Transport in Porous Media (2022) 143:497–525

https://doi.org/10.1007/s11242-022-01781-9

Super‑Resolved Segmentation of X‑ray Images of Carbonate

Rocks Using Deep Learning

Naif J. Alqahtani1 · Yufu Niu1 · Ying Da Wang1 · Traiwit Chung1 · Zakhar Lanetc1 ·
Aleksandr Zhuravljov2 · Ryan T. Armstrong1 · Peyman Mostaghimi1

Received: 28 May 2021 / Accepted: 31 March 2022 / Published online: 28 April 2022
© The Author(s) 2022

Abstract
Reliable quantitative analysis of digital rock images requires precise segmentation and
identification of the macroporosity, sub-resolution porosity, and solid\mineral phases. This
is highly emphasized in heterogeneous rocks with complex pore size distributions such as
carbonates. Multi-label segmentation of carbonates using classic segmentation methods
such as multi-thresholding is highly sensitive to user bias and often fails in identifying low-
contrast sub-resolution porosity. In recent years, deep learning has introduced efficient and
automated algorithms that are capable of handling hard tasks with precision comparable to
human performance, with application to digital rocks super-resolution and segmentation
emerging. Here, we present a framework for using convolutional neural networks (CNNs)
to produce super-resolved segmentations of carbonates rock images for the objective of
identifying sub-resolution porosity. The volumes used for training and testing are based on
two different carbonates rocks imaged in-house at low and high resolutions. We experiment
with various implementations of CNNs architectures where super-resolved segmentation is
obtained in an end-to-end scheme and using two networks (super-resolution and segmenta-
tion) separately. We show the capability of the trained model of producing accurate seg-
mentation by comparing multiple voxel-wise segmentation accuracy metrics, topological
features, and measuring effective properties. The results underline the value of integrating
deep learning frameworks in digital rock analysis.

Keywords Carbonate rocks · Segmentation · Microporosity · Super-resolution · Deep

learning · Digital rock

* Peyman Mostaghimi
[email protected]
1
School of Minerals and Energy Resources Engineering, The University of New South Wales,
Sydney, Australia
2
Institute of Environmental and Agricultural Biology, University of Tyumen, Tyumen, Russia

13
Vol.:(0123456789)
498 N. J. Alqahtani et al.

1 Introduction

Digital Rock Physics (DRP) has emerged as one of the superior technologies for studying
porous media at the pore-scale (Andrä et al. 2013; Blunt et al. 2013; Bultreys et al. 2016a;
Mostaghimi et al. 2017; Wildenschild and Sheppard 2013). DRP integrates high-resolution
X-ray microtomographic imaging (micro-CT) with advanced computational methods for
predicting geomaterial effective properties (Arns et al. 2004; Mostaghimi et al. 2013; Ram-
stad et al. 2010). DRP complements classic and slower laboratory investigations, namely
Conventional Core Analysis (CCA) and Special Core Analysis (SCAL), through fast and
reproducible modeling frameworks (Berg et al. 2017; Raeini et al. 2014). A standard
‘image and model’ framework usually consists of multiple steps including image segmen-
tation. Image segmentation is paramount for accurate pore-scale modeling (Iassonov et al.
2009; Sheppard et al. 2004). Segmentation techniques commonly reported in the litera-
ture for DRP are global and adaptive thresholding (Andrä et al. 2013; Tuller et al. 2013),
watershed algorithms (Garfi et al. 2020; Niu et al. 2020a), and converging active contours
(Sheppard et al. 2004). Thorough reviews of segmentation methods have been presented
and compared for DRP in the literature (Iassonov et al. 2009; Schlüter et al. 2014; Tuller
et al. 2013). However, the downside of all these methods is that they require a certain level
of user judgment and tuning. As a result, the segmentation outcome is highly susceptible to
user bias and experience (Niu et al. 2020a).
Carbonate rocks host more than half of the world’s oil and gas reserves (Harbaugh
1967). Carbonates often exhibit complex multimodal pore systems with sizes ranging from
the nanoscale to the meter scale (cave systems scale) (Biswal et al. 2007). As a result, the
characterization of such geomaterials is often challenging using micro-CT imaging, as the
typical imaging resolution is often in the order of few micrometers (Andrä et al. 2013;
Blunt et al. 2013).
The term microporosity is often used in the literature for different purposes. For
example, the International Union of Pure and Applied Chemistry (IUPAC) defines
a micropore as a pore with a width not exceeding 2 nm. Choquette and Pray (1970)
defined a micropore in carbonate rocks as a pore smaller than 62.5 µm in diameter.
This is while Pittman (1971) described micropores as pores less than one micrometer
in diameter. In recent micro-CT studies, the fraction of the pore space with structures
smaller than the voxel size is termed microporosity or sub-resolution (Lin et al. 2016;
Liu and Mostaghimi 2018; Soulaine et al. 2016). In this study, the latter definition is
used to define pores less than voxel size or pores smaller than 2.68 µm in size. Micr-
oporosity identification is important for flow and reactive transport modeling in porous
media. The identification of sub-resolution porosity using dry images is subjective to
the image quality in terms of the degree of noise and spatial resolution besides the
segmentation algorithm used (Soulaine et al. 2016). Alternative non-invasive methods
include combining micro-CT imaging with two-dimensional Scanning Electron Micros-
copy (SEM) (Othman et al. 2018; Peters 2009), and Differential Imaging (Boone et al.
2014; Knackstedt et al. 2006; Lin et al. 2016; Long et al. 2013). While SEM images pro-
vide unprecedented details of the pore geometry, the framework is time-consuming and
the nature of two-dimensional analysis introduces high uncertainty, especially when it
comes to the three-dimensional connectivity/continuity of the phases. Differential imag-
ing frameworks involve taking X-ray scans of dry and contrast agent-saturated samples
to identify microporosity. The process is perhaps one of the best options to resolve sub-
resolution experimentally, boosting the contrast between the phases and making the

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 499

segmentation task less prone to error. However, the method does not entirely remove
inherited user bias in the segmentation method of choice and requires time-consuming
sample flooding with a contrast agent, image registration and processing.
Deep learning frameworks using Convolutional Neural Networks (CNNs) introduced
fast and robust methods for automated image processing (Ronneberger et al. 2015; Wan
et al. 2014). If trained properly, these networks eliminate the user bias and achieve expert-
level annotations/processing (Niu et al. 2020a). The interest in integrating deep learning
in DRP frameworks is evident in recent years (Wang et al. 2021a) with application to
lithology classification (Alzubaidi et al. 2021), image denoising (Chen et al. 2020; Niu
et al. 2020b), binary and multi-phase segmentation (Ar Rushood et al. 2020; Karimpouli
and Tahmasebi 2019; Wang et al. 2021b), super-resolution (Kamrava et al. 2019; Wang
et al. 2019a, b), and predictive modeling of rock properties (Alqahtani et al. 2020, 2021;
Chung et al. 2020a; Rabbani and Babaei 2021; Rabbani et al. 2020b). Thorough reviews
have been published showing the diverse applications of DL in geoscience and DRP (Tah-
masebi et al. 2020; Wang et al. 2021a). For segmentation problems in DRP, deep learn-
ing frameworks can be used to produce binary and multi-mineral segmentation accord-
ing to the modeling problem. Binary segmentation is suitable for easier problems such as
single-phase non-reactive flow (Ar Rushood et al. 2020; Varfolomeev et al. 2019), while
multi-minerals segmentation is used for more detailed simulations such as reactive trans-
port (Liu et al. 2018). In general, deep learning showed promising results compared to
classic methods such as thresholding. Niu et al. (2020a) investigated the use of LeNet-5, a
CNN architecture, for segmenting 2D images of North Sea Sandstone. The results showed
that CNN segmentation outperformed the classic Otsu thresholding and watershed method
by comparing their physical properties. Karimpouli and Tahmasebi (2019) experimented
the use of SegNet architecture to produce a multi-mineral segmentation of a limited num-
ber of Berea Sandstone images augmented through stochastic reconstruction. The average
reported pixel-wise categorical accuracy for the best model is %96. For super-resolution,
(Wang et al. 2019b) experimented the use of different super-resolution network variants for
obtaining high-resolution 2D images. The study shows the superiority of the performance
of networks to classic methods, such as bicubic interpolation, with a 50–70% reduction in
relative error. Other efforts by (Wang et al. 2019a; Wang et al. 2020a) applied a similar
scheme using 3D images for training to obtain super-resolved volumes. The results show
a better refinement of edge sharpness and reduction of noise compared to other classic
interpolation methods. Kamrava et al. (2019) used a hybrid method of stochastic and deep
learning algorithms to generate super-resolved images of shale formations. The stochastic
reconstruction algorithm is used as an augmentation method for generating many image
realizations that can be used for training. The results show the superiority of the trained
network when the porosity and other metrics are used for comparison with other methods.
Janssens et al. (2020) used real multi-resolution carbonate paired images to obtain more
accurate segmentation that can be used as ground truth (GT) training data. However, the
high-resolution data is not used as-is but downsampled using voxel averaging to obtain
a grid size of similar dimensions to low-resolution data. The results of the study reveal
improvements when computing several physical properties of the medium. All the previ-
ous efforts directed toward super-resolution have used interpolation methods (such as bicu-
bic interpolation) to generate synthetic images pairs as a training dataset. This might not
be ideal for practical applications because synthetically downsampled images do not pos-
sess the same features of real low-resolution images in terms of noise and partial volume
effects. Several efforts have been directed toward reducing this limitation such as the use of
unpaired data for training (Niu et al. 2020b; Zhu et al. 2017). In this work, multi-resolution

