Djamel Benmerzoug
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Djamel Benmerzoug has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
- Hadjer Bechinia
University Abdelhamid Mehri Constantine 2
4 shared publications - Karim Zarour
University of Constantine 2 - Abdelhamid Mehri
4 shared publications Fouzi Lezzar
Constantine 2 University, LIRE Laboratory, Constantine, Algeria
3 shared publications- Nawal Guermouche
LAAS-CNRS, University of Toulouse, INSA, Toulouse, France
3 shared publications - Nawres Khlifa
University of Tunis El Manar
3 shared publications - Nour El Houda Boubaker
Abdelhamid Mehri-Constantine2 University, LIRE Laboratory, Constantine, Algeria
3 shared publications
Latest Publications
- Open Access
- 24 June 2026
- Published by Atlantis Press SARL
- p. 419-429
BreCAI: An AI-Powered Platform for Clinical Decision Support in Breast Cancer Diagnosis
- 24 June 2026
- Published by Atlantis Press SARL
- p. 419-429
Open AccessBreCAI: An AI-Powered Platform for Clinical Decision Support in Breast Cancer Diagnosis
- Open Access
- 14 October 2025
- Published by Kaunas University of Technology (KTU) in Information Technology And Control
- Vol. 54 (3) , 864-884
A TCGAN-Based Real-Time Personalized Motion Guidance System to Reduce Compensatory Movements in Post-Stroke Rehabilitation
Stroke rehabilitation is essential for motor function recovery, yet traditional methods require therapist supervision, which can be costly and inaccessible. Home-based rehabilitation offers an alternative, but without real-time guidance, patients may develop compensatory movements, hindering progress. Existing approaches provide feedback only after exercises are completed, limiting their effectiveness. To address this, we propose a Temporal Conditional Generative Adversarial Network (TCGAN)-based motion generation system that provides real-time skeletal guidance tailored to each patient’s body structure and positioning. By detecting key anatomical landmarks and generating adaptive motion sequences, the system ensures precise movement execution, reducing errors and improving rehabilitation outcomes. Both qualitative and quantitative evaluations confirm the effectiveness of the generated exercises, benefiting from the proposed architecture, improved loss function, optimized training process, and TCGAN hyperparameter tuning. Experimental results show a high degree of similarity between generated and real movements, with a Fréchet Inception Distance (FID) score of 0.87, demonstrating the system’s realism and reliability. This approach enhances patient autonomy and recovery efficiency, offering a more interactive and adaptive rehabilitation experience.
- 14 October 2025
- Published by Kaunas University of Technology (KTU) in Information Technology And Control
- Vol. 54 (3) , 864-884
Open AccessA TCGAN-Based Real-Time Personalized Motion Guidance System to Reduce Compensatory Movements in Post-Stroke Rehabilitation
Stroke rehabilitation is essential for motor function recovery, yet traditional methods require therapist supervision, which can be costly and inaccessible. Home-based rehabilitation offers an alternative, but without real-time guidance, patients may develop compensatory movements, hindering progress. Existing approaches provide feedback only after exercises are completed, limiting their effectiveness. To address this, we propose a Temporal Conditional Generative Adversarial Network (TCGAN)-based motion generation system that provides real-time skeletal guidance tailored to each patient’s body structure and positioning. By detecting key anatomical landmarks and generating adaptive motion sequences, the system ensures precise movement execution, reducing errors and improving rehabilitation outcomes. Both qualitative and quantitative evaluations confirm the effectiveness of the generated exercises, benefiting from the proposed architecture, improved loss function, optimized training process, and TCGAN hyperparameter tuning. Experimental results show a high degree of similarity between generated and real movements, with a Fréchet Inception Distance (FID) score of 0.87, demonstrating the system’s realism and reliability. This approach enhances patient autonomy and recovery efficiency, offering a more interactive and adaptive rehabilitation experience.
