MedGAN: Advanced Medical Image Generation

Overview

MedGAN is a comprehensive framework for generating high-quality synthetic medical images using state-of-the-art Generative Adversarial Networks (GANs). The project focuses on brain tumor MRI scans and includes implementations of multiple cutting-edge GAN architectures optimized for medical imaging applications.

Features

• Multiple GAN Implementations:

  • DCGAN (Deep Convolutional GAN)
  • ProGAN (Progressive Growing of GANs)
  • StyleGAN2 (Style-based Generator with improvements)
  • WGAN (Wasserstein GAN with gradient penalty)

• Web Application Interface:

  • Generate synthetic brain MRI scans
  • Detect tumor types from uploaded MRI images
  • Interactive and user-friendly interface

• Pre-trained Models:

  • Models for three tumor types: Glioma, Meningioma, and Pituitary
  • ViT-based tumor detection model (92% accuracy)

Architecture Performance Comparison

Architecture	Image Quality	Training Stability	Generation Diversity	Training Speed
ProGAN	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐
StyleGAN2	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐
WGAN-GP	⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐
DCGAN	⭐⭐⭐	⭐⭐⭐	⭐⭐	⭐⭐⭐⭐⭐

Getting Started

Prerequisites

• Python 3.9+
• PyTorch 1.9+
• Flask (for web application)
• CUDA-capable GPU (recommended)

Installation

1. Clone the repository:

git clone https://github.com/mozaloom/medgan.git
cd medgan

2. Install required packages:

pip install -r requirements.txt

3. Run the web application:

python app.py

4. Access the web interface at http://localhost:5000

Usage

Web Application

The MedGAN web application offers two primary functionalities:

1. Generate synthetic brain MRI scans:

   • Select tumor type (Glioma, Meningioma, Pituitary)
   • Choose GAN architecture
   • Generate high-quality synthetic MRI images

2. Detect tumor types:

   • Upload brain MRI scans
   • Receive AI-powered tumor classification
   • View detection confidence scores

Check the individual model implementation files for specific training parameters.

Project Structure

medgan/
├── app.py                   # Flask web application
├── medgan/                  # Core GAN implementations
│   ├── dcgan.py
│   ├── progan.py
│   ├── stylegan.py
│   ├── wgan.py
│   └── vit.py
├── models/                  # Pre-trained model weights
├── notebooks/               # Training notebooks
│   ├── dcgan/
│   ├── progan/
│   ├── stylegan/
│   └── wgan/
├── static/                  # Web assets
└── templates/               # HTML templates

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Brain Tumor MRI Dataset from Kaggle
• Research papers implementing the original GAN architectures:
  • DCGAN
  • ProGAN
  • StyleGAN2
  • WGAN