Limited-angle computed tomography with deep image and physics priors

Abstract Computed tomography is a well-established x-ray imaging technique to reconstruct the three-dimensional structure of objects. It has been used extensively in a variety of fields, from diagnostic imaging to materials and biological sciences. One major challenge in some applications, such as i...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Semih Barutcu, Selin Aslan, Aggelos K. Katsaggelos, Doğa Gürsoy
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/42170df63d0145a4ba82f4390d2b4467
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:42170df63d0145a4ba82f4390d2b4467
record_format	dspace
spelling	oai:doaj.org-article:42170df63d0145a4ba82f4390d2b44672021-12-02T17:41:13ZLimited-angle computed tomography with deep image and physics priors10.1038/s41598-021-97226-22045-2322https://doaj.org/article/42170df63d0145a4ba82f4390d2b44672021-09-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-97226-2https://doaj.org/toc/2045-2322Abstract Computed tomography is a well-established x-ray imaging technique to reconstruct the three-dimensional structure of objects. It has been used extensively in a variety of fields, from diagnostic imaging to materials and biological sciences. One major challenge in some applications, such as in electron or x-ray tomography systems, is that the projections cannot be gathered over all the angles due to the sample holder setup or shape of the sample. This results in an ill-posed problem called the limited angle reconstruction problem. Typical image reconstruction in this setup leads to distortion and artifacts, thereby hindering a quantitative evaluation of the results. To address this challenge, we use a generative model to effectively constrain the solution of a physics-based approach. Our approach is self-training that can iteratively learn the nonlinear mapping from partial projections to the scanned object. Because our approach combines the data likelihood and image prior terms into a single deep network, it is computationally tractable and improves performance through an end-to-end training. We also complement our approach with total-variation regularization to handle high-frequency noise in reconstructions and implement a solver based on alternating direction method of multipliers. We present numerical results for various degrees of missing angle range and noise levels, which demonstrate the effectiveness of the proposed approach.Semih BarutcuSelin AslanAggelos K. KatsaggelosDoğa GürsoyNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Semih Barutcu Selin Aslan Aggelos K. Katsaggelos Doğa Gürsoy Limited-angle computed tomography with deep image and physics priors
description	Abstract Computed tomography is a well-established x-ray imaging technique to reconstruct the three-dimensional structure of objects. It has been used extensively in a variety of fields, from diagnostic imaging to materials and biological sciences. One major challenge in some applications, such as in electron or x-ray tomography systems, is that the projections cannot be gathered over all the angles due to the sample holder setup or shape of the sample. This results in an ill-posed problem called the limited angle reconstruction problem. Typical image reconstruction in this setup leads to distortion and artifacts, thereby hindering a quantitative evaluation of the results. To address this challenge, we use a generative model to effectively constrain the solution of a physics-based approach. Our approach is self-training that can iteratively learn the nonlinear mapping from partial projections to the scanned object. Because our approach combines the data likelihood and image prior terms into a single deep network, it is computationally tractable and improves performance through an end-to-end training. We also complement our approach with total-variation regularization to handle high-frequency noise in reconstructions and implement a solver based on alternating direction method of multipliers. We present numerical results for various degrees of missing angle range and noise levels, which demonstrate the effectiveness of the proposed approach.
format	article
author	Semih Barutcu Selin Aslan Aggelos K. Katsaggelos Doğa Gürsoy
author_facet	Semih Barutcu Selin Aslan Aggelos K. Katsaggelos Doğa Gürsoy
author_sort	Semih Barutcu
title	Limited-angle computed tomography with deep image and physics priors
title_short	Limited-angle computed tomography with deep image and physics priors
title_full	Limited-angle computed tomography with deep image and physics priors
title_fullStr	Limited-angle computed tomography with deep image and physics priors
title_full_unstemmed	Limited-angle computed tomography with deep image and physics priors
title_sort	limited-angle computed tomography with deep image and physics priors
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/42170df63d0145a4ba82f4390d2b4467
work_keys_str_mv	AT semihbarutcu limitedanglecomputedtomographywithdeepimageandphysicspriors AT selinaslan limitedanglecomputedtomographywithdeepimageandphysicspriors AT aggeloskkatsaggelos limitedanglecomputedtomographywithdeepimageandphysicspriors AT dogagursoy limitedanglecomputedtomographywithdeepimageandphysicspriors
_version_	1718379699642564608

Limited-angle computed tomography with deep image and physics priors

Ejemplares similares