Abstract:Modeling the three-dimensional (3D) structure from a given 2D image is of great importance for analyzing and studying the physical properties of porous media. As an intractable inverse problem, many methods have been developed to address this fundamental problems over the past decades. Among many methods, the deep learning-(DL) based methods show great advantages in terms of accuracy, diversity, and efficiency. Usually, the 3D reconstruction from the 2D slice with a larger field-of-view is more conducive to simulate and analyze the physical properties of porous media accurately. However, due to the limitation of reconstruction ability, the reconstruction size of most widely used generative adversarial network-based model is constrained to 64^{3} or 128^{3}. Recently, a 3D porous media recurrent neural network based method (namely, 3D-PMRNN) (namely 3D-PMRNN) has been proposed to improve the reconstruction ability, and thus the reconstruction size is expanded to 256^{3}. Nevertheless, in order to train these models, the existed DL-based methods need to down-sample the original computed tomography (CT) image first so that the convolutional kernel can capture the morphological features of training images. Thus, the detailed information of the original CT image will be lost. Besides, the 3D reconstruction from a optical thin section is not available because of the large size of the cutting slice. In this paper, we proposed an improved recurrent generative model to further enhance the reconstruction ability (512^{3}). Benefiting from the RNN-based architecture, the proposed model requires only one 3D training sample at least and generates the 3D structures layer by layer. There are three more improvements: First, a hybrid receptive field for the kernel of convolutional neural network is adopted. Second, an attention-based module is merged into the proposed model. Finally, a useful section loss is proposed to enhance the continuity along the Z direction. Three experiments are carried out to verify the effectiveness of the proposed model. Experimental results indicate the good reconstruction ability of proposed model in terms of accuracy, diversity, and generalization. And the effectiveness of section loss is also proved from the perspective of visual inspection and statistical comparison.

PM-ARNN: 2D-TO-3D Reconstruction Paradigm for Microstructure of Porous Media Via Adversarial Recurrent Neural Network.

3D-PMRNN: Reconstructing Three-Dimensional Porous Media from the Two-Dimensional Image with Recurrent Neural Network

Improved recurrent generative model for reconstructing large-size porous media from two-dimensional images

Reconstruction of three-dimensional porous media using generative adversarial neural networks

Super-Resolution Reconstruction of Porous Media Using Concurrent Generative Adversarial Networks and Residual Blocks

Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network

Slice-to-voxel stochastic reconstructions on porous media with hybrid deep generative model

3D stochastic reconstruction of porous media based on attention mechanisms and residual networks

Stochastic reconstruction of 3D porous media from 2D images using generative adversarial networks

Stochastic reconstruction of porous media based on attention mechanisms and multi-stage generative adversarial network

3D Microstructure Reconstruction and Characterization of Porous Materials Using a Cross-Sectional SEM Image and Deep Learning

Reconstruction of Porous Media from Extremely Limited Information Using Conditional Generative Adversarial Networks.

Three-dimensional Reconstruction of Porous Media by Fusing Multi-Grid Image Features Based on Extended Feature Pyramid Network

Statistical characterization and reconstruction of heterogeneous microstructures using deep neural network

New 2D to 3D Reconstruction of Heterogeneous Porous Media via Deep Generative Adversarial Networks (GANs)

Fast and accurate reconstruction of large-scale 3D porous media using deep learning

An end-to-end three-dimensional reconstruction framework of porous media from a single two-dimensional image based on deep learning

Three-dimensional reconstruction of granular porous media based on deep generative models.

3D-porous-GAN: a high-performance 3D GAN for digital core reconstruction from a single 3D image

3D Reconstruction of Digital Rock Guided by Petrophysical Parameters with Deep Learning

Multi-scale Reconstruction of Porous Media from Low-Resolution Core Images Using Conditional Generative Adversarial Networks