, Yuchao Zhang, Qiming Yu, Haoyi Zhang, Yinan Luan, Haitao Gu, Min |
description | The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, we demonstrate image reconstruction through scattering layers where no optical memory effect exists, by developing a multistage convolutional optical neural network (ONN) integrated with multiple parallel kernels operating at the speed of light. Training this Fourier optics-based, parallel, one-step convolutional ONN with the strong scattering process for direct feature extraction, we achieve memory-less image reconstruction with a field of view enlarged by a factor up to 271. This device is dynamically reconfigurable for ultrafast multitask image reconstruction with a computational power of 1.57 peta-operations per second (POPS). Our achievement establishes an ultrafast and high energy-efficient optical machine learning platform for graphic processing. |
doi_str_mv | 10.1126/sciadv.adn2205 |
format | article |
fullrecord | <record><control><sourceid>proquest_pubme</sourceid><recordid>TN_cdi_pubmedcentral_primary_oai_pubmedcentral_nih_gov_11177939</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3068755260</sourcerecordid><originalsourceid>FETCH-LOGICAL-c282t-1e621d19d53383ca0b02b17fb3e6cb1e887a65040e7ef626db91b414a3a06d533</originalsourceid><addsrcrecordid>eNpVkL1PwzAQxS0EolXpyogysqT47MRJJoQqvqQiBmCOHMdpDa5dbKdV_3tcKKhMv5Pu7r2nh9A54AkAYVdeKN6uJ7w1hOD8CA0JLfKU5Fl5fDAP0Nj7d4wxZIzlUJ2iAS3LIqe0GKKXJ7m0bptq6X3iBQ9BOmXmiVry-Y4bFRZJr4PjHfchEdasre6DsobrxK6CEpFG9u4bYWPdhz9DJx3XXo73HKG3u9vX6UM6e75_nN7MUkFKElKQjEALVRuTlFRw3GDSQNE1VDLRgIwZOctxhmUhO0ZY21TQZJBxyjHbPY3Q9Y_uqm-WshXSxJi6XrkY3m1ry1X9f2PUop7bdQ0ARVHRKipc7hWc_eylD_VSeSG15kba3tcUs1hUThiOpxeHZn8uv1XSL1jXeno</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3068755260</pqid></control><display><type>article</type><title>Memory-less scattering imaging with ultrafast convolutional optical neural networks</title><source>American Association for the Advancement of Science</source><source>PubMed (Medline)</source><creator>Zhang, Yuchao ; Zhang, Qiming ; Yu, Haoyi ; Zhang, Yinan ; Luan, Haitao ; Gu, Min</creator><creatorcontrib>Zhang, Yuchao ; Zhang, Qiming ; Yu, Haoyi ; Zhang, Yinan ; Luan, Haitao ; Gu, Min</creatorcontrib><description>The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, we demonstrate image reconstruction through scattering layers where no optical memory effect exists, by developing a multistage convolutional optical neural network (ONN) integrated with multiple parallel kernels operating at the speed of light. Training this Fourier optics-based, parallel, one-step convolutional ONN with the strong scattering process for direct feature extraction, we achieve memory-less image reconstruction with a field of view enlarged by a factor up to 271. This device is dynamically reconfigurable for ultrafast multitask image reconstruction with a computational power of 1.57 peta-operations per second (POPS). Our achievement establishes an ultrafast and high energy-efficient optical machine learning platform for graphic processing.</description><identifier>ISSN: 2375-2548</identifier><identifier>EISSN: 2375-2548</identifier><identifier>DOI: 10.1126/sciadv.adn2205</identifier><identifier>PMID: 38875337</identifier><language>eng</language><publisher>United States: American Association for the Advancement of Science</publisher><subject>Applied Physics ; Optics ; Physical and Materials Sciences ; SciAdv r-articles</subject><ispartof>Science advances, 2024-06, Vol.10 (24), p.eadn2205</ispartof><rights>Copyright © 2024 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC). 