Selvaraj Mercy Shalinie
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Selvaraj Mercy Shalinie has a total of 152 co-authors. Here are the top co-authors by number of shared publications:
K. Sundarakantham
Department of Computer Engineering, Thia-garajar College of Engineering
14 shared publications- Murugesan Vijayalakshmi
Thiyagarajar College of Engineering
10 shared publications - K. Sundarakantham
Thiagarajar College of Engineering
8 shared publications - Karthick Seshadri
National Institute of Technology Andhra Pradesh
8 shared publications - Nagarathna Ravi
Department of computer science and engineering, Thiagarajar college of engineering, Madurai, India
7 shared publications T. Subbulakshmi
Thiagarajar College of Engineering
6 shared publications
Latest Publications
- Open Access
HQMSGAN: Enhancing Adversarial Robustness Using Hybrid Quantum-Classical Multi-Scale Generative Adversarial Network
The growing vulnerability of computer vision models to adversarial attacks has raised great concerns about the reliability of generative models in security-sensitive applications. Quantum generative models show strong potential for learning complex data distributions. However, the robustness of quantum generative models to adversarial perturbations and the effectiveness of adversarial training using quantum-classical generative architectures remain less explored. In this work, we evaluate quantum and classical generative network-inspired adversarial training to analyse whether quantum state-space representations can enhance adversarial robustness compared with their classical counterparts. We propose HQMSGAN, a novel hybrid quantum-classical generative adversarial network with multiscaling discriminators that denoise the adversarial attacks through adversarial training. During adversarial training, the proposed model performs quantum projections to learn rich nonlinear feature representations using quantum state encoding and performs feature coupling using quantum entanglement in the quantum state space. Thus, it learns the probability distribution of both the clean and adversarially perturbed data and tries to generate pure samples by reducing the loss over iterations in the quantum realm. The experiment is carried out by inducing four different adversarial attacks and defending them by adversarial training using the resources provided by AWS Braket’s quantum simulators in the Noisy Intermediate Scale Quantum (NISQ) era. The results provide favorable performance on metrics such as Structural Similarity Index Measure (SSIM), Peak Signal-to-Noise Ratio (PSNR), Fréchet Inception Distance (FID) score, Learned Perceptual Image Patch Similarity (LPIPS), and classification accuracy after performing adversarial training.
Open AccessHQMSGAN: Enhancing Adversarial Robustness Using Hybrid Quantum-Classical Multi-Scale Generative Adversarial Network
The growing vulnerability of computer vision models to adversarial attacks has raised great concerns about the reliability of generative models in security-sensitive applications. Quantum generative models show strong potential for learning complex data distributions. However, the robustness of quantum generative models to adversarial perturbations and the effectiveness of adversarial training using quantum-classical generative architectures remain less explored. In this work, we evaluate quantum and classical generative network-inspired adversarial training to analyse whether quantum state-space representations can enhance adversarial robustness compared with their classical counterparts. We propose HQMSGAN, a novel hybrid quantum-classical generative adversarial network with multiscaling discriminators that denoise the adversarial attacks through adversarial training. During adversarial training, the proposed model performs quantum projections to learn rich nonlinear feature representations using quantum state encoding and performs feature coupling using quantum entanglement in the quantum state space. Thus, it learns the probability distribution of both the clean and adversarially perturbed data and tries to generate pure samples by reducing the loss over iterations in the quantum realm. The experiment is carried out by inducing four different adversarial attacks and defending them by adversarial training using the resources provided by AWS Braket’s quantum simulators in the Noisy Intermediate Scale Quantum (NISQ) era. The results provide favorable performance on metrics such as Structural Similarity Index Measure (SSIM), Peak Signal-to-Noise Ratio (PSNR), Fréchet Inception Distance (FID) score, Learned Perceptual Image Patch Similarity (LPIPS), and classification accuracy after performing adversarial training.
- 2 September 2026
- Published by Springer Nature
- p. 332-343
Ano-HQNN: Hybrid Quantum Neural Network for Enhanced Anomaly Detection in Surveillance Videos
- 2 September 2026
- Published by Springer Nature
- p. 332-343
Ano-HQNN: Hybrid Quantum Neural Network for Enhanced Anomaly Detection in Surveillance Videos
- 13 January 2026
- Published by Springer Nature
- p. 517-530
Harnessing Quantum Fully Connected Network for Enhancing Object Detection
- 13 January 2026
- Published by Springer Nature
- p. 517-530
Harnessing Quantum Fully Connected Network for Enhancing Object Detection
- 2 October 2025
- Published by Springer Nature
- p. 226-236
CDAC: Collaborative Data Augmentation for the Classification of Chest CT Image
- 2 October 2025
- Published by Springer Nature
- p. 226-236
CDAC: Collaborative Data Augmentation for the Classification of Chest CT Image
- Open Access
- 1 January 2025
- Published by University of Zielona Góra, Poland in International Journal of Applied Mathematics and Computer Science
- Vol. 35 (3) , 535-545
CluM: A clustering-cum-Markov model for resource prediction in a data center
High-end data centers are required to process the user requests and provide them with a better quality of...
- 1 January 2025
- Published by University of Zielona Góra, Poland in International Journal of Applied Mathematics and Computer Science
- Vol. 35 (3) , 535-545
Open AccessCluM: A clustering-cum-Markov model for resource prediction in a data center
High-end data centers are required to process the user requests and provide them with a better quality of...
