Abhijeet-Pitumbur/Xray-Xplorer

Xray Xplorer · CNN, XGBoost & Grad-CAM · COVID-19 Detection in Chest X-Ray Images Using Explainable Boosting Algorithms · Flask Web Application · Python, TensorFlow, Keras · Available in Mauritius # **Xray Xplorer** · CNN, XGBoost & Grad-CAM · COVID-19 Detection in Chest X-Ray Images Using Explainable Boosting Algorithms Welcome to the GitHub repository for "Xray Xplorer", a powerful diagnostic tool birthed from a dissertation project titled "COVID-19 Detection in Chest X-Ray Images using Explainable Boosting Algorithms". The dissertation addresses the urgent demand for transparency in AI-powered diagnostic models. Uniting the robustness of Convolutional Neural Networks (CNNs), the potency of eXtreme Gradient Boosting (XGBoost), and the transparency of Gradient-weighted Class Activation Mapping (Grad-CAM), Xray Xplorer provides precise and interpretable COVID-19 predictions using chest X-ray images. Proven to perform admirably with an accuracy of 94.05% and an F1 score of 94.08%, this tool is ready to contribute to the fight against the pandemic. Xray Xplorer presents a robust interface for healthcare professionals to upload chest X-ray images, returning a diagnostic prediction of whether the image indicates a normal condition, pneumonia (viral or bacterial), or COVID-19. The web application offers a Grad-CAM visualization alongside the prediction, fostering explainability and interpretability - key requirements for trust in AI-driven diagnostic tools. ### View Xray Xplorer ##### Download Repository · GitHub This repository contains the source code for the Flask-based web application, designed to facilitate the deployment and use of the diagnostic AI model trained in the project's Jupyter notebook on Google Colab. ##### View Jupyter Notebook · Google Colab The complete dissertation report, outlining the comprehensive research, methodology, and broader implications, is available for deeper insights into the project. ##### View Dissertation · Google Drive All model files, including individual CNN models and the final hybrid CNN-XGBoost model, are available in the Google Drive linked below. ##### View Model Files · Google Drive ## Prerequisite …