V. AMSAVENI

V.AMSAVENI
Associate Professor,
Dept of EIE,
Noorul Islam Centre for Higher Education,
Kumaracoil
Tamil Nadu, India

RESEARCH INTERESTS

Image Processing, Fuzzy Logic, Soft Computing

Scopus Publications

Enhancing Osteoarthritis Prediction with A Hybrid Deep Learning Model on Knee X-Rays
International Journal of Intelligent Engineering and Systems, 2025
Knee Osteoarthritis (KOA) is a progressive degenerative joint disease marked by the deterioration of articular cartilage and joint inflammation, which severely affects mobility and diminishes quality of life.Timely and precise detection of KOA is essential to enable early intervention and minimize disease progression.Despite advancements in medical imaging, accurate identification and grading of KOA remain challenging due to subtle morphological variations across different stages.To address these limitations, this study proposes a hybrid deep learning framework that combines the strengths of DenseNet-169 and Gated Recurrent Unit (GRU) networks for robust prediction of KOA severity using knee X-ray images.In the proposed model, input X-ray images obtained from the Kaggle repository are preprocessed and resized to a uniform resolution of 224 224 pixels.The dataset is organized into five distinct categories such as healthy, doubtful, minimal, moderate, and severe corresponding to standard KOA grading.DenseNet-169 is employed to extract high-dimensional spatial features, effectively capturing structural details and subtle variations in joint morphology.To further enhance the learning process, GRU layers are incorporated to model contextual and sequential correlations within the extracted features, ensuring better feature representation.Finally, a softmax activation function is used in the classification layer to output multi-class predictions.The proposed DenseNet-169-GRU framework demonstrated exceptional performance, achieving 96.78% accuracy, 96.32% precision, and 96.56% recall.These results highlight its ability to detect fine-grained pathological changes and provide reliable classification across different KOA stages.Overall, the study underscores the potential of hybrid deep learning techniques for improving the accuracy and clinical applicability of KOA diagnosis.
Solanaceae Safeguard: Cnn-Swin Fusion for Precision Disease Management
P Jaferkhan, V. Amsaveni
International Journal of Electrical and Computer Engineering Systems, 2025
Agricultural productivity stands as a cornerstone of India's economy, and enhancing it remains a priority. A pivotal strategy in bolstering agricultural output is the timely identification of diseases. In agriculture, disease detection and management are crucial for ensuring crop health and yield. This study proposes a novel disease detection system for Solanaceae Vegetables utilizing a hybrid deep learning approach. The system integrates SWIN Transformer architecture with Convolutional Neural Networks (CNN) to analyze and classify disease patterns in Solanaceae vegetables. The dataset used for training and evaluation is sourced from Kaggle repository, comprising comprehensive images of diseased and healthy Solanaceae plants. Through extensive experimentation, the proposed hybrid model achieves a remarkable classification accuracy of 96%. The model demonstrated high precision, recall, and F1-scores across most classes, such as Class 0 (0.92, 0.89, 0.91) and Class 14 (0.97, 1.00, 0.99).The system's high accuracy demonstrates its potential as a reliable tool for disease detection and effective management strategies in Solanaceae vegetable cultivation, thereby contributing to enhanced leaf health and productivity.
Beyond the Horizon: Exploring Osteoarthritis Predictive Models
Zareena Jamaluddin, V Amsaveni
2025 1st International Conference on Smart and Intelligent Systems Siscon 2025, 2025
Osteoarthritis (OA) is a prevalent chronic rheumatic condition causing pain and disability worldwide, mostly affecting weight-bearing joints like the knees. The diagnosis relies on criteria such as osteophyte development, cartilage loss, and anatomical abnormalities, with primary and secondary classifications based on various factors. Knee OA, in particular, is a important health concern due to its impact on mobility and quality of life. Cartilage degradation plays a crucial role in the progression of OA, leading to stiffness, pain, and reduced joint function. Traditional radiographic techniques, such as X-rays, are generally used for OA diagnosis, with grading systems like the Kellgren-Lawrence scale helping classify disease severity. However, challenges persist in accurately identifying and grading OA, prompting the exploration of advanced technologies like machine learning (ML) and deep learning (DL). In recent years, DL algorithms have shown promise in automatically diagnosing and tracking OA progression from various imaging modalities, offering personalized risk assessments and guiding treatment decisions. Similarly, ML approaches have demonstrated effectiveness in enhancing OA diagnosis accuracy by leveraging features extracted from knee X-ray images. Hybrid models combining DL, ML, and other techniques have emerged as well, further improving OA detection and classification. This review provides insights into recent advancements in AI-based approaches for knee OA diagnosis, highlighting their potential to revolutionize patient care and outcomes in OA management.
