Loading…
RedCap: residual encoder-decoder capsule network for holographic image reconstruction
A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmen...
Saved in:
Published in: | Optics express 2020-02, Vol.28 (4), p.4876-4887 |
---|---|
Main Authors: | , , |
Format: | Article |
Language: | English |
Citations: | Items that this one cites 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-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3 |
---|---|
cites | cdi_FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3 |
container_end_page | 4887 |
container_issue | 4 |
container_start_page | 4876 |
container_title | Optics express |
container_volume | 28 |
creator | Zeng, Tianjiao So, Hayden K-H Lam, Edmund Y |
description | A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices. |
doi_str_mv | 10.1364/OE.383350 |
format | article |
fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2370532914</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2370532914</sourcerecordid><originalsourceid>FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3</originalsourceid><addsrcrecordid>eNpNkEtLw0AUhQdRbK0u_AMyS12kzjOZuJNSH1AoiF2HycydNppm4kxC8d8bbRVX5y6-c7h8CF1SMqU8FbfL-ZQrziU5QmNKcpEIorLjf_cIncX4RggVWZ6dohFnlNGMqjFavYCd6fYOB4iV7XWNoTHeQkgs_CQ2uo19DbiBbufDO3Y-4I2v_TrodlMZXG31Goa68U3sQm-6yjfn6MTpOsLFISdo9TB_nT0li-Xj8-x-kRjOSJc4ItISUuUyVsrSKZtKq3KpgXKWOqIdlVLmXFvHCWNMgFCpKllKCBNCcOATdL3fbYP_6CF2xbaKBupaN-D7WDCeEclZTsWA3uxRE3yMAVzRhuH18FlQUnxbLJbzYm9xYK8Os325BftH_mrjXxaYbD4</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2370532914</pqid></control><display><type>article</type><title>RedCap: residual encoder-decoder capsule network for holographic image reconstruction</title><source>EZB Electronic Journals Library</source><creator>Zeng, Tianjiao ; So, Hayden K-H ; Lam, Edmund Y</creator><creatorcontrib>Zeng, Tianjiao ; So, Hayden K-H ; Lam, Edmund Y</creatorcontrib><description>A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices.</description><identifier>ISSN: 1094-4087</identifier><identifier>EISSN: 1094-4087</identifier><identifier>DOI: 10.1364/OE.383350</identifier><identifier>PMID: 32121718</identifier><language>eng</language><publisher>United States</publisher><ispartof>Optics express, 2020-02, Vol.28 (4), p.4876-4887</ispartof><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3</citedby><cites>FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3</cites><orcidid>0000-0002-6780-6100 ; 0000-0002-6514-0237 ; 0000-0001-6268-950X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>315,786,790,27957,27958</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/32121718$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Zeng, Tianjiao</creatorcontrib><creatorcontrib>So, Hayden K-H</creatorcontrib><creatorcontrib>Lam, Edmund Y</creatorcontrib><title>RedCap: residual encoder-decoder capsule network for holographic image reconstruction</title><title>Optics express</title><addtitle>Opt Express</addtitle><description>A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices.</description><issn>1094-4087</issn><issn>1094-4087</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><recordid>eNpNkEtLw0AUhQdRbK0u_AMyS12kzjOZuJNSH1AoiF2HycydNppm4kxC8d8bbRVX5y6-c7h8CF1SMqU8FbfL-ZQrziU5QmNKcpEIorLjf_cIncX4RggVWZ6dohFnlNGMqjFavYCd6fYOB4iV7XWNoTHeQkgs_CQ2uo19DbiBbufDO3Y-4I2v_TrodlMZXG31Goa68U3sQm-6yjfn6MTpOsLFISdo9TB_nT0li-Xj8-x-kRjOSJc4ItISUuUyVsrSKZtKq3KpgXKWOqIdlVLmXFvHCWNMgFCpKllKCBNCcOATdL3fbYP_6CF2xbaKBupaN-D7WDCeEclZTsWA3uxRE3yMAVzRhuH18FlQUnxbLJbzYm9xYK8Os325BftH_mrjXxaYbD4</recordid><startdate>20200217</startdate><enddate>20200217</enddate><creator>Zeng, Tianjiao</creator><creator>So, Hayden K-H</creator><creator>Lam, Edmund Y</creator><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0002-6780-6100</orcidid><orcidid>https://orcid.org/0000-0002-6514-0237</orcidid><orcidid>https://orcid.org/0000-0001-6268-950X</orcidid></search><sort><creationdate>20200217</creationdate><title>RedCap: residual encoder-decoder capsule network for holographic image reconstruction</title><author>Zeng, Tianjiao ; So, Hayden K-H ; Lam, Edmund Y</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Zeng, Tianjiao</creatorcontrib><creatorcontrib>So, Hayden K-H</creatorcontrib><creatorcontrib>Lam, Edmund Y</creatorcontrib><collection>PubMed</collection><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><jtitle>Optics express</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Zeng, Tianjiao</au><au>So, Hayden K-H</au><au>Lam, Edmund Y</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>RedCap: residual encoder-decoder capsule network for holographic image reconstruction</atitle><jtitle>Optics express</jtitle><addtitle>Opt Express</addtitle><date>2020-02-17</date><risdate>2020</risdate><volume>28</volume><issue>4</issue><spage>4876</spage><epage>4887</epage><pages>4876-4887</pages><issn>1094-4087</issn><eissn>1094-4087</eissn><notes>ObjectType-Article-1</notes><notes>SourceType-Scholarly Journals-1</notes><notes>ObjectType-Feature-2</notes><notes>content type line 23</notes><abstract>A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices.</abstract><cop>United States</cop><pmid>32121718</pmid><doi>10.1364/OE.383350</doi><tpages>12</tpages><orcidid>https://orcid.org/0000-0002-6780-6100</orcidid><orcidid>https://orcid.org/0000-0002-6514-0237</orcidid><orcidid>https://orcid.org/0000-0001-6268-950X</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 1094-4087 |
ispartof | Optics express, 2020-02, Vol.28 (4), p.4876-4887 |
issn | 1094-4087 1094-4087 |
language | eng |
recordid | cdi_proquest_miscellaneous_2370532914 |
source | EZB Electronic Journals Library |
title | RedCap: residual encoder-decoder capsule network for holographic image reconstruction |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-09-22T11%3A42%3A32IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=RedCap:%20residual%20encoder-decoder%20capsule%20network%20for%20holographic%20image%20reconstruction&rft.jtitle=Optics%20express&rft.au=Zeng,%20Tianjiao&rft.date=2020-02-17&rft.volume=28&rft.issue=4&rft.spage=4876&rft.epage=4887&rft.pages=4876-4887&rft.issn=1094-4087&rft.eissn=1094-4087&rft_id=info:doi/10.1364/OE.383350&rft_dat=%3Cproquest_cross%3E2370532914%3C/proquest_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c320t-f046be68f72b5bf8d65d895ae1326f0af155593adf302224e4868b260024443e3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2370532914&rft_id=info:pmid/32121718&rfr_iscdi=true |