Image-based Chronic Kidney Disease Diagnosis Using 2D Convolutional

Neural Networks in the Context of a Comprehensive Artificial

Intelligence-Driven Healthcare System

FRANK EDUGHOM EKPAR1,2,3

1Founder, Scholars University Ltd, Port Harcourt, NIGERIA

2Department of Computer Engineering, Rivers State University, Port Harcourt, NIGERIA

3Department of Computer Engineering, Topfaith University, Mkpatak, NIGERIA

Abstract: - Reports published by the World Health Organization (WHO) indicate that noncommunicable

diseases (NCDs) including chronic kidney disease (CKD) are among the top ten causes of mortality worldwide.

Accurate and early diagnosis of chronic kidney disease could save lives, ameliorate deleterious effects and

dramatically improve quality of life. This paper presents a system that harnesses convolutional neural networks

(CNNs) that could be incorporated into a comprehensive artificial intelligence (AI)-driven healthcare system

for the automated diagnosis of chronic kidney disease. Utilizing publicly available image datasets featuring

images representing normal kidney states, cysts, tumors and kidney stones split into training and validation

samples, the system achieves an accuracy approximating 97% on the training and validation datasets.

Key-Words: - Chronic Kidney Disease (CKD), Artificial Intelligence (AI), Deep Learning (DL), Convolutional

Neural Network (CNN), Two-dimensional (2D) Convolutional Neural Network (2D CNN), Healthcare System,

CT Image

Received: March 29, 2024. Revised: Agust 25, 2024. Accepted: September 29, 2024. Published: November 14, 2024.

1 Introduction

In addition to the high mortality rate measured in

millions per annum globally, chronic kidney disease

exerts an enormous toll in terms of lost economic

opportunities and physical and psychological

suffering [1] – [2]. Accurate and early diagnosis

could be leveraged to create more efficacious

therapies and management strategies and

consequently save lives and enhance quality of life

indicators and achieve improved health outcomes

across board. Ekpar [3] introduced a comprehensive

artificial intelligence (AI)-driven healthcare system

with a modular design that could accommodate AI

models for the diagnosis of chronic kidney disease.

Previous work has been reported in the literature

featuring the utilization of AI models for the

diagnosis of a wide range of health conditions

including chronic kidney disease [4] – [29] with

varying degrees of success.

This paper presents a system built on

convolutional neural networks (CNNs) trained to

classify diagnostic images. The system employs

publicly available diagnostic CT-radiography

images indicating the normal kidney state as well as

the presence of kidney cysts, stones and tumors [4]

with the possibility of incorporating the resulting

system into a comprehensive artificial intelligence

(AI)-driven healthcare system.

Further integration of genetic and environmental

factors could enhance the utility of the system with

more accurate representations of the circumstances

of the participants. Additionally, locally aggregated

datasets could augment and/or replace the publicly

available datasets currently harnessed to reduce bias

and promote the global relevance of the decisions

that could be supported by the inferences drawn

from the system.

The modular design makes the system amenable

to the incorporation of new modules for the

diagnosis, prediction and management of additional

health conditions and the enhancement of existing

modules on the basis of fresh data.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

135

Volume 4, 2024

2 Materials and Methods

This section outlines how participants could be

recruited for experiments on the basis of ethical

clearances obtained as well as the methods pursued

for the actualization of the system.

2.1 Participant Recruitment

Participants willingly took part in the studies

focused on developing a comprehensive AI-driven

healthcare system, each providing informed consent

for their participation.

2.2 Ethical Clearance

The Health Research Ethics Committee at the

Institute of Biomedical Research, University of

Uyo, approved the studies ethically. All research

complied with relevant ethical and regulatory

standards, and publicly available data was utilized in

line with the licensing terms established by its

creators.

2.3 Methodology

Publicly accessible healthcare datasets can be

improved by integrating data collected from local

experiments and data collection initiatives, which

can be used to train AI models for actionable

predictions based on new data. Sources of public

healthcare datasets include the Centers for Disease

Control, the University of California Irvine Machine

Learning Repository, the American Epilepsy

Society, and Kaggle.

