diagnostics

Article
Deep-Ocular: Improved Transfer Learning Architecture Using
Self-Attention and Dense Layers for Recognition of
Ocular Diseases
Qaisar Abbas 1, * , Mubarak Albathan 1 , Abdullah Altameem 1 , Riyad Saleh Almakki 1 and Ayyaz Hussain 2

1 College of Computer and Information Sciences, Imam Mohammad Ibn Saud Islamic University (IMSIU),
Riyadh 11432, Saudi Arabia; mmalbathan@imamu.edu.sa (M.A.); altameem@imamu.edu.sa (A.A.);
ralmakki@imamu.edu.sa (R.S.A.)
2 Department of Computer Science, Quaid-i-Azam University, Islamabad 44000, Pakistan;
ayyaz.hussain@qau.edu.pk
* Correspondence: qaabbas@imamu.edu.sa; Tel.: +966-5370-14011

Abstract: It is difficult for clinicians or less-experienced ophthalmologists to detect early eye-related

diseases. By hand, eye disease diagnosis is labor-intensive, prone to mistakes, and challenging
because of the variety of ocular diseases such as glaucoma (GA), diabetic retinopathy (DR), cataract
(CT), and normal eye-related diseases (NL). An automated ocular disease detection system with
computer-aided diagnosis (CAD) tools is required to recognize eye-related diseases. Nowadays,
deep learning (DL) algorithms enhance the classification results of retinograph images. To address
these issues, we developed an intelligent detection system based on retinal fundus images. To
create this system, we used ODIR and RFMiD datasets, which included various retinographics of
distinct classes of the fundus, using cutting-edge image classification algorithms like ensemble-based
transfer learning. In this paper, we suggest a three-step hybrid ensemble model that combines a
classifier, a feature extractor, and a feature selector. The original image features are first extracted
using a pre-trained AlexNet model with an enhanced structure. The improved AlexNet (iAlexNet)
Citation: Abbas, Q.; Albathan, M.; architecture with attention and dense layers offers enhanced feature extraction, task adaptability,
Altameem, A.; Almakki, R.S.; interpretability, and potential accuracy benefits compared to other transfer learning architectures,
Hussain, A. Deep-Ocular: Improved making it particularly suited for tasks like retinograph classification. The extracted features are then
Transfer Learning Architecture Using selected using the ReliefF method, and then the most crucial elements are chosen to minimize the
Self-Attention and Dense Layers for feature dimension. Finally, an XgBoost classifier offers classification outcomes based on the desired
Recognition of Ocular Diseases. features. These classifications represent different ocular illnesses. We utilized data augmentation
Diagnostics 2023, 13, 3165. https://
techniques to control class imbalance issues. The deep-ocular model, based mainly on the AlexNet-
doi.org/10.3390/diagnostics13203165
ReliefF-XgBoost model, achieves an accuracy of 95.13%. The results indicate the proposed ensemble
Academic Editors: Joanna model can assist dermatologists in making early decisions for the diagnosing and screening of
Gołȩbiewska and Dominik eye-related diseases.
C. Odrobina

Received: 28 August 2023

Keywords: computer-aided diagnosis; transfer learning; multi-label classification; ocular diseases;
Revised: 7 October 2023 diabetic retinopathy; glaucoma; cataract; XgBoost; AlexNet; deep learning
Accepted: 9 October 2023
Published: 10 October 2023

1. Introduction
A key challenge in healthcare is identifying eye disease using fundus images [1]. Any
Copyright: © 2023 by the authors.
ailment or disorder that impairs the eye’s ability to function normally or negatively affects
Licensee MDPI, Basel, Switzerland.
the eye’s visual acuity is referred to as an ocular disease [2]. Almost everyone experiences
This article is an open access article
distributed under the terms and
visual issues at some point in their lives. Some require the care of a specialist, while others
conditions of the Creative Commons
are minor that do not appear on healthcare claims or may be handled at home [3]. Globally,
Attribution (CC BY) license (https:// fundus problems are the main reason why people go blind. The most prevalent ocular
creativecommons.org/licenses/by/ diseases (AMDs) include age-related macular degeneration (AMD), cataracts, glaucoma,
4.0/). and diabetic retinopathy (DR). By 2030, more than 400 million people will have DR, per

Diagnostics 2023, 13, 3165. https://doi.org/10.3390/diagnostics13203165 https://www.mdpi.com/journal/diagnostics

Diagnostics 2023, 13, x FOR PEER REVIEW 2 of 26

Diagnostics 2023, 13, 3165 2 of 25

glaucoma, and diabetic retinopathy (DR). By 2030, more than 400 million people will have
DR, per related research [4]. These eye conditions are becoming a significant worldwide
health
relatedissue.
research [4]. These eye conditions are becoming a significant worldwide health
issue.Most importantly, the ophthalmic condition is fatal and may leave patients perma-
nently blind.
Most In clinical the
importantly, settings, the early
ophthalmic detection
condition is of these
fatal andillnesses
may leave canpatients
prevent perma-
visual
damage.
nently blind. The number
In clinical of settings,
ophthalmologists
the early and the number
detection of these ofillnesses
patients,can however,
preventare sig-
visual
nificantly
damage. The out number
of proportion. Additionally,and
of ophthalmologists manually
the numberevaluating the fundus
of patients, however, takes
areasignifi-
lot of
time
cantly andoutisofhighly dependent
proportion. on the expertise
Additionally, manually of evaluating
the ophthalmologists.
the fundus takes This makes
a lot ofthor-
time
ough fundus screening more challenging. Therefore, automated
and is highly dependent on the expertise of the ophthalmologists. This makes thorough computer-aided diagnos-
tic methods
fundus are essential
screening for identifyingTherefore,
more challenging. eye problems. This is computer-aided
automated a typical opinion diagnostic
[4].
Globally,
methods there is afor
are essential vast range of prevalence
identifying eye problems. rates of eye
This is a disorders
typical opinionbased [4].
on age, gen-
der, employment,
Globally, therelifestyle,
is a vasteconomic status, cleanliness,
range of prevalence rates of eye habits, and conventions.
disorders based on age,Accord-
gender,
ing to an article
employment, of research
lifestyle, that studied
economic status, individuals
cleanliness, in tropical
habits, andand temperate According
conventions. countries,
to an article of
unavoidable eyeresearch
infections thatare
studied individuals
more common in tropical and
in low-altitude temperatebecause
populations countries, un-
of var-
avoidable
iables eye infections
including are more sunshine,
dust, humidity, common inand low-altitude
other natural populations
elements because of variables
[5]. Additionally,
including
eye disorders dust, humidity,
present sunshine,
differently and other natural
in communities elementsand
in developing [5]. industrialized
Additionally, na- eye
disorders
tions. There present
are actual differently
rates of in communities
ocular morbidityinindeveloping
many developing and industrialized
countries, notablynations.in
Therethat
Asia, are are
actual rates of ocularand
underdiagnosed morbidity
neglected in many developing
[6]. Globally, countries,
285 million notably in
individuals Asia,
are es-
that are to
timated underdiagnosed and neglected
have visual impairments, [6]. Globally,
of which 246 million285 million
have poorindividuals
eyesightare andestimated
39 mil-
to have visual impairments, of which 246 million have poor
lion are blind [7]. The World Health Organization (WHO) estimates that 2.2 billion indi- eyesight and 39 million are
blind [7]. The World Health Organization (WHO) estimates
viduals worldwide have a close-up or distant vision issue [8]. Estimates suggest that halfthat 2.2 billion individuals
worldwide
of have a might
these situations close-up haveor been
distant vision or
avoided issue [8]. Estimates
resolved [9]. The suggest
primarythat halfofofvisual
causes these
situations might have been avoided or resolved [9]. The primary
impairment include cataracts, age-related macular degeneration, glaucoma, diabetic reti- causes of visual impair-
ment include
nopathy, cataracts,
corneal opacity, age-related
trachoma, macular degeneration,
hypertension, glaucoma,
and other diabetic
conditions retinopathy,
[10]. A micro-
corneal opacity, trachoma, hypertension, and other conditions
scopic study was conducted on the prevalence of blindness and visual impairment [10]. A microscopic studyin
was conducted on the prevalence of blindness and visual impairment
Bangladesh. Most of the population lives in rural areas. Over 80% of people who live in Bangladesh. Mostin
of the nowadays
cities populationneed lives medical
in rural areas. Overyet
attention, 80%few of people who live in
ophthalmology cities nowadays
services need
are available.
medical attention,
Although yet few ophthalmology
more businesses provide services services are available.
for blindness, Althoughismore
the incidence businesses
still low [11].
provide services for blindness, the incidence is still low [11].
In practice, there may be a higher risk of vision loss due to the increased prevalence
In practice, there may
of non-communicable be a higher
diseases, includingriskdiabetes
of visionand losssmoking
due to the increased
[12]. prevalence of
Many impoverished
non-communicable
people living in slums diseases,
have poorincluding
mental diabetes and smoking
and physical health [12].
[13]. Many
Givenimpoverished
this, offering
people living in slums have poor mental and physical health [13]. Given this, offering these
these people affordable or accessible complete eye care treatments is critical [14]. Medical
people affordable or accessible complete eye care treatments is critical [14]. Medical image
image analysis is using more and more deep-learning-based techniques [15]. Deep-learn-
analysis is using more and more deep-learning-based techniques [15]. Deep-learning-based
ing-based models have been effective in various tasks, including illness identification [16].
models have been effective in various tasks, including illness identification [16]. One of
One of the most essential steps in minimizing an ophthalmologist’s workload is the auto-
the most essential steps in minimizing an ophthalmologist’s workload is the automated
mated diagnosis of diseases. Without human interaction, illnesses may be detected using
diagnosis of diseases. Without human interaction, illnesses may be detected using deep
deep learning and computer vision technology. Only a small number of these studies were
learning and computer vision technology. Only a small number of these studies were able
able to fully diagnose more than one eye illness [17], even though many of them produced
to fully diagnose more than one eye illness [17], even though many of them produced
encouraging results. To accurately detect diverse eye conditions, more study is required
encouraging results. To accurately detect diverse eye conditions, more study is required to
to examine the many elements of fundus imaging [18]. This study suggests a deep learning
examine the many elements of fundus imaging [18]. This study suggests a deep learning
(DL) technique for recognizing eye illnesses. The most common eye-related diseases, as
(DL) technique for recognizing eye illnesses. The most common eye-related diseases, as
shown
shown in in Figure
Figure 1, 1, are
are briefly
briefly explained
explained below.
below.

