RKRamalakshmi Krishnasamy

Ramalakshmi Krishnasamy

*This is an algorithmically generated author record based on publicly available information.

Ramalakshmi Krishnasamy has a total of 40 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Energy-aware heart disease prediction system using hybrid metaheuristic optimization and deep learning for IoT-enabled healthcare environments

    by Ramalakshmi Krishnasamy, Vinod Kumar Shukla, L. Krishna Kumari, M. Uvaneshwari, B. Mohan, P. Sundaravadivel


  • An Attention-Augmented U-Net Model for Lung Region Segmentation From Chest X-Ray Images

    by K. Ramalakshmi, L. Krishna Kumari

    Lung segmentation from chest X-ray images plays an important role in the early diagnosis of pulmonary diseases such as pneumonia, lung cancer, and COVID-19. Manual segmentation is time-consuming and susceptible to observer variability, creating the need for automated and reliable approaches. This study proposes an Attention-Augmented U-Net model for accurate lung region segmentation. The proposed framework enhances the conventional U-Net architecture by integrating Attention Gates (AGs) into skip connections between encoder and decoder layers, enabling the network to focus on relevant lung regions while suppressing background structures. Publicly available chest X-ray datasets were used for training and evaluation with preprocessing techniques including normalization and data augmentation. Experimental results achieved a Dice coefficient of 0.93 and improved segmentation performance compared with conventional U-Net models. The proposed system may support radiologists in computer-aided diagnosis and collaborative clinical workflows.

  • Enhanced Liver Tumor Classification Using ResNet With Median Filtering

    by K. Ramalakshmi, Vinod Kumar Shukla, P. Sundaravadivel, Jasmine S., L. Krishna Kumari, N. Ismayil Kani, R. Prem Ananth, B. Mohan

    In this study, we propose an enhanced liver tumor classification framework that integrates median filtering as a preprocessing step with a ResNet-based deep learning model. Median filtering effectively reduces noise while preserving tumor boundaries, thereby improving the quality of discriminative features extracted by the network. The liver CT images were first preprocessed through resampling, normalization, and median filtering, followed by data augmentation to increase the robustness of the model. A fine-tuned ResNet-50 architecture was employed to perform multiclass classification of liver tumors. Experimental evaluation on a benchmark liver tumor dataset demonstrated that the proposed approach achieved an overall accuracy of 95.3%, sensitivity of 94.1%, specificity of 96.7%, precision of 95.0%, and F1-score of 94.5%. Furthermore, the area under the ROC curve (AUC) reached 0.972, highlighting the strong discriminative power of the method.

  • Deep Learning Approach Leveraging Optimized Residual Neural Networks for Robust and Early-Stage Alzheimer's Disease Clarification

    by K. Ramalakshmi, M. Uvaneshwari, L. Krishna Kumari, B. Mohan, N. Ismayil Kani, M. Ashok Kumar, R. Santhana Krishnan

    This study presents a deep learning-based framework using the ResNet18 architecture for automatic classification of Alzheimer's disease stages from Magnetic Resonance Imaging (MRI) scans. The datasetcomprises Non Demented, Very Mild Demented, Mild Demented, and Moderate Demented—with both original and augmented images to ensure data diversity and balance. The methodology involves several key stages. First, all MRI images were resized to a uniform resolution, normalized, and augmented. The ResNet18 model, a deep residual convolutional neural network pre-trained on ImageNet, was utilized through transfer learning, where the initial layers were frozen to retain low-level feature extraction, and the final fully connected layers were fine-tuned to adapt to the Alzheimer's dataset. The trained model achieved an outstanding test accuracy of 98.18%, indicating strong generalization and discriminative capability. The classification report demonstrates high reliability with F1-scores of 0.99 (Mild Demented), 1.00 (Moderate Demented), 0.98 (Non Demented), and 0.97 (Very Mild Demented).


  • Hybrid Neural Network Models for Medical Image-Based Disease Detection

    by Ramalakshmi K., Sagarika Kar Chowdhury, Thilagam K., Krishna Kumari L.

    Alzheimer's disease is a progressive neurological disorder characterized by gradual deterioration in memory and cognitive abilities, making timely diagnosis critical for effective management.This study introduces a lightweight deep learning framework that combines a MobileNet backbone with an attention-guided refinement mechanism to enhance feature representation.A Convolutional Block Attention Module is incorporated to selectively emphasize informative channels and spatial regions associated with neurodegeneration. In addition, a Grey Wolf Optimization (GWO) strategy is employed to automatically tune key hyperparameters and improve model convergence. Prior to feature extraction, MRI scans undergo multi-scale enhancement and noise suppression to improve structural clarity. The optimized feature maps are subsequently classified using fully connected layers for multi-stage Alzheimer's detection. The combination of attention-driven learning and metaheuristic optimization offers a practical and efficient solution for early and reliable Alzheimer's disease diagnosis in clinical settings.

