K. Muthamil Sudar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
K. Muthamil Sudar has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Palanigurupackiam Nagaraj
SRM Institute of Science and Technology
6 shared publications - Vaissnave Venkadapathi
SRM Institute of Science and Technology, India
3 shared publications - Chandrakant Sonawane
Symbiosis Centre for Nanoscience and Nanotechnology, Symbiosis International (Deemed University), Pune, India
1 shared publication Choon Kit Chan
Faculty of Engineering and Quantity Surveying, INTI International University, Nilai, Malaysia
1 shared publication- M. Krithik
Mepco Schlenk Engineering College, India
1 shared publication M. Yogeshwaran
Mepco Schlenk Engineering College, Sivakasi, India
1 shared publication
Latest Publications
- 29 July 2026
- Published by IGI Global Scientific Publishing
- p. 287-314
A Deep Learning-Based Framework for Intelligent DDoS Attack Detection
Distributed Denial of Service (DDoS) attacks pose significant threats to the availability and reliability of online services. A deep learning model based on contractive autoencoder (DCAE) is proposed in this paper. The DCAE model is first trained to find the best threshold using which it predicts whether the sample is an attack sample or a normal sample The model has been compared with various other deep learning models based on autoencoder like Long-Short Term Memory (LSTM AE), Undercomplete autoencoders (UCAE) and Basic autoencoders. Three benchmark datasets-NSL-KDD, CIC-IDS2017 and CIC-DDoS2019 have been used for the evaluation of the model. The results shown that the DCAE-based model has outperformed other models in accurately distinguishing the attack and the normal samples. The model is able to distinguish all the attack samples but incorrectly labeled some normal instances which resulted in the decrease of accuracy. The accuracy of the DCAE model ranges from 0.93 to 0.95 for CIC-DDoS2019 datasets whereas 0.95 and 0.93 for NSL-KDD and CIC-IDS2017 respectively.
- 29 July 2026
- Published by IGI Global Scientific Publishing
- p. 287-314
A Deep Learning-Based Framework for Intelligent DDoS Attack Detection
Distributed Denial of Service (DDoS) attacks pose significant threats to the availability and reliability of online services. A deep learning model based on contractive autoencoder (DCAE) is proposed in this paper. The DCAE model is first trained to find the best threshold using which it predicts whether the sample is an attack sample or a normal sample The model has been compared with various other deep learning models based on autoencoder like Long-Short Term Memory (LSTM AE), Undercomplete autoencoders (UCAE) and Basic autoencoders. Three benchmark datasets-NSL-KDD, CIC-IDS2017 and CIC-DDoS2019 have been used for the evaluation of the model. The results shown that the DCAE-based model has outperformed other models in accurately distinguishing the attack and the normal samples. The model is able to distinguish all the attack samples but incorrectly labeled some normal instances which resulted in the decrease of accuracy. The accuracy of the DCAE model ranges from 0.93 to 0.95 for CIC-DDoS2019 datasets whereas 0.95 and 0.93 for NSL-KDD and CIC-IDS2017 respectively.
- 17 July 2026
- Published by IGI Global Scientific Publishing
- p. 259-288
Securing Darknet Web Applications
SQL injection attacks continue to pose a critical threat to web applications since the cybersecurity environment constantly changes, so security requires effective detection and prevention mechanisms. A new second-order SQL injection detection system works using proxy-based steps to improve security for advanced injection techniques. t. Our system adopts two analysis strategies through real-time dynamic execution that mimics live attacks while recording likely source code errors through passive static analysis. The system generates output depending on the given SQL injection points to emulate real-time attack traces efficiently. Integrating static and dynamic execution offers comprehensive system protection through complete control of source code vulnerabilities across all periods of execution. In-transit analysis in the proxy server checks all incoming database queries for anomalies and stops suspicious queries from reaching the database.
- 17 July 2026
- Published by IGI Global Scientific Publishing
- p. 259-288
Securing Darknet Web Applications
SQL injection attacks continue to pose a critical threat to web applications since the cybersecurity environment constantly changes, so security requires effective detection and prevention mechanisms. A new second-order SQL injection detection system works using proxy-based steps to improve security for advanced injection techniques. t. Our system adopts two analysis strategies through real-time dynamic execution that mimics live attacks while recording likely source code errors through passive static analysis. The system generates output depending on the given SQL injection points to emulate real-time attack traces efficiently. Integrating static and dynamic execution offers comprehensive system protection through complete control of source code vulnerabilities across all periods of execution. In-transit analysis in the proxy server checks all incoming database queries for anomalies and stops suspicious queries from reaching the database.
