Conditional Invertible Neural Networks for Medical Imaging

Over recent years, deep learning methods have become an increasingly popular choice for solving tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Alexander Denker, Maximilian Schmidt, Johannes Leuschner, Peter Maass
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	image reconstruction invertible neural networks normalizing flows Photography TR1-1050 Computer applications to medicine. Medical informatics R858-859.7 Electronic computers. Computer science QA75.5-76.95
Acceso en línea:	https://doaj.org/article/0a745086f3d24cb58d64562b47d9acde
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:0a745086f3d24cb58d64562b47d9acde
record_format	dspace
spelling	oai:doaj.org-article:0a745086f3d24cb58d64562b47d9acde2021-11-25T18:03:34ZConditional Invertible Neural Networks for Medical Imaging10.3390/jimaging71102432313-433Xhttps://doaj.org/article/0a745086f3d24cb58d64562b47d9acde2021-11-01T00:00:00Zhttps://www.mdpi.com/2313-433X/7/11/243https://doaj.org/toc/2313-433XOver recent years, deep learning methods have become an increasingly popular choice for solving tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e., low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.Alexander DenkerMaximilian SchmidtJohannes LeuschnerPeter MaassMDPI AGarticleimage reconstructioninvertible neural networksnormalizing flowsPhotographyTR1-1050Computer applications to medicine. Medical informaticsR858-859.7Electronic computers. Computer scienceQA75.5-76.95ENJournal of Imaging, Vol 7, Iss 243, p 243 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	image reconstruction invertible neural networks normalizing flows Photography TR1-1050 Computer applications to medicine. Medical informatics R858-859.7 Electronic computers. Computer science QA75.5-76.95
spellingShingle	image reconstruction invertible neural networks normalizing flows Photography TR1-1050 Computer applications to medicine. Medical informatics R858-859.7 Electronic computers. Computer science QA75.5-76.95 Alexander Denker Maximilian Schmidt Johannes Leuschner Peter Maass Conditional Invertible Neural Networks for Medical Imaging
description	Over recent years, deep learning methods have become an increasingly popular choice for solving tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e., low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.
format	article
author	Alexander Denker Maximilian Schmidt Johannes Leuschner Peter Maass
author_facet	Alexander Denker Maximilian Schmidt Johannes Leuschner Peter Maass
author_sort	Alexander Denker
title	Conditional Invertible Neural Networks for Medical Imaging
title_short	Conditional Invertible Neural Networks for Medical Imaging
title_full	Conditional Invertible Neural Networks for Medical Imaging
title_fullStr	Conditional Invertible Neural Networks for Medical Imaging
title_full_unstemmed	Conditional Invertible Neural Networks for Medical Imaging
title_sort	conditional invertible neural networks for medical imaging
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/0a745086f3d24cb58d64562b47d9acde
work_keys_str_mv	AT alexanderdenker conditionalinvertibleneuralnetworksformedicalimaging AT maximilianschmidt conditionalinvertibleneuralnetworksformedicalimaging AT johannesleuschner conditionalinvertibleneuralnetworksformedicalimaging AT petermaass conditionalinvertibleneuralnetworksformedicalimaging
_version_	1718411661147111424

Conditional Invertible Neural Networks for Medical Imaging

Ejemplares similares