PNEUX-NET FOR PNEUMONIA DETECTION FROM X-RAYS

Pneumonia is a severe respiratory infection that significantly affects pediatric populations worldwide and remains one of the leading causes of mortality among children under five years of age. Early diagnosis is critical for effective treatment; however, interpreting chest X-ray images requires skilled radiologists and considerable clinical experience. In many healthcare settings, especially in developing regions, the shortage of trained radiologists leads to delayed diagnosis and increased medical risks. Recent advances in artificial intelligence and deep learning have enabled automated medical image analysis systems capable of assisting healthcare professionals in disease detection. This project proposes an Explainable Artificial Intelligence (XAI) driven system for pediatric pneumonia detection from chest X-ray images. The system utilizes a transfer learning approach based on the DenseNet121 convolutional neural network architecture to classify X-ray images into pneumonia and normal categories. A curated pediatric chest X-ray dataset is used for training and validation to ensure balanced representation and improved model generalization. To enhance transparency and clinical trust, the proposed system integrates Gradient-weighted Class Activation Mapping (Grad-CAM), which highlights important regions in the X-ray image that contribute to the model’s decision. These visual explanations allow clinicians to better understand the model predictions and verify diagnostic relevance. Additionally, the developed framework includes a secure web-based application that enables users to upload X-ray images and receive automated predictions with confidence scores and visual explanations. Experimental results demonstrate that the proposed model achieves high classification accuracy and reliable performance compared with traditional diagnostic approaches. The system provides a costeffective computer-aided diagnostic tool that can support early pneumonia screening and assist healthcare professionals in improving clinical decision making.