Arunagiri Ramathilagam
Arunagiri Ramathilagam has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Dhinakaran Damodaran
Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, India
3 shared publications - J. Jeno Jasmine
Department of Computer Science and Engineering (AIML) Vel Tech High Tech Dr. Rangarajan Dr. Sakunthala Engineering College Chennai India
2 shared publications - S. Edwin Raja
Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, India
2 shared publications - A. Arulmurugan
Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, India
1 shared publication - Gopalakrishnan Subburayalu
Department of Information Technology, Hindustan Institute of Technology and Science, Chennai, India
1 shared publication - J. J. Jayakanth
SRMIST
1 shared publication
Latest Publications
- 5 February 2026
- Published by IGI Global Scientific Publishing
- p. 293-324
Revolutionizing Health Management Through Federated Learning and AI
The integration of Artificial Intelligence (AI) and Internet of Things (IoT) technologies is transforming healthcare by enabling continuous, personalized health monitoring and proactive medical interventions. However, the sensitive nature of health data presents critical privacy and security challenges. This chapter explores the role of Federated Learning (FL) as a privacy-preserving machine learning paradigm that enables collaborative model training across distributed healthcare data sources without compromising patient confidentiality. We examine the synergy between AI, IoT, and FL in applications such as chronic disease management, remote monitoring, and personalized treatment. Key privacy-enhancing techniques like differential privacy, homomorphic encryption, and secure aggregation are discussed alongside real-world use cases and FL architectures. The chapter also addresses infrastructure enablers including edge computing and 5G/6G connectivity, as well as the challenges of scalability, trust, and adversarial threats.
- 5 February 2026
- Published by IGI Global Scientific Publishing
- p. 293-324
Revolutionizing Health Management Through Federated Learning and AI
The integration of Artificial Intelligence (AI) and Internet of Things (IoT) technologies is transforming healthcare by enabling continuous, personalized health monitoring and proactive medical interventions. However, the sensitive nature of health data presents critical privacy and security challenges. This chapter explores the role of Federated Learning (FL) as a privacy-preserving machine learning paradigm that enables collaborative model training across distributed healthcare data sources without compromising patient confidentiality. We examine the synergy between AI, IoT, and FL in applications such as chronic disease management, remote monitoring, and personalized treatment. Key privacy-enhancing techniques like differential privacy, homomorphic encryption, and secure aggregation are discussed alongside real-world use cases and FL architectures. The chapter also addresses infrastructure enablers including edge computing and 5G/6G connectivity, as well as the challenges of scalability, trust, and adversarial threats.
- Open Access
- 2 May 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (3) , 2198-2208
Tri-level lung cancer classification via deep learning based GoogleNet with computed tomography images
Lung cancer (LC) is one of the most prevalent causes of cancer-related death worldwide. World Health Organization (WHO) classifies LC into two broad histological subtypes: non-small cell lung cancer (NSCLC) which is the cause of about 85% of cases and small cell lung cancer (SCLC) which makes up the remaining 15%. Several issues can influence LC detection including poor image quality, insufficient training data, low-quality image characteristics, and poor tumor localization. To overcome these challenges a novel TRI-level LC classification via deep learning-based GoogleNet with computed tomography (CT) images (TRI-LCNet) approach has been proposed for early-stage LC detection using CT images. Initially, the LC-input images CT are collected from openly accessible datasets. The lung CT images have been preprocessed using a Gaussian star filter (GaSF) to decrease noise, followed by feature extraction using GoogleNet. The extracted LC features are then given into a support vector machine (SVM) which is utilized as a classification tool to distinguish between different classes of LC cases. The TRI-LCNet approach performance was assessed by several metrics: specificity, accuracy, F1 score, and recall. The outcomes show that the suggested method obtains a higher accuracy range of 96.93% for the early identification of LC.