- 10 July 2026
- Published by IGI Global Scientific Publishing
- p. 27-52
Integration of Research Into Undergraduate Curriculum
When considering the integration of research into undergraduate programs, a pedagogical approach can help students go beyond mere technical competencies and towards a holistic learning experience. Coming from Kolb's Experiential Learning Cycle, learning is a construction process where knowledge is born out of real-life experiences. Concrete experiences come from the fieldwork, and thoughtful self-reflection enables them to critique what they have found and encountered. Bloom's Taxonomy is a hierarchical framework for sharpening critical thinking skills and mental abilities. By applying that, teachers can fine-tune research opportunities that progressively push their students. University professors can be role models, instructors and leaders of research projects, and show their students the ropes of the research process. Successful mentorship is all about being supportive, allowing the students to figure things out for themselves. With respect to engineering programs, the use of research-led curricula showcases a very effective blend of theoretical concepts and mentoring.
- 10 July 2026
- Published by IGI Global Scientific Publishing
- p. 27-52
Integration of Research Into Undergraduate Curriculum
When considering the integration of research into undergraduate programs, a pedagogical approach can help students go beyond mere technical competencies and towards a holistic learning experience. Coming from Kolb's Experiential Learning Cycle, learning is a construction process where knowledge is born out of real-life experiences. Concrete experiences come from the fieldwork, and thoughtful self-reflection enables them to critique what they have found and encountered. Bloom's Taxonomy is a hierarchical framework for sharpening critical thinking skills and mental abilities. By applying that, teachers can fine-tune research opportunities that progressively push their students. University professors can be role models, instructors and leaders of research projects, and show their students the ropes of the research process. Successful mentorship is all about being supportive, allowing the students to figure things out for themselves. With respect to engineering programs, the use of research-led curricula showcases a very effective blend of theoretical concepts and mentoring.
- 19 June 2026
- Published by IGI Global Scientific Publishing
- p. 1-26
Securing Metaverse Platforms Using Machine Learning-Based Intrusion Detection Systems
The increasing number of cyber threats requires more sophisticated and intelligent security solutions than the conventional signature-based Intrusion Detection Systems (IDS). Against the background of growing big data, high-speed networks, and dynamically changing attack vectors, Machine Learning (ML) has become a revolutionary technology to augment IDS by allowing systems to learn from past experiences, recognize patterns, and detect previously unseen attacks in real-time. This chapter discusses the application of Machine Learning methods to the development and design of contemporary IDS. When it comes to intrusion detection, ML algorithms such as decision trees, support vector machines, neural networks, k-nearest neighbors, and ensemble methods are really useful and help to identify known and unknown threats. The chapter on intrusion detection takes a look at the different types of learning, supervised, unsupervised, and semi-supervised, and in terms of datasets like NSL-KDD.
- 19 June 2026
- Published by IGI Global Scientific Publishing
- p. 1-26
Securing Metaverse Platforms Using Machine Learning-Based Intrusion Detection Systems
The increasing number of cyber threats requires more sophisticated and intelligent security solutions than the conventional signature-based Intrusion Detection Systems (IDS). Against the background of growing big data, high-speed networks, and dynamically changing attack vectors, Machine Learning (ML) has become a revolutionary technology to augment IDS by allowing systems to learn from past experiences, recognize patterns, and detect previously unseen attacks in real-time. This chapter discusses the application of Machine Learning methods to the development and design of contemporary IDS. When it comes to intrusion detection, ML algorithms such as decision trees, support vector machines, neural networks, k-nearest neighbors, and ensemble methods are really useful and help to identify known and unknown threats. The chapter on intrusion detection takes a look at the different types of learning, supervised, unsupervised, and semi-supervised, and in terms of datasets like NSL-KDD.
