Selorm Adablanu

@uew.edu.gh

Lecturer / Department of ICT EDUCATION
University of Education, Winneba

Selorm Adablanu


                        Download Compact PDF  
https://researchid.co/sadablanu
Selorm Adablanu is a computer scientist at the University of Education, Winneba. He holds a Master of Technology (M.Tech) degree in Computer Science, a Bachelor of Science degree in Computer Science, and a Master of Business Administration (MBA) with a specialization in Information Technology Management. He is currently pursuing a Ph.D. in Computer Science and Engineering.

Selorm has both academic and industry experience, having previously served as a lecturer at the School of Computer Science, Data Link University (Ghana), and as an instructor at All-soft Institute (IBM® software solutions) in India.

His research interests include machine learning, deep learning, computer vision, image processing, medical image processing, and artificial intelligence in education.

His work is driven by the goal of building AI systems that can learn effectively from limited data and make robust, accurate decisions-particularly in applications related to healthcare and education.

EDUCATION

PhD Researcher - Assam down town University - On-going

MBA Information Technology Management - Vignan University, India (2024)

PgD Teaching and Learning in Higher Education - University of Education, Winneba-Ghana (2024)

MTECH Computer Science And Engineering - Rayat Bahra University, India (2019)

Bsc Computer Science - Data Link University, Ghana (2014)

RESEARCH, TEACHING, or OTHER INTERESTS

Computer Science, Artificial Intelligence, Computer Vision and Pattern Recognition
6