13
500 N. J. Alqahtani et al.

scans have been used as-is, where the data has only been cropped and segmented to reveal
the true interpolation/translation improvement offered by deep learning.
Herein, we propose a framework for using CNNs to generate a super-resolved segmen-
tation of carbonates for better identifying sub-resolution porosity. We use a set of multi-
resolution micro-CT scans to create a unique high (HR) and low-resolution (LR) dataset.
While acquiring multi-resolution images is common in super-resolution methods, the pre-
sented approach alleviates the known trade-off between the field of view and spatial reso-
lution. LR images capture a large field of view while HR images show more details and
smaller features (i.e., identifying microporosity). By referring to the term ‘super-resolved’,
we describe the process of translating ‘real’ LR micro-CT domains to grayscale or seg-
mented HR domains using CNNs. The settings of imaging spatial resolution are designed
carefully to resolve sub-resolution in the samples considered as much as possible based on
the laboratory experiment. We utilize high-resolution region of interest scanning to fully
facilitate easy identification of sub-resolution porosity in the HR imaging characterized by
a distinguishable range of intensity values. The LR spatial resolution resolves micropo-
rosity poorly, showing partial volume effects. The HR images are segmented and utilized
as GT for the CNN training. The gray LR and the segmented HR images are utilized for
different deep learning experiments where the main target is to identify sub-resolution
porosity accurately using the LR as an input. We compare the segmentation accuracy using
three metrics: voxel-wise accuracy, geometry-based metrics, and the physical flow charac-
teristics. This framework can be utilized to optimize current frameworks where the only
requirement is to obtain a high-resolution region of interest and train the CNNs to interpo-
late the segmentation to a bigger field of view using the LR images.

2 Materials and Methods

2.1 Materials Description

Two carbonate Images shown in Fig. 1 were considered for training the CNNs. The first
one is Indiana Limestone (ILS) which originates from the Salem formation near Bedford,
Indiana, USA. ILS is mainly monomineralic rock with 98.8% calcite with the rare occur-
rences of quartz (< 1%) and clay minerals (< 1%). The main solid phases seen in ILS are
Allochems, detrital skeletal of marine organisms (i.e., organic detritus), and authigenic cal-
cite cement. Different types of porosities can be distinguished clearly in the HR micro-
CT images including macroporosity (resolved and connected porosity between particles/
grains), microporosity on the outer shells of the ooliths/carbonate grains (Intercrystal
porosity), vugs (isolated or poorly connected pores that are larger than 1/16 mm in diam-
eter), and intra-ooliths porosity (porosity within individual ooliths). The second carbon-
ate is a more heterogeneous Middle Eastern carbonate (MEC) characterized by a variety
of microporous ooliths, skeletal and non-skeletal microporous grains. Calcite accounts for
more than (99% >) with the existence of minerals like aragonite and micrite. Similar poros-
ity systems that can also be distinguished using the HR images are macroporosity, moldic
intra-ooliths porosity (these are formed through selective processes i.e., local dissolution),
vugs, and microporous equant cement. Features identifying porosity systems in carbonates
have been further discussed in the literature (Arns et al. 2018; Cantrell and Hagerty 1999;
Irajian et al. 2016; Ji et al. 2012).

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 501

Fig. 1  (Upper): 1-inch core plug photographs of the Indiana Limestone and the Middle Eastern Carbonate.
(Lower): a showcase of a 2D slice of the Indiana Limestone (Left) showing a macroporosity, b microporos-
ity on the outer shells of the ooliths c vugs, d intra-oolith microporosity, e solid grain. Middle Eastern Car-
bonate (right) showing f macroporosity, g solid ooliths, h microporous equant cement, I intra-grain vugs,
and k microporous ooliths

For the preliminary characterization of the samples, Brine permeability, Helium poros-
ity, and permeability measurement were obtained for a 1-inch in diameter by 2-inch core
plug of each rock (see Fig. 1). Then, smaller 6-mm core plugs were drilled from the bigger
cores for micro-CT imaging. Finally, Mercury Intrusion Capillary Pressure (MICP) tests
were performed on the 6 mm plugs. These tests were performed to assess in (1) choosing
the imaging spatial resolutions and (2) choosing safe thresholds for GT segmentation.

2.2 Methods

2.2.1 Image Acquisition and Processing

Image acquisition was carefully designed in a way that clearly distinguishes a high per-
centage of the microporosity at least with a different shade of gray in the HR images. The
MICP analysis provided in the "Appendix" (Fig. 12) shows a range of micropores below

13
502 N. J. Alqahtani et al.

the imaging resolution in Table 1 with a peak of around 1 µm. We categorize voxels into
pore, solids and micropores. Similar approaches have been presented for identifying micr-
oporosity and characterizing carbonate rocks in the literature (Bultreys et al. 2016b; Ghous
et al. 2007; Soulaine et al. 2016). This leads to the prediction of permeability that is com-
parable with experimental measurements on larger cores as will be discussed in the results
sections. The LR images were obtained with a voxel size four times larger than the HR.
Both carbonate rocks were imaged at the Tyree X-ray facilities at the University of New
South Wales using HeliScan™ micro-CT (Mark I). The system has a Hamamatsu X-ray
tube with a diamond window and a high-quality flatbed detector (3072 × 3072 pixels, 3.75
fps readout rate). The machine has a tube current of 85 μA and a voltage of 80 kV. The
samples were scanned in a double helix trajectory with 2880 projections per revolution and
a filter length of 3 mm in the scanning conditions described in Table 1. The reconstruction
was performed using QMango software (Kingston et al. 2011; Varslot et al. 2011) devel-
oped by the Australian National University. The imaging setup used to obtain the images
with different resolutions is shown in Fig. 2. A rectangular central domain was cropped
from each image and segmented into three phases (pore, solid, and micropores) using two
methods: Otsu’s multi-thresholding algorithm (Otsu 1979) and Avizo Software ™ Seg-
mentation Editor (version 2020.3; FEI Visualization Sciences Group) that uses a combina-
tion of watershed and hysteresis algorithms. Otsu’s multi-thresholding and watershed-hys-
teresis based methods have been utilized in literature for segmenting carbonate into three
phases (Arns et al. 2005; Bultreys et al. 2015; Ji et al. 2012). Due to the size of the data, the
manual creation of segmentation masks is not feasible. So, the ground truth segmentation
of HR data was obtained through Avizo. In the Avizo segmentation editor, a conservative
approach for segmentation was followed by choosing a ‘Safe’ range of intensities for each
phase for the watershed/hysteresis processes. The safe ranges involve first choosing two
cutoff thresholds that are certainly pore space and certainly solid matter. All voxels with
intensities below the lower threshold are pore, and all voxels higher than the upper thresh-
old are solid. Then, a third range for microporosity with upper and lower thresholds laying
between the pore and solid threshold was defined where only obvious microporous textures
were labeled. This was achieved based on a visual inspection of all slices of the volume.
These regions were then initiated as seeds for watershed/hysteresis. The watershed trans-
forms floods unlabeled regions using the image gradient using the Canny method (Canny
1986). The safe thresholding of GT resolved pore space was chosen to have a ‘macro’ pore
space volume fraction equal to the one found in the MICP analysis for each sample (see
"Appendix").
This method for identifying microporosity and other similar methods (Arns et al. 2005;
Bultreys et al. 2015; Ji et al. 2012) can perhaps be plausible for mono-mineralic rocks. In
mono-mineralic rocks, the lower grayscale levels in the solid matrix can be only associated
with the existence of sub-resolution porosity (less dense materials). Also, it is important

Table 1  Scanning conditions of the ILS and MEC samples

Sample name Voxel SIZE (μm) Distance from Exposure Scan duration (h) Image size (voxels)
source (mm) time (sec)

HR MEC 2.68 5.8 0.64 10.5 1520 × 1520 × 4100

LR MEC 10.72 23.2 0.64 5.1 380 × 380 × 1025
HR ILS 2.68 5.8 0.64 9.42 1520 × 1520 × 3552
LR ILS 10.72 23.2 0.64 4.75 380 × 380 × 888

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 503

Fig. 2  The Tyree X-ray micro-CT setup (upper), A schematic showing the imaging procedure used to
obtain a multi-resolution image (middle), and slices of the middle eastern carbonate where the magnifica-
tion showing the resolved microporosity (lower)

to emphasize this approach has limitations whereby the defined solid phase may still pose
a small percentage of porosity at a lower scale (assuming microscale features can be dis-
tinguished). However, the effect of these fine-scale porosities is minor when computing
macroscale effective properties, i.e., permeability. The identification of microporosity in
multi-mineral rocks would require differential imaging frameworks (Boone et al. 2014;
Knackstedt et al. 2006; Lin et al. 2016; Long et al. 2013).