Figure 1.
Figure A visual
1. A visual example
example of
of ocular
ocular eye-related
eye-related diseases
diseases studied.
studied.
Diagnostics 2023, 13, 3165 3 of 25

(a) Diabetic retinopathy: The persistently elevated blood sugar levels caused by diabetes
can damage the microscopic blood vessels (capillaries) that transport oxygen and
nutrients to the retina. Diabetic retinopathy affects approximately one-third of people
with diabetes over 50.
(b) Cataract: A cataract is the clouding of the lens of the eye. Cataracts can eventually
contribute to blindness if left untreated. People with diabetes are more likely to
develop cataracts and experience vision loss at a younger age than those without
the condition.
(c) Glaucoma is a group of conditions that can cause an optic nerve injury. The optic nerve
transmits retinal signals for processing to the brain. Glaucoma is frequently (but not
always) caused by increased intraocular pressure. Diabetes substantially increases the
risk of glaucoma in comparison to the general population. Open-angle glaucoma (also
known as “the sneak thief of sight”) and angle-closure glaucoma (which develops
abruptly and is a medical emergency) are the two most common varieties.
Multilabel categorization has been used as a different strategy [19]. Ocular disease
datasets [20,21] could be more balanced. This imbalance makes it difficult to accurately
identify or classify sickness or even a standard retinograph image. This method is not
recommended for broad classification problems due to its low accuracy. The classification
of ocular illnesses is the goal of this effort. It is not suggested to categorize any disease
using the dataset used in this study since it was very unbalanced. This mismatch caused a
lot of variability throughout training, which left a lot to be desired. The strategy we used to
address the issue was balancing the images of the classes. We balanced the ranks by taking
the same number of pictures from each category and feeding them into a pretrained AlexNet
model rather than utilizing all the images and categorizing all the illnesses simultaneously.
As a result, our study initially balanced the dataset by training the classes on the pretrained
AlexNet architecture with the same amount of data for each category. By selecting an equal
number of images for each type, we first loaded the dataset and the associated image into
the dataset. The AlexNet model used the transfer learning (TL) approach in this research.
The accuracy of each class rose once we correctly balanced the dataset.
We developed a deep-ocular detection system based on retinal fundus images. The
advantage of improving the AlexNet (iAlexNet) architecture by incorporating attention
and dense layers, compared to other transfer learning architectures, lies in its ability to
enhance feature extraction, adaptability, and interpretability for specific tasks. To create
this system, we used ODIR and RFMiD datasets, which included various retinographics
of distinct classes of the fundus, using cutting-edge image classification algorithms like
ensemble-based transfer learning. These models, however, only employ one deep learning
network to identify eye-related diseases; hence, the efficacy of this method still has to be
enhanced. In this paper, we suggest a three-step hybrid ensemble model that combines a
classifier, a feature extractor, and a feature selector. The original image features are first
extracted using a pre-trained AlexNet model with an enhanced structure to overcome
the overfitting and gradient vanishing problems. The extracted features are then selected
using the ReliefF method, and then the most crucial elements are chosen to minimize the
feature dimension. Finally, an XgBoost classifier offers classification outcomes based on the
desired features.

1.1. Research Significance

(1) A deep-ocular detection system is developed in this paper to address the issues of the
multiclass recognition of eye-related diseases based on ensemble model by combing
iAlexNet, ReliefF, and XgBoost.
(2) A preprocessing step is performed by integrating CLAHE and Ben Gram contrast
adjustment techniques. A data augmentation step is also applied to adjust class
imbalance and overfitting problems.
(3) A new model iAlexNet, developed using attention and dense layers, allows the net-
work to focus on important regions within an image. This is particularly beneficial for
Diagnostics 2023, 13, 3165 4 of 25

tasks where certain regions carry critical information (e.g., diseases in medical images).
Attention mechanisms enable the model to assign varying levels of importance to
different parts of an image, leading to improved feature extraction.
(4) The introduction of additional dense layers further refines the features extracted
by the convolutional layers in the iAlexNet model. Dense layers capture high-level
abstractions and relationships among features, enabling the model to learn more
complex patterns and representations.
(5) The improved iAlexNet architecture allows for the fine-tuning of specific layers, mak-
ing it adaptable to a wide range of tasks. By unfreezing selected layers during training,
the model can specialize its learned representations to match the requirements of
the target task. This adaptability is especially valuable when working with limited
labeled data.

1.2. Paper Organization

The remaining sections of this paper are divided into five sections. Section 2 contains
related works. It discusses the current limitations and emphasizes the primary directions
and solutions incorporated into the proposed system to address the current deficiencies.
Section 3 describes in detail the phases and techniques the proposed deep-ocular framework
utilizes. Section 4 describes and presents the results of the various investigations conducted.
Section 5 introduces the discussion and provides a comparative, analytical examination of
the proposed system and other cutting-edge techniques. Section 6 summarizes the paper’s
findings and highlights future research directions.

2. Related Work
Ophthalmology has witnessed significant advancements with the advent of artificial
intelligence (AI) and deep learning [22]. This literature review sheds light on pivotal studies
focusing on these advancements, emphasizing the range of diseases tackled, methodologies
employed, and the results achieved. The state-of-the-art studies are described in Table 1.

Table 1. A comparative study of state-of-the-art techniques to recognize ocular diseases.

Ref. Methodology Results Limitations

[23] CNN for OD localization. Replacement of blue ACC: 99.11% Limited eye diseases with huge
channel with segmented vasculature maps. computational power.
[24] Hybrid DL network with ResNet50 for DR detection ACC: 96% Only detection of DR, and computationally
and grading. expensive.
[25] Single network for multiple binary predictions on ACC: 92% Only detection of DR, and
DR severity. computationally expensive.
[26] VGG19 model for cataract detection using color ACC: 91% Need for expert ophthalmological intervention.
fundus images.
[27] Use of comprehensive OIA-ODIR dataset with nine ACC: 90% Computationally expensive and unbalanced and
CNNs for multi-disease classification. limited datasets.
[28] CNN model for multi-class DED classification. ACC: 81.33% Limited dataset and no generalized solution for
the detection of ocular diseases.
[29] Optimal residual deep neural networks for ACC: <94% Recognize limited categories of eye diseases and
diagnosing DR, glaucoma, and AMD. accuracy is not up to the mark.
[30] VGG16 model for retinal complication diagnosis <94% of ACC Detection of limited diseases and
using retinal fundus images. computationally expensive.
[31] Deep learning for eye disease classification with 90% Only detection of DR and
minimal resolution images. computationally expensive.
Diagnostics 2023, 13, 3165 5 of 25

Table 1. Cont.

Ref. Methodology Results Limitations

[32] Resnet152-based CNN for glaucoma detection. 95% Detection of glaucoma only and
computationally expensive.
[33] Deep learning algorithm with AS-OCT images. ACC < 93% Only detection of DR and
computationally expensive.
[34] EyeDeep-Net, a multi-layer neural network for Significant superiority Used unbalance dataset and model
diagnosing various eye problems. over baseline overfitting, and
computationally expensive.
[35] DCNN optimized with whale optimization 8.1% increase in accuracy Used unbalance dataset and
algorithm for the detection of multiple eye diseases. model overfitting.
[36] DCNet for patient-level multi-label ocular disease Enhanced classification Limited eye diseases with huge
classification with a multi-label soft margin loss. performance computational power.

In the early days of applying AI to ophthalmology, ref. [23] made strides by proposing
a method for optic disc (OD) localization using convolutional neural networks (CNNs).
Their innovative data preparation and two-stage training process adeptly addressed the
class-imbalance issue, demonstrating a remarkable detection rate of 99.11%. Unique to
this approach was the replacement of the less informative blue channel with segmented
vasculature maps, providing richer context for the CNN structures. Continuing on the trail
of diabetic retinopathy (DR), ref. [24] showcased a hybrid DL network to detect and grade
DR severity. With the utilization of ResNet50, they achieved an accuracy of up to 96% on
the IDRiD dataset. The potential of AI in improving clinical outcomes became even more
evident with [25], where a single network was able to predict the severity of DR with high
sensitivity and specificity.
Expanding the AI spectrum, ref. [26] addressed the global challenge of cataracts, a
predominant cause of blindness. They proposed an efficient cataract detection model
using the VGG19 network, which minimized the need for manual retinal feature extraction.
This research marked a critical transition towards automating ophthalmological diagnosis.
Ref. [27] further highlighted the inherent limitations of manual diagnoses due to constrained
medical resources. They released the comprehensive OIA-ODIR dataset and benchmarked
the performance of nine prominent CNNs, suggesting the importance of feature fusion for
multi-disease classification.
Recognizing the importance of timely examinations, ref. [28] explored the application
of AI in Diabetic Eye Disease (DED). They presented a novel model for multi-class DED
classification, achieving an accuracy of 81.33%. Their work established the potential of
optimized deep learning architectures, especially in scenarios with limited data availabil-
ity. Moving ahead, refs. [29,30] concentrated on diagnosing various eye diseases using
advanced deep learning algorithms, achieving significant accuracies and emphasizing the
efficiency of automated systems.
The global shortage of ophthalmologists is a growing concern, and ref. [31] delineated
the promise of AI in addressing this. The study demonstrated the efficacy of deep learning
in classifying eye diseases even with images of minimal resolution. Following this trend,
ref. [32] detailed a unique approach for detecting glaucoma by employing a CNN based on
the Resnet152 model, highlighting the enhanced accuracy potential of machine learning.
Recent advancements have also underscored the vast database potential, as seen
with [33], in which the authors developed an algorithm using a vast collection of AS-OCT
images. Furthermore, ref. [34] introduced EyeDeep-Net, a multi-layer neural network
designed to efficiently diagnose various eye problems, setting a benchmark in the field. Of
note, ref. [35] provided a nuanced approach, deploying a DCNN optimized with a whale
optimization algorithm. With a significant accuracy leap of 8.1%, this model effectively
detected and classified a range of eye diseases, from cataracts to AMD. Concluding this
review, ref. [36] brought forth a groundbreaking patient-level multi-label ocular disease
 baseline
[35] DCNN optimized with whale optimization al- 8.1% increase Used unbalance dataset and model overfitting.
gorithm for the detection of multiple eye dis- in accuracy
eases.
Diagnostics 2023, 13, 3165 6 of 25
[36] DCNet for patient-level multi-label ocular dis- Enhanced Limited eye diseases with huge computational
ease classification with a multi-label soft mar- classification power.
gin loss. performance
classification model, the DCNet. This model, with its three integrated components, pre-
sented a holistic approach to patient diagnostics, further accentuating the role of AI in
3. revolutionizing
Materials andophthalmological
Methods diagnostics.
The overall
3. Materials andsteps
Methodsare represented in Algorithm 1 and a systematic flow diagram of
the proposed system is presented in Figure 2. The deep-ocular system consisted of differ-
The overall steps are represented in Algorithm 1 and a systematic flow diagram of the
entproposed
phases such
systemasispreprocessing, data 2.
presented in Figure augmentation, ansystem
The deep-ocular improved AlexNet
consisted architecture,
of different
feature selection and feature classification. To develop this deep-ocular system,
phases such as preprocessing, data augmentation, an improved AlexNet architecture, fea- we inte-
grated the below
ture selection andsteps,
featureand briefly describe
classification. themthis
To develop in deep-ocular
the subsequent paragraph.
system, we integrated
the below steps, and briefly describe them in the subsequent paragraph.

Input retinographics

Figure 2.2.The
Figure Theproposed
proposed deep-ocular blockdiagram.
deep-ocular block diagram.