- 2 May 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (3) , 2198-2208
Open AccessTri-level lung cancer classification via deep learning based GoogleNet with computed tomography images
Lung cancer (LC) is one of the most prevalent causes of cancer-related death worldwide. World Health Organization (WHO) classifies LC into two broad histological subtypes: non-small cell lung cancer (NSCLC) which is the cause of about 85% of cases and small cell lung cancer (SCLC) which makes up the remaining 15%. Several issues can influence LC detection including poor image quality, insufficient training data, low-quality image characteristics, and poor tumor localization. To overcome these challenges a novel TRI-level LC classification via deep learning-based GoogleNet with computed tomography (CT) images (TRI-LCNet) approach has been proposed for early-stage LC detection using CT images. Initially, the LC-input images CT are collected from openly accessible datasets. The lung CT images have been preprocessed using a Gaussian star filter (GaSF) to decrease noise, followed by feature extraction using GoogleNet. The extracted LC features are then given into a support vector machine (SVM) which is utilized as a classification tool to distinguish between different classes of LC cases. The TRI-LCNet approach performance was assessed by several metrics: specificity, accuracy, F1 score, and recall. The outcomes show that the suggested method obtains a higher accuracy range of 96.93% for the early identification of LC.
- 24 April 2025
- Published by IGI Global Scientific Publishing
- p. 203-230
AI-Driven Preclinical Advances in Nuclear Medicine Radiopharmaceutical Therapy for Prostate Cancer
Radiopharmaceutical therapy represents a growing treatment modality in prostate cancer, and this chapter will focus on how Artificial Intelligence (AI) can be applied to preclinical development of this treatment paradigm. The study, presents a Multi-Modal Attention Enhanced Neural Network (MAENN) to integrate disparate data types—imaging, molecular and pharmacokinetics—to improve predictions of therapy efficacy, determination of optimal dose and the identification of specific biomarkers. Compared to a number of conventional models such as SVMs and CNNs, MAENN maintains the highest accuracy in prediction, sensitivity and data integration capabilities. Although issues like data quality and computational complexity remain, the chapter also describes the scalability, potential for real-time applications, and future applications of MAENN in other cancer therapies, contributing towards personalizing treatment strategies in order to pave the way for precision oncology.
- 24 April 2025
- Published by IGI Global Scientific Publishing
- p. 203-230
AI-Driven Preclinical Advances in Nuclear Medicine Radiopharmaceutical Therapy for Prostate Cancer
Radiopharmaceutical therapy represents a growing treatment modality in prostate cancer, and this chapter will focus on how Artificial Intelligence (AI) can be applied to preclinical development of this treatment paradigm. The study, presents a Multi-Modal Attention Enhanced Neural Network (MAENN) to integrate disparate data types—imaging, molecular and pharmacokinetics—to improve predictions of therapy efficacy, determination of optimal dose and the identification of specific biomarkers. Compared to a number of conventional models such as SVMs and CNNs, MAENN maintains the highest accuracy in prediction, sensitivity and data integration capabilities. Although issues like data quality and computational complexity remain, the chapter also describes the scalability, potential for real-time applications, and future applications of MAENN in other cancer therapies, contributing towards personalizing treatment strategies in order to pave the way for precision oncology.
- 2 December 2024
- Published by IGI Global Scientific Publishing
- p. 213-242
Ensuring Privacy and Security of EHR With Secure Collaborative Transfer Learning (SCTL)
In the ever-evolving landscape of healthcare, preservation of electronic health records (EHRs) has become paramount. The sensitivity of healthcare data, around personal, medical, and financial information, necessitates vigorous privacy and security measures to prevent breaches that could lead to distinctiveness theft, financial loss, and compromised patient care. This chapter discourses the critical challenge of ensuring data privacy and security in healthcare by leveraging advanced generative AI techniques. Specifically, we present and explore Secure Collaborative Transfer Learning (SCTL), a novel algorithm that syndicates the strengths of Transfer Learning and Federated Learning. The key objective is to improve the security and privacy of EHRs while maintaining high model performance and obedience with regulatory standards. The performance assessment of SCTL demonstrates its usefulness in maintaining high levels of data privacy and security without conciliation on model accuracy or operational effectiveness.