2.2.2 Dataset, Network Architectures, Training, and Inference

The dataset used for training is titled ‘Multi-Resolution Complex Carbonates Micro-CT
Dataset (MRCCM)’ and has been published in the Digital rock Portal (https://​www.​digit​
alroc​kspor​tal.​org/​proje​cts/​362) and freely available to be used for future studies. The data
repository comprises multi-resolution raw tomograms and processed volumes of both car-
bonate rocks. The MICP analysis of both samples is included in the repository.

13
504 N. J. Alqahtani et al.

Two different frameworks shown in Fig. 3 were used to obtain the super-resolved seg-
mentation. In one framework we segmented and super-resolved images in one Encoder-
Decoder network (End-to-End super-resolved segmentation). In the second framework,
we applied a 3D Enhanced Deep Residual Network (EDSR) (Lim et al. 2017) to obtain
a super-resolved grayscale image and then segment the super-resolved using a separate
encoder-decoder network. All the network architectures considered are three-dimensional
because 3D CNNs tend to learn geometrical features in 3D space which helps in preserving
topological features, i.e., connectivity, and generally delivers better results compared to 2D
networks (Wang et al. 2021b).
The encoder-decoder networks used in both frameworks are mainly based on
the U-net architecture (Ronneberger et al. 2015). This architecture was originally
designed for 2D image segmentation purposes and has shown excellent performance

(C) EDSR-U-ResNet

Fig. 3  A schematic showing the network architectures used for obtaining super-resolved segmentation. The
top schematic is showing two highlighted architectures where a standard U-net block consists of double
convolutional layers as in (Ronneberger et al. 2015) and b the Residual U-net block (U-ResNet) consists of
triple convolutional layers with skip connections (Lee et al. 2017). The double convolution blocks (high-
lighted in yellow) at the end of the network increase the input domain size four times. The numbers under
each block represent the number of feature maps at each level. In the bottom schematic super-resolved gray-
scale images are obtained using an EDSR network (c), then segmented using a U-ResNet network. The
segmentations obtained from this network is referred to as EDSR-U-ResNet

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 505

for biomedical image segmentation and was later modified to handle volumetric data.
U-net incorporates symmetric skip connections to link shallow features in the encoder
to the equivalent level of the decoder through concatenation/summation. The skip con-
nections improve CNNs performance by alleviating common drawbacks in backpropa-
gation (Rumelhart et al. 1986) such as vanishing gradients in deep CNNs (Drozdzal
et al. 2016). Furthermore, the skip connections also help to preserve and combine local
and well-detailed features, such as edges, with global features without deep supervision
(Panichev and Voloshyna 2019). In this work, we implemented multiple U-net variants
where we compared the use of standard U-Net implementation against Residual U-Net.
Residual U-Net borrows concepts from ResNet where residual blocks are utilized to
achieve better performance.
All networks were trained using Adam optimizer with β1 = 0.9 and β2 = 0.999, and
L2 regularization of 1e-5 and starting learning rate of 2e-4. The learning rate is halved
dynamically during training each time the segmentation evaluation metric reaches a
plateau (no improvement) on a separate validation set. The Sørensen–Dice coefficient
(Dice 1945; Sørensen 1948) is used to compute loss between the network output and
GT. The Sørensen–Dice coefficient is given the following formula:
Sørensen−Dice coefficient(X, Y) = 2 × |X ∩ Y|∕(|X| + |Y|) (1)
where X, Y are two images, and the operator |X| refers to the number of voxels in image X.
The symbol ∩ refers to the intersection between the voxels of the two images. The training
was run at least for 100 K iterations and stopped when the network learning rate falls below
a predefined threshold (1-e6). PyTorch, a deep learning software package, was used to train
the models on an NVIDIA TITAN RTX GPU installed on a PC with Intel I7-8700 CPU
@3.20 GHz and a RAM of 64 GB. The U-net/U-Resnet models were trained with a batch
size of 2 and domain size of 6­ 43 grayscale volumes where the network outputs a domain
size of ­2563 super-resolved segmented volume. For the EDSR-U-ResNet framework, the
images were first upsampled using EDSR and then segmented using domain sizes of ­1283.
In total, 2300 training MEC subvolumes and 1800 ILS training subvolumes were used for
training the presented models.
A brief description of the evaluation metrics considered is reported. These metrics
include voxel-wise accuracy segmentation, topological characterization of each phase,
and measurements of the effective flow properties. Because of the nature of the problem
and the imaging framework and processing, several implications might affect how the
results are assessed. These implications are caused by:

• Registration of the HR and LR images is not exact, a misalignment in the range of

1–2 voxels may affect the voxel-wise metrics results.
• Image quality/noise levels will significantly control the cutoff volume of resolved
features/textures in the HR images, which may be impossible to identify in the LR
images.
• Watershed Transform, the GT segmentation method for HR images, requires the
user to define ‘safe’ thresholds that act as a seed for determining the extent for each
phase. While the method minimizes the effect of user bias in general, it still may
affect the GT segmentation creation. Regardless of this happening as a source of
error, the improvement in segmentation is granted because LR grayscale images are
translated based on the HR images of the same region. This leaves less chance of
erratic segmentation or user misjudgment.

13
506 N. J. Alqahtani et al.

All the numbers reported are computed from the testing set volumes. These volumes are
not used while training or validation.

A. Voxel-wise Accuracy:

We assessed the voxel-wise segmentation accuracy using two segmentation metrics,

namely: the Jaccard similarity index (Jaccard 1912) and the accuracy of voxel classification
using confusion matrices. Jaccard similarity index of phase (P) in two volumes (A) and (B)
can be expressed as:
( ) | /
| | |
Jaccard A(P) , B(P) = |A(P) ∩ B(P) | |A(P) ∪ B(P)| |
| | | | (2)

where A(P) and B(P) indicate the voxels in volumes A and B that are labeled as phase P,
respectively. The operator |A| indicates the number of voxels in volume A. The symbols ∩
and ∪ refer to the intersection and the union between the voxels of the two images, respec-
tively. The confusion charts show the true and predicted voxels percentages of each phase.
The diagonal values of the chart represent truly classified voxels percentage, and the off-
diagonal values represent misclassified voxels percentage. The rate of false positivity and
negativity are also reported outside the confusion matrices.

B. Morphological Measurements

We measured two geometrical properties of each phase which are the porosity and the
specific surface area (SSA). Furthermore, the topological connectivity wassured by com-
puting Euler-Poincare characteristic (EC) for resolved macropores only. The porosity of
each phase was computed through voxel counting. The specific surface area was computed
through the discretization of the Crofton formula (Legland et al. 2011). EC was estimated
using graph portioning where the numbers of vertices (Ѵ), edges (∈), faces (F),. and solids
(L). of the volume were computed. EC can be written as (Legland et al. 2011):
EC = # − V# ∈ + #F − #L (3)
Volume fractions, SSA EC are computed using MatImage, an open-source Matlab
library for image processing (Legland 2021).

C. Flow characteristics and pore networks

For assessing the flow characteristics of the output segmentation, single- and multi-phase
flow simulations were performed. The comparison includes the segmentations oned from
the HR and LR images using Otsu and watershed methods and the trained model Also,
macropore and micropore pore network models were extracted for each image separately
using PoreSpy (Gostick et al. 2019). Coordination numbers and average pores and throats
sizes were compared from the extracted pore network models. The macropore network
model was extracted from the resolved pore phase, and the micropore network model was
extracted from the microporosity phase. The micropore network extraction aimed to give a
general indication about the connectivity of the textures looking microporous and should
not be misinterpreted to indicate the actual connections of unresolved pore space.
For single-phase flow, we signified the importance of assigning conductivity to
the microporosity phase by comparing flow simulations with/without a conductivity

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 507

assigned to the microporosity phase. The permeability was computed using the Pore
Finite Volume Solver (PFVS) (Chung et al. 2019, 2020b; Wang et al. 2019c). PFVS
assigns a conductivity to each voxel according to the proximity of the voxel to the solid
wall and the radius of the inscribed flow channel that the voxel belongs to. For simplic-
ity, the microporosity phase was assigned a voxel conductivity based on the Hagen–Poi-
seuille (Mortensen et al. 2005; Rabbani et al. 2020a; Schmid and Henningson 1994) law
where the permeability (K) of a pore throat is assumed to approximately equal

R2
K≈ (4)
8
for fluid flow with a low Reynolds number where R is the hydraulic throat radius. The con-
tribution of microporosity is known to have lower permeability compared to macropores
by several orders of magnitude (Soulaine et al. 2016). Therefore, the assignment of micr-
oporosity conductivity was based on the average pore throat radius estimated through the
MICP analysis ran on the same cores (See "Appendix"). The conductivity in the micropo-
rous phase is estimated to be 2.45 × ­10−13 ­m2 in the MEC and 1.25 × ­10−13 ­m2 in the ILS.
For the multi-phase flow simulation, relative permeability was computed using
MorphLBM (Wang et al. 2020b). This method applies an accelerated morphologically
coupled multi-phase Lattice Boltzmann Method directly on the macropore space. The
microporosity phase is assumed to be fully saturated with the wetting phase, and the
absolute permeability of macropores is only considered. In the beginning, the fluid
configuration is initialized morphologically and updated after LBM steady-state condi-
tions are reached, with small increments of erosion and dilation to target saturation. The
LBM simulation continues its execution at the same time as the small morphological
increments are updated. Once the target saturation is achieved, the LBM is performed
until the capillary number is stable. Then, the steady-state relative permeability point
is recorded, and another cycle is launched. For the simulation compared in the results
section, imbibition simulations were performed on each segmentation with relaxation
applied every 1000 LBM timestep between morphs. The saturation increments were
set to 5% with a capillarity tolerance of less than 1­ 0−3 per 1000 timesteps. The system
capillary number was held below ­10−5 to mimic capillary-dominated two-phase flow
dynamics. The wettability was set to be uniform at 45 degrees for all solid voxels for
simplicity.