The accuracy and efficiency problems in the multiclass diagnosis of glaucoma (GA),
The accuracy and efficiency problems in the multiclass diagnosis of glaucoma (GA),
diabetic retinopathy (DR), cataract (CT), and normal (NL) eye-related diseases were re-
diabetic retinopathy (DR), cataract (CT), and normal (NL) eye-related diseases were re-
solved with the use of an advanced hybrid model containing feature extraction, feature
solved withand
selection, theclassification
use of an advanced
components. hybrid model
This work containing
suggests feature
a three-step extraction,
hybrid ensemble feature
selection, and classification
model including components.
a feature extractor, feature This work
selector, andsuggests
classifier atothree-step
address thishybrid
issue. ensemble
The
model
retinograph image was preprocessed, and features were first extracted using an upgraded issue.
including a feature extractor, feature selector, and classifier to address this
The retinograph
version of iAlexNetimageas awas preprocessed,
feature andReliefF
extractor. The features werethen
method firstassigned
extracted using an up-
a priority
score to
graded each extracted
version feature.
of iAlexNet as The XgBoost
a feature classifier The
extractor. received the first
ReliefF n features
method then as the
assigned a
input after determining the optimal number of input features (n) using the trial-and-error
priority score to each extracted feature. The XgBoost classifier received the first n features
asapproach.
the inputThe XgBoost
after classifier then
determining output the
the optimal classification
number results
of input based on
features theusing
(n) n chosen
the trial-
features, i.e., the DR, CT, GL, and NL classes of eye-related disease.
and-error approach. The XgBoost classifier then output the classification results based on
the n chosen features, i.e., the DR, CT, GL, and NL classes of eye-related disease.
Algorithm 1: Outline for using the improved iAlexNet architecture for feature extraction,
applying the Relief algorithm for feature selection, and then classifying the information features
using XgBoost classifier
Step 1 Let X be the dataset of eye images. Let Y be the corresponding class labels for the
images. AlexNet(), fAlexNet (x) represents the feature extraction process using the
modified AlexNet. Let F be the set of selected features after applying the Relief
algorithm. Let D be the distribution of the data. Let H be the AdaBoost to predict the
retinograph class (DR, CT, GL, and NL).
Step 2 Load pretrained AlexNet model: load a pretrained AlexNet model: mAlexNet .
Diagnostics 2023, 13, 3165 7 of 25

Algorithm 1: Outline for using the improved iAlexNet architecture for feature extraction,
applying the Relief algorithm for feature selection, and then classifying the information features
using XgBoost classifier
Step 3 Improved AlexNet architecture: load AlexNet base model with ImageNet weights,
base_model, incorporate additional convolutional and attention layers, and add dense
layers to refine features: FeaturesEnhancement(Self, FeatureMaps).
Feature extraction: fAlexNet (x) = m − iAlexNet(mAlexNet , x)
Step 4 Fine-tuning and early stopping: unfreeze selected top layers of the model, recompile
model with reduced learning rate: {model.compile(optimizer = Adam(lr = 0.0001),
loss = ‘categorical_crossentropy’, metrics = [‘accuracy’])}.
Implement early stopping to prevent overfitting: {early_stopping =
EarlyStopping(patience = 3, restore_best_weights = True)}.
Step 5 Apply relief algorithm for feature selection: F = Relief(X,Y).
Step 6 Train gradient boosting classifier on selected features:
gb_classifier.fit(X_train_selected, y} {{train}}).
Step 7 Predict class labels using the classifier: {y_pred =
gb_classifier.predict(X_test_selected)}.
Calculate accuracy by comparing predicted and actual labels: {accuracy =
accuracy_score(y}_{{test}},{y_pred)}.
Step 8 Output: display calculated accuracy and predicted class label for the new retinograph.

3.1. Data Acquisition

The performance of the proposed hybrid system was evaluated based on two publicly
available datasets, i.e., the retinal fundus multi-disease image dataset (RFMiD) [37] and
ODIR (Ocular Disease Intelligent Recognition) [38].
A dataset called ODIR (Ocular Disease Intelligent Recognition) was employed in this
work. This dataset is one of the most extensive public resources on Kaggle for identifying
eye illnesses. Eight classifications of ocular diseases are usually presented. However, we
selected only diabetes, cataracts, glaucoma, and normal eye-related diseases. All of the
photographs for this project were scaled to 224 × 224, as described in Table 2.
The retinal fundus multi-disease image dataset (RFMiD) focuses primarily on diabetic
retinopathy, glaucoma, age-related macular degeneration, and a few additional prevalent
pathologies. This paper aimed to unite the medical-image-analysis community in order to
develop methods for the automatic classification of both common and rare ocular diseases.
The retinal fundus multi-disease image dataset (RFMiD) consists of 3200 fundus images
captured with three distinct fundus cameras and annotated with 46 conditions by two
senior retinal experts. According to the authors’ knowledge, RFMiD is the only publicly
accessible dataset comprising such a wide variety of frequently seen diseases in clinical
settings. This challenge encouraged the development of generalizable models for screening
the retina, in contrast to previous efforts that focused on the detection of specific diseases.
Table 2. Distribution of original datasets obtained from ODIR and RFMiD datasets.

Data Set Images * Size Type

1. ODIR 284 + 16 (PRV) 224 × 224 Glaucoma
500 224 × 224 Diabetic retinopathy
293 + 7 (PRV) 224 × 224 Cataract
900 224 × 224 Normal
2. RFMiD 300 4288 × 2848 to 2144 × 1424 Glaucoma
370 + 130 (ODIR) 4288 × 2848 to 2144 × 1424 Diabetic retinopathy
300 (ODIR) 4288 × 2848 to 2144 × 1424 Cataract
900 4288 × 2848 to 2144 × 1424 Normal
* PRV: Privately collected from different online sources, ODIR: Ocular Disease Intelligent Recognition, RFMiD:
retinal fundus multi-disease image dataset.
Diagnostics 2023, 13, 3165 8 of 25

3.2. Preprocessing and Data Augmentations

The primary phases of the preprocessing stage were image resizing and data enhance-
ment. These steps were performed on each retinograph image. Image Resizing: Every
image was scaled to 224 × 224 pixels. We performed CLAHE and Bengram contrast en-
hancement techniques on every input image of size 224 × 224. This step increased the
contrast of the pixels so that classification accuracy would be increased. Although using
the actual dimensions could be advantageous for learning, it is necessary to resize the input
images to conserve memory and reduce training time.
Each image in the employed dataset contains multiple eye-related diseases from the
retinograph images. Nonetheless, this dataset needs to be more balanced since the number
of standard samples may vastly outnumber those of eye-related disease classes. In the
training dataset, the number of standard images (i.e., negative samples) containing no
eye diseases was 4000. In contrast, the number of images containing specific pathologies,
such as glaucoma, diabetic retinopathy, and cataract, was less than 400. Therefore, data
augmentation techniques should increase the number of images containing rare diseases
and reduce the positive–negative class imbalance. Various transformation techniques
can be applied to the image dataset to expand the dataset, address class imbalances, and
prevent overfitting. We used basic, safe, and manageable augmentation techniques in this
study to preserve labels and encourage the model to learn more general features.
Its propensity to maintain the label after transformation serves as a gauge of a data
augmentation method’s safety. To ensure that each label in the dataset occurs at least
100 times, we augmented images of rare diseases using horizontal and vertical flipping,
rotation, luminance modification, saturation modification, and hue modification. Table 3
details the employed augmentation techniques and their respective parameter values. This
step is visually displayed in Figure 3.

Table 3. Data augmentation technique applied to selected datasets.

Augmentation Method Parameters

Color Enhancement Brightness change [−25% to +25%]
Contrast enhancement CLAHE and Bengrams
Affine Transformation Flipping Vertical, horizontal, and both directions
Rotation −90◦ , +90◦ , −90◦ , +90◦

3.3. Model Architecture

The improved AlexNet (iAlexNet) architecture offers the above advantages. Other
transfer-learning architectures like ResNet, VGG, and Inception can also excel in different
scenarios based on their unique characteristics. It is recommended to experiment with
multiple architectures and analyze their performance in the context of the specific task at
hand. All of the retinograph images were sent into AlexNet for feature extraction, and the
network’s output from a particular layer was considered for classifying eye-related diseases.
Transfer learning allows standard features from trained convolutional neural networks to
be used as input classifiers for imaging tasks, since most computer-aided diagnosis (CAD)
systems and other medical-image interpretation systems cannot train convolutional neural
networks from scratch. The ReliefF method was used to rank the characteristics that were
retrieved in the first phase according to their relevance. The initial features most crucial for
categorization were then chosen through trial and error. The XgBoost model is described in
Section 3. The XgBoost model classifies the filtered features after receiving the previously
selected features as input, producing final classification results.
Diagnostics 2023,
Diagnostics 2023, 13,
13, 3165
x FOR PEER REVIEW 99 of 26
of 25

Figure 3.
Figure 3. A
A visual
visual example
example of
of preprocessing
preprocessing and
and data
data augmentation
augmentation to
to control
control class
class imbalance.
imbalance.

3.3.1. AlexNet
3.3. Model Model for Feature Extraction
Architecture
AlexNet
The improvedis a transfer
AlexNetlearning (TL) algorithm
(iAlexNet) architecturethat can
offers significantly
the above increase computing
advantages. Other
performance using two GPUs for calculations [23]. Since AlexNet
transfer-learning architectures like ResNet, VGG, and Inception can also excel in different is a massive network
with 60 million
scenarios based parameters
on their unique and 650,000 neurons,
characteristics. It itis needs a lot of labeled
recommended examples
to experiment to
with
train [24], which the labeled DR, CAT, GL, and NL image resources
multiple architectures and analyze their performance in the context of the specific task atcannot provide. When
insufficiently
hand. All of the labeled examples
retinograph are available,
images were sent transfer
into AlexNet learning is a practical
for feature and popular
extraction, and the
technique for training deep neural networks. Utilizing every
network’s output from a particular layer was considered for classifying eye-related parameter in a pre-trained
dis-
network as an initialization step might make use of characteristics
eases. Transfer learning allows standard features from trained convolutional neural net- discovered via the use of
enormous datasets. The generated parameters from using these
works to be used as input classifiers for imaging tasks, since most computer-aided diag-layers, primarily utilized
for feature
nosis (CAD) extraction,
systems and can other
aid in medical-image
training convergence. Furthermore,
interpretation systems transfer
cannotlearning may
train convo-
be implemented on regular personal computers, whereas deep network training requires a
lutional neural networks from scratch. The ReliefF method was used to rank the charac-
high-performance GPU and CPU.
teristics that were retrieved in the first phase according to their relevance. The initial fea-
Figure 4 shows the improved AlexNet pretrained model used to recognize DR, CAT,
tures most crucial for categorization were then chosen through trial and error. The
GL and NL ocular diseases. The basic AlexNet is composed of eight layers, categorized
XgBoost model is described in Section 3. The XgBoost model classifies the filtered features
into convolutional, max-pooling, and fully connected layers. The first five layers consist
after receiving the previously selected features as input, producing final classification re-
of convolutional operations, followed by some layers with max-pooling operations. The
sults.
last three layers are fully connected. The network architecture is organized in a way that
involves parallel processing across two GPUs, with the exception of the final layer. To
3.3.1. AlexNet Model for Feature Extraction
develop this improved AlexNet architecture, we added the self-attention and dense layers.
AlexNet
We is a transfer
can represent learning (TL)
the improved algorithm
iAlexNet that canassignificantly
architecture follows: increase compu-
ting performance using two GPUs for calculations [23]. Since AlexNet is a massive net-
# Convolutional layer (1): convolution operation with learnable filters and activa-
work with 60 million parameters and 650,000 neurons, it needs a lot of labeled examples
tion function.
to train [24], which the labeled DR, CAT, GL, and NL image resources cannot provide.
# Max-pooling layer (1): downsampling operation to reduce spatial dimensions.
When insufficiently labeled examples are available, transfer learning is a practical and
# Convolutional Layer (2): another convolution operation with activation.
popular
# technique layer
Max-pooling for training deep neural
(2): additional networks. Utilizing every parameter in a pre-
downsampling.
trained
# network as an
Convolutional initialization
layer step might make
(3): third convolution operationuse of characteristics
with activation. discovered via
the
# use of enormous datasets. The generated parameters
Convolutional layer (4): fourth convolution operation with activation. from using these layers, primar-
ily
# utilized for feature extraction, can aid in training convergence.
Convolutional layer (5): fifth convolution operation with activation. Furthermore, transfer
learning
# may be implemented on regular personal
Max-pooling layer (3): further spatial downsampling. computers, whereas deep network
training
# requires
Flatten layer:aflattening
high-performance
the featureGPU
maps and
to CPU.
a vector for input to fully connected layers.
# Figure 4 shows the
Fully connected improved
layer AlexNet
(1): neurons fullypretrained
connectedmodel to the used to recognize
flattened DR, CAT,
feature vector.
GL
# and NL ocular diseases. The basic AlexNet
Fully connected layer (2): second fully connected layer. is composed of eight layers, categorized
Diagnostics 2023, 13, 3165 10 of 25

Diagnostics 2023, 13, x FOR PEER REVIEW 10 of 26

# Fully connected layer (3): final fully connected layer (output layer).
# Convolutional layer (6): fifth convolution operation with activation.
into convolutional,
# max-pooling,
One self-attention layer. and fully connected layers. The first five layers consist
of
# convolutional operations, followed
Four convolutional layers (10): fifthbyconvolution
some layersoperation
with max-pooling operations. The
with activation.
last
# three layers are fully
One max-pool layer. connected. The network architecture is organized in a way that
involves
# Oneparallel processing
global average poolacross
layer. two GPUs, with the exception of the final layer. To
develop
# this improved
Two dense layers. AlexNet architecture, we added the self-attention and dense layers.

