Computer Engineering, Computer Vision and Pattern Recognition
27
Scopus Publications
Scopus Publications
A Novelty Approach of Usability Testing Method in an E-learning Environment 16th International Conference on Advances in Computing Control and Telecommunication Technologies Act 2025, 2025
IoT-Enabled Personalized Fitness Solutions for Home Workouts using Reinforcement Learning M. Pandi, T. Guhan, T Sivakumar, Aswathy R H Proceedings of the 7th International Conference on Intelligent Sustainable Systems Iciss 2025, 2025 This research examines the integration of Reinforcement Learning (RL) and Internet of Things (IoT) technologies to create intelligent home exercise systems, addressing the increasing need for a virtual workout system. The technique employs RL algorithms to personalize exercises for each person and adapt them dynamically to facilitate achieving fitness goals. It employs IoT devices such as smart workout monitors and portable sensors for tracking human behavior, fingerprints, and ambient variables in real time. To attain optimum fitness outcomes, the proposed smart workout solution employs an automated feedback process through which RL techniques continuously track user interactions and adjust exercise parameters. It can be implemented by people of diverse fitness abilities, ensuring that anyone can discover beneficial workouts. The integration of IoT facilitates uninterrupted connectivity throughout devices, enhancing data transfer and the entire human interface. The device's primary features include continuous performance monitoring, developing individualized training regimens, and the automated adjustment of activity ranges. It examines the safety and protection challenges of collecting and processing private healthcare information in an integrated fitness environment. It aims to enhance digital workouts using RL and IoT adaptive and interactive personal training that transcends traditional static programs. Experimental findings indicate a 15% enhancement in exercise efficiency, a 20% rise in user engagement, and a 12% decrease in fatigue with implementing the proposed IoT-enabled RL fitness model.
Enhancing Customer Experience in Retail for Adaptive Store Layouts with Reinforcement Learning P. Suresh, T Sivakumar, N. Revathy, M. Pandi Proceedings of International Conference on Visual Analytics and Data Visualization Icvadv 2025, 2025 In the competitive retail industry, improving customer experience is essential for increasing sales and cultivating brand loyalty. A crucial determinant of consumer happiness is the shop layout, which impacts customers' navigation, exploration, and interaction with merchandise. Traditional store layouts are often inflexible and do not adjust to evolving customer behaviour, hence diminishing their efficacy in enhancing the shopping experience. This paper examines the use of Reinforcement Learning (RL) to create adaptable shop layouts that adjust based on real-time consumer interactions. Through the analysis of customer behaviour and feedback, RL algorithms may dynamically modify the layout of merchandise, aisles, and promotional areas, therefore optimizing the retail environment for both consumers and merchants. The RL algorithm analyses historical data, progressively optimizing store layouts to correspond with consumer preferences, so improving the entire shopping experience. It provides substantial advantages, such as enhanced consumer involvement, more operational efficiency, and elevated revenue designs. It examines the potential of RL in revolutionizing retail settings in optimizing adaptable shop layouts and its effects on consumer happiness and corporate performance.
An Efficient Authentication and Key Agreement for Vehicular communications Arun Sekar Rajasekaran, A. Obulesu, Azees Maria, T Sivakumar, S. Vimalnath 2025 IEEE International Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation Iatmsi 2025, 2025 Vehicular Ad Hoc Networks (VANETs) are crucial in improving transportation infrastructure by supporting vehicle to infrastructure (V2I) communication. Nevertheless, maintaining secure and efficient interoperability in such environments has always been a big push because of high mobility, and the vulnerability to cyber threats. This paper puts forward a novel authentication and key agreement protocol to be used in V2I communication settings of VANETs taking into consideration security and efficiency factors. The work in the proposal uses a lightweight cryptographic solution to maintain authentication of vehicles and the infrastructure without incurring much computational and communication load. In general, it is possible to use elliptic curve cryptography (ECC), and hash functions to protect the exchange from basic threats like impersonation, replay and so on. In addition, the proposed protocol includes a session key agreement mechanism that builds the secure communication channels in a dynamic way and provides confidentiality and integrity for the data during the exchange of messages. The simulation results show that the proposed scheme has better performance and it has lower computational overhead and communication overhead than the other related protocols.
Dynamic Pricing of Parking Spaces Using Linear Regression for Optimal Revenue Utilization Subodh Panda, Dr.T.Suresh, Raja Thimmarayan, T. Sivakumar, S Sujatha, Karthiga R 2025 2nd Asia Pacific Conference on Innovation in Technology Apcit 2025, 2025 Urban regions face increasing difficulties in parking space administration due to escalating vehicle density and ineffective static pricing strategies. This research introduces a dynamic pricing system based on Linear Regression (LR) aimed at maximizing parking income and use. Data from 5,000 parking transactions across three urban zones were analyzed, including variables such as time, day, weather conditions, event schedules, and occupancy rates. The model archived a R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> value of 0.89, indicating robust prediction performance. Dynamic pricing was tested in a designated area for 30 days, yielding a 15.6% revenue gain and a 23.4% enhancement in space utilization relative to fixed pricing. Peak-hour congestion decreased by 18.2%, indicating improved allocation of parking demand. Real-time pricing adjustments prompted consumers to choose off-peak slots, therefore equilibrating occupancy rates across various time intervals. This methodology illustrates the efficacy of linear models in smart city initiatives, providing an economical, comprehensible, and scalable resolution for municipal parking administration. Subsequent research will investigate the integration of IoT sensors and adaptive learning to improve performance.
