Connectivity-informed drainage network generation using deep convolution generative adversarial networks

Abstract Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. In this study, Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce drainage networks from t...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Sung Eun Kim, Yongwon Seo, Junshik Hwang, Hongkyu Yoon, Jonghyun Lee
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/89c7ad6495e243988933f3e8c66b30bf
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:89c7ad6495e243988933f3e8c66b30bf
record_format	dspace
spelling	oai:doaj.org-article:89c7ad6495e243988933f3e8c66b30bf2021-12-02T14:02:33ZConnectivity-informed drainage network generation using deep convolution generative adversarial networks10.1038/s41598-020-80300-62045-2322https://doaj.org/article/89c7ad6495e243988933f3e8c66b30bf2021-01-01T00:00:00Zhttps://doi.org/10.1038/s41598-020-80300-6https://doaj.org/toc/2045-2322Abstract Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. In this study, Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce drainage networks from the already generated network samples without repetitive long modeling of the stochastic network model, Gibb’s model. In particular, we developed a novel connectivity-informed method that converts the drainage network images to the directional information of flow on each node of the drainage network, and then transforms it into multiple binary layers where the connectivity constraints between nodes in the drainage network are stored. DCGANs trained with three different types of training samples were compared; (1) original drainage network images, (2) their corresponding directional information only, and (3) the connectivity-informed directional information. A comparison of generated images demonstrated that the novel connectivity-informed method outperformed the other two methods by training DCGANs more effectively and better reproducing accurate drainage networks due to its compact representation of the network complexity and connectivity. This work highlights that DCGANs can be applicable for high contrast images common in earth and material sciences where the network, fractures, and other high contrast features are important.Sung Eun KimYongwon SeoJunshik HwangHongkyu YoonJonghyun LeeNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-14 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Sung Eun Kim Yongwon Seo Junshik Hwang Hongkyu Yoon Jonghyun Lee Connectivity-informed drainage network generation using deep convolution generative adversarial networks
description	Abstract Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. In this study, Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce drainage networks from the already generated network samples without repetitive long modeling of the stochastic network model, Gibb’s model. In particular, we developed a novel connectivity-informed method that converts the drainage network images to the directional information of flow on each node of the drainage network, and then transforms it into multiple binary layers where the connectivity constraints between nodes in the drainage network are stored. DCGANs trained with three different types of training samples were compared; (1) original drainage network images, (2) their corresponding directional information only, and (3) the connectivity-informed directional information. A comparison of generated images demonstrated that the novel connectivity-informed method outperformed the other two methods by training DCGANs more effectively and better reproducing accurate drainage networks due to its compact representation of the network complexity and connectivity. This work highlights that DCGANs can be applicable for high contrast images common in earth and material sciences where the network, fractures, and other high contrast features are important.
format	article
author	Sung Eun Kim Yongwon Seo Junshik Hwang Hongkyu Yoon Jonghyun Lee
author_facet	Sung Eun Kim Yongwon Seo Junshik Hwang Hongkyu Yoon Jonghyun Lee
author_sort	Sung Eun Kim
title	Connectivity-informed drainage network generation using deep convolution generative adversarial networks
title_short	Connectivity-informed drainage network generation using deep convolution generative adversarial networks
title_full	Connectivity-informed drainage network generation using deep convolution generative adversarial networks
title_fullStr	Connectivity-informed drainage network generation using deep convolution generative adversarial networks
title_full_unstemmed	Connectivity-informed drainage network generation using deep convolution generative adversarial networks
title_sort	connectivity-informed drainage network generation using deep convolution generative adversarial networks
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/89c7ad6495e243988933f3e8c66b30bf
work_keys_str_mv	AT sungeunkim connectivityinformeddrainagenetworkgenerationusingdeepconvolutiongenerativeadversarialnetworks AT yongwonseo connectivityinformeddrainagenetworkgenerationusingdeepconvolutiongenerativeadversarialnetworks AT junshikhwang connectivityinformeddrainagenetworkgenerationusingdeepconvolutiongenerativeadversarialnetworks AT hongkyuyoon connectivityinformeddrainagenetworkgenerationusingdeepconvolutiongenerativeadversarialnetworks AT jonghyunlee connectivityinformeddrainagenetworkgenerationusingdeepconvolutiongenerativeadversarialnetworks
_version_	1718392092619702272

Connectivity-informed drainage network generation using deep convolution generative adversarial networks

Ejemplares similares