Incorporating local data enhances robustness,

minimizes bias, and fosters inclusivity and global

relevance.

One innovative approach in this project is

combining diagnostic measurements, including

electrocardiographic results, from local experiments

with EEG data from both traditional and novel

advanced three-dimensional multilayer EEG

systems.

For local data collection efforts, the research

has received ethical approval from the relevant

ethics committees overseeing the regions where the

experiments take place. Furthermore, partnerships

have been established with licensed medical doctors

experienced in these areas, who have direct access

to patients and other clinicians in the community.

These doctors will collaborate with the project to

provide anonymized clinical measurements for

validating the AI models.

The trained AI models could be integrated into

a comprehensive healthcare system that offers

clinical decision support to medical practitioners

and facilitates the generation of brain-computer

interfaces (BCIs). This support will be based on

actionable insights derived from new clinical data

provided by medical professionals, aiding in the

early detection, diagnosis, treatment, prediction, and

prevention of various conditions, including diabetes

mellitus, heart disease, stroke, autism, and epilepsy.

This project is dedicated to advancing open

science, reproducibility, and collaboration. As such,

the generated data will be made available in public

repositories like GitHub and Kaggle.

3 Systemic Solution

3.1 System Design and Implementation

The comprehensive healthcare system outlined in

this paper features a modular design, with each

condition (such as chronic kidney disease, heart

disease, diabetes mellitus, stroke, epilepsy, and

autism) assigned to its own module. This structure

allows for future applicability in diagnosing and

predicting additional conditions and facilitates

efficient updates to existing modules with new data.

Brain-computer interface (BCI) modules, such as

those using the motor imagery paradigm, can

process EEG data to generate actionable commands

and other appropriate responses.

The system includes guidelines for adapting

traditional EEG systems to innovative three-

dimensional multilayer EEG systems. These novel

systems, developed by Ekpar [30] – [31], are based

on a conceptual framework that employs

approximations of carefully chosen representative

features of bio-signal sources for characterizing or

manipulating the underlying biological systems.

For each module, robust AI models are

developed and trained using appropriately formatted

data collected as described. These AI models can

integrate genetic, environmental, lifestyle, and other

relevant factors to provide more accurate

representations of participants' circumstances.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

136

Volume 4, 2024

Fig. 1 depicts the design of the system with a

simplified graphical representation of the modules

and other key components of the system.

Fig. 1: System Schematic Design Diagram for the Comprehensive AI-Driven Healthcare Solution and Brain

Computer Interface System. The New Conditions component represents additional health conditions that can be

incorporated into the solution via new modules.

The AI models are developed using the four distinct

approaches listed below. Note that in addition to the

four distinct approaches highlighted herein,

additional approaches (possibly incorporating

modern topological and algebraic methods) could be

adopted and the results (in terms of performance

metrics) compared and contrasted for possible

integration of the models into the system.

1. Direct Use of Large Language Models

(LLMs): Leveraging large language models

(LLMs) like GPT-4 as inference engines,

utilizing the collected data formatted as

multidimensional input vectors. This

process may include fine-tuning the LLM.

2. Prompt Engineering with LLMs:

Applying prompt engineering techniques to

LLMs like Bard and GPT-4 (and their

future iterations) to outline a series of steps

for constructing the AI-based system. The

proposed steps are executed using the

creator's deep expertise in AI, neural

networks, and deep learning, along with

programming in Python and using tools like

TensorFlow, Keras, and other machine

learning and visualization libraries such as

Scikit-learn and Matplotlib.

3. Automated Model Generation: Creating

specific AI models by harnessing the

features of LLMs like Bard and GPT-4

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

137

Volume 4, 2024

through an automated model generation

pipeline.

4. Direct Synthesis of AI Architecture:

Synthesizing a suitable AI architecture

based on the creator's substantial experience

in AI, neural networks, and deep learning,

employing Python, TensorFlow, Keras, and

additional machine learning and

visualization tools.

