Browsing by Author "Steve Adeshina"

Now showing 1 - 15 of 15

A Few-shot custom CNN Model for Retinal Nerve Fibre Layer Thickness Measurement in OCT Images of Epilepsy
(Proc. of International Conference on Artificial Intelligence, Computer, Data Sciences and Applications (, 2024-02-01) Ruqayya Muhammad; Moussa Mahamat Boukar; Steve Adeshina; Senol Dane
This study aims to assess the effectiveness of employing deep learning models for measuring retinal nerve fiber layer (RNFL) thickness in optical coherence tomography (OCT) scans of epilepsy patients. Conventional OCT scan segmentation methods typically rely on supervised learning, demanding substantial data for training and assuming fixed network weights post-training. To mitigate these challenges, we explore the applicability of few-shot learning (FSL) in CNN architectures, allowing dynamic fine-tuning of network weights with minimal additional data. Experimental results demonstrate enhanced segmentation accuracy, with the proposed Few shot Custom CNN achieving a notable 91% accuracy, surpassing both the Custom CNN (86%) and the OCT machine data. This suggests the superiority of the few-shot Custom CNN model in segmentation performance compared to OCT scans.
A framework for Poultry weather control with IoT in sub-Saharan Africa
(15th International Conference on Electronics Computer and Computation (ICECCO 2019), 2019-02-02) Nasiru Afeez; Steve Adeshina; Abdullahi Inci; Moussa Mahamat Boukar
Poultry farming in the sub-Saharan region of Africa is fraught with a lot of challenges among which are high temperature and humidity. In this paper, the authors proposed an Internet of Things (IoT) framework that will help in regulating the various climatic conditions that will help at providing a high yield of poultry products. This framework is aimed at providing proactive and preventive ways to avert or reduce the high mortality rate in a flock of birds as a result of heat stress. IoT which is a connected environment of monitoring sensors with high precision and an accurate decision taken would be presented in managing environmental conditions of poultry house that will gather information, analyze it and effect an action based on the predetermined weather conditions that are suitable for bird’s existence
Age Estimation from Facial Images Using Custom Convolutional Neural Network (CNN)
(International Conference on Frontiers in Academic Research, 2023-02-23) Gilbert George; Steve Adeshina; Moussa Mahamat Boukar
Given that aging is influenced by a variety of factors, including gender, ethnicity, environment, and others, automatic age assessment of facial images is a difficult challenge in computer vision and image analysis. Additionally, a significant amount of data and a laborious training phase are needed to estimate age from facial photos with near accuracy. In this study, we present a custom convolutional neural network-based age estimator that can almost precisely predict age from facial photos. We use the UTK facial image dataset using about 17475 images. We train the model to group the facial images into three groups which are; Child, Teenager and Adult. Compared to similar efforts, our method uses less training data while maintaining a high accuracy of 95%.
Analysis of Bad Roads Using Smart phone
(IEEE, 2019-02-02) Moussa Mahamat Boukar; Steve Adeshina
Developing nations are faced with a lot of bad roads with potholes of different debt ranges, the maintenance and rehabilitation process by government agencies is an ongoing effort that requires periodic bad road inventory to guarantee safety. Bad roads are either identified by government agency’s survey teams or individual who volunteer to report these conditions to the authorities. Our research provided a simple but effective solution to aid in automatically reporting bad roads using smart-phones through measuring the pavement profile based on the vibration of a moving vehicle. In this article, we will explain how we used some a smart-phone in reading the vibration pattern, GPS location, speed and direction of a vehicle that drives through a pothole, these parameters are periodically streamed to a cloud application. We used standard deviation to measure the level of dispersion around a segmented set of streamed vehicle vibration to identify potholes of different sizes, we also used Artificial Intelligence - supervised learning algorithm (classification) to reduce the false positive error rates due to human behaviors. The final results show a distinct vibration levels between small pot-holes, speed bumps and big pot-holes, these values are displayed on map application to visualize the geographical locations of these pot-holes (Google maps)
Analysis of Prostate Cancer DNA Sequences Using Bi-direction Long Short Term Memory Model
(IEEE, 2021-02-02) Yusuf Aleshinloye Abass; Steve Adeshina; Nwojo Nnana Agwu; Moussa Mahamat Boukar
Machine and deep learning-based models are the emerging techniques in addressing prediction problems in biomedical data analysis. DNA sequence prediction is a critical problem that requires huge attention in the biomedical domain. These techniques have been shown to provide better accurate results when compared to the conventional regression-based models. Prediction of the gene sequence that leads to cancerous diseases such as prostate cancer is very crucial. Identifying the most important features in a gene sequence is one of the most challenging tasks and extracting the components of the gene sequence that can give an insight into the kind of mutation in the gene is very important, it will lead to effective drug design and promote the new concept of personalized medicine. In this work we have extracted the exons in the various prostate gene sequence that was used in the experiment, we built a bi-LSTM model using a k-mer encoding for the DNA sequence and one- hot encoding for the class label. The bi-LSTM model was evaluated on different classification metrics. Our experimental results show that the model prediction offers a training accuracy and validation accuracy of 95 percent and 91 percent respectively.