- Open Access
- 14 May 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Deep learning and IoT-based framework for sesame plant identification and weed detection
Automated weed identification in agriculture is a prominent and promising technological intervention as it improves the overall crop productivity and yield. But finding weeds in dense sesame fields is still hard because of occlusion and changes in lighting. This work suggests a hybrid deep learning model that works with IoT to solve these problems. The method uses a Region-based Convolutional Neural Network (R-CNN) model to find sesame plants and weeds. Approximately 1,300 images with YOLO-formatted annotations were used for training and testing. This technique includes making region recommendations, preprocessing images, utilizing a CNN architecture to classify objects, training an SVM, and leveraging CNN features to improve item bounding boxes. The model’s structure is made up of three parts: CNN fine-tuning, CNN + SVM, and bounding box regression. The process of preparing the data included finding and sorting regions of interest. Subsequently, these areas are sorted into groups based on how they overlap with real-world items. After that, the features that were found were used to make accurate bounding boxes around the items that were found, which led to the best classification. The model in question achieved 99.73% classification accuracy with a maximum bounding box IoU accuracy of 0.787 during the forecast procedure. The methodology delineated in this study possesses the potential for application in agricultural sectors for weed removal in farming fields.
- 14 May 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Open AccessDeep learning and IoT-based framework for sesame plant identification and weed detection
Automated weed identification in agriculture is a prominent and promising technological intervention as it improves the overall crop productivity and yield. But finding weeds in dense sesame fields is still hard because of occlusion and changes in lighting. This work suggests a hybrid deep learning model that works with IoT to solve these problems. The method uses a Region-based Convolutional Neural Network (R-CNN) model to find sesame plants and weeds. Approximately 1,300 images with YOLO-formatted annotations were used for training and testing. This technique includes making region recommendations, preprocessing images, utilizing a CNN architecture to classify objects, training an SVM, and leveraging CNN features to improve item bounding boxes. The model’s structure is made up of three parts: CNN fine-tuning, CNN + SVM, and bounding box regression. The process of preparing the data included finding and sorting regions of interest. Subsequently, these areas are sorted into groups based on how they overlap with real-world items. After that, the features that were found were used to make accurate bounding boxes around the items that were found, which led to the best classification. The model in question achieved 99.73% classification accuracy with a maximum bounding box IoU accuracy of 0.787 during the forecast procedure. The methodology delineated in this study possesses the potential for application in agricultural sectors for weed removal in farming fields.
- 25 September 2025
- Published by IGI Global Scientific Publishing
- p. 73-98
Ethical and Social Implications in Quantum-Inspired Soft Computing
Quantum technology has resulted in the emergence of multiple opportunities and difficulties, not only in the realm of computation but also in the complex relationship that occurs between ethics and security. These chances and challenges have been brought about as a result of developing quantum computing (QC). The study also includes a discussion of the implications of these technologies. Additionally, it analyzes the challenges that are involved with the security of the quantum landscape, particularly with regard to quantum communication and cryptography. This is specifically addressed in the context of the quantum landscape. To achieve the synthesis of ethics and security in the quantum world, it is necessary to strike a careful balance between the two perspectives. In order to achieve this equilibrium, responsible research, ethical deployment, and efficient security systems should be given priority. Grasp the ethical and security implications of quantum technologies is becoming increasingly important as these technologies continue to disrupt the digital environment.
- 25 September 2025
- Published by IGI Global Scientific Publishing
- p. 73-98
Ethical and Social Implications in Quantum-Inspired Soft Computing
Quantum technology has resulted in the emergence of multiple opportunities and difficulties, not only in the realm of computation but also in the complex relationship that occurs between ethics and security. These chances and challenges have been brought about as a result of developing quantum computing (QC). The study also includes a discussion of the implications of these technologies. Additionally, it analyzes the challenges that are involved with the security of the quantum landscape, particularly with regard to quantum communication and cryptography. This is specifically addressed in the context of the quantum landscape. To achieve the synthesis of ethics and security in the quantum world, it is necessary to strike a careful balance between the two perspectives. In order to achieve this equilibrium, responsible research, ethical deployment, and efficient security systems should be given priority. Grasp the ethical and security implications of quantum technologies is becoming increasingly important as these technologies continue to disrupt the digital environment.