All processes and tools used in developing the

solution are thoroughly documented to ensure

smooth transfer and reuse of the system. The

performance of the generated AI models is

evaluated and compared using metrics such as

specificity and sensitivity, assessing their suitability

for the challenges presented.

3.2 Convolutional Neural Network

Architecture and Data Processing

Custom-synthesized two-dimensional (2D)

convolutional neural networks (CNNs) were

harnessed in the AI models of the system and

leveraged for multiclass (four-class: normal, cyst,

tumor, stone) image classification for the purpose of

diagnosis of chronic kidney disease from CT-

radiography. Figure 2 illustrates a generalized

depiction of the CNN architecture. The CNNs were

realized via toolsets provided by the TensorFlow

framework coupled with the Keras API in the

Python programming language [32] - [33].

Fig. 2: Generalized Illustration of the Convolutional Neural Network (CNN) Architecture.

The 2D CNN comprised three separate 2D

convolutional blocks each coupled with a 2D max

pooling layer. ReLU activation was utilized. The

first 2D convolutional layer featured 16 filters, the

second featured 32 filters while the third layer

featured 64 filters. The kernel size used was 3. A

preprocessing data augmentation layer

implementing rotation of the images by 18 degrees

was added to prevent overfitting. Z-normalization

was applied via the batch normalization function.

Before flattening and connection to the fully

connected layers, a dropout layer with a dropout rate

of 0.2 was added to the AI model.

Public CT-radiography image datasets [4] were

utilized in training and validation of the AI model.

The dataset comprised a total of 12,446 clinically

validated images spread across four distinct classes

representing normal kidney function (5,077 images),

presence of kidney cyst (3,709 images), presence of

kidney stone (1,377 images) and presence of kidney

tumor (2,283 images). First, the images were

shuffled to ensure balance and then split into

training and validation datasets with the training

dataset being allocated 80% of the data and the

validation dataset being allocated 20% of the data.

Furthermore, a new data partition containing 20% of

the data was generated after random shuffling for

evaluation of the CNN after training and validation.

The images were resized to a width of 180 pixels

and a height of 180 pixels before processing.

Processing proceeded with a batch size of 32.

Figure 3 shows a randomly selected sample of

images from the dataset representing all four

classes: Normal, Cyst, Tumor and Stone.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

138

Volume 4, 2024

Fig. 3: Sample Images from the Dataset Representing All Four Classes: Normal, Cyst, Stone and Tumor.

3.3 Data Availability

The data utilized in this study are available from

Kaggle at

https://www.kaggle.com/datasets/nazmul0087/ct-kidney-

dataset-normal-cyst-tumor-and-stone.

4 Results

The CNN was trained on the training dataset (9957

images or 80% of the original 12,446 images) and

validated on the validation dataset (2,489 images or

20% of the original 12,446 images) over 10 epochs.

Sparse categorical cross entropy loss function was

utilized. The AI model was optimized via the Adam

Optimizer [34] – [35] with a learning rate of 0.001.

Figure 4 illustrates the performance of the AI model

on the training and validation datasets.

As can be seen from Fig. 4, the performance of

the CNN improved over the training and validation

cycles until an accuracy of approximately 97% was

achieved for the training and validation datasets.

Evaluation of the resulting AI model after training

and validation on a randomly selected test dataset

comprising 20% of the original data yielded

comparable results.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

139

Volume 4, 2024

Fig. 4: Performance of the Convolutional Neural Network (CNN) on Training and Validation Datasets.

Implementing the comprehensive AI system

outlined here will provide valuable insights for

clinical decision-making, ultimately saving lives and

enhancing quality of life. It aims to alleviate the

economic, social, psychological, and physical

burdens associated with conditions that can be

predicted, potentially prevented, detected early,

diagnosed, and managed more effectively.

Participating medical doctors and their

colleagues can generate Electronic Health Records

(EHR) that include clinical diagnostic

measurements and EEG data. EEG data may also be

collected during experiments with Brain-Computer

Interfaces (BCIs). This data is gathered in

compliance with ethical approvals and is

anonymized prior to being published in publicly

accessible repositories alongside academic research

articles.