2024 The Authors</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c282t-1e621d19d53383ca0b02b17fb3e6cb1e887a65040e7ef626db91b414a3a06d533</citedby><orcidid>0000-0002-4538-4170 ; 0009-0005-5619-9201 ; 0000-0003-0067-1210 ; 0000-0003-4078-253X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC11177939/pdf/$$EPDF$$P50$$Gpubmedcentral$$Hfree_for_read</linktopdf><linktohtml>$$Uhttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC11177939/$$EHTML$$P50$$Gpubmedcentral$$Hfree_for_read</linktohtml><link.rule.ids>230,314,727,780,784,885,27924,27925,53791,53793</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/38875337$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Zhang, Yuchao</creatorcontrib><creatorcontrib>Zhang, Qiming</creatorcontrib><creatorcontrib>Yu, Haoyi</creatorcontrib><creatorcontrib>Zhang, Yinan</creatorcontrib><creatorcontrib>Luan, Haitao</creatorcontrib><creatorcontrib>Gu, Min</creatorcontrib><title>Memory-less scattering imaging with ultrafast convolutional optical neural networks</title><title>Science advances</title><addtitle>Sci Adv</addtitle><description>The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, we demonstrate image reconstruction through scattering layers where no optical memory effect exists, by developing a multistage convolutional optical neural network (ONN) integrated with multiple parallel kernels operating at the speed of light. Training this Fourier optics-based, parallel, one-step convolutional ONN with the strong scattering process for direct feature extraction, we achieve memory-less image reconstruction with a field of view enlarged by a factor up to 271. This device is dynamically reconfigurable for ultrafast multitask image reconstruction with a computational power of 1.57 peta-operations per second (POPS). Our achievement establishes an ultrafast and high energy-efficient optical machine learning platform for graphic processing.</description><subject>Applied Physics</subject><subject>Optics</subject><subject>Physical and Materials Sciences</subject><subject>SciAdv r-articles</subject><issn>2375-2548</issn><issn>2375-2548</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNpVkL1PwzAQxS0EolXpyogysqT47MRJJoQqvqQiBmCOHMdpDa5dbKdV_3tcKKhMv5Pu7r2nh9A54AkAYVdeKN6uJ7w1hOD8CA0JLfKU5Fl5fDAP0Nj7d4wxZIzlUJ2iAS3LIqe0GKKXJ7m0bptq6X3iBQ9BOmXmiVry-Y4bFRZJr4PjHfchEdasre6DsobrxK6CEpFG9u4bYWPdhz9DJx3XXo73HKG3u9vX6UM6e75_nN7MUkFKElKQjEALVRuTlFRw3GDSQNE1VDLRgIwZOctxhmUhO0ZY21TQZJBxyjHbPY3Q9Y_uqm-WshXSxJi6XrkY3m1ry1X9f2PUop7bdQ0ARVHRKipc7hWc_eylD_VSeSG15kba3tcUs1hUThiOpxeHZn8uv1XSL1jXeno</recordid><startdate>20240614</startdate><enddate>20240614</enddate><creator>Zhang, Yuchao</creator><creator>Zhang, Qiming</creator><creator>Yu, Haoyi</creator><creator>Zhang, Yinan</creator><creator>Luan, Haitao</creator><creator>Gu, Min</creator><general>American Association for the Advancement of Science</general><scope>NPM</scope><scope>7X8</scope><scope>5PM</scope><orcidid>https://orcid.org/0000-0002-4538-4170</orcidid><orcidid>https://orcid.org/0009-0005-5619-9201</orcidid><orcidid>https://orcid.org/0000-0003-0067-1210</orcidid><orcidid>https://orcid.org/0000-0003-4078-253X</orcidid></search><sort><creationdate>20240614</creationdate><title>Memory-less scattering imaging with ultrafast convolutional optical neural networks</title><author>Zhang, Yuchao ; Zhang, Qiming ; Yu, Haoyi ; Zhang, Yinan ; Luan, Haitao ; Gu, Min</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c282t-1e621d19d53383ca0b02b17fb3e6cb1e887a65040e7ef626db91b414a3a06d533</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Applied