Figure 4. A modified AlexNet Architecture (iAlexNet) with Relief method and XgBoost classifier to
recognize eye-related
recognize eye-related diseases.
diseases.

The last represent

We can two layersthe
of improved
the proposed model,architecture
iAlexNet a fully connected layer with 1000 neurons
as follows:
and
o a
Convolutional layer (1): convolution operation with learnablefully
softmax layer, were replaced by our layers, which were two combined
filters layers
and activation
with function.
ten and three nodes (representing the three types of categories: DR, CAT, GL, and NL).
(These layers are shown in Figure 4) The layers improved the original AlexNet architecture.
o Max-pooling layer (1): downsampling operation to reduce spatial dimensions.
The remainder of the original model’s parameters were kept and used as initialization. The
o Convolutional Layer (2): another convolution operation with activation.
pre-trained and transferred networks were then separated, leaving the whole structure
o Max-pooling layer (2): additional downsampling.
in two pieces. The extracted features were effective for classification, and the parameters
o Convolutional layer (3): third convolution operation with activation.
in the pre-trained network had previously been trained on ImageNet with millions of
o Convolutional layer (4): fourth convolution operation with activation.
retinographics. Some settings might have only needed to be slightly changed to modify
o Convolutional layer (5): fifth convolution operation with activation.
them for the new retinographics. For training on a limited dataset, the parameters in a
o Max-pooling layer (3): further spatial downsampling.
transferred network only represent a tiny portion of the complete network.
o Flatten layer: flattening the feature maps to a vector for input to fully connected lay-
An approach to incorporate self-attention into images involves utilizing a CNN for
ers.
extracting feature maps from the initial image. This step is described in Algorithm 2. Fol-
o Fully connected layer (1): neurons fully connected to the flattened feature vector.
lowing this, self-attention is employed on these feature maps. The self-attention mechanism
o be
can Fully connected
employed layer (2):
in different second on
manners fully
theconnected layer.including spatial self-attention
feature maps,
o Fully connected layer
or channel self-attention. (3): final fully connected layer (output layer).
Convolutional
o Spatial layer (6): fifth convolution operation
self-attention encompasses the calculation of attentionwith activation.
weights based on spatial
o One self-attention layer.
correlations among various regions within the feature maps. This enables the model to
o Four convolutional
concentrate selectively on layers (10):areas
distinct fifth convolution
of the image,operation withtheir
considering activation.
spatial interrela-
o One max-pool layer.
tionships. This proves advantageous for tasks like object recognition, where the spatial
o One global
arrangement average
of object pool layer.holds significance.
components
o Two dense layers.
Diagnostics 2023, 13, 3165 11 of 25

Channel self-attention, on the other hand, computes attention weights by assessing

the relationships among different channels within the feature maps. This empowers the
model to focus discriminately on particular channels based on their pertinence to the given
task. Channel self-attention finds utility in tasks such as image classification, where specific
channels might carry greater relevance for certain classes compared to others.
Through a fusion of spatial and channel self-attention, it becomes feasible to capture
both local and global dependencies present in image features. This approach facilitates
targeted emphasis on the most pertinent image areas for a specific task. Empirical evidence
has demonstrated that self-attention yields performance enhancements across various
computer vision tasks, including image classification.

Algorithm 2: FeaturesEnhancement(FeatureMaps): Enhances the features of the given feature

maps using a self-attention mechanism
Input FeatureMaps: The feature maps are extracted from an AlexNet CNN model.
Output: Enhance the features of the given feature maps using a
self-attention mechanism.
Step 1 Extract feature maps using the CNN model: FeatureMaps = Extract-Maps().
Step 2 Compute spatial attention: ComputeSpatialAttention(FeatureMaps): begin.
Repeat for position in FeatureMaps:
AttentionScores = SpatialAttentionScores(position)
WeightedSumFeatures = WeightedSumFeatures(FeatureMaps,
AttentionScores, position)
[End for Loop]
SpatialAttentionMaps SpatialAttention(FeatureMaps)
End function
Step 3 Compute channel attention: ChannelAttention(FeatureMaps): begin.
Repeat for channel in FeatureMaps:
AttentionScores = ChannelAttentionScores(channel)
WeightedChannelFeatures = WeightedChannelFeatures(FeatureMaps,
AttentionScores, channel)
ChannelAttentionFeatureMaps = ChannelAttention(FeatureMaps)
[End for loop]
[end function]
Step 4 CombineAttention(SpatialAttentionFeatureMaps, ChannelAttentionFeatureMaps)
Step 5 EnhancedFeatureMaps = ElementWiseMultiply(SpatialAttentionFeatureMap,
ChannelAttentionFeatureMap)
Step 6 EnhancedFeatureMaps = Combine (SpatialAttentionFeatureMaps,
ChannelAttentionFeatureMaps)
[End of Algorithm]

3.3.2. Feature Selection using Relief

ReliefF, a dimension reduction technique that Kira and Rendell developed, can help
reduce computational complexity, expedite model training, and remove unused character-
istics from a data set. Algorithm 3 shows the steps to select features. The ReliefF model
can be enhanced by increasing the dataset’s noise resistance and adapting it to multi-class
situations by disregarding missing data [24]. ReliefF seeks to identify the correlations and
regularities in the dataset’s properties.

k di f f ( F ,R ,H )
W ( Fl ) = W ( Fl ) − ∑ l i i
m.k +
h j = 1 i (1)
P(c)
∑classs( Ri ) 1− p(class( R )) ∑kj=1 di f f ( Fl , Ri , Mi (c)) /Mi
i

The fundamental actions of ReliefF are given in pseudo-code format in Algorithm 3.

In this study, the ten retrieved characteristics were sorted according to relevance using the
ReliefF method. The training set’s feature data, not the test sets, were the ones that were
utilized. Following extensive testing, we concluded that, to obtain the most significant
Diagnostics 2023, 13, 3165 12 of 25

classification speed and accuracy, just a small number of the most crucial attributes needed
to be employed. The following sections will provide further examples.

Algorithm 3: Features selection using ReliefF method

Input: M learning instances with L features and C classes
Output: the vector w = W [ f 1 , f 2 , f 3 , . . . , f l ]
Step 1 [Initialized] Set all weights W (f_l) = 0, where 1,2,3. . .,l
Step 2 for i = 1 to m do
Step 3 Randomly select an instance Ri
Step 4 Find k nearest histograms h j of Ri
Step 5 Repeat for each class c6= class ( Ri ) do
Step 6 From class c, find k nearest misses m j (c) of Ri
[End for loop]
Step 7 For l = 1 to L, do
Update W(F) by using Equation (1)
[end for loop]
[end for loop]
Step 8 [end]

3.3.3. XgBoost Classifier

Based on statistical learning theory, the author created the supervised learning tech-
nique known as XgBoost [24]. The dataset was split into training and test sets before Xg-
Boost executed the learning procedure. Currently, XgBoost is widely used for classification
tasks in various areas, including text classification, bioinformatics, handwritten-character
recognition, and facial recognition. XgBoost splits the initial classification issue into multi-
classification problems. These steps are mentioned in Algorithm 4. Consequently, when
used for multi-classification, the complexity and difficulty of training rise in steps with
the number of sample categories. For XgBoost, reducing computation and computational
complexity is a known issue that necessitates new research solutions [24]. We suggest using
the ReliefF technique to address this issue and make the sample data less dimensional.
Gradient tree boosting is one of the most effective and widely applied machine learning
(ML) techniques. It excels in various applications and is a highly competent ML technique.
Consequently, it has been demonstrated that it provides state-of-the-art results for various
standard classification, regression, and ranking problems. We discuss the extreme Gradient
Boosting algorithm below.

Algorithm 4: Extreme Gradient Decision-Boosting Machine (xgBoost) algorithm

Input: Extracted feature data x = ( x1 , x2 , . . . , xn ) with labels y and test data xtest
Output: Class labels (glaucoma (GA), diabetic retinopathy (DR), cataract (CT), and normal (NL)).
Process:
Step 1. First, initialize tree as a constant: yit = f 0 = 0 for optimization and parameter defined for
the classifier.
Step 2. Construct XgBoost tree by minimizing the lost function using Equation (2).
The XgBoost() classifier is perfectly trained on the feature samples x = ( x1 , x2 , . . . , xn ).
Step 3. Repeat Step 2 until the model reaches the stop condition.
Step 4. Test samples yit are assigned the class label using the decision function of Equation (2).

f t ( xi ) = argmin L(t) = argminL (y, y + f ( x )) (2)

The parameter settings are as follows for the XgBoost classifier: a learning rate of 0.3,
max_depth of 3, gamma of 4, minchildweight of 1, subsample of 1, samplingmethod of
“uniform”, lambda of 1, alpha of 0, objective of binary: logistic, nestimators of 50, and
nthread of 16.
Diagnostics 2023, 13, 3165 13 of 25

3.4. Fine-Tune Network

Fine-tuning is a technique used in transfer learning (TL), a process in which a pre-
trained neural network is adapted to a new task or dataset. When a neural network is
trained on a large dataset for a specific task, it learns to recognize various features and
patterns within the data. Fine-tuning allows us to leverage this learned knowledge and
adapt the network to perform well on a related but different task. Fine-tuning strikes a
balance between retaining general features learned from the original task and adapting
to the specifics of the new task. It is particularly useful when we have limited data for
the new task, as it helps prevent overfitting by leveraging the knowledge contained in the
pre-trained layers. By freezing and updating layers strategically, fine-tuning allows the
network to adjust its representations to match the new data distribution, ultimately leading
to improved performance on the new task.
In the context of deep learning, the term “fine-tuning” generally refers to two main
steps: freezing and updating layers.
1. In the first step of fine-tuning, we freeze a portion of the pre-trained network’s layers.
These frozen layers are kept fixed and unchanged during the training process. By
doing so, we ensure that the network retains the knowledge it acquired from its
original task, and we focus the training process on adapting the later layers to the
new task.
2. After freezing the initial layers, we introduce new layers on top of the frozen ones.
These new layers are initialized randomly and will be trained from scratch. The
network is then trained on the new dataset using the new layers while keeping
the earlier layers frozen. This step allows the model to learn task-specific features
from the new dataset without drastically altering the knowledge it gained from the
original training.
Fine-tuning is a powerful technique, but it requires the careful consideration of hy-
perparameters, and a good understanding of the domain and the specific tasks involved.
Hence, we performed the following fine-tune steps to the network.
1. Freezing the layers of the pre-trained AlexNet model by setting layer.trainable = False
to prevent them from being updated during the initial training.
2. Compiling the model with a lower learning rate and training it using the new dataset.
3. Unfreezing the top few layers of the model (model.layers [−6:]) to allow them to
be fine-tuned.
4. Recompiling the model with a lower learning rate and continuing training using the
same data generator and new dataset. The EarlyStopping callback was implemented
to prevent overfitting.

