👁️ RetinaScan: Diabetic Retinopathy Detection

A deep learning-powered web application that helps detect and classify Diabetic Retinopathy from retinal fundus images. Built with a FastAPI backend and React frontend.

✨ Key Features

• 5-Stage Classification: Detects and grades Diabetic Retinopathy from No DR to Proliferative DR
• Real-time Analysis: Get instant predictions with confidence scores
• Responsive Design: Works seamlessly on desktop and mobile devices
• Developer Friendly: Well-documented API with Swagger UI
• Easy Deployment: Docker support for hassle-free setup

🛠 Tech Stack

Area	Technologies
Frontend	React 18, TypeScript, Tailwind CSS, Vite
Backend	FastAPI, Python 3.8+
ML/DL	PyTorch, OpenCV
Deployment	Docker, GitHub Actions
CI/CD	GitHub Actions

📸 Screenshots

🚀 Quick Start

Prerequisites

• Python 3.8+
• Node.js 16+
• npm or yarn
• Git

📁 Project Structure

diabetic-retinopathy-detection/
├── backend/                     # Python FastAPI backend
│   ├── main.py                  # FastAPI application entry point
│   ├── model.py                 # Deep learning model implementation
│   ├── preprocess.py            # Image preprocessing utilities
│   ├── requirements.txt         # Python dependencies
│   ├── trained_model.pth        # Trained PyTorch model
│   └── test_model.py            # Model testing script
├── frontend/                    # React frontend
│   ├── public/                  # Static files
│   └── src/
│       ├── components/         # Reusable UI components
│       ├── pages/               # Application pages
│       ├── App.tsx              # Main application component
│       └── main.tsx             # Application entry point
├── .github/                     # GitHub configurations
│   └── workflows/               # GitHub Actions workflows
├── .gitignore                   # Git ignore file
├── docker-compose.yml           # Docker Compose configuration
├── Dockerfile                   # Docker configuration
├── LICENSE                      # Project license
└── README.md                    # Project documentation

💻 Local Development

Backend Setup

1. Clone the repository

   git clone https://github.com/gaurav241102/diabetic-retinopathy-detection.git
   cd diabetic-retinopathy-detection

2. Set up Python environment

   # Create and activate virtual environment
   python -m venv venv
   .\venv\Scripts\activate  # Windows
   # source venv/bin/activate  # Mac/Linux
   
   # Install dependencies
   cd backend
   pip install -r requirements.txt

3. Start the backend server

   uvicorn main:app --reload

   • API Docs: http://localhost:8000/docs
   • Health Check: http://localhost:8000/health

Frontend Setup

1. Install dependencies

   cd ../frontend
   npm install

2. Start the development server

   npm run dev

   • Access at: http://localhost:5173

🤝 Contributing

Contributions are what make the open-source community an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

🙏 Acknowledgments

• APTOS 2019 Blindness Detection for the dataset

---

Built with ❤️ by Gaurav

📬 Contact

Gaurav Prakash - @yourtwitter - [email protected]

Project Link: [https://github.com/gaurav241102/diabetic-retinopathy-detection]

Usage

1. Open the frontend in your browser (default: http://localhost:5173)
2. Upload a retinal fundus image using the interface
3. The system will process the image and display the predicted DR grade

API Endpoints

• POST /predict/ - Accepts an image file and returns the predicted DR grade
  • Request: Form-data with key 'file' containing the image
  • Response: JSON with 'grade' and 'confidence' fields

Notes

• The model expects retinal fundus images in JPG or PNG format
• For best results, use high-quality images with proper lighting and focus
• The system is designed for screening purposes only and should not replace professional medical diagnosis