Pneumonia Detection from Chest X-Ray

This project is a web-based application that leverages a deep learning model to predict whether a given chest X-ray indicates pneumonia. The application is implemented using Streamlit, PyTorch, and other essential libraries to provide an easy-to-use interface for medical image analysis.

---

Table of Contents

1. Project Overview
2. Features
3. Architecture
4. Installation and Usage
5. Model Training
6. Results
7. Future Work
8. Contributing
9. License

---

Project Overview

This project aims to assist healthcare professionals and individuals by providing an AI-driven tool for pneumonia detection using chest X-ray images. The web app accepts an uploaded image, preprocesses it, and outputs a prediction along with a confidence score. It serves as an informational tool and emphasizes that it should not replace professional medical diagnosis.

---

Features

• Upload and Analyze X-Ray Images: Users can upload chest X-ray images in JPEG format for analysis.
• Deep Learning Model: Powered by a convolutional neural network (CNN) based on an extended TinyVGG architecture.
• Real-Time Predictions: Displays the prediction result (Normal or Pneumonia) with a confidence score.
• User-Friendly Interface: Built with Streamlit for intuitive interaction.

---

Architecture

Model Architecture:

The project employs a custom CNN architecture with the following key components:

• Multiple convolutional blocks with ReLU activation and MaxPooling.
• Fully connected layer for classification.
• Grayscale input preprocessing.

Tools and Libraries:

• Streamlit: For building the interactive web application.
• PyTorch: For model loading, inference, and tensor operations.
• Pillow (PIL): For image processing.
• Torchvision: For image transformations.

---

Installation and Usage

Prerequisites:

• Python 3.8 or higher.
• PyTorch (compatible with your hardware and OS).
• Streamlit and other required Python libraries.