• 学术搜索
  • 科研智能体
    • Research Labs
    • AI 阅读
    • AI 文库
    • 深度研究
    • 学者亮点
  • 学术资源
    • AI2000
    • 期刊/会议
    • 学者库
    • 学术API
      MCP&Skill
    • 溯源树
    • 数据集
  • 知识沉淀
    • 学术空间
订阅小程序
旧版功能
aminer vip
开通会员低至0.73元/天
一次搞定AI科研
立即登录
  • English
  • 联系方式
    G

    GRT Institute of Engineering and Technology

    院校
    243论文总数
    1,052引用总数

    GRT Institute of Engineering and Technology is an engineering college in Tiruvallur, Tamil Nadu, India. The college is affiliated to Anna University, Chennai.

    论文量&引用量时间轴

    机构学者

    论文(243)

    年份
    起
    –
    止
    排序
    1Biomedical Image-Based Keratoconus Classification Using Convolutional Transformers and Grey Goose Optimization
    K. Balaji, N. Gobalakrishnan

    Background: Keratoconus (KCN) is a progressive corneal disorder that leads to visual distortion and increases the risk of complications during refractive surgical procedures. Accurate and early diagnosis is critical for effective clinical management. Objective: This work presents a novel deep learning model for keratoconus classification from corneal topographic images. Problem: Conventional diagnostic models are based on corneal topography and biomechanical properties. These processes are time-consuming and prone to variability. Methodology: This work presents an optimization based Deep Learning (DL) model Grey Goose Optimization (GGO) based Convolutional Transformer (CT) model for KCN classification. The model considers the feature extraction capabilities of convolutional neural network (CNN) and the attention mechanism of transformers for enhancing classification accuracy. The GGO is integrated for fine-tuning the hyperparameters of the model and improving its robustness. Findings: The proposed approach is evaluated on a Keratoconus detection dataset and shown superior performance compared to conventional DL models. The outcomes shows the robustness of the suggested approach in correctly categorizing KCN by achieving accuracy of 99.2% and precision of 98.9% Conclusion: The proposed CT-GGO provides the way for automatic and reliable diagnostic tools in ophthalmology. Recommendations: Future research will demonstrate cross-center validation, larger data, and multi-modal inputs for improving model generalization and interpretability.

    2026BIOMEDICAL SIGNAL PROCESSING AND CONTROL(2026)引用:3
    引用
    AI阅读
    加入学术空间
    2Comparative Review of Natural and Modified Amino Acid Crystals: Structural Features and SHG Efficiency with Reference to KDP
    E. Mohanapriya, S. Selvakumari, S. Sankar, N. Kanagathara
    2026Phase Transitions(2026)引用:1
    引用
    AI阅读
    加入学术空间
    3Integrated Dual-Syringe Pump for Precision Simultaneous Infusion in Medical Treatments
    G. Hari Krishnan, T. Sudhakar, Sheeba Santhosh, G. Umashankar

    This work presents a dual-syringe infusion pump with integrated pressure monitoring and syringe presence detection, aimed at improving reliability in controlled fluid administration. The system utilises two independently driven stepper motors, load-cell-based pressure sensing, and IR-based syringe detection managed by an Arduino controller. A complete circuit schematic, firmware control flow, calibration model, and accuracy computation methodology are provided to with delivery error below 2.2%, repeatability characterised by CV values below 3%, and linearity coefficients (R2) above 0.986. The load-cell subsystem enables rapid detection of occlusion events, triggering automatic shutdown within 5 seconds. These results confirm that the proposed design achieves high accuracy, robust safety response, and mechanical repeatability, while remaining lowcost and fully transparent. The system addresses shortcomings in existing open-source dual-pump platforms, which often lack safety-critical sensing and reproducible engineering documentation, and provides a reliable foundation for laboratory-scale infusion research and future development that meets medical compliance standards.

    2026ORIENTAL JOURNAL OF CHEMISTRY(2026)
    引用
    AI阅读
    加入学术空间
    4A Novel Design of Edge-AI Powered Real-Time Health Monitoring System Using Energy-Efficient Embedded Sensors
    R Govindarajan, Radhamani V, K Malathy, M Gopikrishnan, Aravind A R, Prabu M

    The increasing need of health tracking (in remote and resource-limited forecast) in a continuous, real-time fashion has boosted the adoption of Edge Artificial Intelligence (Edge-AI) in wearable healthcare systems. The paper proposes a new design of a real-time health monitoring system running on energy-efficient embedded biomedical sensors and on-device AI inference model executable on STM32 microcontrollers. The system measures multi-modal physiological data- ECG, SpO2, body temperature, motion and local process with a lightweight squeeze net model that is optimized with quantization and pruning. Data transmission is processed through a communication protocol based on LoRa, and it provides connectivity with low power and long range. The system has been evaluated experimentally with 25 participants where it is found to have an average classification accuracy of 96.5 %, inference latency of less than 95 ms and $9.3 \text{mJ} /$ inference energy usage. The proposed solution decreases the transmission volume by 96 %, increases the battery life to 11 days, and decreases emergency alert response time to 1.3 seconds as compared to the traditional cloud-based architectures. Privacy is guaranteed by the system because sensitive data are processed on-site, and the system is scalable enough to be used in the real world. This study supports the possibility of intelligent, low-latency, and sustainable health monitoring systems to be made possible by Edge-AI.

    20262026 International Conference on Innovative Computing, Intelligent Communication and Smart Electrica...(2026)
    引用
    AI阅读
    加入学术空间
    5Improved Deep Convolutional Network-Based Noninvasive Blood Glucose Level Monitoring in Diabetes Using Photoplethysmography Signals
    E. Edith Esther, P. Shobha Rani, N. Kamal, R. Partheepan

    Diabetes is a long-term disease which is caused by the body not being able to utilize the insulin which is produced by pancreas. Noninvasive is an approach that is used to measure the blood glucose levels utilizing photoplethysmography (PPG) signals. However, the existing machine learning models used for glucose level estimation heavily rely on data and suffer from overfitting issue. To overcome this limitation, a deep learning (DL)-based method, an improved deep convolutional neural network (IDCNN), is proposed to estimate the blood glucose levels in noninvasive technique accurately. The proposed IDCNN with several hidden layers learns the information from extracted features efficiently which captures the complex relationships from the data. The features are extracted from the PPG waveforms by empirical mode decomposition (EMD) method, and then, important features are selected based on importance scores. Finally, the selected features are fed into the proposed IDCNN-based evaluation model for blood glucose levels precisely. The performance of proposed IDCNN method is evaluated which attained R2 of 0.79, which represents less error value when compared to existing ML approach such as CatBoost.

    2026Emerging Trends in Communication and Signal Processing(2026)
    引用
    AI阅读
    加入学术空间
    立即登录,查看全部 243 篇论文

    合作机构(100)

    Saveetha Institute of Medical And Technical Sciences合作论文 14
    安那大学合作论文 13
    Chennai Institute of Technology合作论文 12
    SRM Institute of Science and Technology合作论文 11
    Panimalar Engineering College合作论文 11
    Saveetha Engineering College合作论文 11
    Velalar College of Engineering and Technology合作论文 10
    Rajalakshmi Institute of Technology合作论文 9
    Hindustan Institute of Technology and Science合作论文 9
    Sathyabama Institute of Science and Technology合作论文 8

    机构统计