Menwa Alshammeri
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Menwa Alshammeri has a total of 85 co-authors. Here are the top co-authors by number of shared publications:
- Malak Alamri
Department of Computer Science, College of Computer and Information Sciences, Jouf University, Sakaka, Al-Jouf, Saudi Arabia
6 shared publications - Mamoona Humayun
Department of Computing, School of Arts Humanities and Social Sciences, University of Roehampton, London, UK
5 shared publications Abdulwhab Alkharashi
Department of Computer Science, College of Computing and Informatics, Saudi Electronic University, Riyadh, Saudi Arabia
3 shared publications- Ali Alqazzaz
College of Computing and Information Technology, University of Bisha, Bisha, Saudi Arabia
3 shared publications Mamoona Humayun
Department of Computer Science, Nottingham Trent University, Nottingham, UK
3 shared publicationsOthman Alrusaini
Department of Engineering and Applied Sciences, Umm Al-Qura University Makkah, Makkah, Saudi Arabia
3 shared publications
Latest Publications
- Open Access
- 15 August 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Blockchain-enabled intelligent model for energy-constrained IoMT networks with privacy-preserving
- 15 August 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Open AccessBlockchain-enabled intelligent model for energy-constrained IoMT networks with privacy-preserving
- Open Access
- 19 February 2026
- Published by MDPI AG in Diagnostics
- Vol. 16 (4) , 607
Machine Learning Framework for HbA1c Prediction: Data Enrichment, Cost Optimization, and Interpretability Through Stratified Regression and Multi-Stage Feature Selection
Background: Measuring glycated hemoglobin (HbA1c) is essential for assessing long-term glycemic control, yet direct testing remains expensive and underutilized in many large-scale health surveys and resource-constrained settings. This study aims to (i) deliver a highly accurate and interpretable ML model for predicting HbA1c from routinely collected clinical, biochemical, and demographic data, (ii) reduce dependency on extensive laboratory panels by identifying a compact, cost-efficient subset of key predictors, and (iii) establish a transferable, explainable modeling framework applicable across chronic disease biomarkers. Unlike prior HbA1c prediction studies that focus primarily on classification or accuracy-driven models, this work introduces a unified framework for continuous HbA1c regression that jointly integrates cost-oriented feature parsimony, stratified regression validation, and explainability by design. Methods: We aggregated data from the National Health and Nutrition Examination Survey (NHANES) cycles 2007–2020, encompassing 66,148 records and 224 candidate features. We implemented a two-stage feature selection pipeline: Incremental Correlation Selection (ICS) to narrow the variable space, followed by Recursive Feature Elimination with Cross-Validation (RFECV) to isolate the most informative features. Model interpretability was assessed using partial dependence plots and feature importance analysis. Results: The optimal model, LightGBMRegressor with most-frequent imputation, achieved R2 = 0.7161, MAE = 0.334, MSE = 0.304, and MAPE = 5.56%, while using only 40 selected features. Interpretability analysis revealed clinically coherent relationships that align with physiological expectations. Discussion: The proposed framework maintains robust predictive performance while substantially reducing the number of required input features, enabling cost-efficient HbA1c estimation together with transparent, physiologically coherent model insights. By consolidating continuous HbA1c prediction, cost-aware feature selection, stratified evaluation, and explainability within a single pipeline are enhanced. Conclusions: This study advances beyond existing approaches and offers a practical blueprint for scalable biomarker estimation in population health and clinical decision-support applications. Its explainable, efficient, and generalizable design positions it as a strong candidate for clinical decision-support and population-health applications.
