Model-based deep learning framework for accelerated optical projection tomography
Autor(a) principal: | |
---|---|
Data de Publicação: | 2023 |
Outros Autores: | , , , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10400.1/20365 |
Resumo: | In this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. The algorithm is trained using a database of wild-type zebrafish (Danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. Using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a U-Net. The proposed method performs equally well or better than the alternatives. For a highly reduced number of projections, only the U-Net method provides images comparable to those obtained with ToMoDL. However, ToMoDL has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller. |
id |
RCAP_77adba747e9bc645cd2f8ea900e10cca |
---|---|
oai_identifier_str |
oai:sapientia.ualg.pt:10400.1/20365 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Model-based deep learning framework for accelerated optical projection tomographyAlgorithmsIn this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. The algorithm is trained using a database of wild-type zebrafish (Danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. Using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a U-Net. The proposed method performs equally well or better than the alternatives. For a highly reduced number of projections, only the U-Net method provides images comparable to those obtained with ToMoDL. However, ToMoDL has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller.LCF/PR/HR22/00533; Grant agreement no.101094250; WT 203148/Z/16/ZNature PortfolioSapientiaObando, MarcosBassi, AndreaDucros, NicolasMato, GermánCorreia, Teresa2024-02-03T13:24:41Z20232023-01-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articleapplication/pdfhttp://hdl.handle.net/10400.1/20365eng2045-232210.1038/s41598-023-47650-3info:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-02-07T02:01:22Zoai:sapientia.ualg.pt:10400.1/20365Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T02:36:39.370645Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Model-based deep learning framework for accelerated optical projection tomography |
title |
Model-based deep learning framework for accelerated optical projection tomography |
spellingShingle |
Model-based deep learning framework for accelerated optical projection tomography Obando, Marcos Algorithms |
title_short |
Model-based deep learning framework for accelerated optical projection tomography |
title_full |
Model-based deep learning framework for accelerated optical projection tomography |
title_fullStr |
Model-based deep learning framework for accelerated optical projection tomography |
title_full_unstemmed |
Model-based deep learning framework for accelerated optical projection tomography |
title_sort |
Model-based deep learning framework for accelerated optical projection tomography |
author |
Obando, Marcos |
author_facet |
Obando, Marcos Bassi, Andrea Ducros, Nicolas Mato, Germán Correia, Teresa |
author_role |
author |
author2 |
Bassi, Andrea Ducros, Nicolas Mato, Germán Correia, Teresa |
author2_role |
author author author author |
dc.contributor.none.fl_str_mv |
Sapientia |
dc.contributor.author.fl_str_mv |
Obando, Marcos Bassi, Andrea Ducros, Nicolas Mato, Germán Correia, Teresa |
dc.subject.por.fl_str_mv |
Algorithms |
topic |
Algorithms |
description |
In this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. The algorithm is trained using a database of wild-type zebrafish (Danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. Using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a U-Net. The proposed method performs equally well or better than the alternatives. For a highly reduced number of projections, only the U-Net method provides images comparable to those obtained with ToMoDL. However, ToMoDL has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller. |
publishDate |
2023 |
dc.date.none.fl_str_mv |
2023 2023-01-01T00:00:00Z 2024-02-03T13:24:41Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10400.1/20365 |
url |
http://hdl.handle.net/10400.1/20365 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
2045-2322 10.1038/s41598-023-47650-3 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Nature Portfolio |
publisher.none.fl_str_mv |
Nature Portfolio |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799137417674883072 |