- 12 August 2025
- Published by IGI Global Scientific Publishing
- p. 195-216
Integrating Blockchain With Watermarking Systems for Tamper-Proof Attribution in Digital Media
The emergence of digital content has created an instant need for ways of authenticity, ownership, and protection of intellectual property. This chapter elaborates the merging of blockchain with digital watermarking systems for credible attribution of tamper-proof images, videos, and documents. One can add watermark to a digital file. However, a copy of the ownership and copyright information, which is collected through uploading to blockchain's non-deletable and de-centralized ledger, is stored and comes with the entire assurance that the data is safely verifiable. The watermarking methods for use along with blockchain in this system include Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT). Watermark metadata, including cryptographic hash values, timestamps, and the details of the creator, are stored on the blockchain, making it possible to authenticate content easily and to derive its chain of evidence. Smart contracts will take complete charge of operations like licensing for content and royalties payments to ensure clear compliance with usage conversions.
- 12 August 2025
- Published by IGI Global Scientific Publishing
- p. 195-216
Integrating Blockchain With Watermarking Systems for Tamper-Proof Attribution in Digital Media
The emergence of digital content has created an instant need for ways of authenticity, ownership, and protection of intellectual property. This chapter elaborates the merging of blockchain with digital watermarking systems for credible attribution of tamper-proof images, videos, and documents. One can add watermark to a digital file. However, a copy of the ownership and copyright information, which is collected through uploading to blockchain's non-deletable and de-centralized ledger, is stored and comes with the entire assurance that the data is safely verifiable. The watermarking methods for use along with blockchain in this system include Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT). Watermark metadata, including cryptographic hash values, timestamps, and the details of the creator, are stored on the blockchain, making it possible to authenticate content easily and to derive its chain of evidence. Smart contracts will take complete charge of operations like licensing for content and royalties payments to ensure clear compliance with usage conversions.
- 24 June 2025
- Published by IGI Global Scientific Publishing
- p. 337-356
Enhanced DDoS Detection Using Feature Selection and Multi-Model Machine Learning Frameworks
Distributed Denial of Service (DDoS) attacks are among the most pervasive and disruptive threats in today's interconnected digital landscape, capable of overwhelming systems and rendering services unavailable. The effective and timely detection of DDoS attacks has therefore become critical for maintaining the security and availability of networked systems. This chapter presents a holistic approach to DDoS detection based on machine learning techniques, improved by feature selection and the use of multiple predictive models The system uses a diverse ensemble of machine learning models: Random Forest, Support Vector Machine, Gradient Boosting, and Neural Networks to effectively learn the complex patterns typical of DDoS traffic This chapter highlights the potential of integrating feature selection with multi-model machine learning frameworks for efficient and scalable DDoS detection.The proposed framework will be able to counter the evolving nature of DDoS attacks, hence offering resilience and reliability in ever-increasingly complex network environments.
- 24 June 2025
- Published by IGI Global Scientific Publishing
- p. 337-356
Enhanced DDoS Detection Using Feature Selection and Multi-Model Machine Learning Frameworks
Distributed Denial of Service (DDoS) attacks are among the most pervasive and disruptive threats in today's interconnected digital landscape, capable of overwhelming systems and rendering services unavailable. The effective and timely detection of DDoS attacks has therefore become critical for maintaining the security and availability of networked systems. This chapter presents a holistic approach to DDoS detection based on machine learning techniques, improved by feature selection and the use of multiple predictive models The system uses a diverse ensemble of machine learning models: Random Forest, Support Vector Machine, Gradient Boosting, and Neural Networks to effectively learn the complex patterns typical of DDoS traffic This chapter highlights the potential of integrating feature selection with multi-model machine learning frameworks for efficient and scalable DDoS detection.The proposed framework will be able to counter the evolving nature of DDoS attacks, hence offering resilience and reliability in ever-increasingly complex network environments.