- 19 February 2026
- Published by MDPI AG in Diagnostics
- Vol. 16 (4) , 607
Open AccessMachine Learning Framework for HbA1c Prediction: Data Enrichment, Cost Optimization, and Interpretability Through Stratified Regression and Multi-Stage Feature Selection
Background: Measuring glycated hemoglobin (HbA1c) is essential for assessing long-term glycemic control, yet direct testing remains expensive and underutilized in many large-scale health surveys and resource-constrained settings. This study aims to (i) deliver a highly accurate and interpretable ML model for predicting HbA1c from routinely collected clinical, biochemical, and demographic data, (ii) reduce dependency on extensive laboratory panels by identifying a compact, cost-efficient subset of key predictors, and (iii) establish a transferable, explainable modeling framework applicable across chronic disease biomarkers. Unlike prior HbA1c prediction studies that focus primarily on classification or accuracy-driven models, this work introduces a unified framework for continuous HbA1c regression that jointly integrates cost-oriented feature parsimony, stratified regression validation, and explainability by design. Methods: We aggregated data from the National Health and Nutrition Examination Survey (NHANES) cycles 2007–2020, encompassing 66,148 records and 224 candidate features. We implemented a two-stage feature selection pipeline: Incremental Correlation Selection (ICS) to narrow the variable space, followed by Recursive Feature Elimination with Cross-Validation (RFECV) to isolate the most informative features. Model interpretability was assessed using partial dependence plots and feature importance analysis. Results: The optimal model, LightGBMRegressor with most-frequent imputation, achieved R2 = 0.7161, MAE = 0.334, MSE = 0.304, and MAPE = 5.56%, while using only 40 selected features. Interpretability analysis revealed clinically coherent relationships that align with physiological expectations. Discussion: The proposed framework maintains robust predictive performance while substantially reducing the number of required input features, enabling cost-efficient HbA1c estimation together with transparent, physiologically coherent model insights. By consolidating continuous HbA1c prediction, cost-aware feature selection, stratified evaluation, and explainability within a single pipeline are enhanced. Conclusions: This study advances beyond existing approaches and offers a practical blueprint for scalable biomarker estimation in population health and clinical decision-support applications. Its explainable, efficient, and generalizable design positions it as a strong candidate for clinical decision-support and population-health applications.
- 17 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. PP (00983063) , 1
Cybersecurity Framework for Proactive Detection of Image-Based Malware in Healthcare Networks Using Explainable Deep Ensemble Learning and Edge Analytics
The fast-paced digitalization of the healthcare sector (Electronic Health Records (EHRs), Internet of Medical Things (IoMT) devices, medical imaging, and telemedicine) has enabled more efficient patient care but has also increased the attack surface of sophisticated threats. These security concerns are emerging in consumer electronics within smart healthcare ecosystems, which rely on image-based analytics and edge-enabled intelligence. In this study, we propose an edge-enabled cybersecurity framework that integrates deep ensemble learning and edge analytics to preemptively identify image-based malware in healthcare networks and provide low-latency inference at the data source. We introduce VGDense-DeepEnsemble (VGGDenseNet-DeepEnsemble), a novel stacked deep learning model that combines VGG16, VGG19, DenseNet121, and DenseNet169 through ensemble learning with a meta-model to enhance detection performance. Explainable AI (XAI) using Grad-CAM was applied to sample images to interpret the decision-making process. We evaluated the proposed model on the Blended Malware Image Dataset, which comprises 31 malware families with a separate test dataset for final assessment. The proposed model attains 96% overall accuracy with macro precision of 97%, macro recall of 96%, and macro F1 score of 96%. These results exceed the performance of the individual backbones (VGG16: 88%, VGG19: 94%, DenseNet121: 93%, DenseNet169: 94%). As revealed through the ROC and precision-recall analysis, separability and robustness are high across all the classes, and confusion matrices reveal significant decreases in misclassifications in challenging families (e.g., Neshta, Sality, Agent, Injector). Grad-CAM explanations indicate that the predictions are informed by discriminative, malware-relevant structures and not background artifacts. These findings show that transparent edge-based deep ensemble models can be an effective, accurate, explainable defense against image-based malware in healthcare networks.