- 23 October 2024
- Published by Springer Nature
- p. 203-214
Heart Disease Prediction Using Bagging QSVC Algorithm
- 23 October 2024
- Published by Springer Nature
- p. 203-214
Heart Disease Prediction Using Bagging QSVC Algorithm
- Open Access
- 4 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 371-376
SDN Defense: Detection and mitigation of DDoS attack via IoT Network
Internet-of-things is one of the prominent communication technologies in the 21st century. We can connect everyday objects, like baby monitors, thermostats, e-health, etc., Connecting IoT with Software-defined networking is the best approach to security provisioning during network communication. Because the enormous features of SDN are programmable and centralized management, attackers can create multiple security vulnerabilities in SDN that redeem the distributed denial-of-service attack. This security breach causes bandwidth depletion and server resource impoverishment and perplexes benign users. This study proposes and builds a DDoS attack detection and mitigation defense system for SDN to address this issue. Two different defense modules are deployed in the SDN controller: suspicious identification, and mitigation of the malicious traffic flow. The first module of the SDN defense system used for detecting malignant traffic from DDoS attacks which works under a CNN-ELM is an integrated deep learning approach that combines a convolutional neural network with an extreme learning machine. The suspicious flows are identified, whether benign or malignant, by using a hybrid CCN-ELM model, and this process improves the accuracy of the attack detection. The second module of the mitigation strategy identifies the attacker's location by using IP traceback and removes that malicious traffic by transmitting the flow rule from the controller. These two SDN defense modules are evaluated by simulation processes. Finally, the experimental results of the SDN defense system, precisely detect the DDoS attack with an accuracy of 99.85% and efficiently mitigate the malicious traffic flow in real-time.
- 4 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 371-376
Open AccessSDN Defense: Detection and mitigation of DDoS attack via IoT Network
Internet-of-things is one of the prominent communication technologies in the 21st century. We can connect everyday objects, like baby monitors, thermostats, e-health, etc., Connecting IoT with Software-defined networking is the best approach to security provisioning during network communication. Because the enormous features of SDN are programmable and centralized management, attackers can create multiple security vulnerabilities in SDN that redeem the distributed denial-of-service attack. This security breach causes bandwidth depletion and server resource impoverishment and perplexes benign users. This study proposes and builds a DDoS attack detection and mitigation defense system for SDN to address this issue. Two different defense modules are deployed in the SDN controller: suspicious identification, and mitigation of the malicious traffic flow. The first module of the SDN defense system used for detecting malignant traffic from DDoS attacks which works under a CNN-ELM is an integrated deep learning approach that combines a convolutional neural network with an extreme learning machine. The suspicious flows are identified, whether benign or malignant, by using a hybrid CCN-ELM model, and this process improves the accuracy of the attack detection. The second module of the mitigation strategy identifies the attacker's location by using IP traceback and removes that malicious traffic by transmitting the flow rule from the controller. These two SDN defense modules are evaluated by simulation processes. Finally, the experimental results of the SDN defense system, precisely detect the DDoS attack with an accuracy of 99.85% and efficiently mitigate the malicious traffic flow in real-time.
- 30 November 2023
- Published by Springer Nature
- p. 395-402
A Partitioned Task Offloading Approach for Privacy Preservation at Edge
- 30 November 2023
- Published by Springer Nature
- p. 395-402
A Partitioned Task Offloading Approach for Privacy Preservation at Edge
- 26 October 2023
- Published by Springer Nature
- p. 3-16
A Novel Hierarchical Clustering Technique to Analyze Style and Content Factorization During Image Recognition
- 26 October 2023
- Published by Springer Nature
- p. 3-16
A Novel Hierarchical Clustering Technique to Analyze Style and Content Factorization During Image Recognition
- Open Access
- 13 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 17 (3) , 3611-3622
An Efficient and Secure Method for Simultaneous Ownership Transfer of Multiple Mobile Readers
In recent years, radio frequency identification technology has developed rapidly and has been applied in supply chains. Products in supply chains are varied and may belong to several different owners. Ownership transfer requires the consent of a majority of owners. This article proposed a method suitable for simultaneously transferring the ownership of a large number of goods. The method does not require a trusted third party and can securely and efficiently transfer group ownership for multiple owners with multiple tags. The old and new owner groups can be classified as different authorities. This method can be used in mobile readers to transfer one, some, or all tags in a group. It includes mutual authentication between tags, readers, and backend servers, and can ensure that only assigned owners can obtain the tag ownership. The method proved can resist most known attacks, such as secret disclosure attacks and replay attacks. It can also prevent attacks from the dishonest original owners. Finally, through experiments, we compared the proposed method with other many-to-many ownership transfer methods and demonstrated that the proposed method has better security and fewer transmitted messages than other methods.
- 13 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 17 (3) , 3611-3622
Open AccessAn Efficient and Secure Method for Simultaneous Ownership Transfer of Multiple Mobile Readers
In recent years, radio frequency identification technology has developed rapidly and has been applied in supply chains. Products in supply chains are varied and may belong to several different owners. Ownership transfer requires the consent of a majority of owners. This article proposed a method suitable for simultaneously transferring the ownership of a large number of goods. The method does not require a trusted third party and can securely and efficiently transfer group ownership for multiple owners with multiple tags. The old and new owner groups can be classified as different authorities. This method can be used in mobile readers to transfer one, some, or all tags in a group. It includes mutual authentication between tags, readers, and backend servers, and can ensure that only assigned owners can obtain the tag ownership. The method proved can resist most known attacks, such as secret disclosure attacks and replay attacks. It can also prevent attacks from the dishonest original owners. Finally, through experiments, we compared the proposed method with other many-to-many ownership transfer methods and demonstrated that the proposed method has better security and fewer transmitted messages than other methods.