  • Digital Transformation and Technology in Higher Education

    by K. Ramalakshmi, Vinod Kumar Shukla, T. Anjani Devi, N. Shanmugapriya, A. Mohamed Anwar, R. Santhana Krishnan, R. Ranjani

    Digital transformation in higher education is the process of using modern digital technologies to improve how universities and colleges teach, manage, research, and connect with the wider community. This transformation goes beyond simply using computers or online tools. It involves a complete change in how institutions work, how students learn, and how teachers teach. Technologies like artificial intelligence, cloud computing, learning management systems, virtual and augmented reality, and data analytics are becoming central to educational practices and decision-making. The COVID-19 pandemic pushed many institutions to quickly adopt digital solutions, leading to widespread use of online and hybrid learning. As a result, universities have started to build long-term strategies to include digital tools in classrooms, student services, and administration. This chapter discusses the major technologies involved, new teaching methods supported by digital tools, and the effects on students, faculty, and institutions. At the same time, digital transformation brings several challenges.

  • Integrating Deep Learning With Association Rule Mining for Liver Tumor Classification in Medical Imaging

    by K. Ramalakshmi, Vinod Kumar Shukla, P. Sabeena Burvin, N. Ismayil Kani, L. Krishna Kumari

    Liver tumors, whether benign or malignant, present a serious clinical challenge, making accurate classification vital for early detection and effective treatment. This study proposes an AI-driven framework combining advanced image preprocessing with a VGG16-based deep learning model, achieving 98.75% classification accuracy. Median filtering enhances image clarity and preserves key features for deep feature extraction. The classified outputs are then integrated into Association Rule Mining (ARM), converting them into transaction-style datasets to uncover meaningful rules linking tumor types with imaging biomarkers, demographics, and treatment outcomes. This two-stage approach not only ensures high precision in tumor classification but also yields interpretable rules for clinical decision support, helping bridge the gap between black-box AI and real-world medical imaging applications.

  • Harnessing CNN-LSTM Fusion for Enhanced Crime Forecasting and Temporal Pattern Analysis

    by K. Ramalakshmi, A. Mohamed Anwar, L. Krishna Kumari, B. Mohan, R. Ranjani, R. Santhana Krishnan, J. Relin Francis Raj

    This research investigates the efficacy of deep learning models, specifically Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks, in the context of crime rate prediction. By combining the spatial feature extraction capabilities of CNNs with the temporal sequence learning strengths of LSTMs, we developed a hybrid CNN-LSTM model. The l results demonstrate that the CNN-LSTM model significantly surpasses traditional predictive models, including ARIMA and standard neural networks, achieving a remarkable accuracy rate of 97.8%. Furthermore, the incorporation of multiple LSTM layers not only enhanced predictive accuracy but also optimized the training process. These findings highlight the potential of advanced neural network architectures in developing robust predictive systems for crime forecasting. The study concludes with recommendations for future research focused on hyperparameter optimization to improve model performance and reliability further, thereby contributing valuable insights for proactive crime prevention and effective law enforcement strategies.


  • Harnessing Immersive Technologies

    by K. Ramalakshmi, L. Krishna Kumari, B. Mohan, Sravan Kumar Gunturi, S. P. Santhoshkumar, R. Santhana Krishnan, J. Relin Francis Raj

    Both VR and AR are at the forefront of delivering engaging, interactive experiences that are reshaping industries by improving training, operational efficiency, and user engagement through immersive and enhanced environments. These technologies serve as a bridge between the digital and physical realms, providing innovative solutions across a wide range of domains, including education, healthcare, entertainment, and enterprise. The integration of AR and VR is particularly impactful, enabling advanced applications that transform how individuals experience and interact with their surroundings. This study explores the development and capabilities of VR and AR technologies, highlighting their standalone applications and demonstrating how their combination yields powerful outcomes.


  • AI-Driven Curriculum Design for Higher Education Institutions

    by D. Jansirani, A. Mohamed Rizwana, A. Noorjahan Rehana, K. Ramalakshmi, S. T. Gopukumar

    Artificial intelligence (AI) is changing the terrain of higher education institutions in a time of unparalleled technological development, thereby requiring a basic change in curriculum design. This chapter explores innovative strategies for integrating AI into higher education institutional frameworks to prepare learners for an AI driven future. Instead of perceiving AI as a disruptive technology, it is envisioned to be an innovation driver in pedagogy, learning processes, and organizational systems. This section discusses several dimensions of how AI restructures learning. It focuses on understanding AI, learning that is personalized for oneself, blending several subjects, and ongoing skill development. It focuses on developing a growth mindset for students and educators, in which flexibility, ethical thinking, and critical thinking are as crucial as technical skills. Some of the key strategies include curriculum alignment with industry needs, training teachers to use AI tools, and using real-world AI examples in several subjects.