- 17 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. PP (00983063) , 1
Cybersecurity Framework for Proactive Detection of Image-Based Malware in Healthcare Networks Using Explainable Deep Ensemble Learning and Edge Analytics
The fast-paced digitalization of the healthcare sector (Electronic Health Records (EHRs), Internet of Medical Things (IoMT) devices, medical imaging, and telemedicine) has enabled more efficient patient care but has also increased the attack surface of sophisticated threats. These security concerns are emerging in consumer electronics within smart healthcare ecosystems, which rely on image-based analytics and edge-enabled intelligence. In this study, we propose an edge-enabled cybersecurity framework that integrates deep ensemble learning and edge analytics to preemptively identify image-based malware in healthcare networks and provide low-latency inference at the data source. We introduce VGDense-DeepEnsemble (VGGDenseNet-DeepEnsemble), a novel stacked deep learning model that combines VGG16, VGG19, DenseNet121, and DenseNet169 through ensemble learning with a meta-model to enhance detection performance. Explainable AI (XAI) using Grad-CAM was applied to sample images to interpret the decision-making process. We evaluated the proposed model on the Blended Malware Image Dataset, which comprises 31 malware families with a separate test dataset for final assessment. The proposed model attains 96% overall accuracy with macro precision of 97%, macro recall of 96%, and macro F1 score of 96%. These results exceed the performance of the individual backbones (VGG16: 88%, VGG19: 94%, DenseNet121: 93%, DenseNet169: 94%). As revealed through the ROC and precision-recall analysis, separability and robustness are high across all the classes, and confusion matrices reveal significant decreases in misclassifications in challenging families (e.g., Neshta, Sality, Agent, Injector). Grad-CAM explanations indicate that the predictions are informed by discriminative, malware-relevant structures and not background artifacts. These findings show that transparent edge-based deep ensemble models can be an effective, accurate, explainable defense against image-based malware in healthcare networks.
- Open Access
- 27 January 2026
- Published by PeerJ in PeerJ Computer Science
- Vol. 12, e3429
Enhancing micro-architecture optimization using explainable fuzzy neural networks with multi-fidelity reinforcement learning
Micro-architecture design space exploration (DSE) is essential for the optimization of processor performance. However, existing approaches generally fall short in interpretability, preventing designers from comprehending and fine-tuning decisions. In this article, we develop an Explainable Fuzzy Neural Network (FNN) framework coupled with multi-fidelity Reinforcement Learning (RL) for micro-architecture optimization. Instead, this proposed approach is a fuzzy-based one. It induces readable design rules by mimicking a neural network without limiting its flexibility. We employ a multi-fidelity RL strategy to simultaneously enable fast evaluation on analytical models, as well as perform accurate design validation on register transfer level (RTL) simulations to save time and guarantee correctness. Opportune learning of this kind greatly lowers computation costs and retains solutions of suitable quality. The FNN learns interpretable decision rules in an unsupervised fashion, allowing the designer to visualize optimization paths taken and incorporate domain knowledge to guide exploration. Our experiments show that our approach substantially improves the efficiency and accuracy over state-of-the-art DSE methods, yielding near-optimal micro-architectures with a very small sample budget. The interpretability of the framework also enables designers to visually inspect and optimize architectural trade-offs. Nearly closes the gap between black-box optimization and human-guided decision-making, our methodology establishes a solid groundwork for both explainable and efficient micro-architecture DSE, offering a pathway for future work that turns to more transparent and modelled methods of processor design.
- 27 January 2026
- Published by PeerJ in PeerJ Computer Science
- Vol. 12, e3429
Open AccessEnhancing micro-architecture optimization using explainable fuzzy neural networks with multi-fidelity reinforcement learning
Micro-architecture design space exploration (DSE) is essential for the optimization of processor performance. However, existing approaches generally fall short in interpretability, preventing designers from comprehending and fine-tuning decisions. In this article, we develop an Explainable Fuzzy Neural Network (FNN) framework coupled with multi-fidelity Reinforcement Learning (RL) for micro-architecture optimization. Instead, this proposed approach is a fuzzy-based one. It induces readable design rules by mimicking a neural network without limiting its flexibility. We employ a multi-fidelity RL strategy to simultaneously enable fast evaluation on analytical models, as well as perform accurate design validation on register transfer level (RTL) simulations to save time and guarantee correctness. Opportune learning of this kind greatly lowers computation costs and retains solutions of suitable quality. The FNN learns interpretable decision rules in an unsupervised fashion, allowing the designer to visualize optimization paths taken and incorporate domain knowledge to guide exploration. Our experiments show that our approach substantially improves the efficiency and accuracy over state-of-the-art DSE methods, yielding near-optimal micro-architectures with a very small sample budget. The interpretability of the framework also enables designers to visually inspect and optimize architectural trade-offs. Nearly closes the gap between black-box optimization and human-guided decision-making, our methodology establishes a solid groundwork for both explainable and efficient micro-architecture DSE, offering a pathway for future work that turns to more transparent and modelled methods of processor design.