A Deep Learning Approach to Schizophrenia Diagnosis: Leveraging GRU and Attention Mechanisms for Improved EEG Analysis
N. Shan, V. Amsaveni
International Conference on Trends in Engineering Systems and Technologies Ictest 2025 Proceedings, 2025
Schizophrenia (SZ) is a chronic psychological disorder marked by significant instabilities in cognition, emotion, and behavior, affecting millions worldwide. Early and accurate diagnosis is crucial to improve patient outcomes, yet conventional methods, such as clinical interviews and symptom assessment, often lack objectivity and precision, leading to delayed intervention. Recent advances in artificial intelligence, especially deep learning (DL), offer promising diagnostic improvements by analyzing neurophysiological data, such as electroencephalograms (EEG). However, existing EEG-based approaches for SZ diagnosis struggle with inadequate temporal pattern recognition and feature extraction, resulting in suboptimal accuracy. This study proposes a novel hybrid DL model that integrates a Gated Recurrent Unit (GRU) network with an attention mechanism to enhance the diagnosis of SZ by capturing crucial temporal patterns in EEG signals. The method utilizes GRU’s ability to process sequential EEG data, while the attention mechanism selectively focuses on relevant EEG segments, isolating SZ-related neural signatures from noise. The model utilized publicly available EEG dataset from Kaggle, comprising of recordings from SZ patients and healthy controls across multiple electrode sites. The proposed model attained an impressive 98.02% accuracy, with an Area Under the Curve (AUC) of 0.96, 97.53% precision, 96.29% recall, and 96.90% F1-score. These outcomes demonstrate the model’s strong ability to differentiate SZ cases from healthy controls with high sensitivity and specificity. Overall, this study underscores the potential of GRU-attention models in providing an automated, non-invasive, and objective tool for SZ diagnosis, with significant implications for clinical application.
Spectrally Efficient High-Speed MDM-FSO Using Higher-Order Modes
Amina N, R Vaishnav, Parvathy V B, V Amsaveni, Kishan Seby, Nandana K V
International Conference on Trends in Engineering Systems and Technologies Ictest 2025 Proceedings, 2025
Mode division multiplexing is a technique for realizing spectrally efficient communication in upcoming wireless networks. This article proposes a system combining mode division multiplexing (MDM) and free space optics (FSO) to achieve a data rate of 40 Gbps using higher order modes through each channel with Laguerre Gaussian mode (LG mode) or Hermite Gaussian mode (HG mode), each having 8 modes, respectively. Performance analysis is investigated by considering the variations in atmospheric conditions as well as in transmission range. Bit error rate, eye diagram, and Q factor are the parameters used for system evaluation. The results after the simulation report an overall data transmission of 320 Gbps.
Optimized Schizophrenia Detection via EEG: A CNN-LSTM and CNN-GRU Ensemble with SVM
N. Shan, V. Amsaveni
Ssrg International Journal of Electronics and Communication Engineering, 2024
Preserving Solanaceae Crops: Strategies for Detecting and Controlling Diseases
P Jaferkhan, V. Amsaveni
2024 1st International Conference on Trends in Engineering Systems and Technologies Ictest 2024, 2024
Solanaceae vegetables, a vital group of crops including tomatoes, potatoes, peppers, and eggplants, play a significant role in global agriculture and culinary traditions. These crops are essential for our diets and contribute substantially to the economy. However, these plants are prone to numerous ailments, specifically leaf-related diseases, which have the potential to significantly impact the overall well-being and yield of the crops. Detecting and managing these diseases are crucial for ensuring food security and economic stability. In recent years, advanced techniques have been developed to tackle the issue. Machine learning (ML), a subset of artificial intelligence, has revolutionized the field of disease detection. It involves training algorithms to identify disease patterns in images of plant leaves, enabling rapid and accurate diagnosis. Various studies have demonstrated the effectiveness of ML in classifying tomato leaf diseases with high accuracy. Deep learning (DL), a subset of ML, utilizes neural networks to analyze large datasets, thereby improving the precision and efficiency of disease detection. Several DL models have been applied to detect tomato leaf diseases, offering precise and efficient solutions. Hybrid models, which combine traditional ML with DL techniques, have shown promise in improving disease identification and classification. In this study, we delve into a comprehensive exploration of time-honored methods employed for the identification of plant diseases, with a particular focus on leaves, as they serve as critical indicators of a plant's well-being. An in-depth analysis of various conventional techniques is conducted, shedding light on their inherent limitations. These advanced techniques in disease detection offer valuable tools for farmers and researchers, contributing to agricultural sustainability, crop protection, and food security.