- 2 December 2024
- Published by IGI Global Scientific Publishing
- p. 213-242
Ensuring Privacy and Security of EHR With Secure Collaborative Transfer Learning (SCTL)
In the ever-evolving landscape of healthcare, preservation of electronic health records (EHRs) has become paramount. The sensitivity of healthcare data, around personal, medical, and financial information, necessitates vigorous privacy and security measures to prevent breaches that could lead to distinctiveness theft, financial loss, and compromised patient care. This chapter discourses the critical challenge of ensuring data privacy and security in healthcare by leveraging advanced generative AI techniques. Specifically, we present and explore Secure Collaborative Transfer Learning (SCTL), a novel algorithm that syndicates the strengths of Transfer Learning and Federated Learning. The key objective is to improve the security and privacy of EHRs while maintaining high model performance and obedience with regulatory standards. The performance assessment of SCTL demonstrates its usefulness in maintaining high levels of data privacy and security without conciliation on model accuracy or operational effectiveness.
- 1 October 2024
- Published by Taylor & Francis
- p. 145-169
Quantum Information Science
In order to provide novel techniques for processing, storing, and transmitting information, the multidisciplinary field of quantum information science (QIS) integrates ideas from quantum physics [4] with information science. However, due to a number of events, including Peter Shor’s invention of quantum algorithms and the rise in interest in quantum computing, the history of QIS started to expand quickly in the early to middle 1990s. Peter Shor, a physicist and mathematician, developed the first quantum algorithm that could factor big numbers exponentially more rapidly than any classical algorithm in 1994. This finding triggered strong interest in quantum computing’s potential to tackle issues that are currently unsolvable for classical computers. Since then, QIS has advanced due to the development of novel quantum algorithms and enhancements in the hardware used for quantum computing. Understanding complex quantum phenomena like entanglement and superposition and developing novel strategies for utilizing them for information processing and communication are the core goals of QIS. Quantum bits (qubits) can exist in superpositions of 0 and 1 simultaneously, which enables far more efficient computation than classical bits, which can only be in one state or the other at a given time. Superposition is a key concept in quantum computing. In addition to superposition, entanglement, coherence, and quantum algorithms are other important QIS concepts. Entanglement is a key feature of quantum mechanics and has been observed experimentally in a wide range of physical systems, including photons, atoms, and superconducting qubits. Two particles can get entangled with one another through a phenomenon called entanglement in which their quantum states exhibit a high correlation. This indicates that even if two particles are separated by a considerable distance, modifying the state of one can change the state of the other. No matter how far away two separate particles are from one another when they are entangled, measuring the state of one particle immediately gives the state of another particle. It plays a central role in many quantum technologies, including quantum computing for solving complex problems that take classical computers a long time to solve, quantum cryptography for secure communications, quantum sensing for quantum properties of particles to enable more precise measurements, and quantum networking to improve communication and processing speed over both short and long distances. Quantum metrology methods use the principles of quantum mechanics to improve the accuracy of measurements in fields like gravity sensors, timekeeping, and frequency standards. Some of the benefits of QIS are enhanced computing power, improved communication, advanced cryptography, more precise scientific measurements, and better logistics optimization.