- Top Cited PapersOpen Access
- 10 January 2026
- Published by Springer Nature in Cluster Computing
- Vol. 29 (2) , 108
Energy-efficient threat detection in IoT healthcare using AI and blockchain-enhanced fog–cloud architecture
Intrusion detection in Internet of Things (IoT) networks, particularly in healthcare settings, poses critical challenges due to latency constraints, limited resources, and the need for trustworthy auditing in distributed environments. Centralized detection models often fail to deliver timely or scalable responses under real-world IoT conditions. This study proposes a hybrid fog–cloud architecture tailored for healthcare-oriented IoT threat detection, incorporating blockchain-based auditability. The architecture utilizes fog- and cloud-level XGBoost classifiers trained on BoT-IoT and ToN-IoT datasets, with SMOTE applied to mitigate class imbalance. A lightweight blockchain module is integrated at the fog layer to log predictions in real-time for tamper-evident traceability. Simulations were performed using 50 fog-predicted events to evaluate performance, energy usage, and blockchain entropy. The system achieved an average block creation time of under 20 ms with minimal CPU and memory overhead. It also demonstrated robustness against tampering, preserving data integrity. The fog-level model achieved competitive metrics (AUC = 1, F1-score = 98.70%, Accuracy = 99.80%) compared to the cloud model, while outperforming it in terms of response latency and localized decision-making. The proposed blockchain-integrated fog–cloud framework enables secure, low-latency, and scalable threat detection for healthcare IoT systems, offering a promising foundation for privacy-aware edge intelligence.
- 10 January 2026
- Published by Springer Nature in Cluster Computing
- Vol. 29 (2) , 108
Top Cited PapersOpen AccessEnergy-efficient threat detection in IoT healthcare using AI and blockchain-enhanced fog–cloud architecture
Intrusion detection in Internet of Things (IoT) networks, particularly in healthcare settings, poses critical challenges due to latency constraints, limited resources, and the need for trustworthy auditing in distributed environments. Centralized detection models often fail to deliver timely or scalable responses under real-world IoT conditions. This study proposes a hybrid fog–cloud architecture tailored for healthcare-oriented IoT threat detection, incorporating blockchain-based auditability. The architecture utilizes fog- and cloud-level XGBoost classifiers trained on BoT-IoT and ToN-IoT datasets, with SMOTE applied to mitigate class imbalance. A lightweight blockchain module is integrated at the fog layer to log predictions in real-time for tamper-evident traceability. Simulations were performed using 50 fog-predicted events to evaluate performance, energy usage, and blockchain entropy. The system achieved an average block creation time of under 20 ms with minimal CPU and memory overhead. It also demonstrated robustness against tampering, preserving data integrity. The fog-level model achieved competitive metrics (AUC = 1, F1-score = 98.70%, Accuracy = 99.80%) compared to the cloud model, while outperforming it in terms of response latency and localized decision-making. The proposed blockchain-integrated fog–cloud framework enables secure, low-latency, and scalable threat detection for healthcare IoT systems, offering a promising foundation for privacy-aware edge intelligence.
- Open Access
- 1 January 2026
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 89 (3) , 1-10
Lightweight AI-Based Threat Detection and Surveillance Secure Edge-IoT Healthcare Systems
Advances in emerging technologies, such as the Internet of Things (IoT) and artificial intelligence, provide a wide range of real-time applications to the development of smart cities. The physical objects, along with IoT systems, not only provide seamless connectivity to the remote environment but also continuously gather the required data through a sensor-based network. They provide timely information to connected devices, meet their needs, and enhance the flexibility of automated IoT-driven systems. Despite improvements in network maintenance and communication robustness, most existing healthcare systems still encounter challenges in addressing congestion and scalability as data traffic increases. Furthermore, as autonomous activities increase, communication trust among health devices decreases, thereby increasing computational overhead due to privacy concerns and potential attacks. This research proposes an intelligent, edge-enabled machine-learning model for an autonomous environment to identify proactive threats and enhance system trustworthiness by mitigating unauthorized and insecure activities in the healthcare network. The devices are controlled efficiently, with uniform load balancing, using edge computing and heuristic techniques to optimize resource allocation and energy efficiency. It intelligently controls network traffic and reduces on-device computational power by dynamically offloading to selected forwarders, without increasing network consumption or overhead. Moreover, the trusted schemes are designed to increase reliability within a bounded network system and to support the integrity of devices while processing health records at the network edges. Simulation results confirm that the proposed framework achieves improvements in energy overhead, latency, throughput, and response time compared to the RePUF-IoT and AIBS-IoTH schemes.
