Explainable Ai for Covid-19 Detection: A Comparative Analysis of U-Net Architectures with Grad-Cam and Lime

Recent Artificial Intelligence (AI) advancements have significantly enhanced medical image analysis, particularly for COVID-19 diagnosis. Despite improved accuracy, Deep Learning (DL) models' “black-box“ nature raises interpretability concerns crucial for clinical trust. This paper comprehensively combines U-Net architectures with Gradient-weighted Class Activation Mapping (Grad-CAM) and Local Interpretable Modelagnostic Explanations (LIME) for COVID-19 chest X-ray interpretation. Our experimental results demonstrate significant improvements in both accuracy and interpretability. Additionally, we propose potential future enhancements through a Hybrid Attention Module (HAM) that could further improve feature representation and model explainability. The framework is evaluated on COVID-19 chest X-ray datasets, showing promising results for advancing transparent, reliable AI-driven diagnostic solutions.