Scopus Publications

91

Scholar Citations

4

Scholar h-index

3

Scholar i10-index

Scopus Publications

  • 15 Years of optimizers in medical deep learning: A systematic review
    Selorm Adablanu, Utpal Barman, Dulumani Das
    Neuroscience Informatics, 2026
    Optimization algorithms are pivotal in training deep learning (DL) models for medical imaging, determining how efficiently models learn, generalize, and perform across modalities. This systematic review analyzed 69 peer-reviewed studies (2010–2025) on optimizer performance in classification, segmentation, and object detection tasks using MRI, CT, X-ray, ultrasound, histopathology, and ECG data, following PRISMA 2020 guidelines. Adaptive optimizers such as Adam and its variants were most common, offering rapid convergence in CNN-based classification, whereas SGD and momentum-based methods yielded stronger generalization in large-scale segmentation. Emerging techniques—Sharpness-Aware Minimization (SAM), Ranger, and AdamW—improved robustness under domain shift or noisy conditions. Hybrid and metaheuristic optimizers provided marginal accuracy gains but at higher computational cost. Common limitations included inconsistent hyperparameter reporting, limited external validation, and dataset bias toward North American cohorts. Optimizer effectiveness was found to be task- and architecture-dependent: adaptive methods suit small or noisy datasets, while momentum-based and hybrid methods enhance generalization for complex imaging. Future studies should emphasize standardized evaluation, transparent reporting, and diverse data to enable equitable and reproducible deployment of medical AI. • First systematic review of optimization algorithms in medical deep learning (69 studies, 2010–2025). • Adam dominates classification; SAM and Ranger improve robustness in segmentation and detection. • Gaps persist in external validation, hyperparameter sensitivity, and computational efficiency reporting • Optimizer effectiveness varies by task and architecture, requiring context-specific selection. • Recommends standardized benchmarks, transparent reporting, and diverse datasets to improve generalization.
  • Transforming Skin Cancer Detection With AI-Based Convolutional and Transformer Models
    Selorm Adablanu, Utpal Barman, Dulumani Das, Tuward Jade Dweh
    Iradiology, 2026
    Background Skin cancer is a major cause of mortality, and early detection is vital for effective treatment. Diagnosis is challenging because of lesion variability. This study adapts VINCE‐NET, a hybrid deep‐learning model originally designed for stroke detection, to classify melanoma using dermoscopic images. Methods VINCE‐NET combines vision transformer layers for global context, convolutional neural networks for local features, and long short‐term memory for spatial sequence modeling. During preprocessing, Gaussian blur, normalization, and augmentation were applied to reduce noise and handle class imbalance. During training, the public HAM10000 dataset was used in a central processing unit‐only Google Colab environment (12.72 GB random access memory, 107.7 GB disk) with an AdamW optimizer, a batch size of 12, learning‐rate scheduling, and early stopping (patience = 50). VINCE‐NET's performance was compared with those of a convolutional neural networks, long short‐term memory, residual network with 50 layers (ResNet‐50), visual geometry group network with 16 and 19 layers (VGG‐16/19), and densely connected convolutional network with 121 and 201 layers (DenseNet‐121/201) under identical preprocessing conditions. Results VINCE‐NET achieved 94.1% accuracy, 95.5% precision, 90.4% recall, a 92.9% F1‐score, and an area under the receiver operating characteristic curve of 0.98 at a training time of 34,308.42 s. Benchmarks showed that VINCE‐NET outperformed baselines while being computationally efficient. Conclusion VINCE‐NET provides competitive performance for melanoma classification and feasibility in resource‐limited settings. Although promising, VINCE‐NET has not been clinically validated yet. Future work will address resolution, ablation studies, interpretability, and external validation.
  • Feature Extraction and Selection Methods Outperform Machine Learning and Deep Learning Techniques
    Tuward Jade Dweh, Selorm Adablanu
    Feature Selection and Feature Extraction on Omics Data, 2026
  • Comparative Efficacy of Focal and Binary Cross-Entropy Loss in Handling Class Imbalance for Stroke CT Classification
    Selorm Adablanu, Utpal Barman, Dulumani Das
    2026 6th International Conference on Advances in Electrical Computing Communications and Sustainable Technologies Icaect 2026, 2026
    Stroke remains one of the leading causes of death and long-term disability worldwide, demanding rapid and accurate diagnosis through brain CT imaging. Deep learning can assist in stroke detection; however, class imbalance between normal and stroke-positive images presents a major challenge. This study investigates how different loss functions handle this imbalance by comparing Focal Loss and Binary Cross-Entropy (BCE) in training an Xception convolutional neural network on the TEKNOFEST-2021 brain stroke CT dataset comprising 6,653 images. The model trained with Focal Loss achieved 99.08% accuracy, 100% sensitivity, and AUC <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.9996$</tex>, outperforming the BCE model, which reached 98.40% accuracy and AUC = 0.9994. An additional experiment using test-time augmentation (TTA) unexpectedly reduced accuracy to about 86%, indicating that augmentations applied during testing must align with those used in training. The findings demonstrate that Focal Loss effectively improves sensitivity in imbalanced datasets by reducing missed detections without compromising precision. This approach can strengthen the reliability of AI-based stroke screening tools. Future work will focus on adaptive focal parameters and advanced data-balancing techniques to improve generalization across diverse clinical datasets.
  • A novel hybrid adaptive transformer framework with multihead self attention for stroke detection
    Selorm Adablanu, Utpal Barman, Dulumani Das
    Discover Neuroscience, 2025
    Stroke detection using artificial intelligence (AI) continues to show promise in advancing medical imaging diagnostics, particularly in improving accuracy and accelerating clinical decision-making. This study presents VINCE-NETv1, a hybrid deep learning framework that integrates Vision Transformers (ViTs), Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) modules, and a meta-learning component to capture spatial, temporal, and global features from CT stroke images. Designed with architectural efficiency in mind, the model was trained and evaluated separately on three datasets: Near East University Hospital CT, the Core-Penumbra Acute Ischemic Stroke Dataset (CPAISD), and the Kaggle Brain Stroke CT dataset. To assess optimization performance, four optimizers-Adam, AdamW, Lookahead, and RMSProp-were used during training, each demonstrating high classification accuracy, with VINCE-NETv1 achieving up to 99.9% on CPAISD and ~ 100% on the Near East dataset. Generalization claims are limited strictly to datasets with confirmed patient-level separation-specifically the Near East and CPAISD datasets. The Kaggle dataset, lacking metadata for patient disambiguation, was treated solely as an exploratory benchmark. Preliminary interpretability was demonstrated using Grad-CAM visualizations, suggesting alignment with stroke-relevant regions. Approximate confidence intervals were computed from repeated runs to provide early insight into performance robustness, though formal statistical validation and k-fold cross-validation are planned in future work. While VINCE-NETv1 shows architectural potential for scalable deployment and clinical integration, current claims regarding generalization, real-time feasibility, and model interpretability are made cautiously with acknowledgment of existing limitations. Future research will address these aspects to evolve VINCE-NETv1 into a reliable and interpretable clinical decision-support tool.
  • Advancing deep learning for automated stroke detection: a review
    Selorm Adablanu, Utpal Barman, Dulumani Das
    Brain Hemorrhages, 2025
    Stroke remains a leading cause of death and disability worldwide, necessitating improved diagnostic tools for early detection and classification. Machine learning (ML) techniques have shown promise in addressing this critical healthcare challenge by enabling efficient analysis of stroke-related data. However, the lack of standardized datasets, limited real-time clinical applicability, and the complexity of model interpretability hinder broader adoption. This review critically examines 34 research articles published between 2014 and 2025, focusing on traditional ML, deep learning, transfer learning, and hybrid approaches for stroke detection and classification. Key findings highlight that Traditional ML models such as Support Vector Machines (SVM) and Random Forests (RF) have been widely used but show limitations in high-dimensional medical imaging tasks. Conversely, advanced deep learning models, such as EEG-DenseNet and ResNet50, excel in stroke segmentation and classification tasks, while hybrid methods demonstrate potential for improving accuracy through ensemble strategies. The review also underscores the challenges of dataset scarcity, ethical concerns, and integration barriers in clinical settings. Recommendations for future research include developing more representative datasets, advancing explainable AI methods, and exploring real-time implementation frameworks to bridge the gap between research and clinical practice.