- 1 January 2026
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 89 (3) , 1-10
Open AccessLightweight AI-Based Threat Detection and Surveillance Secure Edge-IoT Healthcare Systems
Advances in emerging technologies, such as the Internet of Things (IoT) and artificial intelligence, provide a wide range of real-time applications to the development of smart cities. The physical objects, along with IoT systems, not only provide seamless connectivity to the remote environment but also continuously gather the required data through a sensor-based network. They provide timely information to connected devices, meet their needs, and enhance the flexibility of automated IoT-driven systems. Despite improvements in network maintenance and communication robustness, most existing healthcare systems still encounter challenges in addressing congestion and scalability as data traffic increases. Furthermore, as autonomous activities increase, communication trust among health devices decreases, thereby increasing computational overhead due to privacy concerns and potential attacks. This research proposes an intelligent, edge-enabled machine-learning model for an autonomous environment to identify proactive threats and enhance system trustworthiness by mitigating unauthorized and insecure activities in the healthcare network. The devices are controlled efficiently, with uniform load balancing, using edge computing and heuristic techniques to optimize resource allocation and energy efficiency. It intelligently controls network traffic and reduces on-device computational power by dynamically offloading to selected forwarders, without increasing network consumption or overhead. Moreover, the trusted schemes are designed to increase reliability within a bounded network system and to support the integrity of devices while processing health records at the network edges. Simulation results confirm that the proposed framework achieves improvements in energy overhead, latency, throughput, and response time compared to the RePUF-IoT and AIBS-IoTH schemes.
- Open Access
- 1 January 2026
- Published by Tech Science Press in Computer Modeling in Engineering & Sciences
- Vol. 146 (1) , 1-10
Explainable Ensemble Learning Framework for Early Detection of Autism Spectrum Disorder: Enhancing Trust, Interpretability and Reliability in AI-Driven Healthcare
Artificial Intelligence (AI) is changing healthcare by helping with diagnosis. However, for doctors to trust AI tools, they need to be both accurate and easy to understand. In this study, we created a new machine learning system for the early detection of Autism Spectrum Disorder (ASD) in children. Our main goal was to build a model that is not only good at predicting ASD but also clear in its reasoning. For this, we combined several different models, including Random Forest, XGBoost, and Neural Networks, into a single, more powerful framework. We used two different types of datasets: (i) a standard behavioral dataset and (ii) a more complex multimodal dataset with images, audio, and physiological information. The datasets were carefully preprocessed for missing values, redundant features, and dataset imbalance to ensure fair learning. The results outperformed the state-of-the-art with a Regularized Neural Network, achieving 97.6% accuracy on behavioral data. Whereas, on the multimodal data, the accuracy is 98.2%. Other models also did well with accuracies consistently above 96%. We also used SHAP and LIME on a behavioral dataset for models’ explainability.
- 1 January 2026
- Published by Tech Science Press in Computer Modeling in Engineering & Sciences
- Vol. 146 (1) , 1-10
Open AccessExplainable Ensemble Learning Framework for Early Detection of Autism Spectrum Disorder: Enhancing Trust, Interpretability and Reliability in AI-Driven Healthcare
Artificial Intelligence (AI) is changing healthcare by helping with diagnosis. However, for doctors to trust AI tools, they need to be both accurate and easy to understand. In this study, we created a new machine learning system for the early detection of Autism Spectrum Disorder (ASD) in children. Our main goal was to build a model that is not only good at predicting ASD but also clear in its reasoning. For this, we combined several different models, including Random Forest, XGBoost, and Neural Networks, into a single, more powerful framework. We used two different types of datasets: (i) a standard behavioral dataset and (ii) a more complex multimodal dataset with images, audio, and physiological information. The datasets were carefully preprocessed for missing values, redundant features, and dataset imbalance to ensure fair learning. The results outperformed the state-of-the-art with a Regularized Neural Network, achieving 97.6% accuracy on behavioral data. Whereas, on the multimodal data, the accuracy is 98.2%. Other models also did well with accuracies consistently above 96%. We also used SHAP and LIME on a behavioral dataset for models’ explainability.