- 1 October 2024
- Published by Taylor & Francis
- p. 145-169
Quantum Information Science
In order to provide novel techniques for processing, storing, and transmitting information, the multidisciplinary field of quantum information science (QIS) integrates ideas from quantum physics [4] with information science. However, due to a number of events, including Peter Shor’s invention of quantum algorithms and the rise in interest in quantum computing, the history of QIS started to expand quickly in the early to middle 1990s. Peter Shor, a physicist and mathematician, developed the first quantum algorithm that could factor big numbers exponentially more rapidly than any classical algorithm in 1994. This finding triggered strong interest in quantum computing’s potential to tackle issues that are currently unsolvable for classical computers. Since then, QIS has advanced due to the development of novel quantum algorithms and enhancements in the hardware used for quantum computing. Understanding complex quantum phenomena like entanglement and superposition and developing novel strategies for utilizing them for information processing and communication are the core goals of QIS. Quantum bits (qubits) can exist in superpositions of 0 and 1 simultaneously, which enables far more efficient computation than classical bits, which can only be in one state or the other at a given time. Superposition is a key concept in quantum computing. In addition to superposition, entanglement, coherence, and quantum algorithms are other important QIS concepts. Entanglement is a key feature of quantum mechanics and has been observed experimentally in a wide range of physical systems, including photons, atoms, and superconducting qubits. Two particles can get entangled with one another through a phenomenon called entanglement in which their quantum states exhibit a high correlation. This indicates that even if two particles are separated by a considerable distance, modifying the state of one can change the state of the other. No matter how far away two separate particles are from one another when they are entangled, measuring the state of one particle immediately gives the state of another particle. It plays a central role in many quantum technologies, including quantum computing for solving complex problems that take classical computers a long time to solve, quantum cryptography for secure communications, quantum sensing for quantum properties of particles to enable more precise measurements, and quantum networking to improve communication and processing speed over both short and long distances. Quantum metrology methods use the principles of quantum mechanics to improve the accuracy of measurements in fields like gravity sensors, timekeeping, and frequency standards. Some of the benefits of QIS are enhanced computing power, improved communication, advanced cryptography, more precise scientific measurements, and better logistics optimization.
- 29 September 2024
- Published by Springer Nature in International Journal of Data Science and Analytics
- Vol. 20 (3) , 2851-2870
Secure wireless sensor network transmission using reinforcement learning and homomorphic encryption
- 29 September 2024
- Published by Springer Nature in International Journal of Data Science and Analytics
- Vol. 20 (3) , 2851-2870
Secure wireless sensor network transmission using reinforcement learning and homomorphic encryption
- Open Access
- 19 August 2024
- Published by Springer Nature in Research Square
Revolutionizing Seizure Detection and Monitoring in IoT-Connected Smart Healthcare: Advanced CNN Models and IoT Sensors
In the area of smart healthcare services in IoT-connected smart cities, the integration of cutting-edge CNN models such as DenseNet201, ResNet152V2, and MobileNetV2 with innovative IoT technology for automated seizure detection and monitoring through EEG data analysis. Leveraging a three-layer architecture encompassing device, edge server, and global cloud layers, this study has introduced a novel method to monitor patients' health conditions. The implementation of a cropping training strategy enhances deep learning model efficiency in scenarios with limited data. Notably, the classification accuracy of 2 level CNN models, particularly DenseNet201, reaches remarkable heights, with seizure detection rates ranging from 99.26% to 99.51%, and non-seizure accuracy spanning 99.25% to 99.46%. These models excel in precision for seizure identification, with values between 99.33% and 99.53%, and they exhibit robust recall, particularly 2 level CNN DenseNet201 with 99.24%. The use of Matthews Correlation Coefficient (MCC) further affirms their precision and robustness, with values ranging from approximately 98.53% to 99.19%. This approach, combining advanced CNN models with IoT technology, holds great promise for efficient seizure detection and monitoring, while the assessment of energy consumption and task processing times underlines the importance of model selection and edge server configurations in optimizing system performance.
- 19 August 2024
- Published by Springer Nature in Research Square
Open AccessRevolutionizing Seizure Detection and Monitoring in IoT-Connected Smart Healthcare: Advanced CNN Models and IoT Sensors
In the area of smart healthcare services in IoT-connected smart cities, the integration of cutting-edge CNN models such as DenseNet201, ResNet152V2, and MobileNetV2 with innovative IoT technology for automated seizure detection and monitoring through EEG data analysis. Leveraging a three-layer architecture encompassing device, edge server, and global cloud layers, this study has introduced a novel method to monitor patients' health conditions. The implementation of a cropping training strategy enhances deep learning model efficiency in scenarios with limited data. Notably, the classification accuracy of 2 level CNN models, particularly DenseNet201, reaches remarkable heights, with seizure detection rates ranging from 99.26% to 99.51%, and non-seizure accuracy spanning 99.25% to 99.46%. These models excel in precision for seizure identification, with values between 99.33% and 99.53%, and they exhibit robust recall, particularly 2 level CNN DenseNet201 with 99.24%. The use of Matthews Correlation Coefficient (MCC) further affirms their precision and robustness, with values ranging from approximately 98.53% to 99.19%. This approach, combining advanced CNN models with IoT technology, holds great promise for efficient seizure detection and monitoring, while the assessment of energy consumption and task processing times underlines the importance of model selection and edge server configurations in optimizing system performance.