4. Experimental Results
4.1. Experimental Setup
All deep learning (DL) models involved in the study were pre-trained on ImageNet,
and their training and testing were conducted using the publicly available datasets. This
framework was compiled with CUDA 8.0 and CUDNN 5.1 to enhance performance and
efficiency. The experiments were executed on a high-powered workstation with three Intel
CPUs, and a substantial 64 GB of memory. Ensuring smooth and seamless operations, the
workstation operated on the Windows 2022 Server operating system. This robust hardware
and software setup facilitated the comprehensive training and testing of the deep networks,
ensuring accurate and reliable results in the evaluation process.
In evaluating the performance of multi-label eye-disease classification networks, we
employed five assessment metrics, i.e., accuracy (AC), sensitivity (SE), specificity (SP),
the F1-score (F1), and the AUC, and their average, termed as the Final-score. The F1,
representing the harmonic mean of precision and recall, attains high values only when both
precision and recall are elevated. Given that Kappa and F1 focus on a single threshold, and
considering the probabilistic output of classification networks, the AUC (area under the
 employed five assessment metrics, i.e., accuracy (AC), sensitivity (SE), specificity (SP), the
F1-score (F1), and the AUC, and their average, termed as the Final-score. The F1, repre-
senting the harmonic mean of precision and recall, attains high values only when both
precision and recall are elevated. Given that Kappa and F1 focus on a single threshold,
Diagnostics 2023, 13, 3165 and considering the probabilistic output of classification networks, the AUC (area 14 of 25under
the ROC curve) was utilized to effectively assess multiple thresholds. These five metrics
were computed using the sklearn package, ensuring a thorough and efficient evaluation
process.
ROC curve) was utilized to effectively assess multiple thresholds. These five metrics were
computed
Python’susing the sklearn
pre-trained package,and
iAlexNet ensuring a thorough
transfer learning and efficient
(TL) evaluation
techniques wereprocess.
used to re-
Python’s pre-trained iAlexNet and transfer learning (TL) techniques were
fine the model. Our proposed system accuracy and loss for both training and validation used to
refine the model. Our proposed system
after 300 epochs are shown in Figure 5. accuracy and loss for both training and validation
after 300 epochs are shown in Figure 5.

(a) 5. Proposed deep-ocular model accuracy and loss for

Figure (b)both training and validation after
300 epochs.
Figure 5. Proposed deep-ocular model accuracy and loss for both training and validation after 300
epochs.
The specific fine-tuning technique was the addition of a fully connected layer with
10 neurons between layers 22 and 23 (between fc8 and drop7); the outputs of this layer were
features retrieved for further selection. The original FC8 layer was made up of 1000 neurons,
which can classify 1000 different sorts of images. The number of neurons in this layer was
fixed at 4 because our classification findings only contained four categories: glaucoma (GA),
diabetic retinopathy (DR), cataract (CT), and normal (NL).
After the model structure was modified to better meet the classification aims of this
study, the model was trained to fine-tune the weights using the training set. All of the data
were fed into the model after it had been trained to extract 10 characteristics from a total of
1743 images, including from the test set (521 images) and the training set (1222 images). By
using stochastic gradient descent with momentum (SGDM), the transferred AlexNet was
trained. Table 4 lists the training settings for AlexNet.

Table 4. Parameters used in AlexNet training.

Parameters Value
Model initial learning rate 5 × 10−4
Model learning rate and drop factor 0.1
Regularization term: L2 regularization 1 × 10−4
Max number of epochs 100
Mini batch size 32

4.2. Performance Evaluation

The CNN-based architecture AlexNet was employed in our experiment to evaluate
model performance. The enhanced retinograph image was initially collected from the
dataset in the data sequencing and splitting phase, and the data were then transformed
into train labels and target labels. The training–test split technique from the scikit-learn
package was applied. A total of 70% of the data were used for training and 30% for testing,
dividing the total into a 70:30 ratio. In this part, both models’ performance metrics and
prediction skills are shown. The following are some of the outcomes for each class: During
the process of training our models using the training data, we determined how much our
Diagnostics 2023, 13, 3165 15 of 25

models are learning from the training dataset. The primary objective of training precision is
to extract the hyperparameters and determine whether our models suffer from overfitting
or underfitting. When we completed training our models with the training dataset and
validated their performance in the validation dataset, only then could we proceed to the test
accuracy, which is the final accuracy of our models. When we refer to accuracy in this paper,
we are referring to test accuracy. Sometimes, veracity alone is not sufficient. We could not
conclude that our model was very accurate based solely on its accuracy because, in this
paper, we had to correctly classify both diseases and no diseases. Those who have a disease
were termed “positive” in terms of deep learning, whereas those who do not have a disease
were termed “negative.” Precision provides a distinct picture of how many disease patients
are correctly identified within the entire dataset. Sometimes, even precision is insufficient.
For instance, if the dataset is highly biased towards one target, recall provides the number
of correctly classified true positives, i.e., those individuals who are truly diseased and for
whom our model made a disease prediction. This recall was the most essential metric for
our research project because if we have a low recall, our model may incorrectly predict a
diseased individual.
The F1 score is often referred to as the harmonic mean of precision and recall. If
someone asserts that precision and recall are of equal importance, then he or she should
focus on the F1 score. In this research, the F1 score was our second-highest priority after
the recall. In addition, confusion matrices were utilized in this study’s model evaluation.
The confusion matrix is a common error matrix used to evaluate the efficacy of supervised
learning algorithms. Each column in a confusion matrix represents the expected category,
and the sum of each column represents the total number of expected data points for that
category. The total quantity of data in each row indicates the number of instances of that
category in the data, and each row displays the data’s actual attribution category.
The accuracy of the classification results may reach 95.33% when the top 20 significant
features are used as the XgBoost classifier’s input, and is the greatest value when compared
to other numbers of input features. Even though accuracy can remain constant when only
the first seven characteristics are entered, a larger number of features results in a longer
training period for the model. The top five characteristics provided by the iAlexNet and
ReliefF algorithms were, therefore, found to be the best ones for this application. Table 5
shows the different TL models to classify ocular diseases by using different epochs and
combination of various architectures.
We conducted numerous independent repeated tests to assess the strength of the
proposed model in terms of accuracy, specificity, and sensitivity compared with several
existing models for the accuracy comparison in the following sections. The best feature
number n was chosen for each separate repeat experiment in accordance with the unique
experimental findings, and the value of n was necessarily set to 300. The model’s generality
will not have been impacted, though, because we only needed to know the value of the
ideal feature number n once throughout the model training process.
Four models were compared in the experimental analysis that follows and is addressed
in Section 4. Improved AlexNet, ReliefF, and XgBoost were used in the suggested model
compared to other state-of-the-art deep learning models. Those models were also con-
structed to test the efficacy of the model components suggested in this paper. In Figure 6,
the model accuracy and loss for both training and validation after 300 epochs are shown.
The experimental findings in Table 6 demonstrate that all models had acceptable accuracy.
The proposed model improved on the original AlexNet’s structure in comparison to the
model. According to the results, the two models’ accuracies were comparable, but the im-
proved model (improved AlexNet) may significantly help the overall model perform better
because if model is built using one model rather than another, it must classify 1000 features,
which significantly lengthens the training time without clearly improving accuracy. A
visual result of the proposed system is presented in Figure 6.
Diagnostics 2023, 13, 3165 16 of 25

Table 5. Classify ocular diseases by using different TL models.

Original AlexNet (n = 400)

Classification Accuracy Specificity Sensitivity F-Score
1. Glaucoma 96.864 ± 1.655% 98.244 ± 1.713% 91.656 ± 6.102% 94.704 ± 3.078%
2. Diabetic Retinopathy 95.272 ± 1.835% 92.110 ± 4.613% 97.833 ± 2.290% 94.803 ± 2.389%
3. Cataract 93.528 ± 0.675% 98.145 ± 1.767% 97.611 ± 1.978% 97.854 ± 0.996%
4. Normal 96.528 ± 0.675% 98.145 ± 1.767% 97.611 ± 1.978% 97.854 ± 0.996%
Total 95.832 ± 1.895% 96.167 ± 1.540% 95.700 ± 2.007% 95.787 ± 1.953%
Improved AlexNet (n = 300)
Classification Accuracy Specificity Sensitivity F-score
1. Glaucoma 97.897 ± 0.941% 98.980 ± 0.414% 94.233 ± 3.325% 96.516 ± 1.622%
2. Diabetic Retinopathy 97.648 ± 1.000% 94.656 ± 2.719% 98.833 ± 0.960% 96.676 ± 1.362%
3. Cataract 97.648 ± 1.000% 94.656 ± 2.719% 98.833 ± 0.960% 96.676 ± 1.362%
4. Normal 98.872 ± 0.229% 98.466 ± 1.208% 98.278 ± 1.155% 98.360 ± 0.329%
Total 97.208 ± 0.955% 97.367 ± 0.821% 97.115 ± 1.027% 97.184 ± 0.977%
Improved AlexNet + XgBoost (n = 200)
Classification Accuracy Specificity Sensitivity F-score
1. Glaucoma 98.834 ± 0.436% 98.279 ± 0.555% 97.975 ± 1.262% 98.123 ± 0.712%
2. Diabetic Retinopathy 98.642 ± 0.343% 97.289 ± 1.350% 98.833 ± 0.665% 98.047 ± 0.478%
3. Cataract 98.642 ± 0.343% 97.289 ± 1.350% 98.833 ± 0.665% 98.047 ± 0.478%
4. Normal 98.815 ± 0.217% 98.934 ± 0.546% 97.611 ± 0.644% 98.266 ± 0.320%
Total 98.145 ± 0.404% 98.168 ± 0.384% 98.40 ± 0.426% 98.145 ± 0.410%
The proposed model (Improved AlexNet + ReliefF + XgBoost) (n = 300)
Classification Accuracy Specificity Sensitivity F-score
1. Glaucoma 94.082 ± 0.335% 98.412 ± 0.577% 98.650 ± 0.950% 98.528 ± 0.540%
2. Diabetic Retinopathy 95.082 ± 0.369% 98.302 ± 1.061% 99.056 ± 0.375% 98.674 ± 0.528%
3. Cataract 93.120 ± 0.273% 99.218 ± 0.467% 98.222 ± 0.861% 98.715 ± 0.403%
4. Normal 96.642 ± 0.398% 98.644 ± 0.388% 98.643 ± 0.406% 98.639 ± 0.398%
Total 95.642 ± 0.398% 98.644 ± 0.388% 98.643 ± 0.406% 98.639 ± 0.398%

Table 6. Classification time consumed by models iAlexNet + XgBoost and iAlexNet + ReliefF +
XgBoost (n = 300).

