Research

Papers I've co-authored with collaborators across multiple institutions.

    • ONGOING PUBLICATION
    • Institute of Electrical and Electronics Engineers (IEEE)

    Cross-Dataset Generalization Analysis of Deep Transfer Learning Models for Diabetic Retinopathy Classification

    This research evaluates the robustness of deep transfer learning architectures for diabetic retinopathy classification across different clinical datasets, identifying EfficientNetB2 as the most stable model against domain shift.

    • PUBLISHED
    • Springer, Cham

    Optimizing American Sign Language Recognition with Binarized Neural Networks - A Comparative Study with Traditional Models

    This undergraduate thesis compares the performance of Binarized Neural Networks (BNNs) against traditional models in the context of American Sign Language (ASL) recognition. The results suggest that BNNs are competitive with traditional models while requiring less computational resources.

    • PUBLISHED
    • Copernicus Publications

    HOLOSIE - A System Model for Assessing Carbon Emissions and Balance in Agricultural Systems

    HOLOSIE is a system model for assessing carbon emissions and balance in agricultural systems that simulates the carbon cycle and fluxes in agroecosystems. The model is designed to evaluate the impacts of different management practices, such as crop rotation, fertilization, and irrigation, on the carbon balance of agroecosystems.

    • PUBLISHED
    • SEMARAK ILMU SDN BHD

    Malaysian Sign Language Real-Time Tutorial using CNN Algorithm

    This research aims to develop a real-time Malaysian Sign Language (MSL) tutorial system using Convolutional Neural Networks (CNN) algorithm. The system is designed to provide immediate feedback to users based on their sign language skills.

    • ONGOING PUBLICATION

    An Optimized Deep-Learning Based Pipeline for Recognition of Sign Language from Low-Resolution Thermal Imagery

    This paper presents an optimized deep learning-based pipeline for recognizing sign language from low-resolution thermal imagery, demonstrating improved performance using Binarized Neural Networks (BNNs) and DenseNet121.