5 Conclusion

Convolutional neural networks were harnessed for

the classification of medical images representing

kidney states indicating normal function, the

presence of cysts, the presence of tumors as well as

the presence of kidney stones. The resulting AI

model could be integrated into a chronic kidney

disease diagnosis module within the context of a

comprehensive AI-driven healthcare system. The

system exhibited excellent performance, achieving

an accuracy approximating 97% for the training and

validation datasets. Adopting the comprehensive AI-

powered healthcare system in resource-limited

settings such as low- and middle-income countries

(LMIC) with low doctor-to-patient ratios and

limited healthcare funding could permit a single

medical doctor or healthcare worker to serve up to

ten or more times the usual number of patients,

effectively increasing the doctor-to-patient ratio and

dramatically improving health outcomes and saving

lives without significant additional investments.

Generally, the high accuracy of the system

encourages adoption and utilization of the system

for improved health outcomes in both developed

and developing countries and regions. In the future,

the system could incorporate genetic and lifestyle

factors for a more accurate reflection of the patient’s

circumstances and to facilitate recommendations for

lifestyle modifications that could help prevent

disease.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

140

Volume 4, 2024

References:

[1] World Health Organization (WHO) - Top 10 Causes

of Death: https://www.who.int/news-room/fact-

sheets/detail/the-top-10-causes-of-death. Retrieved

(2024).

[2] World Health Organization (WHO) –

Noncommunicable Diseases:

https://www.who.int/news-room/fact-

sheets/detail/noncommunicable-diseases. Retrieved

(2024).

[3] Ekpar, F. E. A Comprehensive Artificial

Intelligence-Driven Healthcare System, European

Journal of Electrical Engineering and Computer

Science, 8(3), Article 617. (2024).

[4] Islam, N. M., Hassan, M., Hossain, M. K.,

Alam, M. G. R., Uddin, M. Z., Soylu, A. Vision

transformer and explainable transfer learning

models for autodetection of kidney cyst, stone

and tumor from CT-radiography, Scientific

Reports, 12: 11440. (2022).

[5] Nomura, A., Noguchi, M., Kometani, M.,

Furukawa, K., Yoneda, T. Artificial Intelligence

in Current Diabetes Management and

Prediction, Curr Diab Rep. 21(12):61 (2021).

[6] Kumar, Y., Koul, A., Singla, R., Ijaz, M. F.

Artificial intelligence in disease diagnosis: a

systematic literature review, synthesizing

framework and future research agenda, Journal

of Ambient Intelligence and Humanized

Computing 14:8459–8486 (2023).

[7] Ansari, S., Shafi, I., Ansari, A., Ahmad, J.,

Shah, S. I. Diagnosis of liver disease induced by

hepatitis virus using artificial neural network,

IEEE Int Multitopic.

https://doi.org/10.1109/INMIC.2011.6151515

(2011).

[8] Guo, H., Cao, S., Zhou, C., Wu, X., Zou, Y.

Predicting Essential Genes of Alzheimer

Disease based on Module Partition and Gravity-

like Method in Heterogeneous Network, WSEAS

Transactions on Applied and Theoretical

Mechanics, Vol. 17, pp. 158-165. (2022).

[9] Battineni, G., Sagaro, G. G., Chinatalapudi, N.,

Amenta, F. Applications of machine learning

predictive models in the chronic disease

diagnosis, J Personal Med.

https://doi.org/10.3390/jpm10020021 (2020).

[10] Abdar, M., Yen, N., Hung, J. Improving the

diagnosis of liver disease using multilayer

perceptron neural network and boosted decision

tree, J Med Biol Eng 38:953–965 (2018).

[11] Chaikijurajai, T., Laffin, L., Tang, W. Artificial

intelligence and hypertension: recent advances

and future outlook, Am J Hypertens 33:967–974

(2020).