Physics</topic><topic>Optics</topic><topic>Physical and Materials Sciences</topic><topic>SciAdv r-articles</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Zhang, Yuchao</creatorcontrib><creatorcontrib>Zhang, Qiming</creatorcontrib><creatorcontrib>Yu, Haoyi</creatorcontrib><creatorcontrib>Zhang, Yinan</creatorcontrib><creatorcontrib>Luan, Haitao</creatorcontrib><creatorcontrib>Gu, Min</creatorcontrib><collection>PubMed</collection><collection>MEDLINE - Academic</collection><collection>PubMed Central (Full Participant titles)</collection><jtitle>Science advances</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Zhang, Yuchao</au><au>Zhang, Qiming</au><au>Yu, Haoyi</au><au>Zhang, Yinan</au><au>Luan, Haitao</au><au>Gu, Min</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Memory-less scattering imaging with ultrafast convolutional optical neural networks</atitle><jtitle>Science advances</jtitle><addtitle>Sci Adv</addtitle><date>2024-06-14</date><risdate>2024</risdate><volume>10</volume><issue>24</issue><spage>eadn2205</spage><pages>eadn2205-</pages><issn>2375-2548</issn><eissn>2375-2548</eissn><abstract>The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, we demonstrate image reconstruction through scattering layers where no optical memory effect exists, by developing a multistage convolutional optical neural network (ONN) integrated with multiple parallel kernels operating at the speed of light. Training this Fourier optics-based, parallel, one-step convolutional ONN with the strong scattering process for direct feature extraction, we achieve memory-less image reconstruction with a field of view enlarged by a factor up to 271. This device is dynamically reconfigurable for ultrafast multitask image reconstruction with a computational power of 1.57 peta-operations per second (POPS). Our achievement establishes an ultrafast and high energy-efficient optical machine learning platform for graphic processing.</abstract><cop>United States</cop><pub>American Association for the Advancement of Science</pub><pmid>38875337</pmid><doi>10.1126/sciadv.adn2205</doi><orcidid>https://orcid.org/0000-0002-4538-4170</orcidid><orcidid>https://orcid.org/0009-0005-5619-9201</orcidid><orcidid>https://orcid.org/0000-0003-0067-1210</orcidid><orcidid>https://orcid.org/0000-0003-4078-253X</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 2375-2548 |
ispartof | Science advances, 2024-06, Vol.10 (24), p.eadn2205 |
issn | 2375-2548 2375-2548 |
language | eng |
recordid | cdi_pubmedcentral_primary_oai_pubmedcentral_nih_gov_11177939 |
source | American Association for the Advancement of Science; PubMed (Medline) |
subjects | Applied Physics Optics Physical and Materials Sciences SciAdv r-articles |
title | Memory-less scattering imaging with ultrafast convolutional optical neural networks |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T23%3A33%3A10IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_pubme&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Memory-less%20scattering%20imaging%20with%20ultrafast%20convolutional%20optical%20neural%20networks&rft.jtitle=Science%20advances&rft.au=Zhang,%20Yuchao&rft.date=2024-06-14&rft.volume=10&rft.issue=24&rft.spage=eadn2205&rft.pages=eadn2205-&rft.issn=2375-2548&rft.eissn=2375-2548&rft_id=info:doi/10.1126/sciadv.adn2205&rft_dat=%3Cproquest_pubme%3E3068755260%3C/proquest_pubme%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c282t-1e621d19d53383ca0b02b17fb3e6cb1e887a65040e7ef626db91b414a3a06d533%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=3068755260&rft_id=info:pmid/38875337&rfr_iscdi=true |