From Brain Waves to Diagnoses: AI's Role in Schizophrenia Detection
Shan N, V. Amsaveni
2024 1st International Conference on Trends in Engineering Systems and Technologies Ictest 2024, 2024
Brain signals can be represented as numerical vector sets using electroencephalograms (EEGs). These signals are used to estimate a wide range of brain disorders, including schizophrenia, dementia, epilepsy, and Parkinson's. Experts must invest a great deal of time and effort in manually analyzing these signals. Finding the best EEG processing models for clinical use is challenging for researchers. Numerous machine learning-based processing techniques are suggested for these signals. The task of choosing a model for clinical use is made more difficult by the fact that these models have different performance criteria. This study explores the emerging role of Artificial Intelligence (AI) in transforming the landscape of schizophrenia detection through the analysis of brain waves. Leveraging neuroimaging technologies, such as electroencephalography (EEG) and functional magnetic resonance imaging (fMRI), researchers have sought to unveil distinctive patterns in the brain activity of individuals with schizophrenia. The research encompasses studies from the last decade, emphasizing the diverse AI methodologies employed in interpreting brain wave data. Machine learning algorithms, particularly deep learning models, have shown excellent results in identifying subtle aberrations associated with schizophrenia. These algorithms analyze intricate patterns within EEG and fMRI data, revealing hidden insights that traditional diagnostic methods might overlook. As the field advances, collaboration between clinicians, AI researchers, and ethicists becomes imperative to ensure responsible development and deployment of these technologies in clinical settings. This study contributes to the ongoing discourse on the transformative impact of AI in mental health and sets the stage for future research directions in schizophrenia detection.
Application of support vector machine classifier for computer aided diagnosis of brain tumor from MRI
V. Amsaveni, N. Albert Singh, J. Dheeba
Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics, 2015
In this paper a computerized scheme for automatic detection of tumors in brain is examined. Diagnosis of these lesions at the early stage is a very difficult task in normal brain images. The algorithm incorporates steps for preprocessing, feature extraction and classification using brain tumor detection. This paper proposes a supervised machine learning algorithm for detection of tumor. A feature extraction methodology is used to extract the Gabor texture features of the abnormal brain tissues and normal brain tissues prior to classification. Then support vector machine classifier is applied at the end to determine whether the given input data is tumor or non tumor. The detection performance is evaluated using Receiver Operating Characteristic curves. The result shows significantly improves the classification accuracy.
A novel control scheme for a shape memory alloy actuator
Rubin George, V. Amsaveni
2014 International Conference on Control Instrumentation Communication and Computational Technologies Iccicct 2014, 2014
Shape Memory Alloy actuators are commonly termed as smart materials because of its inherent ability to change shape with change in temperature. They have widespread applications in various fields. But due to its highly nonlinear operation, accurate control is difficult to achieve. In this brief, an adaptive neurofuzzy controller is proposed to control an SMA actuator. For the purpose of eliminating output noise and to estimate system states a Kalman filter was employed. From the simulation results it was verified that neurofuzzy control scheme and Kalman filter has successfully implemented.
Intelligent computer aided detection of tumor in MRI brain images using cascaded correlation neural network classifier
V. Amsaveni, N. Albert Singh, J. Dheeba
Applied Mechanics and Materials, 2014
Detection of brain tumor using neural network
V. Amsaveni, N. Albert Singh
2013 4th International Conference on Computing Communications and Networking Technologies Icccnt 2013, 2013
Computer aided detection of tumor in MRI brain images using cascaded correlation neural network
V. Amsaveni, J. Dheeba, N.A. Singh
Iet Seminar Digest, 2013

Publications

Application of Support Vector Machine Classifier for Computer Aided Diagnosis of Brain Tumor from MRI
AMSAVENI.V, ALBERT SINGH .N
Lecture Notes in Computer Science, Volume , Year 2015, Pages
A novel control scheme for a shape memory alloy actuator
AMSAVENI.V, RUBIN GEROGE
International Conference on Control, Instrumentation, Communication and Computational Technologies, ICCICCT 2014, Volume , Year 2014, Pages
Intelligent computer aided detection of tumor in MRI brain images using cascaded correlation neural network classifier
AMSAVENI.V, ALBERT SINGH .N
Applied Mechanics and Materials, Volume , Year 2014, Pages
Computer aided detection of tumor in MRI brain images using cascaded correlation neural network
AMSAVENI.V, ALBERT SINGH .N
IET Seminar Digest, Volume , Year 2013, Pages
Detection of brain tumor using neural network
AMSAVENI.V, ALBERT SINGH .N
International Conference on Computing, Communications and Networking Technologies, ICCCNT 2013, Volume , Year 2013, Pages