Chandrasekar Venkatachalam
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Chandrasekar Venkatachalam has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
A. M. Viswa Bharathy
Jyothishmathi Institute of Technology and Science
2 shared publications- Narender Chinthamu
CEO MahaaAi Group of Companies and International Labs, Dallas Texas, USA
2 shared publications S. Suresh Kumar
2 shared publications- Shanmugavalli Venkatachalam
Department of Computer Science and Engineering, KSR College of Engineering, Tiruchengode, India
2 shared publications - Shanmugavalli Venkatachalam
Manipal Institute of Technology Bengaluru, Manipal Academy of Higher Education, Manipal, India
2 shared publications A. Manikandan
SRM Institute of Science and Technology (Deemed to be University)
1 shared publication
Latest Publications
- Open Access
- 24 April 2026
- Published by BR Publications
- Vol. 1 (01) , 19-26
Article Title
One common consequence of diabetes mellitus is DR, which destroys the retina and can cause permanent blindness if not caught in time. Treatment cannot undo DR, but catching it early and taking steps to prevent further vision loss can help keep eyesight intact. The necessity for trustworthy automated methods is highlighted by the fact that conventional diagnosis, which involves ophthalmologists utilising retinal fundus pictures, is time-consuming, expensive, and susceptible to human mistake. In this research, this survey current deep learning-based techniques for diabetic retinopathy detection and present DualAttTrans, a full-stack network that uses hybrid attention mechanisms to efficiently capture and analyse retinal image features on different scales and domains, without using additional priors or supervision. In order to get representative features for classification, the methodology begins with preprocessing to clean the data. Then, features are extracted. This is built a large-scale benchmark dataset with various retinal pictures at different DR severity levels to validate performance. The experimental results show that DualAttTrans outperforms the state-of-the-art methods with an impressive accuracy of 94.32%. The results show that DualAttTrans is an effective, efficient, and scalable technique for automated diabetic retinopathy detection; it can aid in early screening and protect at-risk patients' eyes from damage. https://doi.org/10.65470/james.2
- 24 April 2026
- Published by BR Publications
- Vol. 1 (01) , 19-26
Open AccessArticle Title
One common consequence of diabetes mellitus is DR, which destroys the retina and can cause permanent blindness if not caught in time. Treatment cannot undo DR, but catching it early and taking steps to prevent further vision loss can help keep eyesight intact. The necessity for trustworthy automated methods is highlighted by the fact that conventional diagnosis, which involves ophthalmologists utilising retinal fundus pictures, is time-consuming, expensive, and susceptible to human mistake. In this research, this survey current deep learning-based techniques for diabetic retinopathy detection and present DualAttTrans, a full-stack network that uses hybrid attention mechanisms to efficiently capture and analyse retinal image features on different scales and domains, without using additional priors or supervision. In order to get representative features for classification, the methodology begins with preprocessing to clean the data. Then, features are extracted. This is built a large-scale benchmark dataset with various retinal pictures at different DR severity levels to validate performance. The experimental results show that DualAttTrans outperforms the state-of-the-art methods with an impressive accuracy of 94.32%. The results show that DualAttTrans is an effective, efficient, and scalable technique for automated diabetic retinopathy detection; it can aid in early screening and protect at-risk patients' eyes from damage. https://doi.org/10.65470/james.2
- 23 November 2025
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 37 (27-28)
A Chaotic Model‐Based Dynamic Resource Allocation and Adaptive Task Offloading With Lightweight Encryption in Multi‐Access Edge Computing
Multi‐Access Edge Computing (MEC) provides a lot of computing with low latency within the nearby portion of the network. Nevertheless, current MEC frameworks have considerable resource management and offloading issues when there are dynamic and uncertain tasks. These are poor flexibility to live with time changes, heavy backhaul traffic during data communication, the use of a static approach of encryption, and a lack of integration with chaotic models on resource achievement. Due to the complexity of encrypting all resources, the developed framework employs a novel MEC paradigm of integrating the chaos modeling of systems to both adaptively distribute resources and dynamically change the task offloading as well as lightweight chaos‐based encryption. In this study, Multi‐Access Edge Computing (MEC) is used for fast task handling near the end user. The present MEC models still face resource and offloading problems when tasks are dynamic or uncertain. In this paper, a new chaotic model‐based method is developed for resource allocation and adaptive task offloading with lightweight encryption. The system applies chaotic sequence features such as initial sensitivity and randomness to decide task distribution among different servers. This helps in balancing the load, reducing delay, and lowering energy use. A secondary chaotic key generator is used to secure data transfer without heavy computation. Simulation results show that the proposed model achieves an average latency of 0.37 s and task processing time of 0.21 s, which are better than the standard MEC methods. Overall, the chaotic idea gives flexibility and safety for real‐time MEC operation.