- 2 October 2025
- Published by Springer Nature
- p. 245-254
A Bilinear Convolutional Neural Network for Arrhythmia Classification on ECG Signals
- 2 October 2025
- Published by Springer Nature
- p. 245-254
A Bilinear Convolutional Neural Network for Arrhythmia Classification on ECG Signals
- 29 September 2025
- Published by Springer Nature in Cluster Computing
- Vol. 28 (14) , 1-27
Heuristic-based DQN for sensitive task-aware partial workflow offloading and migration
- 29 September 2025
- Published by Springer Nature in Cluster Computing
- Vol. 28 (14) , 1-27
Heuristic-based DQN for sensitive task-aware partial workflow offloading and migration
- 12 July 2025
- Published by Springer Nature
- p. 31-43
Fault Prediction in IoT Applications Using Federated Learning Based on Local Differential Privacy
- 12 July 2025
- Published by Springer Nature
- p. 31-43
Fault Prediction in IoT Applications Using Federated Learning Based on Local Differential Privacy
- Open Access
- 27 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 111892-111925
A Comprehensive Survey on Resource Management for IoT Applications in Edge-Fog-Cloud Environments
Resource management in Edge-Fog-Cloud environments has emerged as a fundamental challenge in modern distributed computing systems, primarily due to the heterogeneous nature of underlying infrastructure, highly dynamic workloads, and the need to optimize multiple, often conflicting, Quality of Service (QoS) metrics such as latency, energy consumption, cost, and security. These environments are characterized by their layered architecture and real-time decision-making demands, which require efficient coordination among distributed resources to meet the expectations of both users and applications. Although several surveys in the literature have explored aspects of resource management, they tend to focus narrowly on specific components such as resource allocation, task scheduling, or workload placement, often neglecting the interdependence of these components. Moreover, a considerable number of surveys fail to consider the rapidly evolving landscape of edge and fog computing, where dynamic environmental changes and fluctuating user demands are becoming the norm. To bridge these gaps, this survey provides a comprehensive and up-to-date review of resource management strategies in Edge-Fog-Cloud ecosystems. We adopt a holistic perspective that integrates both combinatorial optimization techniques and learning-based methods, highlighting their roles in addressing different facets of resource management. The survey delves into five key aspects: task scheduling, resource placement, allocation, offloading, and service migration. We propose a structured taxonomy that classifies existing works based on their objectives, applied techniques, QoS decision-making mechanisms, and environmental assumptions. Special attention is given to the synergy between combinatorial and Machine Learning-based approaches, identifying where each excels and under which conditions their integration can yield more robust solutions. Ultimately, this survey outlines the current limitations of the field, uncovers open challenges, and suggests future research directions to enable more intelligent, efficient, and adaptive resource management solutions.
- 27 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 111892-111925
Open AccessA Comprehensive Survey on Resource Management for IoT Applications in Edge-Fog-Cloud Environments
Resource management in Edge-Fog-Cloud environments has emerged as a fundamental challenge in modern distributed computing systems, primarily due to the heterogeneous nature of underlying infrastructure, highly dynamic workloads, and the need to optimize multiple, often conflicting, Quality of Service (QoS) metrics such as latency, energy consumption, cost, and security. These environments are characterized by their layered architecture and real-time decision-making demands, which require efficient coordination among distributed resources to meet the expectations of both users and applications. Although several surveys in the literature have explored aspects of resource management, they tend to focus narrowly on specific components such as resource allocation, task scheduling, or workload placement, often neglecting the interdependence of these components. Moreover, a considerable number of surveys fail to consider the rapidly evolving landscape of edge and fog computing, where dynamic environmental changes and fluctuating user demands are becoming the norm. To bridge these gaps, this survey provides a comprehensive and up-to-date review of resource management strategies in Edge-Fog-Cloud ecosystems. We adopt a holistic perspective that integrates both combinatorial optimization techniques and learning-based methods, highlighting their roles in addressing different facets of resource management. The survey delves into five key aspects: task scheduling, resource placement, allocation, offloading, and service migration. We propose a structured taxonomy that classifies existing works based on their objectives, applied techniques, QoS decision-making mechanisms, and environmental assumptions. Special attention is given to the synergy between combinatorial and Machine Learning-based approaches, identifying where each excels and under which conditions their integration can yield more robust solutions. Ultimately, this survey outlines the current limitations of the field, uncovers open challenges, and suggests future research directions to enable more intelligent, efficient, and adaptive resource management solutions.