Applications of Artificial Intelligence Based Techniques on the Analysis of Chemical Data: a Review
(International Conference on Multidisciplinary Engineering and Applied Science (ICMEAS), 2021-07-15) Chinomso Odimba; Steve Adeshina; Petrus Nzerem
Artificial Intelligence based techniques such as Deep Learning, Machine Learning, Chemometrics have recently begun to replace chemical heuristics. They are promising tools that can be used to gain insight on the characteristics, processes and interactions of a chemical sampleand to a clearer and better understanding of chemical data. The focus of this review paper is on the recent developments on the applications of Artificial Intelligence based techniques for different chemical scenarios of computational chemistry, quantum chemistry, synthetic route design, drug delivery, analysis of spectral data and analytical chemistry.
Comparative studies of machine learning models for predicting higher heating values of biomass
(Institution of Chemical Engineers (IChemE), 2024-06-29) Adekunle Akanni Adeleke; Adeyinka Adedigba; Steve Adeshina; Peter Pelumi Ikubanni; Mohammed S. Lawal; Adebayo Isaac Olosho; Halima S. Yakubu; Temitayo Samson Ogedengbe; Petrus Nzerem; Jude A. Okolie
This study addresses the challenge of efficiently determining the higher heating value (HHV) of biomass, a crucial parameter in large-scale biomass-based energy systems. The conventional method of measuring HHV using an oxygen bomb calorimeter is time-consuming, expensive, and less accessible to researchers, particularly in developing nations. To overcome these limitations, we employed four machine learning (ML) models, namely Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost). These models were developed by using proximate and ultimate analysis parameters as input features. Up to 200 datasets were compiled from literature and used for the ML models. Our results demonstrate the effectiveness of all ML models in accurately predicting the HHV of biomass materials. Notably, the XGBoost model exhibited superior performance with the highest R-squared (R2) values for both training (0.9683) and test datasets (0.7309), along with the lowest root mean squared error (RSME) of 0.3558. Key influential input features identified for HHV prediction include carbon (C), volatile matter (Vm), ash, and hydrogen (H). Consequently, this research provides a reliable alternative for predicting HHV without the need for costly and time-intensive experimental measurements, facilitating broader accessibility in biomass energy research.
Comparison of Transfer Learning Model Accuracy for Osteoporosis Classification on Knee Radiograph
(IEEE, 2022-02-02) Moussa Mahamat Boukar; Steve Adeshina
In terms of financial costs and human suffering, osteoporosis poses a serious public health burden. Reduced bone mass, degeneration of the micro architecture of bone tissue, and an increased risk of fracture are its main skeletal symptoms. Osteoporosis is caused not just by low bone mineral density, but also by other factors such as age, weight, height, and lifestyle. Recent advancement in Artificial Intelligence (AI) has led to successful applications of expert systems that use Deep Learning techniques for osteoporosis diagnosis based on some modalities such as dental radiographs amongst others. This study uses a dataset of knee radiographs (i.e., knee-Xray images) to apply and compare three robust transfer learning model algorithms: GoogLeNet, VGG-16, and ResNet50 to classify osteoporosis. From the statistical analysis and scikit learn python analysis, the accuracy of the GoogLeNet model was 90%, the accuracy of the VGG-16 model was 87% and lastly, the accuracy of the ResNet- 50 model was 83%.