- Open Access
- 18 December 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (24) , 3241
Explainable Cluster-Based Predictive Framework for Early Diagnosis of Autism Spectrum Disorder Using Behavioral Biomarkers
Background/Objectives: Autism Spectrum Disorder (ASD) is a multifaceted neuropsychiatric condition characterized by early behavioral irregularities that often precede formal diagnosis. Timely and precise detection remains a major clinical challenge due to the complexity of behavioral manifestations and the limited accessibility of diagnostic resources. Methods: In this study, we present an explainable machine learning framework for the early diagnosis of ASD using behavioral biomarkers derived from toddler screening data. The framework integrates unsupervised learning (DBSCAN and K-means clustering) to identify latent behavioral patterns, followed by predictive modeling using logistic regression (LR), random forest (RF), and support vector machine (SVM). To ensure transparency and clinical interpretability, a SHAP (SHapley Additive exPlanations) analysis is employed to quantify the contribution of each behavioral feature to the model’s predictions. Results: Experimental evaluations reveal that the RF model achieves the highest accuracy (98.85%), followed by SVM (97.70%) and LR (90.53%). The explainability results highlight meaningful and clinically relevant behavioral indicators associated with ASD risk. Conclusions: The proposed framework not only enhances diagnostic accuracy but also promotes interpretable AI for real-world integration into neuropsychiatric assessment pipelines.
- 18 December 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (24) , 3241
Open AccessExplainable Cluster-Based Predictive Framework for Early Diagnosis of Autism Spectrum Disorder Using Behavioral Biomarkers
Background/Objectives: Autism Spectrum Disorder (ASD) is a multifaceted neuropsychiatric condition characterized by early behavioral irregularities that often precede formal diagnosis. Timely and precise detection remains a major clinical challenge due to the complexity of behavioral manifestations and the limited accessibility of diagnostic resources. Methods: In this study, we present an explainable machine learning framework for the early diagnosis of ASD using behavioral biomarkers derived from toddler screening data. The framework integrates unsupervised learning (DBSCAN and K-means clustering) to identify latent behavioral patterns, followed by predictive modeling using logistic regression (LR), random forest (RF), and support vector machine (SVM). To ensure transparency and clinical interpretability, a SHAP (SHapley Additive exPlanations) analysis is employed to quantify the contribution of each behavioral feature to the model’s predictions. Results: Experimental evaluations reveal that the RF model achieves the highest accuracy (98.85%), followed by SVM (97.70%) and LR (90.53%). The explainability results highlight meaningful and clinically relevant behavioral indicators associated with ASD risk. Conclusions: The proposed framework not only enhances diagnostic accuracy but also promotes interpretable AI for real-world integration into neuropsychiatric assessment pipelines.
- 16 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (9) , 18325-18333
Intent-Based Secure Fault Tolerance Model With Integrated AI for Edge-IoT Networks
The development of real-time applications integrated with intent-based networking (IBN) integrates an Internet of Things (IoT), providing interconnection between heterogeneous devices and physical objects for the formulation of smart cities. These systems provide seamless communication and maintain the adaptive network policies and infrastructure. Many existing schemes have proposed solutions for efficient routing with support from intelligent architectures; however, although most of them overlook the bounded and limited resources of IoT networks, they impose additional overhead while addressing unpredictable communications in IBN-IoT. Furthermore, security and trustworthiness are significant research challenges that must be addressed to prevent data breaches and allow only the use of authentic devices. This research presents a scalable model for an IBN-IoT environment that utilizes edge computing to enable trustworthy and fault-tolerant communication with energy efficiency. First, software-defined networking (SDN) is explored for load balancing and effective resource allocation in 6G-IoT systems. Second, the proposed model explores artificial intelligence techniques to analyze the network environment and predict anomalies in the fault tolerance approach. Finally, data is kept private and maintained in integrity using a private blockchain, providing a more reliable, distributed, autonomous system with minimal overhead. Using synthetic data, the proposed model is validated against quantum genetic algorithm and ant colony optimization (QGA-ACO) and metaheuristic energy-aware routing with optimal DL-based anomaly detection technique (MER-ODLADT) solutions for energy consumption, anomaly detection, and time-to-failure metrics across dynamic scenarios.