- 23 November 2025
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 37 (27-28)
A Chaotic Model‐Based Dynamic Resource Allocation and Adaptive Task Offloading With Lightweight Encryption in Multi‐Access Edge Computing
Multi‐Access Edge Computing (MEC) provides a lot of computing with low latency within the nearby portion of the network. Nevertheless, current MEC frameworks have considerable resource management and offloading issues when there are dynamic and uncertain tasks. These are poor flexibility to live with time changes, heavy backhaul traffic during data communication, the use of a static approach of encryption, and a lack of integration with chaotic models on resource achievement. Due to the complexity of encrypting all resources, the developed framework employs a novel MEC paradigm of integrating the chaos modeling of systems to both adaptively distribute resources and dynamically change the task offloading as well as lightweight chaos‐based encryption. In this study, Multi‐Access Edge Computing (MEC) is used for fast task handling near the end user. The present MEC models still face resource and offloading problems when tasks are dynamic or uncertain. In this paper, a new chaotic model‐based method is developed for resource allocation and adaptive task offloading with lightweight encryption. The system applies chaotic sequence features such as initial sensitivity and randomness to decide task distribution among different servers. This helps in balancing the load, reducing delay, and lowering energy use. A secondary chaotic key generator is used to secure data transfer without heavy computation. Simulation results show that the proposed model achieves an average latency of 0.37 s and task processing time of 0.21 s, which are better than the standard MEC methods. Overall, the chaotic idea gives flexibility and safety for real‐time MEC operation.
- 27 October 2025
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 198 (Pt B) , 111229
Integrating optimal prediction model with hybrid deep learning for coated metal oxide nanoparticle drug delivery systems
- 27 October 2025
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 198 (Pt B) , 111229
Integrating optimal prediction model with hybrid deep learning for coated metal oxide nanoparticle drug delivery systems
- 1 August 2025
- Published by The Electrochemical Society in ECS Journal of Solid State Science and Technology
- Vol. 14 (8) , 081001
Q-Learning Driven Metaheuristic for Tailoring the Solid State Properties of Carbon Nanotube-Graphene Nanocomposites for Automotive and Aerospace Applications
The implementation of deep learning strategies has transformed composite material design and allowed the optimization of carbon fibre with epoxy resin beyond previous limits for greater performance results. Developing dependable structure-property-performance links continues to present a major obstacle because these systems operate through intricate combinations between nanotech advances and manufacturing simulations. This study presents a multi-faceted methodology employing a reinforcement learning (RL) approach to optimize carbon fiber and epoxy resin-based composites, incorporating over 10,000 samples of carbon nanotube, graphene, and polymer matrix combinations, thus aiming to achieve specific performance criteria such as tensile strength and thermal conductivity. Utilizing Langmuir-Blodgett self-assembly and Electrospinning methodologies, the research integrates novel nanotechnology to fabricate and test the designed composites. Combining RL with advanced manufacturing and finite element analysis simulation techniques provides a holistic view of the structure-property-performance relations in composite materials. Through iterative optimization, the expected outcomes indicate a potential improvement in mechanical properties by 15%–25% and thermal performance by 10%–20%. Simulated thermal expansion coefficient distributions show that the proposed model consistently achieves values in the range of 1.8 to 3.0 in terms of frequency and reliability. The outcomes of this research have significant implications for aerospace and automotive sectors.