DEVELOPMENT OF ROAD ANOMALY DATA TRANSMISSION USING ANT COLONY OPTIMIZATION ALGORITHM IN A VEHICLE-TO-VEHICLE COMMUNICATION
(IEEE, 2019-02-02) Muktar Othman; Steve Adeshina; Moussa Mahamat Boukar
This study aim is to design a road anomaly transmission Algorithms using Ant Colony Optimization(ACO) based Technique in a Vehicle-to-Vehicle (V2V) and Vehicle to Infrastructure (V2I) Communication. The developed VACO also uses the features of VANET to find out the optimal path by considering a minimum number of nodes and cost parameters, which provides information related to accidents, speed of neighbouring vehicle and weather to help users in making informed decisions. Vehicle routing protocol based on ACO (VACO) also ensures to mitigate issues by combining the reactive and proactive approach and considers the parameters affecting the Quality of Service (QoS) such as latency, bandwidth, and delivery ratio in evaluating the Algorithms.
Evaluation of Parameter Fine-Tuning with Transfer Learning for Osteoporosis Classification in Knee Radiograph
((IJACSA) International Journal of Advanced Computer Science and Applications, 2022-02-02) Moussa Mahamat Boukar; Steve Adeshina
Osteoporosis is a bone disease that raises the risk of fracture due to the density of the bone mineral being low and the decline of the structure of bone tissue. Among other techniques, such as Dual-Energy X-ray Absorptiometry (DXA), 2D x-ray pictures of the bone can be used to detect osteoporosis. This study aims to evaluate deep convolutional neural networks (CNNs), applied with transfer learning techniques, to categorize specific osteoporosis features in knee radiographs. For objective labeling, we obtained a selection of patient knee x-ray images. The study makes use of the Visual Geometry Group Deep (VGG-16), and VGG-16 with fine-tuning. In this work, the deployed CNNs were assessed using state-of-the-art metrics such as accuracy, sensitivity, and specificity. The evaluation shows that fine-tuning enhanced the VGG-16 CNN's effectiveness for detecting osteoporosis in radiographs of the knee. The accuracy of the VGG-16 with parameter fine-tuning was 88% overall, while the accuracy of the VGG-16 without parameter fine-tuning was 80%.
Hepatitis C Stage Classification with hybridization of GA and Chi2 Feature Selection
(IJCSNS International Journal of Computer Science and Network Security, 2022-02-02) Steve Adeshina; Moussa Mahamat Boukar
In metaheuristic algorithms such as Genetic Algorithm (GA), initial population has a significant impact as it affects the time such algorithm takes to obtain an optimal solution to the given problem. In addition, it may influence the quality of the solution obtained. In the machine learning field, feature selection is an important process to attaining a good performance model; Genetic algorithm has been utilized for this purpose by scientists. However, the characteristics of Genetic algorithm, namely random initial population generation from a vector of feature elements, may influence solution and execution time. In this paper, the use of a statistical algorithm has been introduced (Chi2) for feature relevant checks where p-values of conditional independence were considered. Features with low p-values were discarded and subject relevant subset of features to Genetic Algorithm. This is to gain a level of certainty of the fitness of features randomly selected. An ensembled-based learning model for Hepatitis has been developed for Hepatitis C stage classification. 1385 samples were used using Egyptian-dataset obtained from UCI repository. The comparative evaluation confirms decreased in execution time and an increase in model performance accuracy from 56% to 63%.
Human Detection For Crowd Count Estimation Using CSI of WiFi Signals
(International Conference on Electronics Computer and Computation (ICECCO), 2019-12-01) Omotayo Oshiga; Hussein U. Suleiman; Sadiq Thomas; Petrus Nzerem; Labaran Farouk; Steve Adeshina
We address the problem of crowd estimation in situations such as indoor events using anonymous and non-participatory CSI of WiFi Signals. Observing the great resemblance of Channel State Information (CSI, a finegrained information captured from the received Wi-Fi signal) to texture, we propose a brand-new framework based on statistical mechanics, and relying only on sets of machine learning techniques.In this paper, a framework for crowd count estimation is presented which utilizes Chebyshev filter and SVD to remove background noise in the CSI data, PCA to reduce the dimensionality of the CSI data and spectral descriptors for feature extraction. From the extracted feature, a set of classiffying algorithms are then utilised for training and testing the accuracy of our crowd estimation framework The aim of this framework to effectively and efficiently extract the channel information in WiFi signals across OFDM carriers reflected by the presence of human bodies. From the experiments conducted, we demonstrate the feasibility and efficacy of the proposed framework. Our result depict that our estimation becomes more–rather than less–accurate when the crowd count increases.