RECENT SCHOLAR PUBLICATIONS

  • From Education to Employment: A Deep Learning Approach to Understanding Job Market Trends in Africa
    DK Dake, E Ofori, S Adablanu
    International Journal of Information and Education Technology 16 (4) , 2026
    2026
  • Comparative Efficacy of Focal and Binary Cross-Entropy Loss in Handling Class Imbalance for Stroke CT Classification
    S Adablanu, U Barman, D Das
    2026 Sixth International Conference on Advances in Electrical, Computing … , 2026
    2026
  • Transforming Skin Cancer Detection With AI‐Based Convolutional and Transformer Models
    S Adablanu, U Barman, D Das, TJ Dweh
    iRADIOLOGY 4 (1), 51-62 , 2026
    2026
    Citations: 1
  • Engaging 21st Century Learners and Differentiating Instruction with Multimedia: An Empirical Case Study of the University of Education, Winneba, Ghana
    CS Achulo, S Adablanu, DA Quaye
    International Journal of Computer Applications 975, 8887 , 2026
    2026
  • Feature Extraction and Selection Methods Outperform Machine Learning and Deep Learning Techniques
    TJ Dweh, S Adablanu
    Feature Selection and Feature Extraction on Omics Data, 194-213 , 2026
    2026
  • 15 Years of Optimizers in Medical Deep Learning: A Systematic Review
    S Adablanu, U Barman, D Das
    Neuroscience Informatics, 100249 , 2025
    2025
    Citations: 5
  • A novel hybrid adaptive transformer framework with multihead self attention for stroke detection
    S Adablanu, U Barman, D Das
    Discover Neuroscience 20 (1), 24 , 2025
    2025
    Citations: 1
  • A Thorough Review of AI Developments in Education: Historical Progress, Current Applications, and Future Directions
    S Adablanu, B Ghansah, BB Benuwa, SO Oppong
    Journal of Artificial Intelligence in Education 1 (2), 52-63 , 2025
    2025
  • The Internet of Things in Education: Adoption Patterns and Learning Outcomes from a Ghanaian Case Study
    PK Matey, OA Rubbin, S Adablanu
    International Journal of Computer Applications 187 (51), 32-41 , 2025
    2025
    Citations: 1
  • Advancing deep learning for automated stroke detection: a review
    S Adablanu, U Barman, D Das
    Brain Hemorrhages , 2025
    2025
    Citations: 13
  • Brain Hemorrhages
    S Adablanu, U Barman, D Das
    PRISMA 15, 16 , 2025
    2025
  • Adopting sustainable mobile learning: Investigating long-term integration at UEW with a focus on infrastructure, resources, and institutional support
    S Adablanu, M Offei, A Boateng
    Advances in Mobile Learning Educational Research 4 (2), 1173-1189 , 2024
    2024
    Citations: 4
  • Adversarial Machine Learning for Robust Intrusion Detection Systems
    F Olaoye, P Broklyn, S Adablanu
    EasyChair , 2024
    2024
    Citations: 2
  • The Intersection of Artificial Intelligence and Cybersecurity
    A Olukemi, P Broklyn, S Adablanu
    EasyChair , 2024
    2024
    Citations: 1
  • Homomorphic Encryption for Secure Cloud Computing
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 2
  • Explainable Neural Networks for Interpretable Cybersecurity Decisions
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 1
  • Reinforcement Learning for Adaptive Cybersecurity Policy Optimization
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
  • EasyChair Preprint Adversarial Machine Learning for Cybersecurity Defense
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
  • Blockchain-based Security Solutions for the Internet of Things (IoT)
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 2
  • AI-Powered Diagnostics and Imaging Analysis: Revolutionizing Medical Decision- Making
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 3