- 1 August 2025
- Published by The Electrochemical Society in ECS Journal of Solid State Science and Technology
- Vol. 14 (8) , 081001
Q-Learning Driven Metaheuristic for Tailoring the Solid State Properties of Carbon Nanotube-Graphene Nanocomposites for Automotive and Aerospace Applications
The implementation of deep learning strategies has transformed composite material design and allowed the optimization of carbon fibre with epoxy resin beyond previous limits for greater performance results. Developing dependable structure-property-performance links continues to present a major obstacle because these systems operate through intricate combinations between nanotech advances and manufacturing simulations. This study presents a multi-faceted methodology employing a reinforcement learning (RL) approach to optimize carbon fiber and epoxy resin-based composites, incorporating over 10,000 samples of carbon nanotube, graphene, and polymer matrix combinations, thus aiming to achieve specific performance criteria such as tensile strength and thermal conductivity. Utilizing Langmuir-Blodgett self-assembly and Electrospinning methodologies, the research integrates novel nanotechnology to fabricate and test the designed composites. Combining RL with advanced manufacturing and finite element analysis simulation techniques provides a holistic view of the structure-property-performance relations in composite materials. Through iterative optimization, the expected outcomes indicate a potential improvement in mechanical properties by 15%–25% and thermal performance by 10%–20%. Simulated thermal expansion coefficient distributions show that the proposed model consistently achieves values in the range of 1.8 to 3.0 in terms of frequency and reliability. The outcomes of this research have significant implications for aerospace and automotive sectors.
- Open Access
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 131483-131497
A Dual-Branch Deep Learning Framework Combining Xception and ResNet for Accurate Lung and Colon Cancer Detection
Lung and colon cancers are among the leading causes of cancer-related deaths worldwide. Early detection significantly enhances survival rates, but traditional diagnostic methods, which rely on manual analysis of histopathological images, are labor-intensive, error-prone, and inconsistent. While deep learning has shown promise in automating medical image analysis, existing models often struggle with issues like overfitting, poor generalization, and class imbalance, which limit their clinical effectiveness. This research proposes a hybrid deep learning model combining the strengths of Xception and ResNet architectures. The Xception model excels at feature extraction, while ResNet’s residual connections improve training stability and address issues like gradient vanishing. The model is trained using a large dataset of lung and colon cancer images, with a dynamic learning rate and backpropagation via Stochastic Gradient Descent (SGD) momentum, optimizing performance. Advanced data augmentation techniques, such as rotations and flips, further enhance model generalization. The proposed hybrid model achieves an impressive accuracy of 98.96%, demonstrating high effectiveness in distinguishing between different cancer classes. The model outperforms traditional diagnostic methods, offering a robust and reliable tool for automated cancer detection with significant clinical potential.
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 131483-131497
Open AccessA Dual-Branch Deep Learning Framework Combining Xception and ResNet for Accurate Lung and Colon Cancer Detection
Lung and colon cancers are among the leading causes of cancer-related deaths worldwide. Early detection significantly enhances survival rates, but traditional diagnostic methods, which rely on manual analysis of histopathological images, are labor-intensive, error-prone, and inconsistent. While deep learning has shown promise in automating medical image analysis, existing models often struggle with issues like overfitting, poor generalization, and class imbalance, which limit their clinical effectiveness. This research proposes a hybrid deep learning model combining the strengths of Xception and ResNet architectures. The Xception model excels at feature extraction, while ResNet’s residual connections improve training stability and address issues like gradient vanishing. The model is trained using a large dataset of lung and colon cancer images, with a dynamic learning rate and backpropagation via Stochastic Gradient Descent (SGD) momentum, optimizing performance. Advanced data augmentation techniques, such as rotations and flips, further enhance model generalization. The proposed hybrid model achieves an impressive accuracy of 98.96%, demonstrating high effectiveness in distinguishing between different cancer classes. The model outperforms traditional diagnostic methods, offering a robust and reliable tool for automated cancer detection with significant clinical potential.