[12] Fujita, S., Hagiwara, A., Otsuka, Y., Hori, M.,

Takei, N., Hwang, K. P., Irie, R., Andica, C.,

Kamagata, K., Akashi, T., Kumamaru, K. K.,

Suzuki, M., Wada, A., Abe, O., Aoki, S. Deep

Learning Approach for Generating MRA

Images From 3D Quantitative Synthetic MRI

Without Additional Scans, Invest Radiol

55:249–256 (2020).

[13] Ahuja, R., Sharma, S. C. Exploring Feature

Selection and Classification Algorithms for

Cardiac Arrythmia Disease Prediction, WSEAS

Transactions on Biology and Biomedicine, Vol.

19, pp. 168-175. (2022).

[14] Juarez-Chambi, R. M., Kut, C., Rico-Jimenez, J.

J., Chaichana, L. K., Xi, J., Campos-Delgado,

D. U., Rodriguez, F. J., Quinones-Hinojosa, A.,

Li, X., Jo, J. A. AI-Assisted In Situ Detection of

Human Glioma Infiltration Using a Novel

Computational Method for Optical Coherence

Tomography, Clin Cancer Res 25(21):6329–

6338 (2019).

[15] Nashif, S., Raihan, R., Islam, R., Imam, M. H.

Heart Disease Detection by Using Machine

Learning Algorithms and a Real-Time

Cardiovascular Health Monitoring System,

World Journal of Engineering and Technology

Vol 6, No. 4 (2018).

[16] Chen, P. H. C., Gadepalli, K., MacDonald, R.,

Liu, Kadowaki, S., Nagpal, K., Kohlberger, T.,

Dean, J., Corrado, G. S., Hipp, J. D., Mermel,

C. H., Stumpe, M. C. An augmented reality

microscope with real time artificial intelligence

integration for cancer diagnosis, Nat Med

25:1453–1457 (2019).

[17] Gouda, W., Yasin, R. COVID-19 disease: CT

Pneumonia Analysis prototype by using

artificial intelligence, predicting the disease

severity, Egypt J Radiol Nucl Med 51(1):196

(2020).

[18] Han, Y., Han, Z., Wu, J., Yu, Y., Gao, S., Hua,

D., Yang, A. Artificial Intelligence

Recommendation System of Cancer

Rehabilitation Scheme Based on IoT

Technology, IEEE Access 8:44924–44935

(2020).

[19] Zeynu, S., Patil, S. Prediction of Chronic

Kidney Disease Using Data Mining Feature

Selection and Ensemble Method, WSEAS

Transactions on Information Science and

Applications, Vol. 15, pp. 168-176. (2018).

[20] Chui, C. S., Lee, N. P., Adeoye, J., Thomson,

P., Choi, S. W. Machine learning and treatment

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

141

Volume 4, 2024

outcome prediction for oral cancer, J Oral

Pathol Med 49(10):977–985 (2020).

[21] Koshimizu, H., Kojima, R., Okuno, Y. Future

possibilities for artificial intelligence in the

practical management of hypertension,

Hypertens Res 43:1327–1337 (2020).

[22] Kather, J. N., Pearson, A. T., Halama, N., Jäger,

D., Krause, J., Loosen, S. H., Marx, A., Boor,

P., Tacke, F., Neumann, U. P., Grabsch, H. I.,

Yoshikawa, T., Brenner, H., Chang-Claude, J.,

Hoffmeister, M., Trautwein, C., Luedde, T.

Deep learning microsatellite instability directly

from histology in gastrointestinal cancer, Nat

Med 25:1054–1056 (2019).

[23] Sreeja, M. U., Supriya, M. H. A Deep

Convolutional Model for Heart Disease

Prediction based on ECG Data with Explainable

AI, WSEAS Transactions on Information

Science and Applications, Vol. 20, pp. 254-264.

(2023).

[24] Kwon, J. M., Jeon, K. H., Kim, H. M., Kim, M.

J., Lim, S. M., Kim, K. H., Song, P. S., Park, J.,

Choi, R. K., Oh, B. H. Comparing the

performance of artificial intelligence and

conventional diagnosis criteria for detecting left

ventricular hypertrophy using

electrocardiography, EP Europace 22(3):412–

419 (2020).

