Cloud-Continuum-Based Deep Learning Optimization Framework for Next-Generation Healthcare Data Performance on IoT Platform G. Aravindh, K. P. Sridhar International Journal of Advanced Computer Science and Applications, 2026 The development of healthcare data performance analysis is becoming more driven by the incorporation of intelligent computing paradigms that guarantee real-time, scalable, and personalized feedback for coaches and athletes. However, existing healthcare data analytics systems are challenged with severe issues such as decision-making latency, processing capacity limitations at the edge, data fragmentation, and the inability to integrate across heterogeneous computing environments seamlessly. Athletic data in this scenario refers to a combination of biomechanical factors (motion capture, joint angles, gait patterns), biometric signals (heart rate, oxygen saturation, muscle activity), and sport-specific performance indicators (workload, speed, and acceleration). This paper introduces the Cloud-Continuum-based Deep Learning Optimization Framework (CC-DLOF). This novel architecture leverages the synergistic potential of edge, fog, and cloud computing to provide a dynamic and smart healthcare data performance on an IoT platform. CC-DLOF is a hierarchical continuum architecture, with real-time data gathering and lightweight analytics performed in the edge layer, contextual processing and federated learning in the fog layer, and global intelligence, deep model training, and long-term data storage in the cloud layer. A new Cloud-Fog-Edge Orchestration Device (CFEOD) provides dynamic allocation of computational tasks in terms of latency sensitivity and device capability. At the same time, a blockchain-supported access control is used to maintain data security and privacy. Simulation analysis, done in a simulated training environment that combines with real-world data sets, illustrates the performance of the framework in mitigating latency by 35%, increasing model accuracy by 22%, and boosting system scalability and reliability. CC-DLOF is a revolutionary way in healthcare data technology, leading to smart, responsive, and safe next-generation healthcare data performance on an IoT platform.
AI-Driven AGV with Enhanced YOLO for Smart Garment Handling T. Gowtham, Aravindh. G, Murali. L, Praveen. S, Vishwa. Y, Sachithanantham. P Proceedings of 5th International Conference on Communication Computing and Electronics Systems Iccces 2026, 2026 The rapid expansion of industrial automation, digital warehousing, and intelligent surveillance has increased the demand for autonomous and reliable material-handling systems. This work presents the development of an AI-powered Automated Guided Vehicle (AGV) integrating Enhanced YOLO (E-YOLO)-based object detection, adaptive path navigation, obstacle avoidance, and real-time sensor feedback for efficient and safe mobility. Continuous visual input from an onboard camera is processed using E-YOLO with an adaptive uncertainty-weighted combined loss function, improving detection accuracy, convergence speed, and robustness in dynamic warehouse environments. This enhancement minimizes dispatch delays, reduces manual inefficiencies, and enables intelligent, vision-guided control along customizable pathways. The system is deployed for automated garment handling: it identifies the clothes on hangers, classifies, and routes them, with limited human support when fine alignment is needed. The AGV recognizes garment types, fabric categories, and size labels with high accuracy, constituting a semi-automated workflow that integrates human flexibility with machine precision. Upon classification, the AGV assesses size, material type, and positioning of the hanger to identify the best route toward packaging stations. Continuous communication with sensors and checkpoints allows for congestion-free movement with dynamic path optimization.
Skin Cancer Detection with Severity Analysis Using Machine and Deep Learning Techniques K Deenu, S Kaviya, G Aravindh, D Arokiamercy, G Devadharshini, A Hemambika Proceedings of 5th International Conference on Communication Computing and Electronics Systems Iccces 2026, 2026 This paper describes a novel system for systemically detecting skin cancer and stage of severity using a hybrid approach combining machine learning and deep learningbased approaches. The framework uses a modified DenseNet architecture for the classification of skin lesions, a UNet model for accurate segmentation of the lesion, and a Random Forest (RF) algorithm for analyzing severity of the lesion. DenseNet allows for efficient reuse of features and propagation of gradients while UNet provides accurate extraction of the boundary of the affected area. Transfer learning with ImageNet weights improves convergence of training and accuracy of diagnoses. The integration of these models will support early and accurate diagnosis of benign and malignant skin lesions with high levels of interpretation. It was evaluated against the ISIC dermatascopic dataset and overall classification was 98 % for skin lesions, which outperformed CNN-based approaches in terms of specificity, sensitivity, and F1-score. The results of a fully automated diagnostic sysmtem improves performance, and provides detailed information on severity, ultimately allowing dermatologists to make evidencebased decisions.