Big Data Graph Node Importance Using Page Rank R. Roseline, B. Vasudevan, A. Clementking, S. Rani, T Sivakumar, C. Srinivasan 2024 1st International Conference on Innovations in Communications Electrical and Computer Engineering Icicec 2024, 2024 Improving the efficiency and accuracy of finding important nodes in large networks is the goal of enhancing Big Data graph node significance analysis using PageRank. To deal with the massive and complicated data sets seen in big data settings, it is necessary to modify and enhance the PageRank algorithm. The objective is to provide a more efficient framework that can speed up computation and improve the accuracy of significance ranking in large-scale networks. To control the algorithm's scalability, it is necessary to use state-of-the-art computational methods, such as parallel processing and effective data structures. The expected result is a scalable and reliable system that can quickly and correctly analyze large networks, yielding useful insights for applications like recommendation systems, online search, and social network analysis. The focus is on making PageRank more feasible for use in large data settings while also increasing its computing efficiency and scalability. The Adjacency Matrix of the Graph in a sample of 5 nodes has a value ranging from 15 to 55 in numbers, according to results from synthetic graph data. In another instance of the same dataset, the Initial PageRank Scores range from 0.2-0.15 in numbers, and in yet another instance, the final PageRank Scores for nodes range from 0.177 to 0.242 in numbers.
Blockchain-based Document Verification Scheme for Enhanced Security and Fraud Control Arun Sekar Rajasekaran, Haribabu J, Ramnjaneyulu GV, Sivakumar T, Mohanarathinam A, Velmurugan T 2nd International Conference on Emerging Research in Computational Science Icercs 2024, 2024 In this day and age of digital technology, the verification of documents plays an essential part in a variety of spheres, including the validation of identities, legal contracts, academic credentials, and financial transactions. Traditional document verification methods often face security issues, data tampering, and fraud. This research proposes a Blockchain-Based Document Verification Scheme (BBDVS) to improve document verification security and fraud management. This addresses concerns. The BBDVS uses blockchain technology's immutability and decentralization to create a safe, public document verification platform. Cryptography ensures the integrity, validity, and non-repudiation of checked papers. Smart contracts, blockchain-based self-executing agreements, automate and expedite verification. A prototype system is built and tested to determine BBDVS efficacy. The system simulates various document verification scenarios. Compared to conventional approaches, this technology is more secure, efficient, and reliable. It also shows how the scheme may avoid fraud, decrease costs, and simplify document verification for various businesses. The research found that the BBDVS may significantly improve fraud control and document verification security. This strategy could revolutionize document verification in banking, healthcare, education, and government services. More research is needed to solve the scalability, privacy, and regulatory issues that arise when implementing blockchain-based systems on a large scale.
Mindful Breathing Coaches Using IoT-Integrated AI Systems for Personalized Anxiety Relief P. K. Dhal, T. Sivakumar, B. Vasudevan, V.T. Lakshmi, Baskar Muthaiyan, S. Murugan 2024 1st International Conference on Innovations in Communications Electrical and Computer Engineering Icicec 2024, 2024 The proposed system focuses on personal anxiety alleviation through mindful breathing coaching, which investigates the use of Artificial Intelligence (AI) systems combined with the Internet of Things (IoT). Treatments for anxiety disorders should be easily available because of how common and crippling they are. The proposed solution integrates IoT devices with AI algorithms to provide personalized instructions for mindful breathing exercises and real-time tracking. Using a mixed-methods strategy, quantitative and qualitative user experiences and anxiety levels are analyzed. Anxiety symptoms were significantly reduced, and users were very satisfied with the tailored coaching, according to preliminary studies. There are scalable solutions to individual anxiety requirements that can be provided by integrating IoT and AI technology. It emphasizes the potential of technology to promote well-being and helps advance individualized therapies for mental health. It shows that it is possible to provide effective personalized anxiety reduction. By providing individualized assistance to those dealing with anxiety problems, this method has the potential to completely transform the way mental health treatment is provided.
Artificial Intelligence Techniques and Methodology Available for Lung Cancer Detection P Kanaga Priya, T Sivakumar International Conference on Sustainable Computing and Smart Systems Icscss 2023 Proceedings, 2023 Cancer is a term used to describe the body's aberrant cells growing out of control, which can occur in various body parts. With over a hundred types of cancer known to exist, delays in treatment can lead to severe health issues and, in some cases, loss of life. This study reviews the current methods used to detect lung cancer through image processing. The study suggests that artificial intelligence (AI) is effective in processing large datasets, providing accurate and efficient results for cancer detection. However, implementing these systems on a large scale presents several challenges, such as the capture of images, preparation of images, segmentation, and information management. Classification strategies must also be developed to ensure compatibility with AI. This study examines various imaging techniques and methodologies available for detecting cancer using a computer-aided user interface, which is essential to improve healthcare and meet the growing patient population's needs.
Smart Virtual Trial Room For Apparel Industry Finney Daniel Shadrach, M Santhosh, Sajja Vignesh, S Sneha, T Sivakumar IEEE International Conference on Distributed Computing and Electrical Circuits and Electronics Icdcece 2022, 2022