Influence of Local Additives on Water Based Drilling Mud: A Review
(NJEAS, 2023-09-20) Petrus Nzerem; Khaleel Jakada; Mohammed Shariff; Abdulquddus Ozigi; Ayuba Salihu; Ikechukwu Okafor; Steve Adeshina; Khadijah Ibrahim; Adekunle Akanni Adeleke
This review paper focuses on the use of local additives in water-based drilling mud to reduce environmental impact and improve drilling operations. Drilling mud plays a crucial role in drilling operations by acting as a coolant, carrying drilled particles, stabilizing the wellbore, and preventing wellbore issues. However, poorly formulated drilling mud can lead to drilling difficulties and environmental pollution. Therefore, the exploration of organic-based drilling mud additives, sourced locally in Nigeria, is discussed in this paper.it highlights the potential of transforming organic waste materials, such as rice husk, cassava, corn cobs, and more, into usable products for drilling mud. By utilizing these locally sourced organic materials, the environmental impact of drilling waste can be minimized. The importance of waste management in the oil and gas industry is emphasized, as it is crucial for sustainable drilling practices. The paper further discusses various studies and experiments conducted on the use of local polymers and natural materials as substitutes for imported additives in water-based drilling mud. These materials include cassava starches, agro-waste materials, eco-friendly drilling fluid additives, and various plant-based substances. The performance and effectiveness of these materials are evaluated in terms of viscosity control and fluid loss prevention. The results indicate that many of these local polymers and natural materials can be viable and have shown positive results in terms of improving the rheological and filtration characteristics of the mud, reducing fluid loss, and enhancing overall mud qualities. Earlier studies on the use of agro-waste products as additives in drilling fluids are reviewed in this paper. These studies examine the properties affected by local materials, the type of mud used, and the findings of each study. The economic analysis of using agro-waste materials as drilling additives is also discussed. The conclusion highlights the availability and affordability of agricultural waste materials as potential substitutes for traditional drilling additives, which can help reduce drilling costs. The paper also provides recommendations for future research in this area.
Time Series Analysis and prediction of bitcoin using Long Short Term Memory Neural Network
(International Conference on Electronics Computer and Computation, 2019-02-02) Temiloluwa I. Adegboruwa; Steve Adeshina; Moussa Mahamat Boukar
Bitcoin is the first digital currency that uses decentralization to solve the issue of trust in performing the functions of a digital currency successfully. This digital currency has shown extraordinary growth and intermittent plunge in value and market capitalization over time. This makes it important to understand what determines the volatility of bitcoin and to what extent they are predictable. Long Short Term Memory Neural Networks (LSTM-NN) have recently grown popular for time series prediction systems but there has been no consensus on methods to model time series inputs for LSTMs, this paper proposes the need for this problem to be solved by conducting an experimental research on the efficacy of an LSTM-NN given the form of its time-series input features.
Transfer Learning Model Training Time Comparison for Osteoporosis Classification on Knee Radiograph of RGB and Grayscale Images
(WSEAS TRANSACTIONS on ELECTRONICS, 2022-09-13) Moussa Mahamat Boukar; Steve Adeshina; Senol Dane
In terms of financial costs and human suffering, osteoporosis poses a serious public health burden. Reduced bone mass, degeneration of the microarchitecture of bone tissue, and an increased risk of fracture are its main skeletal symptoms. Osteoporosis is caused not just by low bone mineral density, but also by other factors such as age, weight, height, and lifestyle. Recent advancement in Artificial Intelligence (AI) has led to successful applications of expert systems that use Deep Learning techniques for osteoporosis diagnosis based on some modalities such as dental radiographs amongst others. This study uses a dataset of knee radiographs (i.e., knee-Xray images) to apply and compare the training time of two robust transfer learning model algorithms: GoogLeNet, VGG-16, and ResNet50 to classify osteoporosis. The dataset was split into two subcategories using python opencv library: Grayscale Images and Red Green Blue (RGB) images. From the scikit learn python analysis, the training time of the GoogLeNet model on grayscale images and RGB images was 42minutes and 50 minutes respectively. The VGG-16 model training time on grayscale images and RGB images was 37 minutes and 44 minutes respectively. In addition, to compare the diagnostic performance of the two models, several state-of-the-art neural networks metric was used.