3 Results and Discussion

All the comparisons reported were based on unseen/testing subvolumes of size ­10243 vox-
els for the HR images and ­2563 voxels for the LR images that correspond to a volume of
size 20.6 ­mm3. This volume was considered here only for comparison purposes and might
be subject to further heterogeneity effects at a bigger scale. The methods presented do not
involve upscaling of the physical properties of the medium of interest but rather compares
the segmentation accuracy of the models. As such, representative elementary volume
analysis is not required for this purpose. Overall, 14 segmented volumes were compared
for each of the two carbonate rock types, of which 4 segmented volumes were obtained
from the HR and LR through Otsu and Avizo watershed segmentation methods, and 3 seg-
mented HR volumes obtained through the frameworks described in Sect. (2.2.2).

13
508 N. J. Alqahtani et al.

3.1 Voxel‑wise Accuracy

Voxel-wise accuracy was computed based on a separate testing volume of size 2­ 563 in
LR that were super-resolved to a volume of size ­10243 and compared with the GT seg-
mentation (watershed segmentation) for both carbonate rocks. Both testing volumes are
around 20% of the entire tomogram imaged of each rock. The network frameworks gener-
ate batches of size ­1283 that are stitched together to construct the testing volume. Three
super-resolved segmentation for each rock was segmented and reconstructed, using U-net,
U-Resnet, and EDSR-U-ResNet. The voxel-wise percentage of each phase is shown as a
confusion matrix for ILS and MEC in Fig. 4 for all the networks. Overall, the U-ResNet
scheme provides the best voxel-wise accuracy. The U-net results in second place and is
very close to the performance margins of U-ResNet. EDSR-U-Resnet shows the highest
discrepancy especially in segmenting the MEC rock sample. The same trends are observed
in Fig. 5 as indicated by Jaccard similarity Index where in general U-Resnet tends to per-
form better than the other models. The confusion matrices in Fig. 4 and Jaccard indices in
Fig. 5 show the microporosity phase with the highest margins of error (compared to solid
and pore phases) both in falsely positive and negative classified voxels. Additionally, the
error margins are higher in MEC compared to ILS in general, and this is likely due to the
complex microporous textures in the MEC sample.
A comparison between a region of interest for the different models’ segmentations is
shown in Figs. 6 and 7. The difference maps in Figs. 6 and 7 reveal the misclassified voxels
happen mostly at the boundaries, especially the solid/microporosity boundary. This fact is
also clearly shown by the confusion matrices in Fig. 4. This is expected as the grayscale

Fig. 4  The confusion matrices showing the percentage of a phase in each cell. The rows of a confusion
matrix represent the true class/GT, and the columns represent the predicted class by the network model.
Diagonal and off-diagonal cells represent correctly and incorrectly identified phases classes, respectively.
The row and column summaries shown outside the confusion matrices correspond to the percentages of
false positive and false negative rates for true and predicted classes, respectively. Warmer color codes show
higher error margins

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 509

Fig. 5  Jaccard similarity index of each phase computed for networks’ segmentations as compared to the GT
segmentation

intensities of these boundaries can be impossible to detect, at least visually (see LR images
in Figs. 6 and 7). Overall, the incorrect microporosity segmentation perhaps arises from
two causes, (1) partial volume effects in LR images which might be interpreted as micr-
oporosity by the models and (2) the tendency of the models to smooth out edges of highly
unresolved features/textures. The partial volume effects on the segmentation output are
mostly evident near the grain boundaries characterized by strong blurring effects. The net-
work performance is likely to improve by regularization (Jia et al. 2021) and using more
task-specific loss functions (Caliva et al. 2019).
From the network architecture perspective, the models’ performance ranking was con-
sistent on both rocks where U-Resnet performed best. While the EDSR-U-Resnet frame-
work uses two networks to obtain a super-resolved segmentation as shown in Fig. 3, this
framework poses the highest discrepancy in voxel-wise accuracy. The comparison with
other studied models that utilize end-to-end frameworks (LR to HR segmented image)
might suggest that EDSR super-resolved grayscale images might not preserve the impor-
tant features for accurate segmentation.

3.2 Morphological Comparison

As an addition to the reported voxel-wise metrics of networks segmentation accuracy, the

morphological characteristics of the watershed, Otsu, and the networks segmented volumes
are reported. In this section, the watershed and Otsu segmentation of LR and HR volumes
are included for a broader comparison. This will give an indication of how CNNs can
improve segmentation if compared with other classic methods. The comparison includes
the volume fraction and SSA of each phase, and the Euler number of the effective medium
pore space. These values are reported for the testing volumes of ILS and MEC in Tables 2
and 3, respectively. The measured volume fractions of the trained network segmented vol-
umes clearly provide better results and lower relative difference if compared with the clas-
sic methods of segmenting LR volumes. The network segmentations are mainly erratic in

13
510 N. J. Alqahtani et al.

Fig. 6  1st row: 2D slices of the ILS testing volume (LR and HR). 2nd and 3rd rows: LR (input to networks)
and HR resolution grayscale region of interest as shown in the first row (red square), and the corresponding
segmentation (GT watershed and networks segmentations). 4th row: difference maps where networks are
compared to GT. The blue color corresponds to correctly classified voxels and the pink color to misclassi-
fied phase classes

estimating the microporosity phase as suggested by previous results, however, more con-
forming than Otsu HR segmentation where microporosity is overestimated if compared
to the GT. The end-to-end models (U-net and U-Resnet) estimate the volume fractions
with a relative difference of less than 10%, while the error is found to be up to 22% using
the EDSR-U-Resnet framework. The error margins in estimating the SSA of the differ-
ent segmentations are higher than the volume fractions estimation error margins. However,
the SSA of the network segmentations are again found to be consistent, with a relative
difference of less than 50%, while all the other segmentation showed significant relative
errors (more than 100%) in the estimation of SSA in some phases. The extreme errors

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 511

Fig. 7  1st row: 2D slices of the LR and HR MEC testing volume. 2nd and 3rd rows: LR (input to networks)
and HR resolution grayscale region of interest as shown in the first row (red square), and the corresponding
segmentation (GT watershed and networks segmentations). 4th row: difference maps where networks are
compared to GT. The blue color corresponds to correctly classified voxels and the pink color to misclassi-
fied phase classes

in estimating the SSA perhaps arise from the tendency of the classic methods to create
a microporosity phase at solid grains and pore space boundary. This ‘coating’ effect is a
by-product of strong partial volume effects and the nature of sharp global thresholding in
classic methods. Consequently, this observation is expected in the segmentation of the LR
volumes where partial volume effects are evident.
The topological connectivity of each segmentation macropore space is assessed based
on the measured Euler Characteristic (EC). The results show significant variations in the
computed connectivity of pore space with relative differences of around 80% to GT using
the trained networks. The variation in EC mostly likely arises from strong imaging artifacts
including the translation of partial volume effects to false microporosity. This may cre-
ate undesired connections between solid grains and hence altering the computed EC. An

13
 512

13
Table 2  A set of morphological measurements of the Indiana limestone testing volume reported for different segmentation methods
Indiana Limestone (ILS) Pore Vol. (%) Solid Vol. (%) Micropore Vol. (%) Pore SSA Solid SSA Micropore SSA EC

HR-watershed (GT) 9.17 (0.0%) 77.95 (0.0%) 12.88 (0%) 0.0199 (0.0%) 0.0621 (0.0%) 0.0628 (0.0%) − 8860 (0.0%)
HR-Otsu 9.08 (− 1.0%) 75.36 (− 3.3%) 15.56 (21%) 0.0248 (24.6%) 0.0106 (− 82.9%) 0.1302 (107.3%) − 61,980 (599%)
U-Net 8.86 (− 3.4%) 79.07 (1.4%) 12.07 (− 6%) 0.0159 (− 20.3%) 0.0362 (− 41.7%) 0.0371 (− 40.9%) − 1061 (− 88%)
U-Resnet 8.81 (− 3.9%) 79.45 (1.9%) 11.74 (− 9%) 0.0158 (− 20.6%) 0.0356 (− 42.7%) 0.0354 (− 43.7%) − 989 (− 88%)
EDSR-U-Resnet 8.92 (− 2.7%) 76.60 (− 1.7%) 14.48 (12%) 0.0168 (− 15.5%) 0.0450 (− 27.6%) 0.0465 (− 25.9%) − 4261 (− 51%)
LR-watershed 8.29 (− 9.6%) 84.91 (8.9%) 6.80 (− 47%) 0.0407 (104.7%) 0.0738 (18.8%) 0.0525 (− 16.4%) − 38 (− 99%)
LR-Otsu 8.33 (− 9.2%) 76.85 (− 1.4%) 14.83 (15%) 0.0451 (126.8%) 0.0962 (54.9%) 0.1393 (121.7%) − 573 (− 93%)