Models Feature Extraction Time/s Time for Classifier/s Total/s

iAlexNet + XgBoost 5.896 ± 0.001 0.028 ± 0.000 5.924 ± 0.001
iAlexNet + ReliefF + XgBoost 5.896 ± 0.001 0.022 ± 0.001 5.917 ± 0.001

The proposed model created an XgBoost model to categorize the retrieved features
after using AlexNet to extract the original image’s features. The accuracy of the model was
found to be 95.642 ± 0.398% when the results of the models were compared. The proposed
model also performed better than other architectures in terms of its specificity, sensitivity,
and F-score, demonstrating a significant improvement in performance over the model.
Because the proposed model was an ensemble model that employed XgBoost as a classifier
and AlexNet as a feature extractor, it was superior to the original model. Additionally, the
model we suggested was enhanced based on it. In the model, the AlexNet-extracted features
were further sorted by the ReliefF algorithm, and the trial-and-error method was used to
find the right number of feature inputs to improve the XgBoost classification performance
and, as a result, the model’s overall accuracy. Table 7 also shows that model finished the
classification task faster than the original model did (5.917 ± 0.001 s vs. 5.924 ± 0.001 s),
while the accuracy went up from 95.145 ± 0.404% to 95.642 ± 0.398% (p = 0.05, n = 300).
In terms of model correctness, the proposed model stood out among the models, with an
accuracy rating of 95.642 ± 0.398%.
 racy. A visual result of the proposed system is presented in Figure 6.

Table 6. Classification time consumed by models iAlexNet + XgBoost and iAlexNet + ReliefF +
XgBoost (n = 300).

Models Feature Extraction Time/s Time for Classifier/s Total/s

Diagnostics 2023, 13, 3165 17 of 25
iAlexNet + XgBoost 5.896 ± 0.001 0.028 ± 0.000 5.924 ± 0.001
iAlexNet + ReliefF + XgBoost 5.896 ± 0.001 0.022 ± 0.001 5.917 ± 0.001

Figure
Figure 6. 6.
AA visualdiagram
visual diagram of
of results
results derived
derivedfrom thethe
from proposed model
proposed for the
model fordetection of fourof four
the detection
classes of eye-related diseases.
classes of eye-related diseases.
The proposed model created an XgBoost model to categorize the retrieved features
Table 7. Classification efficiency of proposed deep-ocular (our proposed model (AlexNet + ReliefF +
after using AlexNet to extract the original image’s features. The accuracy of the model was
XgBoost))
found tomodel compared
be 95.642 ± 0.398%to other
whendifferent architectures
the results in terms
of the models of time in seconds.
were compared. The proposed
Using Different Feature Extractors
Models Total Time/s *
AlexNet + ReliefF + XGBOOST 5.917 ± 0.001
InceptionV3 + ReliefF + XGBOOST 6.428 ± 0.001
SqueezeNet + ReliefF + XGBOOST 5.918 ± 0.002
Using Different Feature Selectors
AlexNet + ReliefF + XGBOOST 5.917 ± 0.001
Models Total time/s
AlexNet + PSO + XGBOOST 5.918 ± 0.001
AlexNet + VGG16 + XGBOOST 5.918 ± 0.001
AlexNet + ReliefF + XGBOOST 5.917 ± 0.001
Using Different Final Classifiers
Models Total time/s
AlexNet + ReliefF + XGBOOST 5.917 ± 0.001
AlexNet + ReliefF + Random Forest 7.271 ± 0.001
AlexNet + ReliefF + ELM 6.029 ± 0.001
* The time required for total classification in seconds.
Diagnostics 2023, 13, 3165 18 of 25

In general, the performance of the four compared models got better with each added
element, proving unequivocally that under the suggested model, each component con-
tributes favorably to the ensemble model’s performance improvement. The integrated
model put out in this study can successfully and precisely carry out the identification and
classification of eye-related diseases.
Table 7 presents the classification efficiency of the proposed deep-ocular model, which
integrated AlexNet, ReliefF, and XgBoost, in comparison to various other architectures
based on the total time taken in seconds for classification. It evaluates the performance
using different feature extractors such as InceptionV3 and SqueezeNet combined with
ReliefF and XgBoost. This table illustrates that the total time taken by the proposed model
(AlexNet + ReliefF + XgBoost) was 5.917 s, with a very minimal standard deviation, and
it was compared with other configurations in which the total times varied. Moreover, the
performance of the proposed model with different feature selectors and final classifiers
was evaluated. The models, including AlexNet combined with other algorithms like PSO,
VGG16, Random Forest, and ELM, were compared in terms of the total classification
Diagnostics 2023, 13, x FOR PEER REVIEW 19 time.
of 26
The proposed model consistently showed a lower total time for classification in comparison
with configurations like AlexNet + ReliefF + Random Forest and AlexNet + ReliefF + ELM,
which took 7.271 s and 6.029 s, respectively, for total classification. The results emphasize
as shown in Table 7, there were differences in the time needed for different networks to
the efficiency of the deep-ocular model in terms of classification time. Meanwhile, as
extract features. In addition to accuracy, the AlexNet we used as feature extractor has the
shown in Table 7, there were differences in the time needed for different networks to extract
shortest running time (5.917 ± 0.001 s), which demonstrates the higher efficiency of our
features. In addition to accuracy, the AlexNet we used as feature extractor has the shortest
proposed model (AlexNet + ReliefF + XgBoost).
running time (5.917 ± 0.001 s), which demonstrates the higher efficiency of our proposed
Figure 7 illustrates the confusion matrix derived from the proposed model for the
model (AlexNet + ReliefF + XgBoost).
detection
Figure of7the three main
illustrates the classes of eye-related
confusion matrix deriveddiseases,
fromcataract, diabetic
the proposed retinopathy
model for the
and glaucoma, compared to normal retina. Additionally, as demonstrated in
detection of the three main classes of eye-related diseases, cataract, diabetic retinopathy Table 7, the
proposed
and model
glaucoma, in this study
compared performed
to normal retina.significantly
Additionally, better than the current
as demonstrated models
in Table in
7, the
terms of efficiency, classifying the test set in just 5.917 ± 0.001 s (n = 300),
proposed model in this study performed significantly better than the current models in thanks to its
smallerof neural
terms network
efficiency, construction.
classifying Ourinmodel
the test set can ±
just 5.917 drastically
0.001 s (nreduce
= 300),training
thanks totime
its
while assuring accuracy. The AUC values for our model (given in Figure
smaller neural network construction. Our model can drastically reduce training time while 8) improved
generally,accuracy.
assuring as can be observed.
The In comparison
AUC values for our modelto previous
(given inmethods,
Figure 8) the ensemble
improved model
generally,
suggested in this work can generally identify eye diseases with more
as can be observed. In comparison to previous methods, the ensemble model suggestedefficiency and accu-in
racy.
this work can generally identify eye diseases with more efficiency and accuracy.

(a) (b) (c)

Figure 7.
Figure 7. AAvisual
visualdiagram
diagram of of
thethe
confusion matrix
confusion derived
matrix fromfrom
derived the proposed modelmodel
the proposed for thefor
detec-
the
tion of four classes of eye-related diseases: (a) cataract, (b) diabetic retinopathy, and (c) glaucoma,
detection of four classes of eye-related diseases: (a) cataract, (b) diabetic retinopathy, and (c) glaucoma,
against normal.
against normal.
 (a) (b) (c)
Figure 7. A visual diagram of the confusion matrix derived from the proposed mod
Diagnostics 2023, 13, 3165
tion of four classes of eye-related diseases: (a) cataract, (b) diabetic retinopathy,
19 of 25
and
against normal.

Figure 8. AUC of a proposed model to detect glaucoma and cataract eye-related disease.
Figure 8. AUC of a proposed model to detect glaucoma and cataract eye-related dis
4.3. Feature Visualization through AI Interpretability
Grad-CAM,
4.3. Feature which standsthrough
Visualization for gradient
AI class activation map, is a technique used to
Interpretability
visualize the regions of an image that a deep neural network focuses on when making a
Grad-CAM,
particular prediction. which stands for
This visualization helpsgradient
to provideclass activation
insights map,
into what parts is a techn
of the
input imagethe
visualize contribute
regionsmostoftoantheimage
network’s decision-making
that a deep neural process. By highlighting
network focuses on w
these regions, Grad-CAM offers a better understanding of the features that the network
is recognizing. To integrate Grad-CAM into a model like our deep-ocular system, a few
steps are involved. First, you load your pre-trained model (in this case, using a modified
AlexNet) and the image you want to visualize. The model should have a global average
pooling layer followed by a dense layer for classification. Figure 9 shows the Grad-CAM
visualization of features classified using the proposed deep-ocular system.
The process began by obtaining the index of the predicted class. Then, we calculated
the gradients of the output score for the predicted class with respect to the final convo-
lutional layer’s output. These gradients indicate the importance of different parts of the
convolutional layer’s output in making the specific prediction. The higher the gradient, the
more influential that part of the image is in predicting the chosen class. After obtaining
the gradients, we calculated the heatmap by averaging the gradients along the channels
of the convolutional layer’s output. This heatmap was then normalized and resized to
match the dimensions of the original input image. The heatmap was further enhanced by
applying a color map, often the “jet” color map. Finally, the heatmap was overlaid on the
original image using an additive blending approach. The resulting image shows a visual
representation of where the model’s attention was focused when making the prediction.
 input image contribute most to the network’s decision-making process. By highlighting
these regions, Grad-CAM offers a better understanding of the features that the network is
recognizing. To integrate Grad-CAM into a model like our deep-ocular system, a few steps
are involved. First, you load your pre-trained model (in this case, using a modified
AlexNet) and the image you want to visualize. The model should have a global average
Diagnostics 2023, 13, 3165 20 of 25
pooling layer followed by a dense layer for classification. Figure 9 shows the Grad-CAM
visualization of features classified using the proposed deep-ocular system.

9. A
Figure 9. A visual
visual Grad-CAM
Grad-CAMofofthetheproposed
proposeddeep-ocular
deep-ocularsystem
systemononretinographs, where
retinographs, where(a) figure
repre-
(a) represents
sents diabetes,diabetes,
(b) showsfigure (b) showsand
the glaucoma, the (c)
glaucoma,
indicatesand
the figure (c)eye-related
cataract indicates the cataract eye-re-
diseases.
lated diseases.
5. Discussion
Globally, thebegan
The process prevalence of eye problems
by obtaining the indexvaries
of the greatly according
predicted to age,
class. Then, wegender,
calculated oc-
cupation, lifestyle,
the gradients of theeconomic status,
output score for hygiene, habits,
the predicted andwith
class traditions.
respectAccording
to the finaltoconvolu-
a study
conducted onoutput.
tional layer’s people These
in tropical and temperate
gradients areas,
indicate the unavoidable
importance eye infections
of different parts ofare
themore
con-
common
volutionalinlayer’s
low-altitude
outputpopulations
in making the because of prediction.
specific factors suchThe
as dust,
higherhumidity, sunlight,
the gradient, the
and
moreother naturalthat
influential conditions [5].image
part of the Furthermore, ocular problems
is in predicting the chosenmanifest differently
class. After obtaining in
underdeveloped and developed cultures. Many underdeveloped nations,
the gradients, we calculated the heatmap by averaging the gradients along the channelsparticularly in
Asia, have true rates of ocular morbidity that are underdiagnosed and
of the convolutional layer’s output. This heatmap was then normalized and resized toneglected [6]. A
total of 285 million people are thought to have visual impairments worldwide,
match the dimensions of the original input image. The heatmap was further enhanced by of which
246 million have poor vision and 39 million are blind [7]. According to the World Health
Organization (WHO), 2.2 billion people worldwide suffer from nearsightedness or far-
sightedness [8]. According to estimates, 50% of these circumstances might have been
prevented or resolved [9]. Cataracts, age-related macular degeneration, glaucoma, diabetic
retinopathy, corneal opacity, trachoma, hypertension, and other disorders are among the
main causes of visual impairment [10].
Deep learning-based algorithms are increasingly being used in medical-picture anal-
ysis [15]. Several tasks, including identifying illnesses, have shown the effectiveness of
deep-learning-based models [16]. Automated illness diagnosis is one of the most crucial
Diagnostics 2023, 13, 3165 21 of 25

