Untrained Neural Network Priors for Inverse Imaging Problems: A Survey

<p dir="ltr">In recent years, advancements in machine learning (ML) techniques, in particular, deep learning (DL) methods have gained a lot of momentum in solving inverse imaging problems, often surpassing the performance provided by hand-crafted approaches. Traditionally, analytical...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Adnan Qayyum (16875936) (author)
مؤلفون آخرون: Inaam Ilahi (16904676) (author), Fahad Shamshad (16904679) (author), Farid Boussaid (5334431) (author), Mohammed Bennamoun (6883301) (author), Junaid Qadir (16494902) (author)
منشور في: 2022
الموضوعات:
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
_version_ 1864513560974983168
author Adnan Qayyum (16875936)
author2 Inaam Ilahi (16904676)
Fahad Shamshad (16904679)
Farid Boussaid (5334431)
Mohammed Bennamoun (6883301)
Junaid Qadir (16494902)
author2_role author
author
author
author
author
author_facet Adnan Qayyum (16875936)
Inaam Ilahi (16904676)
Fahad Shamshad (16904679)
Farid Boussaid (5334431)
Mohammed Bennamoun (6883301)
Junaid Qadir (16494902)
author_role author
dc.creator.none.fl_str_mv Adnan Qayyum (16875936)
Inaam Ilahi (16904676)
Fahad Shamshad (16904679)
Farid Boussaid (5334431)
Mohammed Bennamoun (6883301)
Junaid Qadir (16494902)
dc.date.none.fl_str_mv 2022-09-05T00:00:00Z
dc.identifier.none.fl_str_mv 10.1109/tpami.2022.3204527
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/Untrained_Neural_Network_Priors_for_Inverse_Imaging_Problems_A_Survey/24056355
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Machine learning
Inverse problems
Neural networks
Imaging
Task analysis
Image reconstruction
Noise measurement
Deep learning
Inverse imaging problems
Untrained neural networks priors
dc.title.none.fl_str_mv Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p dir="ltr">In recent years, advancements in machine learning (ML) techniques, in particular, deep learning (DL) methods have gained a lot of momentum in solving inverse imaging problems, often surpassing the performance provided by hand-crafted approaches. Traditionally, analytical methods have been used to solve inverse imaging problems such as image restoration, inpainting, and superresolution. Unlike analytical methods for which the problem is explicitly defined and the domain knowledge is carefully engineered into the solution, DL models do not benefit from such prior knowledge and instead make use of large datasets to predict an unknown solution to the inverse problem. Recently, a new paradigm of training deep models using a single image, named untrained neural network prior (UNNP) has been proposed to solve a variety of inverse tasks, e.g., restoration and inpainting. Since then, many researchers have proposed various applications and variants of UNNP. In this paper, we present a comprehensive review of such studies and various UNNP applications for different tasks and highlight various open research problems which require further research.</p><h2>Other Information</h2><p dir="ltr">Published in: IEEE Transactions on Pattern Analysis and Machine Intelligence<br>License: <a href="https://creativecommons.org/licenses/by/4.0/legalcode" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/tpami.2022.3204527" target="_blank">https://dx.doi.org/10.1109/tpami.2022.3204527</a></p>
eu_rights_str_mv openAccess
id Manara2_7350f17f8dab7ab75eeb5fb9f987c436
identifier_str_mv 10.1109/tpami.2022.3204527
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/24056355
publishDate 2022
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Untrained Neural Network Priors for Inverse Imaging Problems: A SurveyAdnan Qayyum (16875936)Inaam Ilahi (16904676)Fahad Shamshad (16904679)Farid Boussaid (5334431)Mohammed Bennamoun (6883301)Junaid Qadir (16494902)Information and computing sciencesMachine learningInverse problemsNeural networksImagingTask analysisImage reconstructionNoise measurementDeep learningInverse imaging problemsUntrained neural networks priors<p dir="ltr">In recent years, advancements in machine learning (ML) techniques, in particular, deep learning (DL) methods have gained a lot of momentum in solving inverse imaging problems, often surpassing the performance provided by hand-crafted approaches. Traditionally, analytical methods have been used to solve inverse imaging problems such as image restoration, inpainting, and superresolution. Unlike analytical methods for which the problem is explicitly defined and the domain knowledge is carefully engineered into the solution, DL models do not benefit from such prior knowledge and instead make use of large datasets to predict an unknown solution to the inverse problem. Recently, a new paradigm of training deep models using a single image, named untrained neural network prior (UNNP) has been proposed to solve a variety of inverse tasks, e.g., restoration and inpainting. Since then, many researchers have proposed various applications and variants of UNNP. In this paper, we present a comprehensive review of such studies and various UNNP applications for different tasks and highlight various open research problems which require further research.</p><h2>Other Information</h2><p dir="ltr">Published in: IEEE Transactions on Pattern Analysis and Machine Intelligence<br>License: <a href="https://creativecommons.org/licenses/by/4.0/legalcode" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/tpami.2022.3204527" target="_blank">https://dx.doi.org/10.1109/tpami.2022.3204527</a></p>2022-09-05T00:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1109/tpami.2022.3204527https://figshare.com/articles/journal_contribution/Untrained_Neural_Network_Priors_for_Inverse_Imaging_Problems_A_Survey/24056355CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/240563552022-09-05T00:00:00Z
spellingShingle Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
Adnan Qayyum (16875936)
Information and computing sciences
Machine learning
Inverse problems
Neural networks
Imaging
Task analysis
Image reconstruction
Noise measurement
Deep learning
Inverse imaging problems
Untrained neural networks priors
status_str publishedVersion
title Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
title_full Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
title_fullStr Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
title_full_unstemmed Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
title_short Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
title_sort Untrained Neural Network Priors for Inverse Imaging Problems: A Survey
topic Information and computing sciences
Machine learning
Inverse problems
Neural networks
Imaging
Task analysis
Image reconstruction
Noise measurement
Deep learning
Inverse imaging problems
Untrained neural networks priors