[25] Khan, M. A. An IoT Framework for Heart

Disease Prediction Based on MDCNN

Classifier, IEEE Access 8:34717–34727 (2020).

[26] Bora, N., Gutta, S., Hadaegh, A. Using machine

learning to Predict Heart Disease, WSEAS

Transactions on Biology and Biomedicine, Vol.

19, pp. 1-9. (2022).

[27] Oikonomou, E. K., Williams, M. C., Kotanidis,

C. P., Desai, M. Y., Marwan, M.,

Antonopoulos, A. S., Thomas, K. E., Thomas,

S., Akoumianakis, I., Fan, L. M., Kesavan,

S., Herdman, L., Alashi, A., Centeno, E. H.,

Lyasheva, M., Griffin, B. P., Flamm, S. D.,

Shirodaria, C. Sabharwal, N., Kelion, A.,

Dweck, M. R., Van Beek, E. J. R., Deanfield, J.,

Hopewell, J. C., Neubauer, S., Channon, K.

M., Achenbach, S., Newby, D. E., Antoniades,

C. A novel machine learning-derived

radiotranscriptomic signature of perivascular fat

improves cardiac risk prediction using coronary

CT angiography, Eur Heart J 40(43):3529–

3543 (2019).

[28] Sabottke, C. F., Spieler, B. M. The Effect of

Image Resolution on Deep Learning in

Radiography, Radiology: Artificial Intelligence

Vol. 2. No. 1, 2:e190015 (2020).

[29] Mohammed, A. N., Albagul, A., Ahmad, M. M.

Automated Alzheimer’s Disease Diagnosis

using Convolutional Neural Networks and

Magnetic Resonance Imaging, WSEAS

Transactions on Signal Processing, Vol. 19, pp.

118-127. (2023).

[30] Ekpar, F. E. System for Nature-Inspired Signal

Processing: Principles and Practice, European

Journal of Electrical Engineering and

Computer Science, 3(6), pp. 1-10, (2019).

[31] Ekpar, F. E. A Novel Three-dimensional

Multilayer Electroencephalography Paradigm,

Fortune Journal of Health Sciences, 7(3).

(2024).

[32] Abadi, M., Barham, P., Chen, J., Chen, Z.,

Davis, A., Dean, J., Devin, M., Ghemawat, S.,

Irving, G., Isard, M., Kudlur, M., Levenberg, J.,

Monga, R., Moore, S., Murray, D. G., Steiner,

B., Tucker, P., Vasudevan, V., Warden, P.,

Wicke, M., Yu, Y., Zheng, X. TensorFlow: A

System for Large Scale Machine Learning,

Proceedings of the 12th USENIX Symposium on

Operating Systems Design and Implementation

(OSDI ’16). (2016).

[33] Pang, B., Nijkamp, E., Wu, Y. N. Deep

Learning With TensorFlow: A Review, Journal

of Educational and Behavioral Statistics. Vol.

45, Iss. 2. (2019).

[34] Kingma, D. P., Ba, J. L. Adam: A Method

for Stochastic Optimization, International

Conference on Learning Representations

(ICLR) (2015).

[35] Zhang, Z. Improved Adam Optimizer for

Deep Neural Networks, IEEE/ACM 26th

International Symposium on Quality of

Service (IWQoS) (2018).

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

142

Volume 4, 2024

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

Frank Edughom Ekpar took charge of all aspects of

this work including conceptualization, design,

implementation, experiment design, execution and

administration, data gathering (including from

public sources) and data analysis and processing as

well as manuscript preparation.

Conflict of Interest

The author has no conflicts of interest to declare that

are relevant to the content of this article.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US

Sources of Funding for Research Presented in a

Scientific Article or Scientific Article Itself

No funding was received for conducting this study.

MOLECULAR SCIENCES AND APPLICATIONS

DOI: 10.37394/232023.2024.4.13

Frank Edughom Ekpar

E-ISSN: 2732-9992

143

Volume 4, 2024