Percentages inside brackets show the relative differences with respect to ground truth segmentation values. EC was calculated for macropore space only
N. J. Alqahtani et al.
 Table 3  A set of morphological measurements of the Middle Eastern Carbonate testing volume reported for different segmentation methods
Middle Eastern Car- Pore Vol. (%) Solid Vol. (%) Micropore Vol. (%) Pore SSA Solid SSA Micropore SSA EC
bonate (MEC)

HR-watershed (GT) 17.81 (0.0%) 69.53 (0.0%) 12.66 (0.0%) 0.0434 (0.0%) 0.1288 (0.0%) 0.1066 (0.0%) − 30,806 (0.0%)
HR-Otsu 16.23 (− 8.9%) 63.76 (− 8.3%) 20.01 (58.0%) 0.0512 (18.1%) 0.1698 (31.8%) 0.2178 (104.3%) − 115,448 (274%)
U-Net 17.90 (0.5%) 69.37 (− 0.2%) 12.72 (0.5%) 0.0345 (− 20.4%) 0.0812 (− 37.0%) 0.0594 (− 44.3%) − 5877 (− 80%)
U-Resnet 17.97 (0.9%) 70.15 (0.9%) 11.88 (− 6.2%) 0.0351 (− 19.0%) 0.0806 (− 37.4%) 0.0575 (− 46.0%) − 5979 (− 80%)
EDSR-U-Resnet 16.79 (− 5.7%) 67.68 (− 2.7%) 15.53 (22.6%) 0.0320 (− 26.2%) 0.0803 (− 37.7%) 0.0727 (− 31.8%) − 2205 (− 92%)
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks…

LR-watershed 14.27 (− 19.9%) 72.20 (3.8%) 13.53 (6.9%) 0.0757 (74.7%) 0.1364 (5.9%) 0.1183 (11.0%) − 515 (− 98%)
LR-Otsu 12.67 (− 28.8%) 67.12 (− 3.5%) 20.20 (59.5%) 0.0751 (73.1%) 0.1458 (13.2%) 0.2165 (103.1%) − 1520 (− 95%)

Percentages inside brackets show the relative difference with respect to ground truth segmentation values. EC was calculated for macropore space only
513

13
514 N. J. Alqahtani et al.

example of this segmentation error can be seen in the EDSR-U-ResNet segmentation in

Fig. 7. Also, there is always a limitation on the signal/representation that can be resolved as
pore space in some parts of the LR images causing connectivity loss in the output network
segmentation. Irrespectively, the network segmentations improve the connectivity meas-
ured when compared to the LR watershed and Otsu segmentation as shown in Tables 2 and
3. It is also anticipated that global thresholding (Otsu method) creates many isolated holes
and solids, and redundant loops, hence the high error margins.

3.3 Macro‑ and Micropore Networks, Single‑ and Multi‑phase Flow Analyses

The flow features of the segmented volumes are probably the most crucial measures for
assessing the accuracy of the super-resolved segmentation. Hence, pore and micropore net-
works, single- and multi-phase flow simulations are analyzed. Macro- and micropore net-
works are extracted ‘separately’ for the different segmentations considered. The statistics
of these networks and the single-phase permeability values are reported in Tables 4 and 5
for the ILS and MEC samples, respectively.
The coordination number (i.e., the average number of pore throats connected to a pore
body) is computed to estimate the connectivity from the macropore and micropore net-
works. The results show more consistent trends when computing the coordination number
of the macropore compared to the micropore networks. For the macropore networks, the
LR segmentations show lower coordination numbers compared to the GT, this is perhaps
due to the loss of tight pore throats during segmentation. However, the CNNs segmentation
seems to preserve similar connections showing minor differences in the computed coor-
dination number. For the micropore networks, the CNNs segmentation seems to overesti-
mate the connections in general. This is mainly because the adherence to boundaries of the
super-resolved segmentation is prone to relatively high errors. The average pore sizes, pore
throat lengths and diameters show in general a similar trend, where the CNNs improve the
computed statistics. Moreover, it is also generally observed that the ILS results are again
more conforming compared to MEC because of the lower microporosity in ILS.
For single-phase flow, the permeability in Tables 4 and 5 is computed with and without
assigning conductivity to the micropore phase. The addition of microporosity conductiv-
ity increases the computed permeability, as both rocks pose relatively high percentages of
unresolved porosity. However, the contribution to the computed permeability might not be
as significant to the macropore phase. In any case, the CNNs segmentation show more
accurate permeability values compared to the simulation ran on LR segmented images.
More interestingly, U-net and U-ResNet specifically present more accurate permeability
values than the HR Otsu segmentation with and without microporous conductivity. Look-
ing over all the results and metrics, it might be concluded that U-ResNet and U-net pro-
vide the best and second-best results, respectively. The permeability of the ground truth
and networks show in general a good agreement with the experimental Helium permeabil-
ity (Klingenberg-corrected) on bigger cores of the ILS (221 mD) and MEC (1092 mD)
samples.
For further comparison, multi-phase flow experiments are run on the U-ResNet, LR,
and HR watershed segmentations of both rocks using MorphLBM. The secondary imbi-
bition experiments are simulated on the testing volumes macropore space using the Aus-
tralian National Computational Infrastructure (NCI) supercomputer Gadi. The micropo-
rosity is assumed to be fully saturated with the wetting phase. The simulation of the LR
watershed segmentation did not converge to a solution. The reason is likely to be the low

13
 Table 4  Statistics of the macro- and micropore networks and the PFVS single-phase permeability for the ILS testing sample using different segmentation methods
Indiana Limestone (ILS) Coordination Avg. Pore Size (m) Avg. Throat Length (m) Avg. Throat Diameter (m) Kmacro (mD) Kmacro + micro
Number (mD)

HR-watershed (GT) 2.5/2.1 2.4E-5/1.5E-5 9.8E-5/1.2E-4 2.2E-5/1.2E-5 139 175

HR-Otsu 2.6/1.6 2.2E-5/1.0E-5 1.0E-4/2.0E-4 1.9E-5/7.1E-6 91 129
U-Net 2.7/3.1 3.0E-5/2.1E-5 9.4E-5/8.0E-5 2.6E-5/1.7E-5 147 180
U-Resnet 2.6/3.2 2.9E-5/2.1E-5 9.5E-5/8.1E-5 2.6E-5/1.8E-5 142 174
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks…

EDSR-U-Resnet 2.1/2.9 2.6E-5/2.2E-5 9.5E-5/8.0E-5 2.6E-5/1.7E-5 130 176

LR-watershed 1.7/1.3 7.6E-5/5.1E-5 2.6E-4/2.9E-4 4.7E-5/3.3E-5 35 41
LR-Otsu 2.1/1.6 7.2E-5/1.0E-5 2.8E-4/2.0E-4 3.9E-5/7.1E-6 67 104
515

13
 516

13
Table 5  Statistics of the macro- and micropore networks and the PFVS single-phase permeability for the MEC testing sample using different segmentation methods
Middle eastern carbonate Coordination Avg. pore size (m) Avg. throat length (m) Avg. throat diameter (m) Kmacro (mD) Kmacro + micro
(MEC) number (mD)

HR-watershed (GT) 2.5/1.4 2.0E-5/9.8E-6 9.2E-5/1.4E-4 2.0E-5/6.7E-6 960 1007

HR-Otsu 2.7/2.2 2.0E-5/9.8E-6 9.3E-5/2.8E-4 1.6E-5/5.8E-6 490 560
U-Net 2.9/2.8 2.7E-5/1.8E-5 9.1E-5/7.9E-5 2.3E-5/1.2E-5 1125 1411
U-Resnet 2.8/2.7 2.7E-5/1.8E-5 9.1E-5/7.8E-5 2.3E-5/1.2E-5 1114 1348
EDSR-U-Resnet 2.3/2.6 2.7E-5/1.9E-5 9.2E-5/8.4E-5 2.5E-5/1.2E-5 1133 1216
LR-watershed 1.9/3.3 1.9E-5/4.3E-5 7.3E-5/7.5E-4 1.1E-5/2.3E-5 751 819
LR-Otsu 2.0/1.1 1.7E-5/5.1E-5 7.4E-5/4.2E-4 9.4E-6/2.6E-5 411 499
N. J. Alqahtani et al.
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 517

connectivity of the pore space and the narrow flow paths. The results from the GT and
U-ResNet show a good match on both MEC and ILS samples. The computed relative per-
meability curves in Fig. 8 show similar saturation endpoints for water (wetting phase) and
oil (non-wetting phase). The topology of the non-wetting phase during pore desaturation is
observed to be analogous in both GT and U-ResNet segmentation (see Fig. 9). The differ-
ence maps visualized in Fig. 9e show minor differences in the non-wetting phase distribu-
tion. These differences happen mainly: (1) at the boundary of grains/fluids interface (seg-
mentation differences) and (2) the filling/desaturation of smaller pores, which perhaps arise
from differences in the overall pore topology.

