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.
Highlights
Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation
Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce many drainage networks from the drainage network samples already generated by the stochastic network generation model, Gibb’s model
DCGANs have the promising potential for quickly generating similar network topology, as many previous studies have already shown that it could generate similar images and patterns very well
Summary
Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. Data acquisition and hydrologic analysis of actual drainage networks require time-consuming processes To overcome these difficulties in analyzing real drainage networks, statistical description of network topology has been utilized to generate drainage networks that can be used to assess the effect of drainage network topology on runoff[3,4]. It takes a relatively long time to generate a number of large enough networks for meaningful statistical evaluation because the Gibbs’ model has to consider all possible flow directions at each node of the n etwork[11]. Among these techniques, generative adversarial networks (GANs) have a deep generative framework that can effectively learn a probability distribution of training sample data and generate realistic samples from the given distribution without explicitly modeling the probability density function[14,15]. GANs have demonstrated remarkable results in image synthesis, image translation, data augmentation, and image/data/topology reconstruction[16,17,18]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
