VRVinoth Rathinam

Vinoth Rathinam

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

Vinoth Rathinam has a total of 25 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Design of a Fault Tolerance System for Industrial Controllers Utilizing Fractional Ensemble Learning

    by Vinoth Rathinam, Valarmathi Krishnasamy, K. C. Rajheshwari, Mohamed Shakeel Pethuraj, S. Baskar

    Smart manufacturing systems integrate technological platforms like cloud computing and computational paradigms to facilitate maintenance, monitoring, and automation. Operations and automation control based on devices or controllers depend on their connectivity and the quantity of machines they manage. This leads to further issues of task failures caused by controller interruptions. This paper introduces a Fault Tolerance Computable (FTC) method utilizing Fractional Learning (FL) to address this issue. The suggested computational approach uses ensemble learning via two fractional ensemble methods: stacking and boosting. During the stacking process, the ratio of the preceding maximum fault tolerance rate is set equal to the current failure rate. This equalization is optimized from median tolerance to a high rate via recurrent training. The training commences in the stacking state utilizing the previously provided controller logs. The boosting state provides an alternative to the existing job allocations and machine schedules derived from the equalization factor obtained in the preceding state. The controller’s machine-control output facilitates swapping at different time intervals. The process is initiated between two consecutive operational states of the industrial controller to optimize task completion rates. Findings: FTC-FL improves task allocation, completion, and tolerance rate by 10.79%, 11.88%, and 9.86%, respectively, with 11.9% fewer failures.

  • Open Access

    A Hybrid Quantum-Inspired Deep Learning Framework with Bio-Inspired Optimization for Cardiovascular Disease Prediction

    by Vinoth Rathinam, Valarmathi K, Madhumathi A, Lalitha S D

    The prognostication of cardiovascular diseases is paramount for the facilitation of early detection and enhancement of patient prognoses. It introduced a novel hybrid deep learning architecture that amalgamates Convolutional Neural Networks (CNN), Quantum Convolutional Neural Networks (Q-CNN), Long Short-Term Memory (LSTM) networks, Quantum-Inspired Long Short-Term Memory (Q-LSTM) models, Denoising Autoencoders (DAE), and Transformer Encoder–Decoder frameworks. The quantum models were innovatively structured by integrating unitary transformations and Hilbert space representations within traditional deep learning paradigms. Hyperparameter optimization, including learning rate, hidden unit count, dropout rates, and batch size, was executed utilizing the Greylag Goose Optimization (GGO) algorithm, which was meticulously chosen after initial benchmarking against conventional optimization techniques. These models underwent training and validation processes on a meticulously curated clinical dataset encompassing both demographic and clinical attributes, with preprocessing measures implemented to rectify missing data and address class imbalances. Among the array of assessed models, the GGO-optimized Q-LSTM exhibited superior performance, attaining an accuracy of 98.05% (95% CI: 96.8–99.2%), a precision of 1.00, a recall of 98.96%, an F1-score of 97.95%, and an AUC-ROC of 0.980. The DAE demonstrated an accuracy of 97.08% alongside an AUC-ROC of 0.989. Future research endeavors will focus on external validation and statistical significance testing to evaluate model generalization. Additionally, considerations regarding model interpretability through SHAP analysis and the practical deployment aspects (e.g., integration with Electronic Health Records) are thoroughly examined. This investigation underscores the assertion that the integration of deep learning methodologies, quantum-inspired modeling, and bio-inspired optimization strategies can markedly enhance predictive analytics for cardiovascular disease identification, while concurrently underscoring the critical importance of model interpretability and rigorous validation

  • Open Access

    An Optimized Theory-Guided CNN Framework with Siberian Tiger Algorithm for Breast Cancer Image Analysis

    by Vinoth Rathinam, Sasireka Rajendran, Valarmathi Krishnasamy, Vimala Mannarsamy

    Medical imaging technologies play a vital role in diagnosing and identifying breast cancer, where initial and accurate detection is essential to improve patient outcomes. However, existing methods face challenges in noise reduction, effective segmentation, and precise classification. This study proposes a novel framework, Breast cancer image classification is performed using a Theory-Guided Convolutional Neural Network (TCNN) enhanced through optimization with the Siberian Tiger Optimization (STO) algorithm, forming the proposed BCI-TCNN-STO model. Input images from the MIAS dataset are pre-processed using the Time-domain Real-valued Generalized Wiener Filter (TRG-WF) for noise reduction, segmented with Localized Sparse Incomplete Multi-View Clustering (LSIMC) to extract Regions of Interest, and classified using a Theory-guided CNN enhanced with STO for improved precision. Experimental evaluation demonstrates that the proposed method consistently outperforms existing techniques in terms of accuracy and error rates. These results highlight the potential of BCI-TCNN-STO as an effective tool to aid radiologists in the early diagnosis of breast cancer and clinical decision-making.