4 Conclusions

The segmentation/identification of pore space, sub-resolution porosity, and the solid matrix
are vital for reliable digital rock analysis frameworks. Deep learning workflows involving
the utilization of CNNs to enhance the segmentation of grayscale images improve the over-
all outcome. CNNs can work in an end-to-end scheme to super-resolve and segment raw
X-ray images, without any interference from the user. This reduces the user prejudice asso-
ciated with classic segmentation methods which often require user input. Two CNNs train-
ing configurations were considered to super-resolve and segment grayscale images into
pore, solids, and micropores. Firstly, U-Net and U-ResNet were trained in an end-to-end
manner to super-resolve and segment images in one network. Secondly, EDSR-U-Resnet
was trained to super-resolve grayscale images at once then segment the image (two dif-
ferent networks). The output segmentation of all the CNNs frameworks shows relatively
consistent voxel-wise accuracy compared to the GT segmentation. The U-ResNet dis-
played the best performance with Jaccard indices of 0.92, 0.83, and 0.57 for solid, pore,
and micropore phases, respectively. U-Net shows very close voxel-wise accuracy margins
to U-ResNet (with less than 1% difference in Jaccard score). In general, the highest error
margins are observed in the identification of the microporosity phase. This is perhaps due

ILS: GT vs. U-RESNET MEC: GT vs. U-RESNET

Kro - GT Krw - GT Kro - U-ResNet Krw - U-ResNet Kro - GT Krw - GT Kro - U-ResNet Krw - U-ResNet

1 1
0.9 0.9
0.8 0.8
Relave Permeability
Relave Permeability

0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1
Weng Phase Saturaon Weng Phase Saturaon

Fig. 8  Secondary Imbibition Relative Permeability Curves of (left) the Indiana limestone and (right) Mid-
dle Eastern Carbonate. The ground truth and U-ResNet segmentation are compared. Matching curves and
relative permeability endpoints with minimal differences are observed

13
518 N. J. Alqahtani et al.

Fig. 9  Visualizations of the imbibition simulation on the Middle Eastern Carbonate cubic sample at wetting
phase saturation of (Swp = 0.43). Upper: 3D visuals of the non-wetting phase distribution (red) in the a GT
and the b U-ResNet segmentation. Lower: 2D slices showing the non-wetting phase distribution (yellow) of
c the GT and d U-ResNet segmentations. The difference map in e shows most of the discrepancies happen
in the filling of smaller pores and the boundaries of solid and fluids (segmentation differences)

to the effects of the noise (partial volume effects) and extreme sub-resolution in the input
LR images, making it impossible to identify microporosity by the CNNs (or even judging
visually). The results also show the phases volume fractions of network segmentation are
more conforming than using only LR segmentation or HR Otsu segmentation as compared
to GT. The same trends in terms of network consistency are observed in measuring spe-
cific surface area, however, with higher relative errors (up to 46% using U-ResNet). The
connectivity of the pore space as measured using EC number shows also high relative dif-
ferences, when network and GT segmentations are compared (up to 92%). Regardless, the
network segmentations show lower relative error compared to HR Otsu, LR watershed and
Otsu segmentations (see Tables 2 and 3).
Additionally, macro- and micropore networks comparisons with GT show better
results in terms of connectivity, pores, and pore throats sizes for CNNs segmentation.
The same outcome trends are observed in single-phase permeability and relative perme-
ability curves. Overall, the end-to-end training frameworks are found to be superior to
using two networks for super-resolution and segmentation confirming the suitability of
end-to-end learning to perform more complex tasks. The reason is likely to be the loss
of important features that distinguish the different rock phases to obtain precise segmen-
tation when upsampling LR images using super-resolution network. This leads to a gen-
eral conclusion that end-to-end CNNs training for X-ray imaging super-resolution and
processing promise a lot of improvements to current DRP frameworks. Similar applica-
tions to this study can be applied based on single or multiple rock types, where image
acquisition includes LR imaging capturing high field of view and HR region of interest
imaging capturing more explicit details of pore geometry. Also, the CNNs methods pre-
sented here do not necessarily highlight all potential improvements that can be gained.

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 519

The CNNs frameworks presented only show a general workflow for improving the accu-
racy in segmentation and modeling. The interpolation of the presented methods on other
rocks type can be established by adding subvolumes from the medium of interest to the
current dataset and commencing training in a transfer learning scheme. This eliminates
the need for training the models from the ground up. The results obtained with more
sophisticated CNNs architectures, training data and ad hoc strategies are anticipated to
boost the outcome accuracy. CNNs architectures such as HRNet-OCR (Tao et al. 2020),
and EfficientNet (Tan and Le 2019) are few examples of the active research of improv-
ing network design and performance. The automation of DRP using CNNs would also
benefit from including important physical properties, i.e., permeability, as a component
in the loss function for optimizing the CNNs performance.

Appendix

The MICP tests are performed using P ­ OREMASTER® by Quantachrome instruments on

the 6 mm core plugs after X-ray imaging. For the MICP experiment analysis, a methodol-
ogy presented by (Buiting and Clerke 2013; Clerke et al. 2008) is followed. This method
fits multiple Thomeer hyperbolas (Thomeer 1983, 1960) to quantify the different pore sys-
tems in carbonates using superposition. This enables the identification of the porosity of
macro- and micropores systems and their pore throat distribution. The pore volume frac-
tion of each pore is determined based on the volume of mercury injected (%BVocc). The
pore volume fraction obtained from the MICP for the macropore system is used to choose
a safe threshold to complete the ground truth (watershed) segmentation. The pore throats
distribution is estimated based on the minimum entry pressure (Pd) from Thomeer Hyper-
bola. The conductivity of the micropore system is then estimated based on the average
pore throats for the single-phase permeability simulation. In the below figures, we show
the Thomeer Hyperbolas (Figs. 10 and 11) and the pore throat distribution (Fig. 12) for the
ILS and MEC samples.

13
520 N. J. Alqahtani et al.

Thomeer Capillary Pressure Model (ILS Sample)

Total Porosity: 16.38 Macro Pore System: G1: 0.20 Pd1: 4.50 BV1: 9.17 Closure Corr: 0.40 Micro Pore
System: G2: 0.40 Pd2: 150.00 BV2: 7.22
100000

10000

1000

Pd (psia)
100

%Bvocc
10 (Experiment)

THOMEER
BV1
(MACRO)
1 THOMEER
100.00 10.00 1.00 0.10 BV2 (MICRO)

BV1+BV2
%BV occ

Fig. 10  The estimation of porosity contribution of each pore system in the ILS samples using Thomeer
Hyperbola

Thomeer Capillary Pressure Model (MEC Sample)

Total Porosity: 28.81 Macro Pore System: G1: 0.30 Pd1: 3.50 BV1: 17.84 Micro Pore SysteM: G2: 0.30
Pd2: 90.00 BV2: 10.97
100000

10000

1000
P (psia)

100

%Bvocc
(Experiment)
10
THOMEER
BV1 (MACRO)

THOMEER
1 BV2 (MICRO)
100.00 10.00 1.00 0.10
%BV occ BV1+BV2

Fig. 11  The estimation of porosity contribution of each pore system in the MEC samples using Thomeer
Hyperbola

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 521

Pore Throat Distribuon

0.16

0.14
MEC
Incremental Pore Volume (ΔV/V)

ILS
0.12

0.10

0.08

0.06

0.04

0.02

0.00
0.001 0.01 0.1 1 10 100 1000
Pore Throat Diameter (µm)

Fig. 12  Pore throat distribution of the ILS and MEC samples based on the MICP test

Acknowledgements The authors acknowledge the Tyree X-ray CT Facility, a UNSW network laboratory
funded by the UNSW Research Infrastructure Scheme, for the acquisition of the 3D micro-CT images.

Funding Open Access funding enabled and organized by CAUL and its Member Institutions. Not
applicable.

Data Availability Original microcomputed tomograms, mercury intrusion capillary pressure data and CNNs
training volumes will be made available on (https://​www.​digit​alroc​kspor​tal.​org/​proje​cts/​362).

Code Availability The code was developed using open-source libraries (Pytroch and Numpy) and can be
requested by contacting the authors.

Declarations
Conflict of interest The authors declare that there is no conflict of interest or competing interests involved in
the subject matter or materials discussed in this manuscript.

Consent to Participate Not applicable.

Consent for Publication Not applicable.

Ethical Approval Not applicable.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com-
mons licence, and indicate if changes were made. The images or other third party material in this article
are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

13
522 N. J. Alqahtani et al.