advances in reducing an ophthalmologist’s workload. Deep learning and computer vision

techniques can be used to identify illnesses without the need for human input. Even though
several of these studies have yielded hopeful findings, only a limited number of them have
been able to fully diagnose more than one eye disease [17]. More research is needed to
analyze the different components of fundus imaging in order to accurately detect a variety
of eye disorders [18]. This study suggests using deep learning (DL) to identify various
eye diseases.
Another tactic has been to use multilabel classification [19]. The databases on ocular
diseases [37,38] are very uneven. Due to this imbalance, it is challenging to correctly detect
or categorize disease or even a typical retinograph image. Due to its poor accuracy, this
approach is not suggested for situations involving broad classification. The objective of
this effort was to categorize ocular diseases. Given how unevenly distributed the dataset
utilized in this study was, it is not advised to categorize any diseases using it. Because of this
mismatch a lot of variety during training was undesirable. By balancing the representations
of the classes, we were able to address this problem. We created a technique for deep-ocular
detection using retinal fundus images. We made this system by using state-of-the-art
image classification techniques, such as ensemble-based transfer learning, on the ODIR and
RFMiD datasets, which have many images of different classes of the fundus. To diagnose
eye-related disorders, the current models only use one deep learning network; therefore,
the method’s effectiveness still has to be improved. In this study, we proposed a hybrid
ensemble model with three steps that included a classifier, a feature extractor, and a feature
selector. In order to avoid gradient vanishing and overfitting, the original picture features
were first retrieved using a pre-trained AlexNet model with an improved structure. The
most important features were then selected from the retrieved features using the ReliefF
method in order to reduce the feature dimension. An XgBoost classifier gave classification
results based on the selected features at the end.
The training–test split technique from the scikit-learn package was applied. A total
of 70% of the data were used for training and 30% for testing, dividing the total into a
70:30 ratio. When the XgBoost classifier was fed the top 20 significant features, which was
the highest value when compared to other numbers of input features, the accuracy of the
classification results could reach 95.33%, as shown in Figure 7. Even though the accuracy
can remain constant when only the first seven characteristics are entered, a larger number
of features necessitates a longer training period for the model. Therefore, it was determined
that the top five features offered by the AlexNet and ReliefF algorithms were the most
appropriate for this application. All of the models had respectable accuracy, as seen in
Table 4. The model had a structure that was superior to the original AlexNet. The results
show that the accuracies of the two models were comparable, but the proposed model
(improved AlexNet) may significantly help a model perform better, since if a model is built
using one model rather than another model, it must classify 1000 features, significantly
increasing the training time without clearly improving accuracy.
When comparing the performance of the proposed iAlexNet + ReliefF + XgBoost
classifier to the studies listed in Table 1, several considerations were observed. The proposed
system boasted an accuracy (Acc) of 95.642%, a sensitivity (SE) of 98.644%, and a specificity
(SP) of 98.643%, with an F1 Score of 98.639 and an area under the curve (AUC) of 95.
These metrics collectively offer a comprehensive insight into the model’s robustness and
reliability. In terms of accuracy, the proposed system aligned competitively with several
studies noted in Table 1. For instance, its accuracy was in the same range as studies [24,32],
which reported accuracies around 96% and 95%, respectively. However, it did not surpass
the accuracy of the methodology reported in [23], which achieved an impressive 99.11%
accuracy. Nevertheless, the proposed model outperformed those in other studies like [26–31],
whose accuracies fell below 95%. Evaluating on the basis of sensitivity and specificity, the
proposed model demonstrates high performance, with both metrics exceeding 98%. This
indicates a solid ability to correctly identify positive and negative cases, showcasing the
model’s effectiveness.
Diagnostics 2023, 13, 3165 22 of 25

The integrated module successfully and precisely identified and classified eye-related
diseases. The proposed model (improved AlexNet, ReliefF, and XgBoost) had the highest
accuracy value, demonstrating its excellence. The feature extraction component had the
greatest impact on the model’s accuracy, with InceptionV3, SqueezeNet, and AlexNet
being the most accurate. Finding a suitable network for feature extraction is crucial for
successful classification. This study also compared the performance of five existing models
with the new model to confirm its usefulness. The proposed model had the best overall
accuracy rate (95.642%), reducing the training time and ensuring the accuracy, as reported
in Table 6 and Figure 7. The ensemble model improves our overall efficiency and accuracy
in identifying eye diseases compared to previous methods.
Utilizing the ODIR and RFMiD datasets, our approach integrated cutting-edge image
classification techniques through an ensemble-based transfer learning approach. Central
to our proposal was a novel three-step hybrid ensemble model, encompassing a classi-
fier, feature extractor, and feature selector. Our approach harnessed the potential of an
improved AlexNet architecture, featuring attention and dense layers, to enhance feature
extraction, adaptability, interpretability, and accuracy. This architecture demonstrated supe-
rior performance compared to other transfer learning approaches, especially for retinograph
classification tasks.
Moreover, our framework incorporates the ReliefF feature selection method to retain
the most salient features and reduce dimensionality. The subsequent classification using
XgBoost is capable of accurately identifying different ocular diseases. By integrating data
augmentation techniques, we effectively address class-imbalance concerns. Notably, our
deep-ocular model, predominantly built upon the AlexNet-ReliefF-XgBoost ensemble,
achieved an impressive accuracy of 95.13%. This underscores the potential of our proposed
approach to significantly aid dermatologists in making informed decisions for the early
diagnosis and screening of eye-related diseases, ultimately contributing to improved patient
care and outcomes.

5.1. Advantages of Proposed System

We developed an intelligent detection system for the early diagnosis of various eye-
related diseases using retinal fundus images. This is a significant contribution, as automated
systems can greatly aid clinicians and less-experienced ophthalmologists in identifying
such diseases accurately and efficiently. The deep-ocular system’s approach incorporates
deep learning (DL) algorithms to enhance classification results, which is a common and
effective strategy for image-based medical diagnosis. The use of ensemble-based transfer
learning (TL), combining a classifier, a feature extractor, and a feature selector, demonstrates
a comprehensive approach to improving classification performance. The following are the
key steps in the deep-ocular system:
(1) We used the ODIR and RFMiD datasets, which provide a diverse set of retinal fun-
dus images representing various classes, including glaucoma, diabetic retinopathy,
cataract, and normal eyes.
(2) We utilized a basic pre-trained AlexNet model and improved it through a self-attention
mechanism and dense layers as an effective feature extractor. This model is known for
its effectiveness in image classification tasks. Enhancements to the model’s structure
were made to address overfitting and gradient vanishing issues. Extracted features
from the images served as inputs for subsequent steps.
(3) The ReliefF method was applied to select relevant features and reduce dimensionality.
This is crucial for improving model efficiency and reducing the risk of overfitting.
(4) An XgBoost classifier was employed for the final classification task. XgBoost is a
popular machine learning algorithm that excels in tabular data and has been effective
in various domains, including medical diagnosis.
(5) To tackle class imbalance issues, we employed data augmentation techniques. This
helped ensure that the model was exposed to a balanced representation of different
classes, preventing it from being biased toward the majority class.
Diagnostics 2023, 13, 3165 23 of 25

(6) The proposed ensemble model, named “Deep-ocular,” achieved an impressive accu-
racy of 95.13%. This accuracy suggests that our system is highly capable of accurately
classifying various ocular diseases.
(7) The high accuracy achieved by our model indicates its potential to be a valuable tool
for dermatologists and healthcare professionals, aiding them in making early and
accurate decisions for diagnosing and screening eye-related diseases.

5.2. Limitations of Proposed System

This work addresses a critical need in the field of ophthalmology by providing an
automated system that can assist medical professionals in identifying eye diseases more
effectively. The integration of deep learning, ensemble methods, and data augmentation
showcases a comprehensive approach to developing a robust diagnostic tool. However,
there are some limitations of the proposed deep-ocular system, as described in Table 8.

Table 8. Limitations of proposed system, when further developing and deploying the deep-ocular
system in clinical or research settings.

Limitation Description
The system’s performance might degrade when applied to datasets from different sources or
Limited Generalization populations due to differences in imaging quality, patient demographics, and diseases. In addition,
other ocular diseases such as hypertension should be considered.
The system’s performance heavily relies on the availability and diversity of high-quality labeled
Data Availability
retinal fundus images for training. Limited data could hinder its effectiveness.
Despite enhancements, overfitting might still occur due to the complexity of deep learning models.
Overfitting
Regularization techniques might be needed to mitigate this issue.
The system’s applicability is limited to eye diseases that can be diagnosed using retinal fundus
Limited to fundus images
images. Conditions that require other types of tests will not be detected.
Training and utilizing deep learning models can demand significant computational resources, which
Computational Resources
might limit the system’s accessibility in resource-constrained settings.

6. Conclusions
In conclusion, the challenges faced by clinicians and less-experienced ophthalmologists
in identifying early eye-related diseases have underscored the need for automated ocular
disease detection systems. Manual diagnosis is riddled with labor-intensive efforts and
the potential for errors, magnified by the diversity of ocular conditions such as glaucoma,
diabetic retinopathy, cataract, and normal cases. In response, this study presents a compre-
hensive solution in the form of an intelligent detection system for retinal fundus images.
Leveraging the power of deep learning algorithms, particularly the iAlexNet architecture
enhanced with attention and dense layers, offers a paradigm shift in feature extraction,
adaptability, interpretability, and accuracy within the realm of retinograph classification.
To realize this solution, this study harnessed the potential of ensemble-based transfer
learning, establishing a three-step hybrid ensemble model that seamlessly integrated a clas-
sifier, a feature extractor, and a feature selector. The initial phase involved feature extraction
utilizing a pre-trained AlexNet model, tailored with attention and dense layers to heighten
its performance. This improved architecture, iAlexNet, offers distinct advantages over
other transfer learning models, catering to the unique demands of retinograph classification
tasks. The feature selection phase employs ReliefF to identify and retain relevant features,
while the dimensionality is minimized for computational efficiency. Ultimately, a robust
XgBoost classifier culminates in disease classification, yielding differentiated outcomes for
various ocular conditions.
The efficacy of the proposed model was evidenced by the achievement of a remarkable
95.13% accuracy in the deep-ocular model. These promising results underscore the potential
of this ensemble approach, affirming its viability as a valuable tool for aiding healthcare
professionals in the timely diagnosis and screening of eye-related diseases. By combining
Diagnostics 2023, 13, 3165 24 of 25

cutting-edge techniques in deep learning, feature selection, and classification, this study
contributes to a pivotal advancement in ocular disease detection and management.