The Litter Coach: Smart Waste Management System R. Vishnu Vardhan, G. Aravindh, C. S. Karthik, P. Gowtham Vignesh, M. Gunasekaran Proceedings of 10th International Conference on Communication and Electronics Systems Icces 2025, 2025
Comparative Evaluation of Diabetic Retinopathy Diagnosis Using Fundus Imaging Techniques R. Vishnu Vardhan, Aravindh G, Abinaya P, N. Kavyasree, M.S.Naga Anna Poorni Proceedings 3rd International Conference on Artificial Intelligence and Machine Learning Applications Healthcare and Internet of Things Aimla 2025, 2025 People can get a variety of diseases, some of which have no known treatment. One or both eyes may be affected by diabetic retinopathy (DR), a condition that can cause vision issues and possibly irreversible blindness. This is a very complicated problem. However, as was the case in other research conducted at the AIIMS, early detection can greatly reduce the burden of vision loss with the appropriate treatment and precautions. Applying advanced models with various picture Kaggle datasets and suggesting a convolutional neural network (CNN) model of comparable performance is the primary aim of this research. Interestingly, our suggested model had a 80.46% accuracy rate on kaggle dataset. Finding potential enhancements and resolving the issues these approaches face are the goals.
Sequential Deep Learning Approach for Simultaneous Detection of Alzheimer's and Parkinson's Diseases using Recurrent Neural Networks B Saranraj, G Aravindh, L Murali, M Manoj Kumar, K Naveenkumar, M Lokesh 4th International Conference on Automation Computing and Renewable Systems Icacrs 2025 Proceedings, 2025 Alzheimer’s disease (AD) and Parkinson's disease (PD) are chronic progressive neurodegenerative diseases that significantly diminish the quality of life and require diagnosis in time to be controlled. We introduce a new hybrid deep learning-based system in this paper that utilizes RNN, LSTM, and GRU networks in the automatic detection of AD and PD using the brain MRI images. Although the majority of the literature addresses the problem as the task of image classification with the help of CNNs, our study uses a hybrid model based on RNNs, treating image rows as temporal sequences, thereby leveraging the sequential relationships between spatially distributed features, leveraging the temporal relationships between spatially distributed measurements. More importantly, six individual validation checks in the proposed system enable enhanced brain MRI validation in a reliable analysis of a medical image. Findings show the best performance to the current CNN-based methodologies, with the high accuracy of multi-class classification-normal, Alzheimer's, Parkinson's, and dual condition. The system has advanced confidence thresholding processes and provides detailed diagnostic reasoning.
A Hybrid Ensemble Model for Brain Tumor Detection and Classification: Enhancing Diagnostic Accuracy L Murali, G Aravindh, R Vishnu Vardhan, P. K. Nivethitha, M. Pavatharani, V. Santhiya 4th International Conference on Automation Computing and Renewable Systems Icacrs 2025 Proceedings, 2025 Brain tumors are serious ailments resulting from an uncontrollable increase in cell growth in the brain, making them difficult to cure. The Magnetic Resonance Imaging (MRI) is a well-known and commonly used diagnosis device for brain tumors due to its excellent quality images. Traditional diagnosing methods are often time-sensitive, inaccurate, and dangerous; thus, they are a challenge to health care. This automated medical imaging analysis with Artificial Intelligence (AI), especially using deep learning, has highlighted a disruption in focus towards accurate and automated methods. This paper presents a Computer-Aided Diagnosis (CAD) model for intelligent brain tumor detection with deep learning-based methods. More specifically, the CAD model uses the Exception and Vision Transformer (ViT) model to detect and define brain tumors as Grade II, Grade III, and Grade IV. There are many metrics used to measure the performance of high dimensional data, including accuracy, sensitivity, precision, specificity and F1-score. Based on the experimental results of these models from an MRI-large dataset, both the Xception and ViT can detect brain tumors fast and accurately.
Optimal Control of Solar PV Fed EV Charging Station Using PSO Algorithm R. Chandrasekaran, S. Karthikkumar, A. Sheela, Pragaspathy Subramani, Aravindh. G, Rajkumar P 2023 3rd International Conference on Advances in Electrical Computing Communication and Sustainable Technologies Icaect 2023, 2023
An Efficient Finger Gesture Recognition System Using Image M. Arun Kumar, S. Jayachithra, G. Aravindh, M. Bhuvaneswari Proceedings of the 4th International Conference on Electronics Communication and Aerospace Technology Iceca 2020, 2020
Study on Hand Gesture Recoginition by using Machine Learning A. Mohanarathinam, K.G. Dharani, R. Sangeetha, G. Aravindh, P. Sasikala Proceedings of the 4th International Conference on Electronics Communication and Aerospace Technology Iceca 2020, 2020
Automatic Detection of Brain Tumor Using Deep Learning Algorithms R. Sangeetha, A. Mohanarathinam, G. Aravindh, S. Jayachitra, M. Bhuvaneswari Proceedings of the 4th International Conference on Electronics Communication and Aerospace Technology Iceca 2020, 2020
Algorithm and implementation of distributed canny edge detector on FPGA Arpn Journal of Engineering and Applied Sciences, 2015