- 28 February 2025
- Published by IGI Global Scientific Publishing
- p. 325-340
Application of Machine Unlearning Techniques to Enhance the Performance and Adaptability of DDoS Attack Detection Models
Traditional detection models may be accurate at first, but they quickly lose effectiveness as patterns of attacks change. They therefore require frequent and resource-intensive retraining. To achieve this, the authors have adopted several machine unlearning techniques, such as incremental unlearning, selective forgetting, recurrent unlearning, and adversarial unlearning, which allow a model to forget outdated information dynamically and learn from the new relevant traffic patterns. They assessed the performance of the models in terms of accuracy, precision, recall, F1-score, AUC-ROC, Matthews correlation coefficient (MCC), and adaptability measures such as the unlearning rate and time taken for retraining. The results show that detection accuracy was better in the models using machine unlearning, from 95.6% to 97.2%, with fewer false positives and negatives. A notable improvement in AUC-ROC scores was observed, from 0.89 to 0.94, and a 35% reduction in retraining time, which indicates the model's efficiency in adapting to changing threats. Moreover, the adversarial unlearning integration improved the model's robustness against evasive attack tactics by 30%. The hybrid approach, combining machine unlearning with ensemble methods, further elevated performance metrics to an AUC-ROC of 0.96 and balanced precision and recall rates over 95%. This chapter highlights the potential of machine unlearning in creating more robust, efficient, and adaptive DDoS detection systems, which have tremendous implications for real-time cybersecurity defense in dynamic network environments.
- 28 February 2025
- Published by IGI Global Scientific Publishing
- p. 325-340
Application of Machine Unlearning Techniques to Enhance the Performance and Adaptability of DDoS Attack Detection Models
Traditional detection models may be accurate at first, but they quickly lose effectiveness as patterns of attacks change. They therefore require frequent and resource-intensive retraining. To achieve this, the authors have adopted several machine unlearning techniques, such as incremental unlearning, selective forgetting, recurrent unlearning, and adversarial unlearning, which allow a model to forget outdated information dynamically and learn from the new relevant traffic patterns. They assessed the performance of the models in terms of accuracy, precision, recall, F1-score, AUC-ROC, Matthews correlation coefficient (MCC), and adaptability measures such as the unlearning rate and time taken for retraining. The results show that detection accuracy was better in the models using machine unlearning, from 95.6% to 97.2%, with fewer false positives and negatives. A notable improvement in AUC-ROC scores was observed, from 0.89 to 0.94, and a 35% reduction in retraining time, which indicates the model's efficiency in adapting to changing threats. Moreover, the adversarial unlearning integration improved the model's robustness against evasive attack tactics by 30%. The hybrid approach, combining machine unlearning with ensemble methods, further elevated performance metrics to an AUC-ROC of 0.96 and balanced precision and recall rates over 95%. This chapter highlights the potential of machine unlearning in creating more robust, efficient, and adaptive DDoS detection systems, which have tremendous implications for real-time cybersecurity defense in dynamic network environments.
- 17 February 2023
- Published by IGI Global Scientific Publishing
- p. 19-36
Detection of DDoS Attack Using Machine Learning Techniques in Software Defined Networking
The software-defined network (SDN) has recently developed as a network paradigm due to its high network programmability and flexibility, which can overcome the difficulty in traditional networks by detaching the control plane from the data plane. Following the controller's decision in the control plane, the data plane will transfer the packets. This unified administration will make it possible to see the network architecture as a whole in an abstract way. The controller is exposed to a significant threat if control is lost due to its centralised structure. Data plane resources are attacked by the attacker by focusing on switches that support OpenFlow. DDoS attacks damage network performance by overloading the SDN controller and network links, depleting the victim's bandwidth, and flooding the server with massive amounts of data. To address this issue, the authors employ statistical-based and machine learning-based techniques in SDN controller to inspect the new incoming flows.
- 17 February 2023
- Published by IGI Global Scientific Publishing
- p. 19-36
Detection of DDoS Attack Using Machine Learning Techniques in Software Defined Networking
The software-defined network (SDN) has recently developed as a network paradigm due to its high network programmability and flexibility, which can overcome the difficulty in traditional networks by detaching the control plane from the data plane. Following the controller's decision in the control plane, the data plane will transfer the packets. This unified administration will make it possible to see the network architecture as a whole in an abstract way. The controller is exposed to a significant threat if control is lost due to its centralised structure. Data plane resources are attacked by the attacker by focusing on switches that support OpenFlow. DDoS attacks damage network performance by overloading the SDN controller and network links, depleting the victim's bandwidth, and flooding the server with massive amounts of data. To address this issue, the authors employ statistical-based and machine learning-based techniques in SDN controller to inspect the new incoming flows.