- 26 February 2025
- Published by Springer Nature in Journal of Grid Computing
- Vol. 23 (1) , 1-34
Optimizing Workflow Offloading and Migration under Timed Constraints in Fog and Cloud Computing
- 26 February 2025
- Published by Springer Nature in Journal of Grid Computing
- Vol. 23 (1) , 1-34
Optimizing Workflow Offloading and Migration under Timed Constraints in Fog and Cloud Computing
- 5 February 2025
- Published by Springer Nature
- p. 240-253
MobileViT-Based Approach for Arrhythmia Classification Using ECG Scalograms
- 5 February 2025
- Published by Springer Nature
- p. 240-253
MobileViT-Based Approach for Arrhythmia Classification Using ECG Scalograms
- Open Access
- 2 February 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (1) , 20211-20221
A Multidimensional Approach for Formal Modeling and Analyzing Medical Cyber-Physical Systems
The combination of integrated software controlling devices, networking capabilities, and sensing/actuation technologies in Medical Cyber-Physical Systems (M-CPS) highlights some specific research challenges. The major challenge is to formally ensure the confidentiality of the data or resources they handle. This study tackles this problem by proposing a formal approach that combines CA-BRS (Control Agent and Bigraphical Reactive Systems) and BPMN (Business Process Model Notation) to specify and analyze CPS in general, while respecting several dimensions. The structural dimension of the CPS, representing the space (physical and cyber entities) in which agents exist and interact, is defined with BRS. Control agents constitute the virtual dimension and observe and control the physical and cyber entities of their environment. The complex and adaptive behavior of CPS (behavioral dimension) is defined through several types of rules, each managing a possible evolution of a CPS component (physical, cyber, or virtual). Two distinctive perspectives are associated with the semantic interpretation of these rules: the states perspective and the activities perspective. This study focuses on the activities perspective that specifies the behavior of control agents with a BPMN activity diagram. This highlights how these two models (CA-BRS and BPMN) complement each other to assist the designer in defining formal models for CPS. Additionally, it reveals how to provide the CA-BRS model with means to control unauthorized access to an electronic health record system.
- 2 February 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (1) , 20211-20221
Open AccessA Multidimensional Approach for Formal Modeling and Analyzing Medical Cyber-Physical Systems
The combination of integrated software controlling devices, networking capabilities, and sensing/actuation technologies in Medical Cyber-Physical Systems (M-CPS) highlights some specific research challenges. The major challenge is to formally ensure the confidentiality of the data or resources they handle. This study tackles this problem by proposing a formal approach that combines CA-BRS (Control Agent and Bigraphical Reactive Systems) and BPMN (Business Process Model Notation) to specify and analyze CPS in general, while respecting several dimensions. The structural dimension of the CPS, representing the space (physical and cyber entities) in which agents exist and interact, is defined with BRS. Control agents constitute the virtual dimension and observe and control the physical and cyber entities of their environment. The complex and adaptive behavior of CPS (behavioral dimension) is defined through several types of rules, each managing a possible evolution of a CPS component (physical, cyber, or virtual). Two distinctive perspectives are associated with the semantic interpretation of these rules: the states perspective and the activities perspective. This study focuses on the activities perspective that specifies the behavior of control agents with a BPMN activity diagram. This highlights how these two models (CA-BRS and BPMN) complement each other to assist the designer in defining formal models for CPS. Additionally, it reveals how to provide the CA-BRS model with means to control unauthorized access to an electronic health record system.