Medical Image Classification for Pneumonia Detection Using Deep Learning

Respiratory illnesses like pneumonia continue to be a major threat to global health, particularly among children and the elderly. Timely detection directly influences patient outcomes, yet conventional diagnostic methods remain slow and heavily dependent on specialist availability. This paper introduces an end-to-end AI-powered web application that automates the analysis of chest X-ray images using a VGG16-based Convolutional Neural Network (CNN). The model is trained on a four-class dataset covering Normal, Bacterial Pneumonia, Viral Pneumonia, and COVID-19 cases, reaching approximately 86% accuracy on a held-out test set. To bridge the gap between raw model output and clinical understanding, the system incorporates the Google Gemini API, which converts prediction probabilities into straightforward clinical narratives. A React.js front end lets healthcare workers upload images and instantly view labelled results together with Grad-CAM heatmaps that visually indicate the affected lung regions. All data — user profiles, uploaded scans, and generated reports — are persisted in a MySQL database. The outcome is a fast, interpretable, and scalable diagnostic assistant that can meaningfully support early pneumonia screening, even in low-resource or rural healthcare environments.