References
Alqahtani, N.J., Alzubaidi, F., Armstrong, R.T., Swietojanski, P., Mostaghimi, P.: Machine learning for pre-
dicting properties of porous media from 2d X-ray images. J. Pet. Sci. Eng. 184, 106514 (2020)
Alqahtani, N.J., Chung, T., Wang, Y.D., Armstrong, Ryan T., Swietojanski, P., Mostaghimi, P.: Flow-Based
Characterization of Digital Rock Images Using Deep Learning. SPE J. 26, 1800–1811 (2021). https://​
doi.​org/​10.​2118/​205376-​PA
Alzubaidi, F., Mostaghimi, P., Swietojanski, P., Clark, S.R., Armstrong, R.T.: Automated lithology classifi-
cation from drill core images using convolutional neural networks. J. Pet. Sci. Eng. 197, 107933 (2021)
Andrä, H., et al.: Digital rock physics benchmarks—Part I: imaging and segmentation. Comput. Geosci. 50,
25–32 (2013)
Ar Rushood, I., et al.: Segmentation of X-ray images of rocks using deep learning. In: SPE Annual Techni-
cal Conference and Exhibition (2020)
Arns, C.H., Knackstedt, M.A., Pinczewski, W.V., Martys, N. (2004b). Virtual permeametry on microtomo-
graphic images. J. Pet. Sci. Eng. 45, 41–46. https://​doi.​org/​10.​1016/j.​petrol.​2004.​05.​001
Arns, C.H., et al.: Pore scale characterization of carbonates using X-ray microtomography. SPE-191379-PA,
10(04): 475–484 (2005)
Arns, C.H., et al.: Pore-type partitioning for complex carbonates: effective versus total porosity and applica-
tions to electrical conductivity. In: SPWLA 59th Annual Logging Symposium (2018)
Berg, C.F., Lopez, O., Berland, H.: Industrial applications of digital rock technology. J. Pet. Sci. Eng. 157,
131–147 (2017)
Biswal, B., Øren, P.E., Held, R.J., Bakke, S., Hilfer, R.: Stochastic multiscale model for carbonate rocks.
Phys. Rev. E 75(6), 061303 (2007)
Blunt, M.J., et al.: Pore-scale imaging and modelling. Adv. Water Resour. 51, 197–216 (2013)
Boone, M.A., et al.: 3D mapping of water in oolithic limestone at atmospheric and vacuum saturation using
X-ray micro-CT differential imaging. Mater. Charact. 97, 150–160 (2014)
Buiting, J.J.M., Clerke, E.A.: Permeability from porosimetry measurements: derivation for a tortuous and
fractal tubular bundle. J. Petrol. Sci. Eng. 108, 267–278 (2013)
Bultreys, T., Van Hoorebeke, L., Cnudde, V.: Multi-scale, micro-computed tomography-based pore network
models to simulate drainage in heterogeneous rocks. Adv. Water Resour. 78, 36–49 (2015)
Bultreys, T., De Boever, W., Cnudde, V.: Imaging and image-based fluid transport modeling at the pore
scale in geological materials: a practical introduction to the current state-of-the-art. Earth Sci. Rev.
155, 93–128 (2016a)
Bultreys, T., et al.: Investigating the relative permeability behavior of microporosity-rich carbonates and
tight sandstones with multiscale pore network models. J. Geophys. Res. Solid Earth 121(11), 7929–
7945 (2016b)
Caliva, F., Iriondo, C., Morales Martinez, A., Majumdar, S., Pedoia, V.: Distance map loss penalty term for
semantic segmentation. (2019) arXiv:​1908.​03679
Canny, J.: A computational approach to edge detection. IEEE Trans. Pattern Anal. Mach. Intell. PAMI 8(6),
679–698 (1986)
Cantrell, D.L., Hagerty, R.M.: Microporosity in arab formation carbonates, Saudi Arabia. Geoarabia 4(2),
129–154 (1999)
Chen, H., et al.: Super-resolution of real-world rock microcomputed tomography images using cycle-con-
sistent generative adversarial networks. Phys. Rev. E 101(2), 023305 (2020)
Choquette, P.W., Pray, L.C.: Geologic nomenclature and classification of porosity in sedimentary carbon-
ates. AAPG Bull. 54(2), 207–250 (1970)
Chung, T., Da Wang, Y., Armstrong, R.T., Mostaghimi, P.: CNN-PFVS: integrating neural network and
finite volume models to accelerate flow simulation on pore space images. Transp. Porous Media
135(1), 25–37 (2020a)
Chung, T., Wang, Y.D., Armstrong, R.T., Mostaghimi, P.: Voxel agglomeration for accelerated estimation of
permeability from micro-CT images. J. Pet. Sci. Eng. 184, 106577 (2020b)
Chung, T., Wang, Y.D., Armstrong, R.T., Mostaghimi, P.: Approximating permeability of microcomputed-
tomography images using elliptic flow equations. SPE-191379-PA 24(03), 1154–1163 (2019)
Clerke, E.A., et al.: Application of Thomeer Hyperbolas to decode the pore systems, facies and reservoir
properties of the Upper Jurassic Arab D Limestone, Ghawar field, Saudi Arabia: a “Rosetta Stone”
approach. GeoArabia 13(4), 113–160 (2008)
Dice, L.R.: Measures of the amount of ecologic association between species. Ecology 26(3), 297–302 (1945)
Drozdzal, M., Vorontsov, E., Chartrand, G., Kadoury, S., Pal, C.: The importance of skip connections in
biomedical image segmentation. In: Carneiro, G., et al. (eds.) Deep learning and data labeling for med-
ical applications, pp. 179–187. Springer International Publishing, Cham (2016)

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 523

Garfi, G., John, C.M., Berg, S., Krevor, S.: The sensitivity of estimates of multiphase fluid and solid proper-
ties of porous rocks to image processing. Transp. Porous Media 131(3), 985–1005 (2020)
Ghous, A., et al.: 3D Characterisation of Microporosity in Carbonate Cores. In: SPWLA Middle East
Regional Symposium (2007)
Gostick, J., Khan, Z., Tranter, T., Kok, M., Agnaou, M., Sadeghi, M., Jervis, R. PoreSpy: A python toolkit
for quantitative analysis of porous Media Images. JOSS 4, 1296 (2019). https://​doi.​org/​10.​21105/​joss.​
01296.
Harbaugh, J.W.: Chapter 7 carbonate oil reservoir rocks. In: Chilingar, G.V., Bissell, H.J., Fairbridge, R.W.
(eds.) Developments in sedimentology, pp. 349–398. Elsevier, Amsterdam (1967)
Iassonov, P., Gebrenegus, T., Tuller, M.: Segmentation of X-ray computed tomography images of porous
materials: A crucial step for characterization and quantitative analysis of pore structures, Water Resour.
Res. 45, W09415 (2019). https://​doi.​org/​10.​1029/​2009W​R0080​87.
Irajian, A.-A., Bazargani-Guilani, K., Mahari, R., Solgi, A.: Porosity and rock-typing in hydrocarbon reser-
voirs case study in upper member of dalan formation in kish gas field, South of Zagros, Iran. Open J.
Geol. 6(06), 399 (2016)
Jaccard, P.: The distribution of the flora in the alpine zone.1. New Phytol. 11(2), 37–50 (1912)
Janssens, N., Huysmans, M., Swennen, R.: Computed tomography 3D super-resolution with generative
adversarial neural networks: implications on unsaturated and two-phase fluid flow. Materials 13(6),
1397 (2020)
Ji, Y., Baud, P., Vajdova, V., Wong, T.-F.: Characterization of pore geometry of indiana limestone in relation
to mechanical compaction. Oil Gas Sci. Technol. Rev. IFP Energ. Nouv. 67(5), 753–775 (2012)
Jia, F., Liu, J., Tai, X.-C.: A regularized convolutional neural network for semantic image segmentation.
Anal. Appl. 19(01), 147–165 (2021)
Kamrava, S., Tahmasebi, P., Sahimi, M.: Enhancing images of shale formations by a hybrid stochastic and
deep learning algorithm. Neural Netw. 118, 310–320 (2019)
Karimpouli, S., Tahmasebi, P.: Segmentation of digital rock images using deep convolutional autoencoder
networks. Comput. Geosci. 126, 142–150 (2019)
Kingston, A., Sakellariou, A., Varslot, T., Myers, G., Sheppard, A.: Reliable automatic alignment of tomo-
graphic projection data by passive auto-focus. Med. Phys. 38(9), 4934–4945 (2011)
Knackstedt, M.A., Arns, C.H., Ghous, A., Sakellariou, A., Senden, T.J., Sheppard, A.P., Sok, R.M., Aver-
dunk, H., Pinczewski, W.V., Padhy, G.S., Ioannidis, M.A.: 3D imaging and flow characterisation of the
pore space of carbonate core samples. Paper presented at the international symposium of the society of
core analysts, Trondheim (2006)
Lee, K., Zung, J., Li, P., Jain, V., Seung, H.S.: Superhuman accuracy on the SNEMI3D connectomics chal-
lenge. (2017) arXiv:​1706.​00120
Legland, D., Kiêu, K., Devaux, M.-F.: computation of minkowski measures on 2D and 3D binary images.
Image Anal. Stereol. (2011). https://​doi.​org/​10.​5566/​ias.​v26.​p83-​92
Legland, D.: imMinkowski (2021)
Lim, B., Son, S., Kim, H., Nah, S., Lee, K.M.: Enhanced deep residual networks for single image super-
resolution. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops
(CVPRW), pp. 1132–1140 (2017)
Lin, Q., Al-Khulaifi, Y., Blunt, M.J., Bijeljic, B.: Quantification of sub-resolution porosity in carbonate
rocks by applying high-salinity contrast brine using X-ray microtomography differential imaging.
Adv. Water Resour. 96, 306–322 (2016)
Liu, M., Mostaghimi, P.: Reactive transport modelling in dual porosity media. Chem. Eng. Sci. 190,
436–442 (2018)
Liu, M., Shabaninejad, M., Mostaghimi, P.: Predictions of permeability, surface area and average dis-
solution rate during reactive transport in multi-mineral rocks. J. Pet. Sci. Eng. 170, 130–138 (2018)
Long, H., et al.: Multi-scale imaging and modeling workflow to capture and characterize microporosity
in sandstone. In: International Symposium of the Society of Core Analysts, Napa Valley, Califor-
nia, USA (2013)
Mortensen, N.A., Okkels, F., Bruus, H.: Reexamination of Hagen-Poiseuille flow: shape dependence of
the hydraulic resistance in microchannels. Phys. Rev. E 71(5), 057301 (2005)
Mostaghimi, P., Blunt, M.J., Bijeljic, B.: Computations of absolute permeability on Micro-CT images.
Math. Geosci. 45(1), 103–125 (2013)
Mostaghimi, P., et al.: Cleat-scale characterisation of coal: an overview. J. Nat. Gas Sci. Eng. 39, 143–
160 (2017)
Niu, Y., Mostaghimi, P., Shabaninejad, M., Swietojanski, P., Armstrong, R.T.: Digital rock segmenta-
tion for petrophysical analysis with reduced user bias using convolutional neural networks. Water
Resour. Res. 56(2), e2019WR026597 (2020a)

