Sivashankari Rajadurai

Dr. Sivashankari Rajadurai


Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 February 2024 in Diagnostics
Verified
Journal Article
Open access
PDF file
SRSivashankari Rajadurai
Sivashankari Rajadurai
KUMARESAN PERUMALKUMARESAN PERUMAL
KUMARESAN PERUMAL
MIMuhammad Fazal Ijaz
Muhammad Fazal Ijaz
Chiranji Lal ChowdharyChiranji Lal Chowdhary
Chiranji Lal Chowdhary
Malignant lymphoma, which impacts the lymphatic system, presents diverse challenges in accurate diagnosis due to its varied subtypes—chronic lymphocytic leukemia (CLL), follicular lymphoma (FL), and mantle cell lymphoma (MCL). Lymphoma is a form of cancer that begins in the lymphatic system, impacting lymphocytes, which are a specific type of white blood cell. This research addresses these challenges by proposing ensemble and non-ensemble transfer learning models employing pre-trained weights from VGG16, VGG19, DenseNet201, InceptionV3, and Xception. For the ensemble technique, this paper adopts a stack-based ensemble approach. It is a two-level classification approach and best suited for accuracy improvement. Testing on a multiclass dataset of CLL, FL, and MCL reveals exceptional diagnostic accuracy, with DenseNet201, InceptionV3, and Xception exceeding 90% accuracy. The proposed ensemble model, leveraging InceptionV3 and Xception, achieves an outstanding 99% accuracy over 300 epochs, surpassing previous prediction methods. This study demonstrates the feasibility and efficiency of the proposed approach, showcasing its potential in real-world medical applications for precise lymphoma diagnosis.

14 August 2017 in Arabian Journal for Science and Engineering
Verified
Journal Article
Closed access
PDF file
SRSivashankari Rajadurai
Sivashankari Rajadurai
BVB. Valarmathi
B. Valarmathi

6 November 2025
Verified
Conference Paper
Open access
PDF file
SRSivashankari Rajadurai
Sivashankari Rajadurai
MMM. Mohanasudha
M. Mohanasudha
In the revolution of the medical field, Biomedical Natural Language Processing (BioNLP) plays a major role in extracting meaningful entities from unstructured medical text. With the rise of advanced NLP technologies, medical entity extraction has become an essential task for understanding and organizing complex clinical language. The evolution of deep learning models, particularly transformer-based architectures, has enabled machines to interpret biomedical content with remarkable accuracy. Amid these advancements, the Hybrid BioBERT–BiLSTM model stands out as an intelligent and efficient framework for medical entity extraction — the process of identifying entities such as diseases, drugs, symptoms, and anatomical terms from clinical narratives. And a rule set is prepared to convert the unstructured medical data into a structured data. BioBERT, a domain-specific adaptation of BERT trained on biomedical literature, effectively captures the semantic meaning of medical terms, while BiLSTM enhances contextual understanding by processing information from both directions of a sentence. Using BIO-tagged datasets derived from medical transcription samples, this hybrid system accurately labels terms such as “asthma” as a Disease or “aspirin” as a Drug, thereby improving the automation and accuracy of medical data analysis. The model’s performance evaluation yielded a precision of 81.45%, recall of 82.23%, and an F1-score of 81.63%, demonstrating its robust capability in biomedical entity recognition

6 November 2025
Verified
Conference Paper
Open access
PDF file
TGT. Gnana
T. Gnana
SRSivashankari Rajadurai
Sivashankari Rajadurai
Parkinson’s disease (PD) is a chronic neurological disease that affects millions of people worldwide characterized by movement impairments, cognitive disability, and decrease the quality of life. PD complexity and the severity of its symptoms demand robust and early diagnostic methodologies and the advancement of new treatment approaches. This survey paper provides a comprehensive and breakthroughs in Parkinson's disease research focusing on innovative diagnostic biomarkers, advanced non-invasive diagnostic techniques and the transformative role of artificial intelligence in enhancing early detection and intervention strategies. The machine learning models such as CNN, DNN, LSTM, ResNet, PCNN, ALSTM and the optimization methods are GDABC, MASS, QMFO and Swarm are highlighted. The multimodal method used to identify the voice,text,sensor, image or video and Giat data. The paper is also analysed various multimodal techniques useful for classification. The review is identified as a valuable resource for clinicians, researchers and expert to understand the evolving of PD detection and management.

6 November 2025
Verified
Conference Paper
Open access
PDF file
MMM. Mohanasudha
M. Mohanasudha
SRSivashankari Rajadurai
Sivashankari Rajadurai
Medical records contain a wealth of information about a patient’s health, treatments, and clinical history; however, much of this data is recorded in free-text format, making it difficult to analyse and apply directly. Discharge summaries, which provide a detailed overview of a patient’s hospital stay including diagnoses, treatments, medications, and follow-up instructions frequently contain domain-specific terminology, abbreviations, and inconsistent grammar, further complicating automated analysis. Extracting meaningful information from these summaries can help healthcare systems better understand patient conditions, support clinical decision-making, and facilitate medical research. Various approaches have been proposed to address these challenges. The SR classifier effectively extracts fine-grained relationships, while methods such as Three-way Inter-Annotator Agreement (IAA) use annotation tools for keyword-based relation identification. Deep learning techniques, including Bi-LSTM, CRF, and CNN models, have been employed to capture sequential patterns, entity structures, and relationships between medical entities. Modern transformer-based models further enhance these capabilities by learning complex contextual representations. By combining these methods, unstructured medical text can be transformed into structured, actionable knowledge, improving patient care and enabling data-driven healthcare analytics.

24 September 2026 in Frontiers in Artificial Intelligence
Verified
Journal Article
Open access
PDF file
TAT. Gnana Abinaya
T. Gnana Abinaya
SRSivashankari Rajadurai
Sivashankari Rajadurai
Introduction: Parkinson's Disease (PD) is a neurological disorder that worsens over time and is marked by interhemispheric asymmetry, dopaminergic degeneration, and structural changes in the brain. Methods: This paper presents a novel approach: AGMNet- Adaptive Gated Multimodel Network, which is an explainable gated multimodal deep learning framework that combines structural magnetic resonance imaging (MRI), dopamine transporter (DaT) imaging, asymmetry-aware representations, and biomarker features for robust PD classification. The Grad-CAM method is incorporated with the proposed AGMNet framework to improve the model interpretability by identifying image regions that influence classification decisions. The performance of AGMNet framework was evaluated using five-fold cross-validation and compared with established deep learning architectures, including VGG16, VGG19, InceptionV3, and Xception. Results: The proposed AGMNet has achieved 0.8318 of accuracy, 0.8278 of balanced accuracy, sensitivity of 0.8556, specificity of 0.8000, AUC of 0.711, and Matthews correlation coefficient (MCC) of 0.6458. The proposed AGMNet model demonstrated competitive classification performance relative to the evaluated baseline architectures. The combination of DaT imaging, asymmetry-aware features, and biomarker representations in the proposed model, contributed more strongly to the fused representation than MRI features alone. Discussion: The findings indicate that adaptive multimodal fusion in the AGMNet framework can improve the integration of complementary structural, functional, asymmetry-related, and biomarker information for PD classification. The learned gate weights further demonstrate the relevance of functional and asymmetry-aware information within the proposed framework, while Grad-CAM provides additional insight into model decision-making. Therefore, the AGMNet framework offers a potentially useful and interpretable framework for multimodal PD classification.