MOST CITED SCHOLAR PUBLICATIONS

  • Artificial intelligence in education: Trends, opportunities and pitfalls for institutes of higher education in Ghana
    AMA Nsoh
    International Journal of Computer Science and Mobile Computing (IJCSMC) , 2023
    2023
    Citations: 33
  • Multimodal Deep Learning for Integrated Cybersecurity Analytics (No. 14011)
    K Potter, D Stilinski, S Adablanu
    EasyChair , 2024
    2024
    Citations: 19
  • Advancing deep learning for automated stroke detection: a review
    S Adablanu, U Barman, D Das
    Brain Hemorrhages , 2025
    2025
    Citations: 13
  • 15 Years of Optimizers in Medical Deep Learning: A Systematic Review
    S Adablanu, U Barman, D Das
    Neuroscience Informatics, 100249 , 2025
    2025
    Citations: 5
  • Adopting sustainable mobile learning: Investigating long-term integration at UEW with a focus on infrastructure, resources, and institutional support
    S Adablanu, M Offei, A Boateng
    Advances in Mobile Learning Educational Research 4 (2), 1173-1189 , 2024
    2024
    Citations: 4
  • AI-Powered Diagnostics and Imaging Analysis: Revolutionizing Medical Decision- Making
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 3
  • Adversarial Machine Learning for Robust Intrusion Detection Systems
    F Olaoye, P Broklyn, S Adablanu
    EasyChair , 2024
    2024
    Citations: 2
  • Homomorphic Encryption for Secure Cloud Computing
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 2
  • Blockchain-based Security Solutions for the Internet of Things (IoT)
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 2
  • Fasttracking Healthcare Services for Students through the Design of a Hospital Information System
    H Techie-Menson, D Danso Essel, G Kudjo Bada, S Adablanu
    Journal of Education and Practice 2 , 2022
    2022
    Citations: 2
  • Transforming Skin Cancer Detection With AI‐Based Convolutional and Transformer Models
    S Adablanu, U Barman, D Das, TJ Dweh
    iRADIOLOGY 4 (1), 51-62 , 2026
    2026
    Citations: 1
  • A novel hybrid adaptive transformer framework with multihead self attention for stroke detection
    S Adablanu, U Barman, D Das
    Discover Neuroscience 20 (1), 24 , 2025
    2025
    Citations: 1
  • The Internet of Things in Education: Adoption Patterns and Learning Outcomes from a Ghanaian Case Study
    PK Matey, OA Rubbin, S Adablanu
    International Journal of Computer Applications 187 (51), 32-41 , 2025
    2025
    Citations: 1
  • The Intersection of Artificial Intelligence and Cybersecurity
    A Olukemi, P Broklyn, S Adablanu
    EasyChair , 2024
    2024
    Citations: 1
  • Explainable Neural Networks for Interpretable Cybersecurity Decisions
    K Potter, S Adablanu, D Stilinski
    EasyChair , 2024
    2024
    Citations: 1
  • Review on Automatic Smart Car Parking System
    S Adablanu
    International Journal of Computer Science and Mobile Computing 11 (8), 79-83 , 2022
    2022
    Citations: 1
  • From Education to Employment: A Deep Learning Approach to Understanding Job Market Trends in Africa
    DK Dake, E Ofori, S Adablanu
    International Journal of Information and Education Technology 16 (4) , 2026
    2026
  • Comparative Efficacy of Focal and Binary Cross-Entropy Loss in Handling Class Imbalance for Stroke CT Classification
    S Adablanu, U Barman, D Das
    2026 Sixth International Conference on Advances in Electrical, Computing … , 2026
    2026
  • Engaging 21st Century Learners and Differentiating Instruction with Multimedia: An Empirical Case Study of the University of Education, Winneba, Ghana
    CS Achulo, S Adablanu, DA Quaye
    International Journal of Computer Applications 975, 8887 , 2026
    2026
  • Feature Extraction and Selection Methods Outperform Machine Learning and Deep Learning Techniques
    TJ Dweh, S Adablanu
    Feature Selection and Feature Extraction on Omics Data, 194-213 , 2026
    2026