- 9 July 2024
- Published by Springer Nature in Wireless Personal Communications
- Vol. 137 (2) , 671-684
Customizable Fuzzy-Neuro Inference System Attack Detection Based on Trust for Mobile Wireless Sensor Networks
- 9 July 2024
- Published by Springer Nature in Wireless Personal Communications
- Vol. 137 (2) , 671-684
Customizable Fuzzy-Neuro Inference System Attack Detection Based on Trust for Mobile Wireless Sensor Networks
- 4 January 2023
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 65 (5) , 1963-1988
Selection of browsers for smartphones: a fuzzy hybrid approach and machine learning technique
- 4 January 2023
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 65 (5) , 1963-1988
Selection of browsers for smartphones: a fuzzy hybrid approach and machine learning technique
- 26 January 2021
- Published by Wiley in International Journal of Communication Systems
- Vol. 34 (5) , e4746
Workflow scheduling in cloud environment using a novel metaheuristic optimization algorithm
Summary: Workflow scheduling is the most focused research issue in the on‐demand clouds where the user satisfaction like cost and bandwidth is more difficult. Several research works have been conducted earlier towards performing reliable workflow scheduling with the aim of reducing cost or execution time. However, those works lack to produce better result by compromising any attributes for attaining the goal. The existing work lacks from the security where the tasks might get corrupted during execution. To resolve this problem, an Enhanced Artificial Fish Swarm Algorithm (EAFSA)‐based IaaS Cloud Partial Critical Path (IC‐PCP) Replication and Hyper Elliptic Curve Cryptography (EAFSAIPR with HECC) is proposed. The main goal is to perform better workflow scheduling which can complete the task execution before deadline given by the users. This is done by predicting the early start time and latest finish time using EAFSA algorithm, so that task replication can be made to meet the soft deadline constraint. The task authentication is done efficiently using HEEC algorithm, so that corruption from malicious users can be avoided. The task replication is done securely using the cryptographic algorithm. The proposed EAFSAIPR with HECC algorithm uses idle time of provisioned resources to replicate workflow tasks optimally. The proposed EAFSAIPR algorithm scheduler focused to ensure the lowest cost while serving a deadline set by the user. The experimental results show that the scheduler can find good schedules of deadlines being met and reduces the total execution time of applications as the budget available for replication increases.
- 26 January 2021
- Published by Wiley in International Journal of Communication Systems
- Vol. 34 (5) , e4746
Workflow scheduling in cloud environment using a novel metaheuristic optimization algorithm
Summary: Workflow scheduling is the most focused research issue in the on‐demand clouds where the user satisfaction like cost and bandwidth is more difficult. Several research works have been conducted earlier towards performing reliable workflow scheduling with the aim of reducing cost or execution time. However, those works lack to produce better result by compromising any attributes for attaining the goal. The existing work lacks from the security where the tasks might get corrupted during execution. To resolve this problem, an Enhanced Artificial Fish Swarm Algorithm (EAFSA)‐based IaaS Cloud Partial Critical Path (IC‐PCP) Replication and Hyper Elliptic Curve Cryptography (EAFSAIPR with HECC) is proposed. The main goal is to perform better workflow scheduling which can complete the task execution before deadline given by the users. This is done by predicting the early start time and latest finish time using EAFSA algorithm, so that task replication can be made to meet the soft deadline constraint. The task authentication is done efficiently using HEEC algorithm, so that corruption from malicious users can be avoided. The task replication is done securely using the cryptographic algorithm. The proposed EAFSAIPR with HECC algorithm uses idle time of provisioned resources to replicate workflow tasks optimally. The proposed EAFSAIPR algorithm scheduler focused to ensure the lowest cost while serving a deadline set by the user. The experimental results show that the scheduler can find good schedules of deadlines being met and reduces the total execution time of applications as the budget available for replication increases.