13
524 N. J. Alqahtani et al.

Niu, Y., Wang, Y.D., Mostaghimi, P., Swietojanski, P., Armstrong, R.T.: An innovative application of
generative adversarial networks for physically accurate rock images with an unprecedented field of
view. Geophys. Res Lett. 47(23), e2020GL089029 (2020b)
Othman, F., Yu, M., Kamali, F., Hussain, F.: Fines migration during supercritical CO2 injection in sand-
stone. J. Nat. Gas Sci. Eng. 56, 344–357 (2018)
Otsu, N.: A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern.
9(1), 62–66 (1979)
Panichev, O., Voloshyna, A.: U-net based convolutional neural network for skeleton extraction. In: Pro-
ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops,
pp. 0–0 (2019)
Peters, C.A.: Accessibilities of reactive minerals in consolidated sedimentary rock: an imaging study of
three sandstones. Chem. Geol. 265(1), 198–208 (2009)
Pittman, E.D.: Microporosity in carbonate rocks. AAPG Bull. 55(10), 1873–1878 (1971)
Rabbani, A., Babaei, M.: Image-based modeling of carbon storage in fractured organic-rich shale with
deep learning acceleration. Fuel 299, 120795 (2021)
Rabbani, A., Babaei, M., Javadpour, F.: A triple pore network model (T-PNM) for gas flow simulation in
fractured, micro-porous and meso-porous media. Transp. Porous Media 132(3), 707–740 (2020a)
Rabbani, A., Babaei, M., Shams, R., Wang, Y.D., Chung, T.: DeePore: a deep learning workflow for
rapid and comprehensive characterization of porous materials. Adv. Water Resour. 146, 103787
(2020b)
Raeini, A.Q., Blunt, M.J., Bijeljic, B.: Direct simulations of two-phase flow on micro-CT images of
porous media and upscaling of pore-scale forces. Adv. Water Resour. 74, 116–126 (2014)
Ramstad, T., Øren, P.-E., Bakke, S.: Simulation of two-phase flow in reservoir rocks using a lattice
boltzmann method. SPE-191379-PA 15(04), 917–927 (2010)
Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmenta-
tion. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) Medical Image Computing and
Computer-Assisted Intervention – MICCAI 2015, pp. 234–241. Springer International Publishing,
Cham (2015)
Rumelhart, D.E., Hinton, G.E., Williams, R.J.: Learning representations by back-propagating errors.
Nature 323(6088), 533–536 (1986)
Schlüter, S., Sheppard, A., Brown, K., Wildenschild, D.: Image processing of multiphase images
obtained via X-ray microtomography: a review. Water Resour. Res. 50(4), 3615–3639 (2014)
Schmid, P.J., Henningson, D.S.: Optimal energy density growth in Hagen-Poiseuille flow. J. Fluid Mech.
277, 197–225 (1994)
Sheppard, A.P., Sok, R.M., Averdunk, H.: Techniques for image enhancement and segmentation of
tomographic images of porous materials. Phys. A 339(1), 145–151 (2004)
Sørensen, T.J.: A method of establishing groups of equal amplitude in plant sociology based on similar-
ity of species content and its application to analyses of the vegetation on Danish commons. I kom-
mission hos E. Munksgaard, København (1948)
Soulaine, C., Gjetvaj, F., Garing, C., Roman, S., Russian, A., Gouze, P., Tchelepi, H.: The impact of sub-
resolution porosity of X-ray microtomography images on the permeability. Transp. Porous Media
113, 227–243 (2016)
Tahmasebi, P., Kamrava, S., Bai, T., Sahimi, M.: Machine learning in geo- and environmental sciences:
from small to large scale. Adv. Water Resour. 142, 103619 (2020)
Tan, M., Le, Q.: Efficientnet: rethinking model scaling for convolutional neural networks In: International
Conference on Machine Learning. PMLR, pp. 6105–6114 (2019)
Tao, A., Sapra, K., Catanzaro, B.: Hierarchical multi-scale attention for semantic segmentation. (2020)
arXiv preprint arXiv:​2005.​10821.
Thomeer, J.H.M.: Introduction of a pore geometrical factor defined by the capillary pressure curve. SPE-
58046-JPT 12(03), 73–77 (1960)
Thomeer, J.H.: Air permeability as a function of three pore-network parameters. SPE-58046-JPT 35(04),
809–814 (1983)
Tuller, M., Kulkarni, R. and Fink, W. (2013). Segmentation of X-Ray CT Data of Porous Materials: A
Review of Global and Locally Adaptive Algorithms. In Soil–Water–Root Processes: Advances in
Tomography and Imaging (eds S.H. Anderson and J.W. Hopmans). https://​doi.​org/​10.​2136/​sssas​pecpu​
b61.​c8
Varfolomeev, I., Yakimchuk, I., Safonov, I.: An application of deep neural networks for segmentation of
microtomographic images of rock samples. Computers 8(4), 72 (2019)
Varslot, T., Kingston, A., Myers, G., Sheppard, A.: High-resolution helical cone-beam micro-CT with theo-
retically-exact reconstruction from experimental data. Med. Phys. 38(10), 5459–5476 (2011)

13
Super‑Resolved Segmentation of X‑ray Images of Carbonate Rocks… 525

Wan, J., et al.: Deep learning for content-based image retrieval: a comprehensive study. In: Proceedings
of the 22nd ACM international conference on Multimedia. Association for Computing Machinery,
Orlando, Florida, USA, pp. 157–166 (2014)
Wang, Y., Teng, Q., He, X., Feng, J., Zhang, T.: CT-image of rock samples super resolution using 3D convo-
lutional neural network. Comput. Geosci. 133, 104314 (2019a)
Wang, Y.D., Armstrong, R.T., Mostaghimi, P.: Enhancing resolution of digital rock images with super reso-
lution convolutional neural networks. J. Pet. Sci. Eng. 182, 106261 (2019b)
Wang, Y.D., Chung, T., Armstrong, R.T., McClure, J.E., Mostaghimi, P.: Computations of permeability of
large rock images by dual grid domain decomposition. Adv. Water Resour. 126, 1–14 (2019c)
Wang, Y.D., Armstrong, R.T., Mostaghimi, P.: Boosting resolution and recovering texture of 2D and 3D
Micro-CT images with deep learning. Water Resour. Res. 56(1), e2019WR026052 (2020)
Wang, Y.D., et al.: Accelerated computation of relative permeability by coupled morphological and direct
multiphase flow simulation. J. Comput. Phys. 401, 108966 (2020)
Wang, Y.D., Blunt, M.J., Armstrong, R.T., Mostaghimi, P.: Deep learning in pore scale imaging and mod-
eling. Earth-Sci. Rev. 215, 103555 (2021a)
Wang, Y.D., Shabaninejad, M., Armstrong, R.T., Mostaghimi, P.: Deep neural networks for improving phys-
ical accuracy of 2D and 3D multi-mineral segmentation of rock micro-CT images. Appl. Soft Comput.
104, 107185 (2021b)
Wildenschild, D., Sheppard, A.P.: X-ray imaging and analysis techniques for quantifying pore-scale struc-
ture and processes in subsurface porous medium systems. Adv. Water Resour. 51, 217–246 (2013)
Zhu, J.-Y., Park, T., Isola, P., Efros, A.A.: Unpaired image-to-image translation using cycle-consistent adver-
sarial networks. (2017) arXiv:​1703.​10593.

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

13