- 29 April 2025
- Published by Springer Nature in Journal of The Institution of Engineers (India): Series B
- Vol. 107 (1) , 81-94
A Novel Predict Earliest Finish Time Algorithm with Flower Pollination Algorithm for Effective Communication in Ad-Hoc Network
- 29 April 2025
- Published by Springer Nature in Journal of The Institution of Engineers (India): Series B
- Vol. 107 (1) , 81-94
A Novel Predict Earliest Finish Time Algorithm with Flower Pollination Algorithm for Effective Communication in Ad-Hoc Network
- 1 April 2025
- Published by The Electrochemical Society in ECS Journal of Solid State Science and Technology
- Vol. 14 (4) , 047007
Cognitive Edge-Driven Data Handling across Distributed Sensor Nodes for Robust Communication Networks
For applications such as smart cities, environmental monitoring, and industrial automation in critical infra, sensor networks are important, but they are confronted with serious issues regarding scalability, energy consumption, and reliable communication. To overcome these challenges, an adaptive smart sensor node communication framework (AS-SNCF) is proposed, which is an intelligent and efficient communication framework. A hexagonal grid deployment that provides uniform coverage is employed by AS-SNCF, and the nodes are divided into primary nodes (cluster heads) and secondary nodes (data collectors). It employs advanced quantum-inspired self-organizing clustering for dynamic cluster formation and bio-inspired neural routing for adaptive and energy-efficient multi-hop communication. The framework is based on event-driven sensing wherein nodes are activated only when critical thresholds are crossed, and lightweight data compression is applied to minimize transmission loads. Energy minimization and communication interference reduction are achieved through adaptive power control, and data handling based on priorities, whereby essential data is reliably sent with the support of self-healing system allowing the system to tolerate failure. AS-SNCF demonstrates a 45% increase in energy efficiency, a 38% reduction in latency, and a 42% improvement in scalability, making it a robust and efficient solution for complex and resource-constrained environments.
- 1 April 2025
- Published by The Electrochemical Society in ECS Journal of Solid State Science and Technology
- Vol. 14 (4) , 047007
Cognitive Edge-Driven Data Handling across Distributed Sensor Nodes for Robust Communication Networks
For applications such as smart cities, environmental monitoring, and industrial automation in critical infra, sensor networks are important, but they are confronted with serious issues regarding scalability, energy consumption, and reliable communication. To overcome these challenges, an adaptive smart sensor node communication framework (AS-SNCF) is proposed, which is an intelligent and efficient communication framework. A hexagonal grid deployment that provides uniform coverage is employed by AS-SNCF, and the nodes are divided into primary nodes (cluster heads) and secondary nodes (data collectors). It employs advanced quantum-inspired self-organizing clustering for dynamic cluster formation and bio-inspired neural routing for adaptive and energy-efficient multi-hop communication. The framework is based on event-driven sensing wherein nodes are activated only when critical thresholds are crossed, and lightweight data compression is applied to minimize transmission loads. Energy minimization and communication interference reduction are achieved through adaptive power control, and data handling based on priorities, whereby essential data is reliably sent with the support of self-healing system allowing the system to tolerate failure. AS-SNCF demonstrates a 45% increase in energy efficiency, a 38% reduction in latency, and a 42% improvement in scalability, making it a robust and efficient solution for complex and resource-constrained environments.
- 19 March 2025
- Published by Springer Nature in Neural Computing and Applications
- Vol. 37 (17) , 11527-11553
Brain tumor diagnosis using modified DenseNet121 architecture with adaptive learning rate and callback mechanism
- 19 March 2025
- Published by Springer Nature in Neural Computing and Applications
- Vol. 37 (17) , 11527-11553
Brain tumor diagnosis using modified DenseNet121 architecture with adaptive learning rate and callback mechanism
- 17 October 2024
- Published by Springer Nature
- p. 105-120
An Efficient Quadrature LEACH Routing Protocol with Enhanced FODPSO Optimization in WSN
- 17 October 2024
- Published by Springer Nature
- p. 105-120
An Efficient Quadrature LEACH Routing Protocol with Enhanced FODPSO Optimization in WSN