- 16 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (9) , 18325-18333
Intent-Based Secure Fault Tolerance Model With Integrated AI for Edge-IoT Networks
The development of real-time applications integrated with intent-based networking (IBN) integrates an Internet of Things (IoT), providing interconnection between heterogeneous devices and physical objects for the formulation of smart cities. These systems provide seamless communication and maintain the adaptive network policies and infrastructure. Many existing schemes have proposed solutions for efficient routing with support from intelligent architectures; however, although most of them overlook the bounded and limited resources of IoT networks, they impose additional overhead while addressing unpredictable communications in IBN-IoT. Furthermore, security and trustworthiness are significant research challenges that must be addressed to prevent data breaches and allow only the use of authentic devices. This research presents a scalable model for an IBN-IoT environment that utilizes edge computing to enable trustworthy and fault-tolerant communication with energy efficiency. First, software-defined networking (SDN) is explored for load balancing and effective resource allocation in 6G-IoT systems. Second, the proposed model explores artificial intelligence techniques to analyze the network environment and predict anomalies in the fault tolerance approach. Finally, data is kept private and maintained in integrity using a private blockchain, providing a more reliable, distributed, autonomous system with minimal overhead. Using synthetic data, the proposed model is validated against quantum genetic algorithm and ant colony optimization (QGA-ACO) and metaheuristic energy-aware routing with optimal DL-based anomaly detection technique (MER-ODLADT) solutions for energy consumption, anomaly detection, and time-to-failure metrics across dynamic scenarios.
- Open Access
- 30 October 2025
- Published by PeerJ in PeerJ Computer Science
- Vol. 11, e3197
GRIFD: a graded region-wise dissection and cross-pooling RNN framework for precise diabetic retinopathy detection in fundus images
Early and precise detection of diabetic retinopathy (DR) is essential in averting vision impairment. This work introduces graded region wise inspection and feature dissection (GRIFD), a novel deep learning architecture designed explicitly for fine-grained DR detection from fundus images. GRIFD enforces a graded region of interest (ROI) dissection approach, which methodically segments fundus images into diagnostic blocks to promote local lesion visibility. A recurrent neural network (RNN) is used to model sequential dependencies between these ROI blocks, aiming to mimic the progressive development of DR symptoms. We also introduce a cross-pooling refinement procedure to eliminate inter-region inconsistencies and improve feature continuity. Experiments on the Mendeley and EyePACS benchmarks show that GRIFD obtains 94.35% accuracy, 96.05% precision, and 94.91% specificity on Mendeley, all of which surpass baseline models such as convolutional neural network (CNN), long short-term memory (LSTM), Transformer, and some published DR detection models. The introduced framework presents a lightweight and robust method for DR screening and can potentially facilitate real-time clinical practice.
- 30 October 2025
- Published by PeerJ in PeerJ Computer Science
- Vol. 11, e3197
Open AccessGRIFD: a graded region-wise dissection and cross-pooling RNN framework for precise diabetic retinopathy detection in fundus images
Early and precise detection of diabetic retinopathy (DR) is essential in averting vision impairment. This work introduces graded region wise inspection and feature dissection (GRIFD), a novel deep learning architecture designed explicitly for fine-grained DR detection from fundus images. GRIFD enforces a graded region of interest (ROI) dissection approach, which methodically segments fundus images into diagnostic blocks to promote local lesion visibility. A recurrent neural network (RNN) is used to model sequential dependencies between these ROI blocks, aiming to mimic the progressive development of DR symptoms. We also introduce a cross-pooling refinement procedure to eliminate inter-region inconsistencies and improve feature continuity. Experiments on the Mendeley and EyePACS benchmarks show that GRIFD obtains 94.35% accuracy, 96.05% precision, and 94.91% specificity on Mendeley, all of which surpass baseline models such as convolutional neural network (CNN), long short-term memory (LSTM), Transformer, and some published DR detection models. The introduced framework presents a lightweight and robust method for DR screening and can potentially facilitate real-time clinical practice.