Pneumonia Detection

A deep learning system for detecting pneumonia in chest X-rays using PyTorch.

Project Structure

pneumonia_detection/
├── data/                      # Data directory
│   └── chest_xray/            # X-ray image dataset
│       ├── train/             # Training images
│       ├── val/               # Validation images
│       └── test/              # Test images
├── models/                    # Saved model weights
│   └── pneumonia_model_v1.pth # Pre-trained model weights
├── pneumonia_detection/       # Package source code
|   ├── __init__.py            # Package initialisation
│   ├── dataset.py             # Data loading module
│   ├── paths.py               # Path management utilities
│   ├── pneumonia_net.py       # Model architecture definition
│   └── predict.py             # Prediction script
├── scripts/                   # Additional scripts
│   ├── __init__.py            # Package initialisation
│   ├── test.py                # Script for evaluating the model
│   └── train.py               # Script for training the model
├── tests/                     # Unit tests
│   ├── __init__.py            # Test package initialisation
│   ├── test_dataset.py        # Tests for the dataset module
│   ├── test_paths.py          # Tests for the path management utilities
│   ├── test_pneumonia_net.py  # Tests for the model architecture
│   └── test_system.py         # System tests
├── .gitignore                 # Git ignore file
├── LICENCE                    # Licence file
├── pyproject.toml             # Package configuration
├── README.md                  # Project documentation
├── requirements-dev.txt       # Development dependencies
└── requirements.txt           # Project dependencies

Installation

pip install -e .

Usage

Training

python scripts/train.py

Prediction

Using the installed package:

python -m pneumonia_detection.predict path/to/image.jpeg

Or directly:

predict path/to/image.jpeg

Docker Usage

This guide explains how to build and run the Pneumonia Detection project using Docker.

Building the Docker Image

docker build -t pneumonia-detection .

Running the Docker Container

For predictions using the CLI

To analyze images from your local machine, you need to mount the directory containing the images:

docker run -it --rm -v ./data:/data pneumonia-detection pneumonia_detection.predict /data/image.jpeg

Using GPU support

docker run -it --rm --gpus all -v ./data:/data pneumonia-detection pneumonia_detection.predict /data/image.jpeg

Dataset

This project uses the Chest X-Ray Images (Pneumonia) dataset from Kaggle, which is available under the CC BY 4.0 license.