Author Contributions: Conceptualization, Q.A., M.A., A.A., R.S.A. and A.H.; data curation, Q.A.,
R.S.A. and A.H.; formal analysis, Q.A., M.A., A.A., R.S.A. and A.H.; funding acquisition, M.A. and
A.A.; investigation, Q.A., M.A., A.A. and R.S.A.; methodology, Q.A., M.A., R.S.A. and A.H.; project
administration, Q.A. and M.A.; resources, M.A., A.A. and A.H.; software, Q.A., M.A., A.A., R.S.A.
and A.H.; supervision, A.H.; validation, Q.A., M.A., A.A. and R.S.A.; visualization, R.S.A. and A.H.;
writing—original draft, Q.A., M.A., A.A., R.S.A. and A.H.; writing—review and editing, Q.A., M.A.,
A.A., R.S.A. and A.H. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported and funded by the Deanship of Scientific Research at Imam
Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-RP23079).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The Python code is provided on GitHub (https://github.com/Qaisar2
56/Deep-ocular) to better understand the deep-ocular system.
Acknowledgments: This work was supported and funded by the Deanship of Scientific Research at
Imam Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-RP23079).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Abbas, Q.; Qureshi, I.; Yan, J.; Shaheed, K. Machine learning methods for diagnosis of eye-related diseases: A systematic review
study based on ophthalmic imaging modalities. Arch. Comput. Methods Eng. 2022, 29, 3861–3918. [CrossRef]
2. Sajid, M.Z.; Qureshi, I.; Abbas, Q.; Albathan, M.; Shaheed, K.; Youssef, A.; Hussain, A. Mobile-HR: An Ophthalmologic-Based
Classification System for Diagnosis of Hypertensive Retinopathy Using Optimized MobileNet Architecture. Diagnostics 2023,
13, 1439. [CrossRef] [PubMed]
3. Qureshi, I.; Abbas, Q.; Yan, J.; Hussain, A.; Shaheed, K.; Baig, A.R. Computer-Aided Detection of Hypertensive Retinopathy
Using Depth-Wise Separable CNN. Appl. Sci. 2022, 12, 12086. [CrossRef]
4. Guan, H.; Zhang, Y.; Wang, Z.; Chen, X.; Xue, J.; Ding, Y. The prevalence and county-level socioeconomic correlates of visual impairment
and glasses ownership among rural students in Western China. Front. Public Health 2023, 11, 1069793. [CrossRef] [PubMed]
5. Qureshi, I.; Ma, J.; Abbas, Q. Diabetic retinopathy detection and stage classification in eye fundus images using active deep
learning. Multimed. Tools Appl. 2021, 80, 11691–11721. [CrossRef]
6. Katibeh, M.; Pakravan, M.; Yaseri, M.; Pakbin, M.; Soleimanizad, R. Prevalence and causes of visual impairment and blindness in
central Iran; the Yazd eye study. J. Ophthalmic Vis. Res. 2015, 10, 279–285. [PubMed]
7. Ivleva, S.; Hajimajeed, M.; Kartashova, O.; Mikhaylov, A. A comparative study of regional variations in common visual
impairments among school-age children. Electron. J. Gen. Med. 2023, 20, em504. [CrossRef]
8. Ruamviboonsuk, P.; Tiwari, R.; Sayres, R.; Nganthavee, V.; Hemarat, K.; Kongprayoon, A.; Webster, D.R. Real-time diabetic
retinopathy screening by deep learning in a multisite national screening programme: A prospective interventional cohort study.
Lancet Digit. Health 2022, 4, e235–e244. [CrossRef] [PubMed]
9. Virgili, G.; Parravano, M.; Petri, D.; Maurutto, E.; Menchini, F.; Lanzetta, P.; Lucenteforte, E. The association between vision
impairment and depression: A systematic review of population-based studies. J. Clin. Med. 2022, 11, 2412. [CrossRef]
10. Müller, K.; Engel, C.; Loitsch, C.; Stiefelhagen, R.; Weber, G. Traveling more independently: A study on the diverse needs and
challenges of people with visual or mobility impairments in unfamiliar indoor environments. ACM Trans. Access. Comput.
TACCESS 2022, 15, 13. [CrossRef]
11. Ting, D.S.; Peng, L.; Varadarajan, A.V.; Keane, P.A.; Burlina, P.M.; Chiang, M.F.; Schmetterer, L.; Pasquale, L.R.; Bressler, N.M.;
Webster, D.R.; et al. Deep learning in ophthalmology: The technical and clinical considerations. Prog. Retin. Eye Res. 2019,
72, 100759. [CrossRef] [PubMed]
12. Fan, R.; Alipour, K.; Bowd, C.; Christopher, M.; Brye, N.; Proudfoot, J.A.; Zangwill, L.M. Detecting Glaucoma from Fundus
Photographs Using Deep Learning without Convolutions: Transformer for Improved Generalization. Ophthalmol. Sci. 2023,
3, 100233. [CrossRef]
13. He, J.; Li, C.; Ye, J.; Qiao, Y.; Gu, L. Self-speculation of clinical features based on knowledge distillation for accurate ocular disease
classification. Biomed. Signal Process. Control. 2021, 67, 102491. [CrossRef]
14. Zang, P.; Hormel, T.T.; Hwang, T.S.; Bailey, S.T.; Huang, D.; Jia, Y. Deep-Learning–Aided Diagnosis of Diabetic Retinopathy,
Age-Related Macular Degeneration, and Glaucoma Based on Structural and Angiographic OCT. Ophthalmol. Sci. 2023, 3, 100245.
[CrossRef] [PubMed]
Diagnostics 2023, 13, 3165 25 of 25

15. Roy, A.G.; Conjeti, S.; Karri, S.P.K.; Sheet, D.; Katouzian, A.; Wachinger, C.; Navab, N. ReLayNet: Retinal layer and fluid
segmentation of macular optical coherence tomography using fully convolutional networks. Biomed. Opt. Express 2017, 8,
3627–3642. [CrossRef]
16. Lee, C.S.; Tyring, A.J.; Deruyter, N.P.; Wu, Y.; Rokem, A.; Lee, A.Y. Deep-learning based, automated segmentation of macular
edema in optical coherence tomography. Biomed. Express 2017, 8, 3440–3448. [CrossRef]
17. Karri, S.P.K.; Chakraborty, D.; Chatterjee, J. Transfer learning based classification of optical coherence tomography images with
diabetic macular edema and dry age-related macular degeneration. Biomed. Opt. Express 2017, 8, 579–592. [CrossRef]
18. Oda, M.; Yamaguchi, T.; Fukuoka, H.; Ueno, Y.; Mori, K. Automated Eye Disease Classification Method from Anterior Eye Image
Using Anatomical Structure Focused Image Classification Technique. In Medical Imaging 2020: Computer-Aided Diagnosis; SPIE:
Houston, TX, USA, 2020; Volume 11314, pp. 991–996.
19. Eperjesi, F.; Fowler, C.W.; Kempster, A.J. Luminance and chromatic contrast effects on reading and object recognition in low
vision: A review of the literature. Ophthalmic Physiol. Opt. 1995, 15, 561–568. [CrossRef]
20. Ocular Disease Recognition Dataset. Available online: https://www.kaggle.com/andrewmvd/ocular-disease-recognition-odir5k
(accessed on 12 January 2022).
21. Pachade, S.; Porwal, P.; Thulkar, D.; Kokare, M.; Deshmukh, G.; Sahasrabuddhe, V.; Mériaudeau, F. Retinal fundus multi-disease
image dataset (rfmid): A dataset for multi-disease detection research. Data 2021, 6, 14. [CrossRef]
22. Tayal, A.; Gupta, J.; Solanki, A.; Bisht, K.; Nayyar, A.; Masud, M. DL-CNN-based approach with image processing techniques for
diagnosis of retinal diseases. Multimed. Syst. 2022, 28, 1417–1438. [CrossRef]
23. Meng, X.; Xi, X.; Yang, L.; Zhang, G.; Yin, Y.; Chen, X. Fast and effective optic disk localization based on convolutional neural
network. Neurocomputing 2018, 312, 285–295. [CrossRef]
24. Reddy, V.P.C.; Gurrala, K.K. Joint DR-DME classification using deep learning-CNN based modified grey-wolf optimizer with
variable weights. Biomed. Signal Process. Control. 2022, 73, 103439. [CrossRef]
25. Gulshan, V.; Peng, L.; Coram, M.; Stumpe, M.C.; Wu, D.; Narayanaswamy, A.; Venugopalan, S.; Widner, K.; Madams, T.;
Cuadros, J.; et al. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus
photographs. JAMA 2016, 316, 24–35. [CrossRef]
26. Khan, M.S.M.; Ahmed, M.; Rasel, R.Z.; Khan, M.M. Cataract detection using convolutional neural network with VGG-19 model.
In Proceedings of the 2021 IEEE World AI IoT Congress (AIIoT), Virtual, 10–13 May 2021; pp. 0209–0212.
27. Li, N.; Li, T.; Hu, C.; Wang, K.; Kang, H. A benchmark of ocular disease intelligent recognition: One shot for multi-disease detection.
In Benchmarking, Measuring, and Optimizing: Proceedings of the Third BenchCouncil International Symposium, Bench 2020, Virtual Event,
15–16 November 2020; Revised Selected Papers 3; Springer International Publishing: Cham, Switzerland, 2021; pp. 177–193.
28. Sarki, R.; Ahmed, K.; Wang, H.; Zhang, Y.; Wang, K. Convolutional neural network for multi-class classification of diabetic eye
disease. EAI Endorsed Trans. Scalable Inf. Syst. 2021, 9, 23–38. [CrossRef]
29. Chea, N.; Nam, Y. Classification of Fundus Images Based on Deep Learning for Detecting Eye Diseases. Comput. Mater. Contin.
2021, 67, 412–426. [CrossRef]
30. Sarki, R.; Ahmed, K.; Wang, H.; Zhang, Y. Automated detection of mild and multi-class diabetic eye diseases using deep learning.
Health Inf. Sci. Syst. 2020, 8, 32. [CrossRef]
31. Triwijoyo, B.K.; Sabarguna, B.S.; Budiharto, W.; Abdurachman, E. Deep learning approach for classification of eye diseases based
on color fundus images. In Diabetes and Fundus OCT; Elsevier: Amsterdam, The Netherlands, 2020; pp. 25–57.
32. Badah, N.; Algefes, A.; AlArjani, A.; Mokni, R. Automatic eye disease detection using machine learning and deep learning
models. In Pervasive Computing and Social Networking: Proceedings of ICPCSN 2022, Salem, India, 3–4 March 2022; Springer Nature:
Singapore, 2022; pp. 773–787.
33. Elsawy, A.; Eleiwa, T.; Chase, C.; Ozcan, E.; Tolba, M.; Feuer, W.; Abdel-Mottaleb, M.; Shousha, M.A. Multidisease deep learning
neural network for the diagnosis of corneal diseases. Am. J. Ophthalmol. 2021, 226, 252–261. [CrossRef] [PubMed]
34. Sengar, N.; Joshi, R.C.; Dutta, M.K.; Burget, R. EyeDeep-Net: A multi-class diagnosis of retinal diseases using deep neural
network. Neural Comput. Appl. 2023, 35, 10551–10571. [CrossRef]
35. Chellaswamy, C.; Geetha, T.S.; Ramasubramanian, B.; Abirami, R.; Archana, B.; Bharathi, A.D. Optimized Convolutional Neural
Network based Multiple Eye Disease Detection and Information Sharing System. In Proceedings of the 2022 6th International
Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India, 25–27 May 2022; pp. 1105–1113. [CrossRef]
36. He, J.; Li, C.; Ye, J.; Qiao, Y.; Gu, L. Multi-label ocular disease classification with a dense correlation deep neural network. Biomed.
Signal Process. Control 2020, 63, 102167. [CrossRef]
37. Ocular Disease Intelligent Recognition through Deep Learning Architectures. Available online: https://jordicorbilla.github.io/
ocular-disease-intelligent-recognition-deep-learning/ (accessed on 12 January 2023).
38. Guergueb, T.; Akhloufi, M.A. Ocular diseases detection using recent deep learning techniques. In Proceedings of the Annual
International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), Virtual, 1–5 November 2021; pp. 3336–3339.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.