  • Open Access

    Tri-level lung cancer classification via deep learning based GoogleNet with computed tomography images

    by Vinoth Rathinam, Ramathilagam Arunagiri, Valarmathi Krishnasamy, Sasireka Rajendran

    Lung cancer (LC) is one of the most prevalent causes of cancer-related death worldwide. World Health Organization (WHO) classifies LC into two broad histological subtypes: non-small cell lung cancer (NSCLC) which is the cause of about 85% of cases and small cell lung cancer (SCLC) which makes up the remaining 15%. Several issues can influence LC detection including poor image quality, insufficient training data, low-quality image characteristics, and poor tumor localization. To overcome these challenges a novel TRI-level LC classification via deep learning-based GoogleNet with computed tomography (CT) images (TRI-LCNet) approach has been proposed for early-stage LC detection using CT images. Initially, the LC-input images CT are collected from openly accessible datasets. The lung CT images have been preprocessed using a Gaussian star filter (GaSF) to decrease noise, followed by feature extraction using GoogleNet. The extracted LC features are then given into a support vector machine (SVM) which is utilized as a classification tool to distinguish between different classes of LC cases. The TRI-LCNet approach performance was assessed by several metrics: specificity, accuracy, F1 score, and recall. The outcomes show that the suggested method obtains a higher accuracy range of 96.93% for the early identification of LC.

  • Open Access

    A unique YOLO-based gated attention deep convolution network-Lichtenberg optimization algorithm model for a precise breast cancer segmentation and classification

    by Vinoth Rathinam, Sasireka Rajendran, Valarmathi Krishnasamy

    A novel you only look once (YOLO)-based gated attention deep convolution network (GADCN) classification algorithm is developed and utilized in this present study for the detection of breast cancer. In this framework, contrast enhancement-based histogram equalization is applied initially to produce the normalized breast image with reduced noise artifacts. Then, the breast region is accurately segmented from the preprocessed images with low complexity and segmentation error using the YOLO-based attention network model. To diagnose breast cancer with better accuracy, the GADCN model is used to predict the exact class of image (i.e., benign or malignant). During classification, the activation function is optimally computed with the use of the Lichtenberg optimization algorithm (LOA). It aids in achieving improved classification performance with little complexity in training and assessment. The significance of the present study includes the use of a unique, YOLO-based GADCN-LOA model that helps in the prediction of breast cancer with higher accuracy. It was observed that the model exhibited 99% accuracy for the datasets utilized. In addition, the selected model outperforms well with sensitivity, specificity, precision, and F1-score. Hence the proposed model could be exploited for the diagnosis of breast cancer at an early stage to enable preventive care.

  • A Novel Method for Analyzing the Performance of Free Space Optical Communication in WDM Using EDFA

    by Karuppasamy Periyasamy, Vinoth Rathinam, Karthikeyan Ganesan, Meenakumari Ramachandran, Suresh Muthusamy, Ravita Lamba, Hitesh Panchal, Maheswaran Shanmugam, Suji Prasad Sudarsanan Nair Jalajakumari, Rajyalakshmi Kottapalli


  • Ultrasound Kidney Images with IKNN-Dependent FPGA Abnormality Classification

    by R. Vinoth, R. Sasireka

    Ultrasound imaging is commonly used to diagnose internal anomalies. Imaging for abnormality detection is a challenging process in today’s world. Even though there is an advancement in technology, tele-radiographers face difficulty in the accurate diagnosis of abnormalities. In order to resolve this issue, tele-radiology has paved a new way for doctors around the world to access the Internet to share the radiological images from one location to another. But frequent online access is one of the bottleneck issues. In order to overcome this drawback, Computer Assisted Diagnosis (CAD) is preferred in this proposed study and it uses VIRTEX-6 FPGA to clearly identify abnormality in the platform and also manual control is minimized in this condition. The proposed algorithm includes five steps: pre-processing, segmentation, feature extraction, selection and classification. The classification is performed using the Iterative K-Nearest Neighbor (IKNN) classifier based on the selected features. Unlike popular KNN, the proposed IKNN algorithm performs the similarity measurement on selective neighbors for a number of times where the number of neighbors has been dynamically selected at each iteration. Also, at each iteration, the method would select a subset of features in a random way. For the features selected and with the neighbors selected, the method computes the similarity value of Hist-sim which is being measured according to the features selected from the histogram features where the method computes the Haralick similarity with the features selected from the Haralick features. Using the features selected, the method computes the value of cumulative class drive similarity (CCDS). At each iteration the class with maximum similarity is selected and finally, the class being selected for the most number of times is selected as a result of classification. This improves the performance of classification. While comparing with the existing algorithms such as Support Vector Machine (SVM) with the linear, Radial Basis Function (RBF) and polynomial kernels, greater accuracy is achieved via IKNN classification. The specificity is found to be 95, 80 and 75 for normal, cystic and stone kidneys.


  • Segmentation and Detection of Tumor in MRI images Using CNN and SVM Classification

    by R. Vinoth, Chunchu Venkatesh

    Among cerebrum tumors, gliomas are the most well-known what's more, forceful, prompting a short future in their most noteworthy review. In this way, treatment arranging is a key stage to move forward the oncological patients personal satisfaction. Magnetic resonance imaging (MRI) is a Broadly utilized imaging system to survey these tumors, yet the vast measure of information created by MRI avoids manual division in a sensible time, constraining the utilization of exact quantitative estimations in the clinical practice. In this paper, we propose a programmed division strategy based on Convolutional Neural Networks (CNN). The kernels are used for the purpose of classification. Here, likewise researched the utilization of force standardization as a pre-preparing step, which in spite of the fact that not regular in CNN-based division strategies, demonstrated together with information enlargement to be extremely viable for mind tumor division in MRI pictures. The extension of the work is done by calculating certain parameters of the image. Detecting the accurate tumour cells where high density of area is infected Also calculating the feature of cells. Calculating the features provide us the depth of infection i.e stage of infection. SVM classification is performed with the calculated parameters. Extraction and detection of tumour from MRI scan images of the brain is done by using MATLAB tool.